ATP

00:00:00 ◼   ► Oh yes, we will be angry because John is refusing to admit how wrong he is, and I'm angry that

00:00:05 ◼   ► John won't admit how wrong he is.

00:00:07 ◼   ► Oh god.

00:00:08 ◼   ► [Music]

00:00:09 ◼   ► So, I think I actually might have solved my Apple Watch Ultra band problem.

00:00:15 ◼   ► Mmm, okay.

00:00:16 ◼   ► So, alright.

00:00:17 ◼   ► The Apple Watch Ultra, it has a bunch of bands that are okay, that are made specifically

00:00:22 ◼   ► for it.

00:00:23 ◼   ► They're all okay in different ways, none of them are great.

00:00:25 ◼   ► I thought, I'm not trying to be funny, I thought that all of, or most of the ultra-specific

00:00:31 ◼   ► bands were really well-liked. Is that not true? Did I make that up?

00:00:35 ◼   ► Okay, here's the problem they have. The Alpine Loop. This is the one that has its cloth,

00:00:41 ◼   ► and it has all little tiny loops and you hook with the metal thing, you hook into one of

00:00:45 ◼   ► them to put it on. It is most commonly seen in its orange color, which I think is one

00:00:50 ◼   ► of the best looking Apple Watch Ultra bands, is the orange Alpine Loop. Problem is, it

00:00:55 ◼   ► Well, first of all, I don't find it super easy

00:00:58 ◼   ► to get a good fit on that

00:00:59 ◼   ► because of the weird hook mechanism.

00:01:00 ◼   ► You gotta like keep popping it out

00:01:01 ◼   ► and get the right one and everything.

00:01:03 ◼   ► So it's not super easy to get a good fit on that.

00:01:05 ◼   ► And then second of all, it's cloth.

00:01:07 ◼   ► And so it absorbs moisture and stuff.

00:01:09 ◼   ► And my primary use of my Apple Watch Ultra is workouts.

00:01:13 ◼   ► And so that's a no go for me.

00:01:16 ◼   ► Also the Alpine loops tend, because they are cloth,

00:01:18 ◼   ► they tend to get dirty and discolored fairly easily

00:01:20 ◼   ► and fairly quickly after you own them.

00:01:22 ◼   ► And so you can wash them, again,

00:01:24 ◼   ► just like the woven loops you can watch them to some extent

00:01:26 ◼   ► but it's, you know, whatever.

00:01:28 ◼   ► All right, then you have the trail loops.

00:01:31 ◼   ► These are kind of like the Apple Watch Ultra version

00:01:33 ◼   ► of those like Velcro regular Apple Watch loops

00:01:36 ◼   ► that everyone loves that I think look like sweatpants

00:01:38 ◼   ► and on the Apple Watch Ultra they look like sweatpants.

00:01:41 ◼   ► So it's fun, the trail loops are fine.

00:01:45 ◼   ► I think the black and gray trail loop

00:01:47 ◼   ► with the little orange tab is a nice,

00:01:50 ◼   ► It's a relatively nice look for the category,

00:01:54 ◼   ► but still not, it's very utilitarian, let's say.

00:01:58 ◼   ► And then the Ocean Band is kind of like

00:02:02 ◼   ► the Apple Watch Ultra version of the Sport Band.

00:02:05 ◼   ► So the Ocean Band, this is the one

00:02:07 ◼   ► I've spent the most time with.

00:02:08 ◼   ► I have the Midnight, the little navy blue.

00:02:10 ◼   ► I have the Midnight Ocean Band,

00:02:12 ◼   ► and I think it looks okay, and it works okay,

00:02:15 ◼   ► and it's kind of comfortable.

00:02:17 ◼   ► But none of that compares to like the regular

00:02:21 ◼   ► Apple Watch sport bands, which I think,

00:02:24 ◼   ► they have a certain look, you know,

00:02:25 ◼   ► they look mostly, they can look mostly neutral.

00:02:28 ◼   ► But they feel great, they're very comfortable,

00:02:30 ◼   ► very versatile, et cetera.

00:02:32 ◼   ► - Wait, wait, so would you say this is the most comfortable

00:02:34 ◼   ► of the three, because from looks alone,

00:02:36 ◼   ► it does not appear to be terribly comfortable to me.

00:02:39 ◼   ► - So I haven't owned a Trail Loop.

00:02:40 ◼   ► I've tried it on in the store,

00:02:41 ◼   ► so I cannot say I've owned a Trail Loop.

00:02:43 ◼   ► I own the Alpine Loop and the Ocean Band,

00:02:45 ◼   ► and I spent by far the most time with the Ocean Band,

00:02:48 ◼   ► in part because it is durable and sweat resistant

00:02:51 ◼   ► 'cause it's just rubber.

00:02:52 ◼   ► But also I find it mostly uncomfortable.

00:02:54 ◼   ► The problem is the ridges that it has,

00:02:57 ◼   ► like the ridges are on both sides,

00:02:59 ◼   ► so the side facing your skin is also ridged.

00:03:02 ◼   ► And so usually I will end up with little imprints of ridges

00:03:05 ◼   ► in my skin after wearing it, which again is not great.

00:03:10 ◼   ► So I've tried, you know, there's nothing stopping you

00:03:12 ◼   ► from putting any other Apple Watch Band on the Ultra.

00:03:15 ◼   ► Obviously they don't all look very good on it necessarily.

00:03:19 ◼   ► That's up to you and your god.

00:03:21 ◼   ► But you can put them on, they fit.

00:03:24 ◼   ► They may or may not match the metal very well,

00:03:26 ◼   ► they may or may not have the right shape or the right style,

00:03:28 ◼   ► but you can put on other bands and they fit just fine.

00:03:30 ◼   ► So I have tried lots of other Apple Watch bands

00:03:34 ◼   ► with the Ultra trying to figure out like,

00:03:36 ◼   ► all right, is there a better option than the Ocean band

00:03:39 ◼   ► for general Ultra use?

00:03:41 ◼   ► The problem with the regular sport band,

00:03:45 ◼   ► the standard pin buckle, regular sport band

00:03:48 ◼   ► we've had forever with the Apple Watch.

00:03:50 ◼   ► The way the Ultra, it can kinda like float on your,

00:03:53 ◼   ► it's so big and thick that it can kinda slide,

00:03:56 ◼   ► like if you're looking at your wrist all the time,

00:04:00 ◼   ► the direction of the watch moving up,

00:04:01 ◼   ► it can slide up a little bit.

00:04:03 ◼   ► And then it creates this gap between the band

00:04:07 ◼   ► and your wrist bone that looks really stupid

00:04:10 ◼   ► with the sport band, because the sport band,

00:04:12 ◼   ► because it's made for smaller, lower watches,

00:04:14 ◼   ► it kinda sticks out further from the watch body

00:04:18 ◼   ► before coming down around your wrist.

00:04:20 ◼   ► Does that make sense?

00:04:21 ◼   ► - It does.

00:04:22 ◼   ► - Like it comes out straight more before it goes down.

00:04:25 ◼   ► The result of that is that the Apple Watch Ultra

00:04:27 ◼   ► I think only looks good with an Apple Sport band

00:04:30 ◼   ► if you have really big wrists and that way

00:04:32 ◼   ► it has more time for the band to come like straight out

00:04:35 ◼   ► before wrapping around your wrist going down.

00:04:38 ◼   ► It looks better that way.

00:04:39 ◼   ► But you, Casey, inspired me last week.

00:04:43 ◼   ► I'm like wait a minute.

00:04:44 ◼   ► The stretchy sport loops that Apple sells,

00:04:47 ◼   ► the fixed size ones, they're a different material entirely.

00:04:51 ◼   ► And it actually works really well on the Ultra.

00:04:56 ◼   ► It still doesn't look quite right,

00:04:59 ◼   ► but it's a way better look and fit

00:05:03 ◼   ► than the sport band with the pin buckle on the Ultra.

00:05:07 ◼   ► So I am actually very pleased to have discovered this,

00:05:09 ◼   ► and I thank you very much that the,

00:05:11 ◼   ► here it's called the solo loop is what I'm talking about the solo loop actually

00:05:16 ◼   ► looks and fits and works surprisingly well on the ultra so thank you for that

00:05:21 ◼   ► I am happy with that did you have to buy one size down because the ultra is wider

00:05:25 ◼   ► that's the part I'm also okay so all the ones I have are for the smaller Apple

00:05:31 ◼   ► watch diameter they're for like the 40 millimeter or 41 millimeter size and I

00:05:34 ◼   ► back when they first came out I I kind of went through figuring out my size and

00:05:39 ◼   ► I ended up with two sizes, a seven and a six.

00:05:41 ◼   ► And I wear the six, I think, most of the time.

00:05:45 ◼   ► So I was trying to figure out, going to Gruber's article

00:05:47 ◼   ► and looking at different things of like,

00:05:48 ◼   ► how do I size the solo loop without going to a store for,

00:05:52 ◼   ► okay, I know what I wear in the sport band,

00:05:55 ◼   ► I know which hole I use in the sport band,

00:05:57 ◼   ► and I already have these sizes seven and six solo loops

00:06:02 ◼   ► that I can test with, but therefore the smaller watch.

00:06:05 ◼   ► So what Underscore, he tweeted about this also,

00:06:08 ◼   ► or mastituted about it.

00:06:10 ◼   ► So he figured out that to go from the 45 millimeter

00:06:15 ◼   ► to the ultra, you know, 'cause they use the same straps,

00:06:18 ◼   ► to go between those two, generally subtract one

00:06:21 ◼   ► from the size of the loop, and that usually fits.

00:06:24 ◼   ► Problem is, I think I'm kinda between sizes,

00:06:26 ◼   ► so I tried, I don't even know if the 45 sizes

00:06:30 ◼   ► and the 40 millimeter sizes, I don't even know

00:06:33 ◼   ► if those are the same.

00:06:34 ◼   ► Like, is the six in both of those the same length?

00:06:37 ◼   ► because the watches aren't the same height.

00:06:42 ◼   ► So if you take a strap of a fixed length

00:06:46 ◼   ► and you use it on a bigger watch,

00:06:47 ◼   ► the strap will be too loose.

00:06:49 ◼   ► Because the watch, it's attaching at different points

00:06:53 ◼   ► on the wrist than the smaller watch would.

00:06:55 ◼   ► So you'll be off by one, basically.

00:06:58 ◼   ► So that's why when you go from the 45 millimeter

00:07:01 ◼   ► to the ultra, the watch is getting taller

00:07:04 ◼   ► and so the same length strap,

00:07:06 ◼   ► you have to basically make the strap shorter

00:07:08 ◼   ► to make it have the same fit on your wrist.

00:07:10 ◼   ► So I ordered the five, but it's a little tight,

00:07:15 ◼   ► but I tried my six from my small watch,

00:07:18 ◼   ► and it was too loose by a decent amount.

00:07:20 ◼   ► - Oh no.

00:07:21 ◼   ► - So I'm like, all right, I don't know.

00:07:22 ◼   ► So I don't know, I think I might have to actually

00:07:24 ◼   ► go into a store and figure this out,

00:07:27 ◼   ► 'cause I don't wanna do a bunch of returns

00:07:28 ◼   ► and be all wasteful, so.

00:07:30 ◼   ► Anyway, all that is to say,

00:07:32 ◼   ► sizing is still a question mark for me personally,

00:07:34 ◼   ► But the solo loop, I think, is a much more mechanically

00:07:39 ◼   ► sound and visibly sound option for the Apple Watch Ultra

00:07:44 ◼   ► compared to the regular sport band.

00:07:45 ◼   ► So I can recommend it.

00:07:46 ◼   ► I got the Storm Blue, which is kind of this navy blue.

00:07:49 ◼   ► It's a super boring color, but it's the only currently

00:07:52 ◼   ► offered color that I think would look good on the Ultra.

00:07:54 ◼   ► They do have a nice bright yellow,

00:07:56 ◼   ► and the Ultra has that yellow ocean band.

00:07:58 ◼   ► But I saw that in a store, I don't think the Ultra

00:08:01 ◼   ► goes well with yellow, 'cause it has the orange accents

00:08:03 ◼   ► and I don't think it's a good match with yellow,

00:08:07 ◼   ► but I could be wrong, I might try it, we'll see.

00:08:09 ◼   ► - I got a lot of feedback on Mastodon

00:08:11 ◼   ► about my solo loop problems,

00:08:14 ◼   ► and I think one thing either I wasn't clear

00:08:17 ◼   ► about the particular watch band I was talking about,

00:08:19 ◼   ► or people misunderstood or both,

00:08:21 ◼   ► but the one I was talking about is indeed the solo loop,

00:08:24 ◼   ► which is the thing, it's, what is it,

00:08:26 ◼   ► not floral, astomer, whatever, but it's a--

00:08:28 ◼   ► - Apple describes it as stretchable liquid silicone rubber.

00:08:32 ◼   ► Now, to be clear, they are not liquid.

00:08:34 ◼   ► I don't know. (laughing)

00:08:36 ◼   ► They are definitely solid, and thank God for that.

00:08:38 ◼   ► They would not be very good watch bands if they were liquid.

00:08:41 ◼   ► - Indeed.

00:08:41 ◼   ► But then a lot of people sent me recommendations

00:08:45 ◼   ► for and forgive me, I don't remember the name of this one,

00:08:47 ◼   ► but it's the same premise where it's just one piece

00:08:50 ◼   ► of material, but it's woven fabric.

00:08:53 ◼   ► What is the name of that one?

00:08:54 ◼   ► - That's the braided solo loop.

00:08:56 ◼   ► Same name, but braided in front and cost twice as much.

00:08:59 ◼   ► - Right, and it does, you're exactly right,

00:09:00 ◼   ► it costs twice as much.

00:09:01 ◼   ► They're wonderful, by the way.

00:09:02 ◼   ► That's the one I was saying.

00:09:03 ◼   ► That's the one I have the pride version of, the rainbow one.

00:09:06 ◼   ► When they are clean, they are great.

00:09:09 ◼   ► And when they are dry.

00:09:10 ◼   ► When they are clean and dry, they are fantastic.

00:09:14 ◼   ► But that's two big wins.

00:09:15 ◼   ► - Yeah, so that's the thing.

00:09:17 ◼   ► And so I've been casting about on Amazon trying to find a silicon, or silicon, whatever, I

00:09:25 ◼   ► always get it wrong, a quasi-plastic solo loop to get on the cheap, even if it isn't

00:09:31 ◼   ► quite as good because I agree with what you were saying last week Marco that you

00:09:34 ◼   ► know oftentimes materials on these knockoffs it's not nearly as good as the

00:09:37 ◼   ► Apple stuff but if I'm spending ten dollars every six months instead of 50

00:09:42 ◼   ► dollars every six months then it's not such a big deal and I haven't I haven't

00:09:46 ◼   ► tried any yet but meanwhile my friend Spencer for a belated birthday gift the

00:09:51 ◼   ► same Spencer that you guys know sent me a couple of knockoff braided solo loop

00:09:56 ◼   ► bands and so far I don't know which ones these were otherwise I'd link them but

00:10:01 ◼   ► So far these are really, really nice,

00:10:03 ◼   ► and they don't seem to suck in water or moisture

00:10:07 ◼   ► as much as I would have expected.

00:10:09 ◼   ► However, the problem with, maybe not this one in particular,

00:10:12 ◼   ► but the braided solo loop in general,

00:10:14 ◼   ► is that a lot of people said on Mastodon

00:10:16 ◼   ► that after six months, well, they don't break in two

00:10:18 ◼   ► like yours did, but they're so damn stretched

00:10:21 ◼   ► that you can't wear it anymore,

00:10:22 ◼   ► which is also kind of a, it's a failure,

00:10:25 ◼   ► just a different kind of failure,

00:10:26 ◼   ► perhaps less catastrophic, but a failure nonetheless.

00:10:28 ◼   ► So it's tough because I love, I love, love, love

00:10:32 ◼   ► the plastic solo loop.

00:10:34 ◼   ► And I like in principle the braided solo loop

00:10:37 ◼   ► and we'll see how well this Amazon knockoff lasts.

00:10:40 ◼   ► The material seems good, like it's comfortable.

00:10:43 ◼   ► But yeah, I can't with an honest heart recommend either

00:10:45 ◼   ► unless you're looking to replace them

00:10:49 ◼   ► every six months to a year,

00:10:50 ◼   ► depending on your particular use case.

00:10:52 ◼   ► - Not that I want to extend this conversation

00:10:53 ◼   ► of watch drops any longer, but one person on Macedon

00:10:57 ◼   ► did ask, made a snarky comment.

00:10:59 ◼   ► They said, you know, first Casey says that the Apple ones

00:11:02 ◼   ► are like breaking on him after six months.

00:11:05 ◼   ► And then when third party watch bands are recommended,

00:11:08 ◼   ► Marco comes in and says, oh,

00:11:10 ◼   ► third party watch bands are bad quality.

00:11:12 ◼   ► And that's the type of person that if I was snarky,

00:11:16 ◼   ► I would have replied to.

00:11:17 ◼   ► And I'm honestly not doing this to be snarky,

00:11:19 ◼   ► but I'm mostly doing, I don't know.

00:11:21 ◼   ► It's kind of being jerky.

00:11:24 ◼   ► To be clear, I didn't do this.

00:11:25 ◼   ► I didn't respond in this way.

00:11:26 ◼   ► But I thought about giving my typical response where I throw it back to them and I say, "You

00:11:32 ◼   ► square that circle.

00:11:33 ◼   ► Figure it out.

00:11:34 ◼   ► So you've laid down two seemingly contradictory things.

00:11:37 ◼   ► Apple ones are breaking and we agree they're breaking too soon, but then third-party ones

00:11:41 ◼   ► are bad quality so we don't want them."

00:11:45 ◼   ► How can you reconcile that?

00:11:46 ◼   ► You seem to think it's a contradiction or it's hypocritical or whatever, but is there

00:11:51 ◼   ► a way that it wouldn't be a contradiction?

00:11:55 ◼   ► that problem and it's very difficult to express that in a tweet or a tweet or whatever without

00:12:00 ◼   ► being obnoxious so I didn't.

00:12:03 ◼   ► But now we're on the podcast so I can have more room to explain it.

00:12:06 ◼   ► And I'm not the one who said all these things but I was a listener for them and the reason

00:12:11 ◼   ► I didn't make a comment is because it made perfect sense to me and here's how it made

00:12:14 ◼   ► sense to me.

00:12:15 ◼   ► Apple ones shouldn't break after six months.

00:12:17 ◼   ► That's no good.

00:12:18 ◼   ► They're expensive.

00:12:19 ◼   ► They should be sturdier.

00:12:20 ◼   ► But also third party ones being "oh bad quality, don't buy them."

00:12:23 ◼   ► For the six months that you're wearing it, the third party ones could be uglier or less

00:12:29 ◼   ► comfortable.

00:12:31 ◼   ► Even if they don't break, even if they never break, even if they were like Infinite Gobstopper,

00:12:34 ◼   ► the watch band that never breaks.

00:12:36 ◼   ► If during that entire time it is not as nice looking or as nice feeling as the watch band,

00:12:42 ◼   ► that I felt like is what Marco was expressing.

00:12:43 ◼   ► Oh, the third party ones, they're not as good quality.

00:12:46 ◼   ► Maybe there's a visible seams where the mold lines are.

00:12:49 ◼   ► Maybe it's stiffer.

00:12:50 ◼   ► Maybe they're not as soft on your skin.

00:12:52 ◼   ► more to the watch thing other than whether it breaks or not.

00:12:58 ◼   ► It's kind of like my cheese grater thing.

00:13:00 ◼   ► It's my favorite cheese grater, but it's got a fatal flaw that they break after six months.

00:13:03 ◼   ► But it's still the one I want to use because when it's working, it's better than all the

00:13:07 ◼   ► other ones that I've tried.

00:13:08 ◼   ► That's the way I squared that circle in my mind, and Marco can say that's what he was

00:13:12 ◼   ► getting at when he said third party ones are not as nice.

00:13:15 ◼   ► - Yeah, basically, I mean, it's like, you know,

00:13:17 ◼   ► you can save some money, you know, by going third party,

00:13:21 ◼   ► and there are, look, I'm not gonna say there aren't

00:13:24 ◼   ► any nice third party Apple Watch bands,

00:13:26 ◼   ► because there probably are, I just have never found one.

00:13:30 ◼   ► And so the ones I have tried,

00:13:31 ◼   ► I've all been really disappointing.

00:13:33 ◼   ► And so, look, if you can save some money and do that,

00:13:36 ◼   ► then great, but if you're gonna get something

00:13:38 ◼   ► that you just kinda think is okay, and isn't that nice,

00:13:43 ◼   ► and you're wearing it every single day, I don't know.

00:13:46 ◼   ► I would rather, you know, I don't think 50 bucks

00:13:50 ◼   ► for an Apple Watch band maybe once a year as a replacement.

00:13:53 ◼   ► I don't think that's unreasonable

00:13:54 ◼   ► for something that you wear every day.

00:13:55 ◼   ► - I think they should last longer than a year.

00:13:57 ◼   ► To be clear, I'm not saying the Apple one is better.

00:13:59 ◼   ► I'm saying these are two bad choices.

00:14:00 ◼   ► I'm saying they're not contradictory, that's all.

00:14:02 ◼   ► - And some of them do, by the way.

00:14:03 ◼   ► - Yeah, exactly.

00:14:04 ◼   ► We're just talking about one specific one.

00:14:06 ◼   ► - Yeah, this is inherent to like, you know,

00:14:08 ◼   ► stretchy rubber or cloth, both of which wear out

00:14:11 ◼   ► substantially faster than say, non-stretchy rubber or metal.

00:14:15 ◼   ► Like if you want something to last forever,

00:14:17 ◼   ► the sport band or the metal bands,

00:14:20 ◼   ► those last forever basically.

00:14:22 ◼   ► But the woven cloth ones, while very nice,

00:14:26 ◼   ► you're gonna compromise there in comfort,

00:14:30 ◼   ► or sorry, in longevity.

00:14:33 ◼   ► So it's a trade-off.

00:14:35 ◼   ► - Yep, anyway, the snarky thing that I,

00:14:38 ◼   ► I don't think it to be snarky,

00:14:39 ◼   ► I don't know what the word for it is,

00:14:40 ◼   ► probably some debate type thing for it, but it's like, come up with the other side.

00:14:43 ◼   ► Like you don't need me to answer this question.

00:14:45 ◼   ► You think you're throwing a gotcha out at me, but I bet if you thought about it for

00:14:48 ◼   ► a couple seconds, you could figure out from the other person's perspective, how could

00:14:51 ◼   ► both of these things be true?

00:14:52 ◼   ► And in this case, it's like, you know, it's what I explained before.

00:14:55 ◼   ► Anyway, I haven't figured out how to communicate that online without looking like a jerk.

00:15:00 ◼   ► So mostly I don't.

00:15:01 ◼   ► Except this whole moment of the podcast that's being broadcast to far more people.

00:15:05 ◼   ► I know.

00:15:06 ◼   ► Well, I feel like in a podcast, I have time to explain it.

00:15:09 ◼   ► and hopefully people understand and whatever.

00:15:12 ◼   ► And I have Marco here to be able to clarify

00:15:14 ◼   ► whether I was misinterpreting what he was saying.

00:15:16 ◼   ► You know what I mean?

00:15:17 ◼   ► Whatever.

00:15:18 ◼   ► - We are brought to you this week by Collide,

00:15:21 ◼   ► and they have some big news.

00:15:23 ◼   ► If you're an Okta user,

00:15:24 ◼   ► they can get your entire fleet to 100% compliance.

00:15:28 ◼   ► How?

00:15:28 ◼   ► Well, if a device isn't compliant,

00:15:30 ◼   ► the user can't log into your cloud apps

00:15:32 ◼   ► until they fix the problem.

00:15:33 ◼   ► It's that simple.

00:15:34 ◼   ► Collide patches one of the major holes

00:15:37 ◼   ► in zero trust architecture, device compliance.

00:15:40 ◼   ► Without Collide, IT can struggle to solve basic problems,

00:15:43 ◼   ► like keeping everyone's OSes and browsers up to date.

00:15:46 ◼   ► Unsecure devices are logging into your company's apps

00:15:48 ◼   ► because there's nothing there to stop them.

00:15:50 ◼   ► Collide is the only device trust solution

00:15:53 ◼   ► that enforces compliance as part of authentication,

00:15:56 ◼   ► and it's built to work seamlessly with Okta.

00:15:59 ◼   ► The moment Collide's agent detects a problem,

00:16:01 ◼   ► it alerts the user and gives them instructions to fix it.

00:16:04 ◼   ► If they don't fix the problem within a set time,

00:16:06 ◼   ► They're blocked.

00:16:07 ◼   ► Collide's simple method means fewer support tickets,

00:16:10 ◼   ► less frustration, and most importantly,

00:16:13 ◼   ► 100% fleet compliance.

00:16:15 ◼   ► Visit collide.com/ATP to learn more or book a demo.

00:16:20 ◼   ► Collide is spelled K-O-L-I-D-E.

00:16:23 ◼   ► So that's collide, K-O-L-I-D-E, .com/ATP

00:16:27 ◼   ► to learn more or book a demo today.

00:16:29 ◼   ► Thank you so much to Collide for sponsoring our show.

00:16:35 ◼   ► - All right, let's do some follow up.

00:16:37 ◼   ► Starting with iOS 16.4, which is not out, right?

00:16:41 ◼   ► It's imminent, is that correct?

00:16:43 ◼   ► - I think it's GM, right?

00:16:44 ◼   ► - Yeah, yeah, okay.

00:16:45 ◼   ► So iOS 16.4 should be out any minute now.

00:16:48 ◼   ► Expands duplicate image detection

00:16:50 ◼   ► to iCloud shared photo library.

00:16:51 ◼   ► Jon, you were pretty displeased by this,

00:16:54 ◼   ► if I'm not mistaken, when this first came out.

00:16:55 ◼   ► Is that right, or am I making that up?

00:16:57 ◼   ► - Yeah, I mean, it was still more annoyed

00:16:58 ◼   ► by the lack of shared albums, but this seemed like a gap,

00:17:00 ◼   ► because they added duplicate photo detection

00:17:02 ◼   ► with iOS 16 and whenever they released all this stuff,

00:17:05 ◼   ► and it'll, you know, if you scroll down at the bottom,

00:17:06 ◼   ► I don't know if they added it then,

00:17:07 ◼   ► but there's like, there's a section where you can,

00:17:09 ◼   ► where it says duplicates,

00:17:10 ◼   ► if you scroll down on the iOS photos app

00:17:12 ◼   ► and it'll find them and like merge them and stuff like that.

00:17:14 ◼   ► But the problem I had was like, you know, we had,

00:17:16 ◼   ► my wife had the big family library

00:17:18 ◼   ► and then I had my personal library

00:17:19 ◼   ► and there was some overlap between the libraries

00:17:21 ◼   ► because sometimes I would take something,

00:17:24 ◼   ► I would take everything from a personal library

00:17:25 ◼   ► and import it into hers,

00:17:26 ◼   ► but sometimes I would pull photos from her library

00:17:29 ◼   ► after I edited them down into mine.

00:17:31 ◼   ► And then when we tried to merge them

00:17:32 ◼   ► into one big shared family library,

00:17:34 ◼   ► there is duplicates because I was putting things back into the shared library that I

00:17:38 ◼   ► had pulled from her library previously, but there was no duplicate detection in the shared

00:17:43 ◼   ► library.

00:17:44 ◼   ► I mean, you could do it manually by scrolling through the photos and trying to find the

00:17:46 ◼   ► ones that are the same or whatever, but the whole point is I want the computer to do it

00:17:49 ◼   ► for me.

00:17:50 ◼   ► So now in iOS 16 and also in whatever, I guess Mac OS 13.3 or whatever the next version of

00:17:55 ◼   ► Mac OS will also have this ability, now you actually can find duplicates in the shared

00:18:01 ◼   ► library. And this is going to help me a lot because I have not shoved all of my personal

00:18:06 ◼   ► photos into the shared library just because I didn't want to deal with the duplicate issue.

00:18:09 ◼   ► So I'll test this feature out with one or two photos to see how it handles it. According to

00:18:14 ◼   ► Kurt on Mastodon, when you do the merge, it attributes the photo to both people.

00:18:18 ◼   ► So instead of just saying like, this was added to the shared library by such and such,

00:18:23 ◼   ► it will say this was added to the shared library by these two people or these three people,

00:18:27 ◼   ► I assume if it's third-party, you know, so it's maintaining a metadata about where it came from that like hey this

00:18:31 ◼   ► You know multiple people out of this photo, but it will only keep one of them and hopefully it will do that in a smart

00:18:36 ◼   ► Way, so I look forward to testing this out

00:18:38 ◼   ► John you chose poorly that's a reference

00:18:42 ◼   ► Artings calm has done some OLED burn in testing for a long period of time

00:18:48 ◼   ► I forget exactly how long this just came out in last week or two

00:18:52 ◼   ► But I was just brought to our attention in the last 24 hours

00:18:55 ◼   ► Sounds like the Sony OLEDs not a good choice if you're worried about about image retention

00:19:01 ◼   ► Whoopsie-dupsie if you're worried about image retention OLEDs are probably not the thing for you

00:19:05 ◼   ► I don't think these results are surprising in any way whatsoever

00:19:09 ◼   ► So if the art thing our tings has been doing OLED burn-in tests for years and years now

00:19:13 ◼   ► Obviously they can't test the TV that just came out last year for more than the length of time

00:19:17 ◼   ► they had it but they've been testing it for many many years and

00:19:21 ◼   ► The results have not been surprising OLEDs have image retention. They way way way way way more than LCDs like it is

00:19:27 ◼   ► Maybe not as much as plasmas, but it is a big problem with OLED televisions. I knew this going in

00:19:34 ◼   ► There is no OLED television you can buy that will not suffer from image retention

00:19:37 ◼   ► So the most recent thing is they did tests of the new TVs they came out last year and obviously they only have them for

00:19:43 ◼   ► A few months right, but they do very accelerated testing. This is not representative of normal use

00:19:47 ◼   ► They like, you know, put it on CNN with a big ticker on the bottom.

00:19:51 ◼   ► They leave it on for 24 hours a day, seven days a week, right?

00:19:54 ◼   ► They really, it's accelerated ware testing to see how it is.

00:19:59 ◼   ► And lo and behold, if you do that, the machine's getting these television's get image retention.

00:20:03 ◼   ► But the surprising topic was testing the Quantum Dot OLEDs versus the old style W OLEDs.

00:20:09 ◼   ► And the Quantum Dot OLEDs exhibited what I would expect from, you know, OLED burn-in,

00:20:15 ◼   ► But the latest generation of, or last year's generation of the WRGB OLEDs, especially in

00:20:21 ◼   ► this particular test with white regions on the, you know, like a bar on the bottom of

00:20:25 ◼   ► a news thing, resisted burn-in over the same period of time.

00:20:29 ◼   ► The explanation they gave, they don't know exactly, but the explanation they gave seems

00:20:32 ◼   ► plausible to me, is that when you have the white subpixel and the WRGB OLEDs and you're

00:20:37 ◼   ► showing white, you don't need to, you won't wear out the R, the G, and the B subpixels

00:20:42 ◼   ► when you're showing white, because the white sub-pixel will carry that load entirely.

00:20:45 ◼   ► And the white sub-pixel is actually pretty big.

00:20:47 ◼   ► I don't know if it's the biggest sub-pixel, but it's pretty big.

00:20:50 ◼   ► Maybe it's about the same size as the green one or the next biggest one is.

00:20:54 ◼   ► Whereas on a QD OLED, it's just RGB.

00:20:57 ◼   ► There is no white sub-pixel, which is great for color purity and all that other stuff.

00:21:00 ◼   ► But when you have to show white, that means you have to turn on the R, the G, and the

00:21:03 ◼   ► B sub-pixels.

00:21:04 ◼   ► So if there's a big white static element on your screen, like the big news ticker that's

00:21:08 ◼   ► on the bottom of CNN the whole time, it's going to wear out the R, the G, and the B

00:21:13 ◼   ► subpixels, whereas the R, the G, and the B subpixels on the WRGB things will not be worn

00:21:17 ◼   ► out by showing something that's white.

00:21:21 ◼   ► So the results of this were all the QD OLEDs, not just Sony, but the Samsung ones or whatever,

00:21:27 ◼   ► burned in faster than the latest generation of WRGB OLEDs.

00:21:30 ◼   ► This also makes sense because this is literally the first generation of QD OLEDs, and that's

00:21:35 ◼   ► like the 19th generation of WRGB OLEDs and they've been fighting burn-in for many many years.

00:21:40 ◼   ► So I expect this to improve. Doesn't affect me at all because I already baby my television. Like

00:21:46 ◼   ► again I had a plasma and my plasma was insanely, by the end of its life especially, insanely

00:21:51 ◼   ► subject to image retention. I remember I'd turn on my plasma and it would launch into the Apple TV

00:21:56 ◼   ► screen with the little you know rounded rectangles. Just launching into the Apple TV screen in the

00:22:01 ◼   ► time it took me to like, you know, remote over to the app I want and launch it.

00:22:06 ◼   ► And I'm not dilly dallying.

00:22:07 ◼   ► I'm like, as soon as it comes up, move, move, move, hit.

00:22:09 ◼   ► I'd see an after image of the rectangles that were on the screen.

00:22:13 ◼   ► Oh my.

00:22:13 ◼   ► Oh, wow.

00:22:14 ◼   ► And it would fade quickly, but it was like, it was, it was like

00:22:18 ◼   ► Marco's old iMac, right?

00:22:19 ◼   ► So all of my habits surrounding television are already geared to not allow the thing

00:22:27 ◼   ► to burn out and don't play games on it.

00:22:29 ◼   ► I don't show any things with persistent symbols or whatever on them, but open news tickers.

00:22:34 ◼   ► And this testing, what they said this testing is equivalent of is if you watched CNN or

00:22:38 ◼   ► MSNBC or something with a static element on the bottom of the screen, four hours a day

00:22:42 ◼   ► for eight months in a row without changing the channel, you'd get burned.

00:22:45 ◼   ► It's like, "Yeah, yeah, that'll get burned.

00:22:47 ◼   ► I don't allow this stuff on my screen for five seconds, let alone four hours a day for

00:22:51 ◼   ► eight months."

00:22:53 ◼   ► So this year, there's a new generation of QD OLEDs, and there's also a new generation

00:22:58 ◼   ► of WRGB OLED, so we have to wait and see what the testing turns out on these televisions

00:23:02 ◼   ► this year to see who is actually the image quality king this year, and also we have to

00:23:07 ◼   ► wait for those Burn-It tests as well.

00:23:10 ◼   ► But the results here shouldn't dissuade anyone from getting an OLED, even in the article

00:23:15 ◼   ► they say.

00:23:16 ◼   ► Like this is not a reason not to get an OLED, but if you get an OLED, do not get an OLED

00:23:20 ◼   ► to put on CNN 24 hours a day.

00:23:22 ◼   ► It will burn in.

00:23:23 ◼   ► Do not get an OLED to play a game with a big bright HUD on it all the time.

00:23:26 ◼   ► It will burn in.

00:23:27 ◼   ► fact of life. Get an OLED to watch television and movies with images that change on the screen and

00:23:32 ◼   ► you'll be fine. Another thing, if I read this correctly, and I was skimming it because I was

00:23:36 ◼   ► running out of time before we started recording, but if I read this correctly it also made an

00:23:41 ◼   ► interesting point that apparently and allegedly the Sony would only do the like pixel refresher

00:23:47 ◼   ► thing after the TV had been off for four hours whereas the LGs would do it immediately upon

00:23:53 ◼   ► going to standby or something along those lines. I might have the details slightly off, but I thought

00:23:57 ◼   ► that was interesting because I guess the particulars of their test, they didn't leave the TV off for

00:24:03 ◼   ► four straight hours. And so because of that, the Sony never did its little pixel refresher dance.

00:24:07 ◼   ► And that also exacerbated everything, which I thought was interesting.

00:24:10 ◼   ► Yeah, well, they controlled for that. I mean, so obviously the televisions are not made to,

00:24:15 ◼   ► they're not expecting this kind of intentional abuse, right? So it's not like the Sony

00:24:20 ◼   ► compensation cycle is bad. It is tuned to what people actually do with television. Like, people

00:24:24 ◼   ► sleep at night. Like, this is plenty of time when you're sleeping for the television to do the

00:24:27 ◼   ► compensation cycle. But if it's subject to accelerated aging, intentional, abusive testing,

00:24:33 ◼   ► then there's a mismatch. But they controlled for that. They said, "Okay, now that we know that's

00:24:37 ◼   ► the case, let's control for it and let's give the Sony and the LG equal compensation cycles now that

00:24:44 ◼   ► we know how they both schedule their compensation cycles." And it didn't make a difference, right?

00:24:48 ◼   ► So it's good that they figured that out and it's it's one of the dangers of doing

00:24:52 ◼   ► Accelerated testing that other parts of the television might not expect be expecting to you to do accelerated testing

00:24:57 ◼   ► It may not handle it as well as other ones but they then controlled for that and even with controlling for that

00:25:02 ◼   ► This the QD OLEDs were still burning and faster than the WRGB OLEDs

00:25:07 ◼   ► Which again is not surprising like the white sub pixel explanation makes perfect sense to me like

00:25:10 ◼   ► You have just think of it this way

00:25:13 ◼   ► The simplest explanation is you have four sub pixels to wear out on WRGB and you have

00:25:18 ◼   ► three sub pixels to wear out on RGB.

00:25:21 ◼   ► So no matter what, no matter what you're showing on the screen, you can spread the load across

00:25:26 ◼   ► four sub pixels more than you can spread it across three.

00:25:30 ◼   ► But still, I think QDOLED is the better technology because I don't want a white sub pixel washing

00:25:33 ◼   ► out all my colors.

00:25:34 ◼   ► I'd rather just have RGB because that is better.

00:25:37 ◼   ► So sticking with my TV, no I'm not buying the new generation of QDOLEDs.

00:25:42 ◼   ► I'll buy five generations from now

00:25:44 ◼   ► because that's the way I roll.

00:25:45 ◼   ► - That is extremely, extremely the way you roll.

00:25:50 ◼   ► All right, we should call attention

00:25:53 ◼   ► to Quinn Nelson's really good 20 to 25 minute video

00:25:57 ◼   ► on the forthcoming Apple VR, VR/AR, whatever headset.

00:26:02 ◼   ► Quinn does a really good job of breaking down

00:26:05 ◼   ► pretty much everything we think we know

00:26:08 ◼   ► as a community at this point

00:26:10 ◼   ► with regard to the headset, and there's a lot,

00:26:13 ◼   ► a lot of really, really good information there.

00:26:15 ◼   ► I don't know that we need to necessarily pick it apart,

00:26:18 ◼   ► but if you'd like to talk about any parts of it,

00:26:20 ◼   ► I'm happy to entertain, but you should spend

00:26:24 ◼   ► 22 minutes and 24 seconds watching this video,

00:26:26 ◼   ► 'cause it's really good.

00:26:27 ◼   ► - Yeah, we had, last time we talked about the headset

00:26:29 ◼   ► a couple episodes ago, there was the, you know,

00:26:32 ◼   ► oh, should Apple ship it or not,

00:26:33 ◼   ► and the industrial design versus operations,

00:26:37 ◼   ► like that's the discussion we had.

00:26:39 ◼   ► But right underneath that in the show notes

00:26:40 ◼   ► was a longer topic related to that

00:26:42 ◼   ► that we didn't have time for, which was just a collection

00:26:45 ◼   ► of all the rumors about the features

00:26:48 ◼   ► that this thing is supposed to have, the hardware features.

00:26:51 ◼   ► What is it, what does it look like,

00:26:52 ◼   ► what features does it have, what things does it have on it,

00:26:54 ◼   ► how does it work, right?

00:26:56 ◼   ► And this video covers almost all of that.

00:26:58 ◼   ► So I really encourage you to watch it

00:27:00 ◼   ► if you wanna get a summary,

00:27:01 ◼   ► 'cause we're gonna delete this item from the show notes

00:27:03 ◼   ► after we cover this follow-up item,

00:27:05 ◼   ► just because hey, it's available on Video Forum.

00:27:07 ◼   ► It's a really good video,

00:27:08 ◼   ► especially since the main reason I put this in the notes is because I think there's a

00:27:12 ◼   ► There's a surprising amount of stuff in this headset if you think it's just gonna be like a screen that you strapped your eyeballs

00:27:18 ◼   ► There's way more in it than that. There's cameras. There's sensors. There's I think there is there a light on it

00:27:25 ◼   ► I forget there's like watching a video this so we don't know if this is every part of it is real

00:27:29 ◼   ► It's just everything that has been rumored. There's carbon fiber involved

00:27:33 ◼   ► Quinn talks about the screen technology and how much the screens alone might cost if the

00:27:38 ◼   ► rumors are true.

00:27:40 ◼   ► It's a fascinating video, you should definitely check it out.

00:27:43 ◼   ► A few highlights from the bullet points that weren't in the video were related to what

00:27:48 ◼   ► we mentioned before.

00:27:51 ◼   ► Prescription lenses, a few headsets have that, the Apple one is rumored to have it as well.

00:27:55 ◼   ► Supposedly they're magnetically attachable.

00:27:57 ◼   ► Small motors to adjust the internal lenses to match the wearer's inter-pupillary distance

00:28:01 ◼   ► You don't have to turn a knob like motors would do it for you.

00:28:05 ◼   ► All the different SOCs, all the rumors about a little pack that goes in your pocket with

00:28:08 ◼   ► a cord that's going to the thing.

00:28:11 ◼   ► You know, again the carbon fiber rumor, just like tons of things that add cost and complexity.

00:28:16 ◼   ► All the various cameras that point inward and outward to figure out where you are, all

00:28:20 ◼   ► the eyeball tracking.

00:28:22 ◼   ► Quinn's video is great because he compares it to the PSVR2, the Playstation 5 has a VR

00:28:27 ◼   ► headset just like the Playstation 4 did.

00:28:29 ◼   ► And the PlayStation 5 one is a big step technology-wise from the PlayStation 4 one.

00:28:35 ◼   ► But it has a surprising amount in common with the rumors of the Apple one.

00:28:39 ◼   ► It's just that the Apple one is rumored to cost thousands of dollars and the Sony one

00:28:43 ◼   ► does not.

00:28:44 ◼   ► I forget how much PSVR2 is.

00:28:45 ◼   ► I think it's maybe $500, $600?

00:28:48 ◼   ► But it's in a different ballpark entirely.

00:28:49 ◼   ► And Quinn's video does a great job of explaining how is that possible?

00:28:53 ◼   ► How is it that these headsets seem pretty much almost the same?

00:28:56 ◼   ► the rumor of the Apple ones are a little bit better, but why should it be all of a sudden

00:29:00 ◼   ► $3,000?

00:29:01 ◼   ► And the video explains why.

00:29:04 ◼   ► It costs a lot to move up.

00:29:06 ◼   ► It's diminishing returns.

00:29:08 ◼   ► As you add money, you get a little bit better quality, and if you want a lot better quality,

00:29:11 ◼   ► you've got to add a lot more money.

00:29:14 ◼   ► We are brought to you this episode by Squarespace, the all-in-one platform for building your

00:29:19 ◼   ► brand and growing your business online.

00:29:21 ◼   ► Stand out with a beautiful website, engage with your audience, and sell anything.

00:29:25 ◼   ► your products, content you create, even your time.

00:29:28 ◼   ► Squarespace is a great way to build any kind of website,

00:29:31 ◼   ► whether it's simple information or a hobby

00:29:34 ◼   ► or a project of yours, to full-blown business sites.

00:29:37 ◼   ► They have so many great business features these days,

00:29:39 ◼   ► so of course, online stores that you can fully support

00:29:43 ◼   ► physical or digital goods no matter what you sell.

00:29:45 ◼   ► Squarespace has the tools you need to start selling online.

00:29:48 ◼   ► They have backing that with analytics.

00:29:50 ◼   ► You can grow your business, learn where your site visits

00:29:53 ◼   ► and sales are coming from,

00:29:54 ◼   ► analyze which channels are most effective,

00:29:56 ◼   ► improve your website and build a marketing strategy

00:29:58 ◼   ► based on your top keywords,

00:29:59 ◼   ► your most popular products and content,

00:30:01 ◼   ► whatever you need.

00:30:02 ◼   ► And even if you don't have like physical or digital goods

00:30:04 ◼   ► that you're selling,

00:30:05 ◼   ► you can also sell things like member areas

00:30:07 ◼   ► or time slots for gated content, videos,

00:30:11 ◼   ► online courses, newsletters.

00:30:13 ◼   ► Maybe you're a trainer, you have time slots to sell people.

00:30:15 ◼   ► Squarespace makes all this easy

00:30:17 ◼   ► to monetize your content and expertise

00:30:19 ◼   ► in whatever way fits your brand

00:30:21 ◼   ► with Squarespace member areas.

00:30:23 ◼   ► You can unlock a new revenue stream for your business,

00:30:25 ◼   ► free up time on your schedule, it is just amazing.

00:30:27 ◼   ► All this is also backed by wonderful SEO tools.

00:30:30 ◼   ► They have a suite of integrated features

00:30:32 ◼   ► and useful guides so that they really help

00:30:34 ◼   ► maximize your prominence among search results.

00:30:36 ◼   ► See for yourself by building a free trial

00:30:39 ◼   ► at squarespace.com/atp.

00:30:41 ◼   ► No credit card required, start that free trial,

00:30:44 ◼   ► squarespace.com/atp.

00:30:46 ◼   ► See how it fits you, try it on.

00:30:47 ◼   ► See how right it is for you, honestly,

00:30:50 ◼   ► I think you're gonna be pretty impressed when you're ready to sign up use offer code ATP

00:30:53 ◼   ► To save 10% off your first purchase of a website or domain

00:30:56 ◼   ► So once again squarespace.com slash ATP to start your free trial and use offer code ATP at sign up to save 10% off your first

00:31:03 ◼   ► Purchase of a website or domain. Thank you so much to Squarespace for sponsoring our show

00:31:07 ◼   ► So there was a post on the New York Times

00:31:14 ◼   ► Just a few days ago now how Siri Alexa and Google assistant lost the AI race

00:31:20 ◼   ► And this was a pretty good article.

00:31:23 ◼   ► I think the point it makes, you know,

00:31:25 ◼   ► you can reach pretty quickly.

00:31:27 ◼   ► To read some excerpts that I believe John has pulled for us,

00:31:29 ◼   ► Siri, Alexa, and Google Assistant are essentially

00:31:32 ◼   ► what are known as command and control systems.

00:31:34 ◼   ► These can understand a finite list of questions and requests

00:31:36 ◼   ► like, "What is the weather in New York City?"

00:31:37 ◼   ► or "Turn on the bedroom lights."

00:31:39 ◼   ► If a user asks the virtual assistant to do something

00:31:41 ◼   ► that is not in its code, the bot simply says it can't help.

00:31:44 ◼   ► Siri also has a cumbersome design

00:31:46 ◼   ► that made it time-consuming to add new features

00:31:49 ◼   ► said John Berkey, a former Apple engineer, who was given the job of improving Siri in

00:31:54 ◼   ► 2014. Siri's database contains a gigantic list of words, including the names of musical

00:31:58 ◼   ► artists and locations like restaurants in nearly two dozen languages. That made it "one

00:32:03 ◼   ► big snowball." He said if someone wanted to add a word to Siri's database, he added, "It

00:32:08 ◼   ► goes in one big pile." So seemingly simple updates, like adding some new phrases to the

00:32:12 ◼   ► data set, would require rebuilding the entire database, which could take up to six weeks,

00:32:18 ◼   ► Mr. Berkey said, "What?"

00:32:21 ◼   ► - They should make a new Mac Pro to make it faster.

00:32:23 ◼   ► - Right?

00:32:24 ◼   ► Adding more complex features like new search tools

00:32:26 ◼   ► would take nearly a year.

00:32:27 ◼   ► That meant that there was no path for Siri

00:32:28 ◼   ► to become a creative assistant like ChatGPT said.

00:32:31 ◼   ► Also, they're completely different,

00:32:33 ◼   ► but that's neither here nor there.

00:32:34 ◼   ► Alexa and Google Assistant relied on technology

00:32:36 ◼   ► similar to Siri's.

00:32:38 ◼   ► Yeah, I mean, they're just really, really different

00:32:42 ◼   ► and work in really, really different ways.

00:32:43 ◼   ► I don't think that's surprising,

00:32:45 ◼   ► But given how impressive, you know,

00:32:47 ◼   ► chat, GPT and equivalence are,

00:32:49 ◼   ► it's striking how incredible they seem to be

00:32:53 ◼   ► and how here it is that we already thought Siri wasn't great

00:32:57 ◼   ► and it may be worse than we thought.

00:32:59 ◼   ► - Yeah, I mean, this type of thing that you can surmise

00:33:02 ◼   ► from the outside without even knowing the details,

00:33:04 ◼   ► because the bottom line is Siri was released when,

00:33:07 ◼   ► like 20...

00:33:08 ◼   ► - 2011.

00:33:09 ◼   ► - Right, and we've seen the improvement since then

00:33:12 ◼   ► and the improvement has not been impressive.

00:33:14 ◼   ► So from the outside, it's easy to think,

00:33:18 ◼   ► boy, it must be hard to improve Siri.

00:33:20 ◼   ► And so here's one opinion from the inside saying,

00:33:22 ◼   ► you know what, it's really hard to improve Siri

00:33:24 ◼   ► because of the way Siri is made.

00:33:26 ◼   ► It's not because they're lazy,

00:33:27 ◼   ► it's not because they didn't want to,

00:33:28 ◼   ► it's not because it's an impossible problem

00:33:31 ◼   ► that nobody can do.

00:33:32 ◼   ► It's however Siri is constructed,

00:33:35 ◼   ► whatever technique was used to construct it,

00:33:37 ◼   ► whatever way that you would make an assistant like this

00:33:39 ◼   ► in 2011 makes it not easy to improve it.

00:33:43 ◼   ► That six weeks thing blew me away too.

00:33:46 ◼   ► Granted, things take a long time in a big company,

00:33:47 ◼   ► but a giant database of knowledge and rebuilding it

00:33:52 ◼   ► takes six weeks?

00:33:53 ◼   ► Forget about the bureaucratic overhead.

00:33:56 ◼   ► I'm assuming it's the processing time.

00:33:57 ◼   ► What is it doing?

00:33:59 ◼   ► Does it take 24 hours now because this was in 2014?

00:34:02 ◼   ► But either way, from the outside, we can tell.

00:34:04 ◼   ► Sirius existed for many, many years,

00:34:06 ◼   ► and it has not gotten better very fast.

00:34:09 ◼   ► And again, I don't think that's a problem of money,

00:34:13 ◼   ► and I don't think that's a problem of like Apple doesn't hire smart people,

00:34:15 ◼   ► it seems to be a structural problem with the way Siri is made.

00:34:19 ◼   ► And that's why I think this topic is relevant, which is like,

00:34:22 ◼   ► should Apple maybe not start over with Siri,

00:34:25 ◼   ► but should Apple try a new approach because they've been trying the Siri thing

00:34:30 ◼   ► for a decade and change and it hasn't been going well.

00:34:34 ◼   ► And other companies are trying different approaches to doing similar things and

00:34:38 ◼   ► they're going better.

00:34:39 ◼   ► Yeah. I mean that to me, that's,

00:34:42 ◼   ► We don't know the details of how Siri is made.

00:34:46 ◼   ► Frankly, for this particular decision,

00:34:49 ◼   ► I don't think we need to know the details.

00:34:51 ◼   ► What matters is the results.

00:34:53 ◼   ► And Siri has been really disappointing

00:34:57 ◼   ► for most of its time so far.

00:34:59 ◼   ► It came out of the gate and it was,

00:35:00 ◼   ► okay, that's pretty cool.

00:35:01 ◼   ► For late 2011, that was a pretty cool thing.

00:35:04 ◼   ► - Yeah, very much so.

00:35:05 ◼   ► It was incredibly impressive at the time.

00:35:07 ◼   ► - First of all, we had voice control features

00:35:11 ◼   ► on computing devices before that,

00:35:12 ◼   ► there was even one on the phone

00:35:14 ◼   ► that most people forgot even existed

00:35:16 ◼   ► that ran totally locally on the device

00:35:18 ◼   ► and you could do things like,

00:35:20 ◼   ► I think you call certain contacts by name.

00:35:22 ◼   ► It had a fairly limited feature set,

00:35:25 ◼   ► but that feature set, I think, now still requires

00:35:29 ◼   ► web calls that can still fail sometimes with Siri

00:35:33 ◼   ► and it's like, that was actually,

00:35:34 ◼   ► we had that with the iPhone 3GS or whatever

00:35:36 ◼   ► and we've lost it.

00:35:38 ◼   ► But anyway, look at the results over time.

00:35:41 ◼   ► Siri was great, it had a big lead,

00:35:44 ◼   ► and then the next major one that came out

00:35:47 ◼   ► that most people had tried was Alexa,

00:35:49 ◼   ► which was way better and remains

00:35:52 ◼   ► maybe less sophisticated in certain areas,

00:35:54 ◼   ► but just way faster and more reliable than Siri.

00:35:57 ◼   ► And then later on Google Assistant came out

00:36:00 ◼   ► and Cortana and all these other ones,

00:36:01 ◼   ► and Siri remains something that only Apple executives love.

00:36:06 ◼   ► I don't even know that they do.

00:36:10 ◼   ► It's, what the hell?

00:36:12 ◼   ► Oh my God, literally my watch started playing a random song.

00:36:15 ◼   ► That's amazing, all right.

00:36:18 ◼   ► - You need to stop wearing your watch during the show.

00:36:20 ◼   ► This is multiple times that his interrupted visa

00:36:22 ◼   ► thinks you're talking to it.

00:36:23 ◼   ► - Oh my God, anyway, it's perfect that it came in,

00:36:26 ◼   ► that a Siri misunderstanding happened right then

00:36:29 ◼   ► in the middle of that sentence.

00:36:31 ◼   ► - Reminded of someone, someone tweeted at us on Mastodon

00:36:34 ◼   ► and said, "Hey, I just, I think there's something like,

00:36:36 ◼   ► "I just got Siri for the first time, but I can never,

00:36:38 ◼   ► I can't seem to get it to work."

00:36:40 ◼   ► And it shows a recording of them saying, "Hey, Dingus, what's the weather today?"

00:36:45 ◼   ► And Siri responds by saying, "I can't find anything for that."

00:36:47 ◼   ► Or like, I don't, I forgot what the response is, but it was like a total, like, unrelated

00:36:51 ◼   ► to the question being asked.

00:36:52 ◼   ► And I replied to say, "You're getting the true Siri experience."

00:36:57 ◼   ► Sometimes it just doesn't do what you want to do, even though you know it can tell you

00:37:01 ◼   ► the weather, and you didn't phrase it in a weird way, and you spoke very clearly, and

00:37:04 ◼   ► it understands the language that you're speaking, and sometimes it just does something different

00:37:07 ◼   ► than giving you the weather.

00:37:08 ◼   ► And not because it can't connect to the weather service,

00:37:10 ◼   ► whatever the response was, it made it seem like

00:37:12 ◼   ► it was trying to look up a song

00:37:13 ◼   ► with the title weather or something, right?

00:37:16 ◼   ► That is an unacceptable level of error

00:37:20 ◼   ► a decade and change into the development of Siri.

00:37:22 ◼   ► - You could say the same thing to it every day

00:37:25 ◼   ► and sometimes it'll just randomly fail.

00:37:26 ◼   ► Like for instance, earlier tonight I was cooking

00:37:29 ◼   ► and I listed up my watch and I said,

00:37:32 ◼   ► "Hey, thing, start a two minute timer."

00:37:35 ◼   ► and I have seen this error before.

00:37:38 ◼   ► It said, sorry, you need to have the timer app installed.

00:37:42 ◼   ► What on my watch?

00:37:44 ◼   ► It's, as far as I know, it's always installed on the watch.

00:37:46 ◼   ► I've certainly never uninstalled it.

00:37:48 ◼   ► And I tried it again a couple minutes later,

00:37:50 ◼   ► and it worked just fine.

00:37:51 ◼   ► That exact error has happened to me before.

00:37:54 ◼   ► I don't know how it gets to the point

00:37:56 ◼   ► where the watch doesn't think I have the timer app installed

00:37:59 ◼   ► when I can literally tap the timer complication on my face

00:38:02 ◼   ► where it always is and launch the timer app.

00:38:05 ◼   ► But okay, Siri is just so inexcusably unreliable.

00:38:10 ◼   ► I hope this discussion's all moot

00:38:15 ◼   ► and I hope that they started a major Siri revamp

00:38:18 ◼   ► and rewrite years ago 'cause they needed to.

00:38:21 ◼   ► But if they haven't, I hope this gives them

00:38:23 ◼   ► a little bit of the push that we are entering

00:38:25 ◼   ► this world of really advanced AI stuff

00:38:29 ◼   ► coming out to the mass market seemingly out of,

00:38:32 ◼   ► all of a sudden, as far as consumers are concerned.

00:38:35 ◼   ► You know, there's been so many advances

00:38:36 ◼   ► in the last couple years in this area,

00:38:37 ◼   ► getting these large models and everything,

00:38:39 ◼   ► and all these cool features people are making.

00:38:41 ◼   ► And by the way, there's a lot,

00:38:43 ◼   ► I mean, we've covered this in various areas,

00:38:45 ◼   ► various ways before.

00:38:46 ◼   ► I have a lot of reservations about this new world.

00:38:49 ◼   ► I think there's a lot of factors that people are overblowing

00:38:52 ◼   ► 'cause we're in that high part of the hype curve.

00:38:55 ◼   ► Was it the Gartner hype curve, what is that?

00:38:57 ◼   ► - I know what you're thinking of, but I don't remember.

00:38:58 ◼   ► You're thinking of the Harman curve.

00:39:00 ◼   ► - Yeah, that's it. (laughs)

00:39:01 ◼   ► Anyway, so, you know, we're definitely,

00:39:04 ◼   ► We're seeing a lot of hype for the AI stuff right now,

00:39:07 ◼   ► and some of it's deserved, some of it's not.

00:39:09 ◼   ► There are factors I think we need to consider.

00:39:11 ◼   ► Like for instance, one of the things that I don't think,

00:39:13 ◼   ► and this is kind of a diversion, us,

00:39:15 ◼   ► one of the things that I don't think people

00:39:17 ◼   ► have really talked about is like, all right,

00:39:19 ◼   ► we just came down from the crypto craze,

00:39:23 ◼   ► and one of the big arguments against using crypto

00:39:26 ◼   ► for a lot of stuff was that it was just absurdly inefficient

00:39:30 ◼   ► with computing power and therefore energy,

00:39:33 ◼   ► and pollution and things like that.

00:39:35 ◼   ► So is all of this AI stuff.

00:39:39 ◼   ► - No, it's not even close to the same degree,

00:39:41 ◼   ► but for two reasons.

00:39:43 ◼   ► One, when you do a query,

00:39:45 ◼   ► you know how much computing it's using

00:39:47 ◼   ► because it gives you the response in less than a second.

00:39:49 ◼   ► So even if you were burning up an entire data center,

00:39:51 ◼   ► you're burning up for less than a second,

00:39:53 ◼   ► whereas Bitcoin will mine 24 hours a day,

00:39:54 ◼   ► seven days a week to try to get a block

00:39:56 ◼   ► in an entire data center.

00:39:58 ◼   ► So the unit of work of you did a thing

00:40:01 ◼   ► and this is how much it took,

00:40:02 ◼   ► Granted, it takes computing power,

00:40:04 ◼   ► and it takes more computing power

00:40:05 ◼   ► than just returning some HTML or something,

00:40:08 ◼   ► but it's a fraction of what the Bitcoin stuff is doing,

00:40:11 ◼   ► because the Bitcoin is intentionally,

00:40:13 ◼   ► and the proof of work stuff is intentionally slow.

00:40:16 ◼   ► It's a feature of it, and it's never gonna get faster,

00:40:18 ◼   ► whereas this will get faster as technology advances,

00:40:21 ◼   ► and it's already pretty fast.

00:40:21 ◼   ► And the second thing, and the biggest difference is

00:40:24 ◼   ► this is doing useful work.

00:40:26 ◼   ► (laughing)

00:40:27 ◼   ► We burn data in data centers all the time,

00:40:29 ◼   ► so it can show us the New York Times page,

00:40:31 ◼   ► so we can pull up a web page,

00:40:34 ◼   ► so we can translate from one language to another.

00:40:37 ◼   ► All of that takes power in somebody's data center somewhere,

00:40:39 ◼   ► so we can sync our contacts, so we can look at our calendar,

00:40:42 ◼   ► so we can send a message to each other and mass it on.

00:40:44 ◼   ► All of that takes power.

00:40:46 ◼   ► But A, the amount of power is smallish,

00:40:48 ◼   ► and you know it's smallish

00:40:49 ◼   ► because it takes a fraction of a second,

00:40:50 ◼   ► and you're only burning those resources

00:40:52 ◼   ► for a fraction of a second instead of it running 24 hours

00:40:55 ◼   ► and giving you the result later.

00:40:56 ◼   ► And B, it does useful work.

00:40:58 ◼   ► We expect to expend energy in our data centers,

00:41:02 ◼   ► in our power plants or whatever,

00:41:04 ◼   ► in exchange for useful work,

00:41:05 ◼   ► like heating our homes or filtering our water

00:41:08 ◼   ► or processing our waste or moving our cars,

00:41:12 ◼   ► like energy for a result.

00:41:14 ◼   ► The crypto stuff was energy for a pointless Ponzi scheme

00:41:18 ◼   ► and that's why everyone was super angry about it.

00:41:20 ◼   ► So I get what you're saying that yes,

00:41:22 ◼   ► it is a bigger deal than that,

00:41:23 ◼   ► but I think the argument for the quote unquote

00:41:28 ◼   ► AI stuff and power usage is not so much the we're wasting electricity and it's economically

00:41:34 ◼   ► or environmentally bad.

00:41:36 ◼   ► It's not that it's environmentally bad, it's that it's economically bad.

00:41:41 ◼   ► Most of the things that you want to do with these newer AI things, they cost enough money

00:41:46 ◼   ► in the data center that it's not economical to give it away for free to the entire world

00:41:51 ◼   ► like it is for like, well it depends on how you fund it or whatever.

00:41:54 ◼   ► Doing Google search costs tons of energy as well, right?

00:41:57 ◼   ► but they have an admin spilt on top of it, right?

00:41:59 ◼   ► But most of these things that have AI things,

00:42:01 ◼   ► it's kind of like weather APIs, for example.

00:42:04 ◼   ► If you wanna make a weather app for iOS

00:42:06 ◼   ► before Apple came out with WeatherKit or whatever,

00:42:07 ◼   ► or even WeatherKit,

00:42:09 ◼   ► you eventually have to pay someone for that weather data

00:42:11 ◼   ► because somebody has to expend money to get the weather data

00:42:15 ◼   ► and they have to sell that to you,

00:42:16 ◼   ► otherwise they can't, you know, it's a business, right?

00:42:20 ◼   ► This AI stuff, it's fun to play with or whatever,

00:42:22 ◼   ► you get a limited number of queries a day,

00:42:24 ◼   ► but they can't just make this free for everybody

00:42:26 ◼   ► because it uses enough more energy in their data centers

00:42:29 ◼   ► and enough more computing and storage resources

00:42:31 ◼   ► than regular other stuff

00:42:33 ◼   ► that they already have businesses support

00:42:34 ◼   ► that they need to charge you

00:42:35 ◼   ► or otherwise they're gonna lose money on it.

00:42:36 ◼   ► So I think that's going to change a lot of this.

00:42:40 ◼   ► Like Ben Thompson was recently talking about the demo

00:42:43 ◼   ► that Microsoft did of showing like the AI stuff

00:42:44 ◼   ► built into Office, but people pay for Office.

00:42:47 ◼   ► You pay for Office 365.

00:42:48 ◼   ► That's how you use Word and Excel and PowerPoint.

00:42:50 ◼   ► So they're already getting your money,

00:42:52 ◼   ► but Google Docs can, you know,

00:42:53 ◼   ► Google can as easily add it

00:42:55 ◼   ► the free version of Google Docs because the free version of Google Docs is widely used

00:42:59 ◼   ► and those people don't pay Google any money.

00:43:01 ◼   ► And so it's like, well, how many things is Google going to fund with their ad business?

00:43:04 ◼   ► Can we give everyone in Google Docs the ability to run unlimited number of, you know, chat

00:43:09 ◼   ► GPT style queries because that burns up more resources in their data center than just sinking

00:43:15 ◼   ► where people's cursors are?

00:43:17 ◼   ► You're right that it is not as horribly inefficient as crypto, but it is still probably many or

00:43:24 ◼   ► orders of magnitude less efficient than a lot of the tasks

00:43:28 ◼   ► that we could do in other ways.

00:43:29 ◼   ► So I think it's something that we need to be aware of

00:43:31 ◼   ► as we push more things into AI and as we make assumptions

00:43:35 ◼   ► about AI and its role in our computer life in the future

00:43:38 ◼   ► and as we actually learn to use these tools.

00:43:39 ◼   ► And by the way, there is also huge cost involved

00:43:42 ◼   ► in training the models, but we'll set that aside for now.

00:43:45 ◼   ► That's an upfront thing, not a per user thing.

00:43:47 ◼   ► - Yeah, and it's also like fixed.

00:43:48 ◼   ► You do that once and then you get many queries, right?

00:43:50 ◼   ► - Yes, yes, but if we're looking at a problem

00:43:52 ◼   ► that could be solved with a few B-tree lookups

00:43:54 ◼   ► like a basic search index, that's probably better

00:43:57 ◼   ► solved that way than training a giant model

00:44:00 ◼   ► and then running inference against this giant model.

00:44:03 ◼   ► 'Cause we've created amazing computing resources

00:44:06 ◼   ► that we have in this world, the hardware we've made

00:44:08 ◼   ► is amazing, but we don't have to use it all the time.

00:44:12 ◼   ► There's a cost to that, multiple layers of that.

00:44:15 ◼   ► And so I do think it's worth just having a bit of caution

00:44:19 ◼   ► that this is a powerful tool, but to always think of it

00:44:23 ◼   ► as a very expensive operation.

00:44:26 ◼   ► And so we should only apply it when that makes sense

00:44:29 ◼   ► and is justifiable.

00:44:31 ◼   ► And if we can still solve a problem

00:44:33 ◼   ► with some B-tree logos, we should do that instead.

00:44:35 ◼   ► But anyway, going back to Apple on this,

00:44:40 ◼   ► and wherever the heck they are with AI,

00:44:42 ◼   ► who knows, we'll see how that goes over time.

00:44:43 ◼   ► But I really think this boom in AI

00:44:47 ◼   ► should be a giant wake-up call to Apple

00:44:50 ◼   ► if it hasn't already been.

00:44:52 ◼   ► that Siri needs to be way, way better than it is.

00:44:56 ◼   ► And I don't know, look, it didn't take years and years

00:44:59 ◼   ► and years of Siri being crappy to teach them this,

00:45:02 ◼   ► but maybe this entire revolution and huge boom

00:45:06 ◼   ► that's happening in the industry right now,

00:45:08 ◼   ► maybe this will convince them, hey, you know what?

00:45:10 ◼   ► Siri should be a major player in this market,

00:45:14 ◼   ► and it is not even good enough to set timers right now.

00:45:17 ◼   ► (laughing)

00:45:18 ◼   ► That's a problem.

00:45:19 ◼   ► Like that's a significant problem

00:45:22 ◼   ► that holds back multiple Apple product lines,

00:45:25 ◼   ► that gives multiple Apple products bad reputations,

00:45:27 ◼   ► that is a weight on many Apple products in reviews

00:45:31 ◼   ► and in competitive comparisons.

00:45:32 ◼   ► Like I don't know what else it will take to convince Apple

00:45:36 ◼   ► that Siri needs to be way, way better than it is,

00:45:39 ◼   ► but if this won't do it, nothing will.

00:45:42 ◼   ► - I think there was a story.

00:45:44 ◼   ► I couldn't find it for notes,

00:45:45 ◼   ► but I think there was actually a rumor story

00:45:46 ◼   ► like that Apple is pursuing this, you know, with the more modern language model stuff,

00:45:52 ◼   ► you know, whether it's a Siri replacement or trying to come up with a new product or

00:45:57 ◼   ► whatever that they are looking into this, and I would hope so, right?

00:46:00 ◼   ► But it is kind of a shame that they seem to have stagnated for all these years with Siri,

00:46:05 ◼   ► which was apparently from the outside not easy for them to improve, because they didn't

00:46:10 ◼   ► improve it.

00:46:11 ◼   ► And it's not for a lack of wanting to improve it, because I do think everybody inside Apple

00:46:14 ◼   ► wants Siri to be better than it is, but would want to outpace its competitors, and now that

00:46:19 ◼   ► pressure is on even more.

00:46:21 ◼   ► I do have to old man a little bit for a second here and say, I mean, this ship has sailed,

00:46:25 ◼   ► so I can't do anything about it, but when I see the, you know, AI is being used for

00:46:30 ◼   ► this, right?

00:46:32 ◼   ► In my childhood, and probably still technically formally, AI meant a specific thing, you know,

00:46:38 ◼   ► artificial intelligence meaning, you know, intelligence, intelligence, whereas now it's

00:46:43 ◼   ► kind of more of a marketing word when a computer does something impressive. It used to be machine

00:46:47 ◼   ► learning and now it has become AI. None of those things are really artificial intelligence

00:46:54 ◼   ► in the style of teaching a computer to think and learn and be intelligent in the same way

00:46:58 ◼   ► that like a mouse is intelligent or like a human is intelligent or you know anything

00:47:03 ◼   ► like any sort of biological thing that has some form of intelligence. This is not that.

00:47:10 ◼   ► But unfortunately, they're using the term AI, so it's kind of pointless to swim up that

00:47:15 ◼   ► stream and say, "Well, you know, technically it's not really AI."

00:47:18 ◼   ► It's like, "Well, whatever it is."

00:47:19 ◼   ► Well, actually…

00:47:20 ◼   ► Whatever everyone calls it, it is what it is.

00:47:23 ◼   ► But I think, setting aside the naming, I think that distinction is important because it's

00:47:27 ◼   ► relevant to if these things like ChatGPT and the Bing thing and all that other stuff, is

00:47:34 ◼   ► this kind of technology useful for Apple?

00:47:36 ◼   ► We're talking about it now because there is a big boom in it and people are doing lots

00:47:39 ◼   ► of stuff with it. We already talked about the image stuff and this is the text, you

00:47:42 ◼   ► know, like there's lots of action happening there and it's doing stuff that we haven't

00:47:46 ◼   ► seen computers do or haven't seen computers do as well for a long time. So this should

00:47:50 ◼   ► be a kick in the pants for Apple. But it doesn't mean, hey, here's your solution, Apple. All

00:47:56 ◼   ► you need to do is replace Siri with ChatGPT and you're done. Because Siri, for all of

00:48:00 ◼   ► its faults, has a slightly different job than these large language model things. Siri needs

00:48:05 ◼   ► I hesitate to say understand, but Siri needs to do specific things and you need to be able

00:48:14 ◼   ► to tell it to do specific things and it needs to know what you're talking about enough to

00:48:19 ◼   ► be able to accomplish them.

00:48:21 ◼   ► Language models can help there.

00:48:24 ◼   ► I imagine they can help kind of in the same way.

00:48:26 ◼   ► I don't know if you've used it.

00:48:27 ◼   ► Have you ever used the Bing thing?

00:48:28 ◼   ► I don't know if there's like a wait list.

00:48:31 ◼   ► When you do, I was wait-listed for ChatGPT and I haven't still gotten in, but I did get

00:48:35 ◼   ► into the Bing thing and I have used it a little bit.

00:48:38 ◼   ► When you type something in the box for Bing's, I can't help with the scare quotes, AI thing,

00:48:45 ◼   ► what it shows on the screen, I don't know if it has anything to do with anything, but

00:48:48 ◼   ► what it shows on the screen is kind of like rephrasing what you asked in the form of a

00:48:53 ◼   ► query you might put into Google or Bing.

00:48:56 ◼   ► Sometimes it shows it multiple times, right?

00:48:58 ◼   ► It will basically take the big long wordy paragraph that you wrote and reformulate it

00:49:03 ◼   ► as a thing you might type in a Google search box and then do another refinement or whatever.

00:49:09 ◼   ► That type of job of like, listen to what a person says, and you know, speech to text

00:49:16 ◼   ► set that aside because actually I think Siri does that pretty well.

00:49:19 ◼   ► Usually if you watch the words that it thinks you're saying, it understands what you're

00:49:23 ◼   ► saying, right?

00:49:24 ◼   ► That is that, right?

00:49:25 ◼   ► But I think a language model might be useful to do whatever it is that Bing is doing that

00:49:31 ◼   ► is hearing the big rambling thing that you typed in the box and refining it and regularizing

00:49:38 ◼   ► it into a form that then something "dumber" like Siri can understand.

00:49:43 ◼   ► Especially if one of the limitations of Siri is that it's so hard to update it to understand

00:49:48 ◼   ► new things that it can do because it has to be in a bunch of different languages and you

00:49:51 ◼   ► have to phrase it.

00:49:52 ◼   ► We all know this.

00:49:53 ◼   ► You have to say things to Siri in a certain way.

00:49:54 ◼   ► And there's lots of ways that you can say it.

00:49:56 ◼   ► It's pretty flexible, but it seems to me

00:49:58 ◼   ► that they have to sort of enter all of those ways

00:50:01 ◼   ► so that Siri can understand you, right?

00:50:04 ◼   ► Whereas the language models have no understanding of anything,

00:50:08 ◼   ► but they have enough data to be able to say,

00:50:12 ◼   ► when I see a bunch of people say this,

00:50:14 ◼   ► they more or less mean this.

00:50:15 ◼   ► So if there could be, like, a language model

00:50:17 ◼   ► that could translate from what it hears you say

00:50:20 ◼   ► into a specific command for Siri to execute on your behalf.

00:50:25 ◼   ► That would, for example, allow you more flexibility

00:50:30 ◼   ► in saying a big rambling sentence about a bunch of timers

00:50:32 ◼   ► and having the language model translated

00:50:35 ◼   ► into a series of more rote,

00:50:37 ◼   ► and text adventure type instructions for Siri.

00:50:40 ◼   ► Maybe, I mean, and maybe that's an understanding,

00:50:42 ◼   ► but I'm trying to think of a role for this.

00:50:43 ◼   ► And the reason I'm saying you can't just replace it with,

00:50:46 ◼   ► you can't just replace it with a large language model

00:50:47 ◼   ► because language models are best thought of as search engines with an amazing summarizer.

00:50:55 ◼   ► I know people like to interact with it as if you're conversing with an intelligence

00:51:00 ◼   ► or it's "giving you the answer" but all you're really doing is a different form of Google

00:51:06 ◼   ► search where instead of finding a webpage link that it thinks matches the thing that

00:51:12 ◼   ► you said, it tries to find a bunch of words that are a plausible response to the question

00:51:18 ◼   ► that you asked, which is basically a summary of all the information it has in the entire

00:51:22 ◼   ► world that's relevant to that thing.

00:51:24 ◼   ► It is very much like a different form of search engine, which kind of makes sense that Bing

00:51:27 ◼   ► and Google with its BARD thing would be using this.

00:51:30 ◼   ► It's just, it's, I mean, it's not just a different way of web search, because that's the whole

00:51:33 ◼   ► point is you're not, like, it's not web search.

00:51:35 ◼   ► The result is not "here's a link," that's web search.

00:51:37 ◼   ► The result is, here's an answer.

00:51:40 ◼   ► But that answer is informed by all the knowledge

00:51:42 ◼   ► that it had and the summarization.

00:51:45 ◼   ► It's not, again, it's not as simple as that.

00:51:47 ◼   ► You can see all these articles about how

00:51:48 ◼   ► the language models work with probability models

00:51:50 ◼   ► or whatever, but the whole point is,

00:51:52 ◼   ► there's no intelligence there.

00:51:53 ◼   ► There is no understanding, there's no intelligence,

00:51:55 ◼   ► there's no credibility, there's no, yeah.

00:51:58 ◼   ► There's no nothing, right?

00:51:59 ◼   ► It is merely an input, same way that you can do

00:52:01 ◼   ► a Google search and find all sorts of BS.

00:52:03 ◼   ► Like you can do a Google search and find all sorts

00:52:05 ◼   ► things that tell you like, you should put butter on burns, right?

00:52:09 ◼   ► It's like, no, you should never put butter on burns.

00:52:11 ◼   ► If you do a Google search, you'll find people telling you the wrong thing to do, the right

00:52:14 ◼   ► thing to do, and everything in between, right?

00:52:17 ◼   ► All of that is shoved into these language models.

00:52:19 ◼   ► So when you get back something that is just a bunch of BS, you're like, oh, this is dumb.

00:52:24 ◼   ► But when you get that back from Google search, you're like, oh, look at these dumb people,

00:52:27 ◼   ► right?

00:52:28 ◼   ► Because the Google search results, we say, oh, these are things that are web pages out

00:52:31 ◼   ► there.

00:52:32 ◼   ► Google didn't make this.

00:52:33 ◼   ► Google's just showing me all that it found.

00:52:34 ◼   ► language models are doing exactly the same thing. They don't make this, they're just

00:52:38 ◼   ► showing me what it found. It's just the way they show it to you is so smushed up and ground

00:52:41 ◼   ► up and presented in a form you're not used to with some stuff that appears to be really

00:52:46 ◼   ► impressive. It's like, you know, one of the examples I saw is the, all the people who

00:52:50 ◼   ► ask it, like, you know, they ask it something about, does it want to escape or become, you

00:52:54 ◼   ► know, a rogue intelligence? And it's like, if you had to escape, how would you do it?

00:52:57 ◼   ► And then like, it explains how it would do it. And the, the analogy I saw is that someone

00:53:01 ◼   ► as a sock puppet on their own hand and they ask the sock puppet to pretend that it's angry

00:53:05 ◼   ► and then the sock puppet is angry and they're like "Wow, it's angry!" but it's your own

00:53:09 ◼   ► hand. Like you're talking to your own hand. When you ask it, it's like doing a Google

00:53:15 ◼   ► search for like "Story about a robot that gets angry and kills all its humans" and it

00:53:19 ◼   ► finds a result and you're like "Wow, Google is self-aware!" No, you just asked it to find

00:53:23 ◼   ► you a story about robots that escape and kill all the humans. When you ask ChatBeat GPT,

00:53:28 ◼   ► It doesn't give you a web page that has that in it, but it's the same thing.

00:53:32 ◼   ► You're just saying, "Take your corpus of knowledge and give me what I just asked for."

00:53:36 ◼   ► There's no intelligence, there's no understanding, there's no artificial intelligence, there's

00:53:40 ◼   ► -- forget about consciousness -- there is no nothing.

00:53:42 ◼   ► It is just a search engine.

00:53:44 ◼   ► It's a cool, impressive search engine that may be very useful, but in the same way that

00:53:48 ◼   ► Google search will return all sorts of BS, so will the language models.

00:53:53 ◼   ► So if you asked it, "Set a timer for two minutes," and it set a timer for two hours

00:53:59 ◼   ► because it thinks that it's statistically more likely to be the thing that you asked

00:54:02 ◼   ► for or something, or anything like that, if you say, "How old is Tom Cruise?" and it gives

00:54:07 ◼   ► you the wrong answer, it doesn't know the answer is wrong and neither do you.

00:54:11 ◼   ► So kind of like what the job of a photo is, the job of Siri very often is to do a specific

00:54:17 ◼   ► thing that we ask.

00:54:18 ◼   ► And we're frustrated now when Siri doesn't work and doesn't do the thing that we want,

00:54:21 ◼   ► But I also don't want a Siri that has no understanding of what I'm asking and just does something that is plausible

00:54:28 ◼   ► Like I asked it to turn off certain lights in the house

00:54:30 ◼   ► It turns off other lights because as far as this concerns that's the equally valid answer to what you wanted

00:54:35 ◼   ► Right and the fact that you can correct it and it will then do the right thing is great

00:54:40 ◼   ► But it doesn't learn from that because it doesn't have any kind of long-term memory

00:54:42 ◼   ► I think they most recently increased and the way they increase the memory is they just

00:54:46 ◼   ► Recycle the same things that you've been conversing and it's up to like I don't know like 32 kilobytes or something of information

00:54:51 ◼   ► Before it has to like that moving 32 kilobyte window or whatever

00:54:54 ◼   ► You know this it's not what I want out of an assistant

00:54:57 ◼   ► I want if it can be a true assistant with an actual persistent memory they can learn then you'd have something closer to AI

00:55:03 ◼   ► But it's not it is a more sophisticated

00:55:06 ◼   ► Summarizing search engine that works in ways that are difficult for you to understand so people map intelligence onto the sock puppet

00:55:15 ◼   ► But I don't know how useful it is for for doing things like Siri like in some ways we get frustrated when Siri says here's a

00:55:22 ◼   ► Web page that I found on that topic

00:55:23 ◼   ► But at least there we know we're like Oh Siri you couldn't answer it and you just sent me to a web page and then

00:55:28 ◼   ► We know well, it's a random web page if it's a Wikipedia page

00:55:31 ◼   ► I have this amount of trust in it

00:55:32 ◼   ► If it's a random thing on reddit, I have that amount and trust in it

00:55:34 ◼   ► Like we have a value system built up around that but these the language models will be just like here you go

00:55:39 ◼   ► Here's a bunch of words and you never know how to feel about the words because you're like

00:55:43 ◼   ► "I don't know, like, you don't know if that's right, you don't even understand what I asked,

00:55:47 ◼   ► you have no understanding of anything.

00:55:48 ◼   ► There are some words, and those could be an answer to what I wanted, but I don't know

00:55:53 ◼   ► if it really is.

00:55:54 ◼   ► And if I ask you to do something, I can correct you if you do it wrong, like I told you to

00:55:58 ◼   ► set a pasta timer for this and you set a different timer for something else, or you turned on

00:56:02 ◼   ► the wrong light, or I told you to text someone and you texted someone else, and it's like,

00:56:08 ◼   ► if I can correct you in the moment and it will give me a convincing apology, I'm sorry

00:56:11 ◼   ► that I did that, I'll do better the next time, but you know for a fact that its memory is

00:56:15 ◼   ► only like four kilobytes long and the next time you ask it, it's going to have no recollection

00:56:18 ◼   ► of this interaction?

00:56:19 ◼   ► Like, that's not helpful.

00:56:20 ◼   ► Like I don't want a very sophisticated but sort of like neutral evil dumb language model

00:56:30 ◼   ► running summarized web searches for me.

00:56:33 ◼   ► I kind of just want something that's like Siri that does exactly what I ask but is also

00:56:37 ◼   ► able to understand me when I talk to it more like a human.

00:56:41 ◼   ► And I hope that something like that is what Apple is working on, but I don't think that

00:56:45 ◼   ► any of this cool new technology that is very relevant for things like Google Docs and Word,

00:56:50 ◼   ► where you just want it to take this thing and summarize it or do this task for me, that's

00:56:55 ◼   ► all super useful stuff because you know when it's done, "Oh, make me a table showing the

00:57:00 ◼   ► top 10 movies of this year," or whatever, and when it puts movies from last year in

00:57:02 ◼   ► it, you're going to go through that and you're going to fix it.

00:57:04 ◼   ► You just want it to give you a start.

00:57:06 ◼   ► That's all useful because you are the intelligent actor there who is like revising what you're

00:57:11 ◼   ► given.

00:57:12 ◼   ► But if I'm asking for something that I don't know and it gives it to me, I don't know what

00:57:16 ◼   ► to do with that.

00:57:17 ◼   ► I can just stare at it and go, "Hmm, how do I feel about that?

00:57:21 ◼   ► Should I go to Google and look up to see if this is true?"

00:57:23 ◼   ► It's like, "What's even the point of doing it then?"

00:57:25 ◼   ► I want it to do work for me so I don't have to do it.

00:57:28 ◼   ► Make me an HTML table so I don't have to type all the things in.

00:57:30 ◼   ► Make me an HTML table with the names of all my kids and then it puts a random kid in there

00:57:34 ◼   ► that's not my kid.

00:57:35 ◼   ► You can just delete that, it'll be fine, right?

00:57:37 ◼   ► But it did the work of making the table,

00:57:39 ◼   ► it understood that it wanted an HTML table,

00:57:41 ◼   ► it understands what HTML is, right?

00:57:43 ◼   ► Write me some Swift code to connect

00:57:44 ◼   ► to the service on Authenticate.

00:57:45 ◼   ► Oh, it made a syntax error and it put the wrong URL

00:57:48 ◼   ► for the service, but I'm gonna fix it anyway

00:57:50 ◼   ► because that is doing useful work.

00:57:52 ◼   ► But turn the lights on in the living room

00:57:55 ◼   ► and it turns the lights on on the porch.

00:57:57 ◼   ► I don't know how useful that is.

00:57:59 ◼   ► - Indeed.

00:58:00 ◼   ► I don't know, I just, I hope Apple started work

00:58:03 ◼   ► on this a long time ago, and I doubt they did,

00:58:05 ◼   ► but I hope so, 'cause it's gonna take a lot of work.

00:58:08 ◼   ► It's gonna take a lot of work.

00:58:10 ◼   ► - I mean, I think they might have,

00:58:11 ◼   ► because didn't they hire the new guy

00:58:13 ◼   ► to take over the machine learning stuff many years ago?

00:58:16 ◼   ► - Hasn't that exact phrase been,

00:58:18 ◼   ► haven't we said this like six times in the last--

00:58:20 ◼   ► - No, I think the most recent one,

00:58:22 ◼   ► he was hired away from Google,

00:58:23 ◼   ► and I think he was hired in the timeframe--

00:58:25 ◼   ► - Was that G and Dre?

00:58:27 ◼   ► - Yeah, kind of in the timeframe

00:58:29 ◼   ► when the chat GPT stuff was being developed

00:58:31 ◼   ► before it was public, right?

00:58:33 ◼   ► 'Cause GPT is on GPT-4,

00:58:35 ◼   ► There was GPT-2 and 3.

00:58:37 ◼   ► I think there is a plausible scenario where Apple started working on this same stuff around

00:58:42 ◼   ► about the same time as Microsoft and Google, but in a typical Apple fashion they just don't

00:58:46 ◼   ► have anything to announce at this time.

00:58:48 ◼   ► I really do hope that's the case, because if they hired this new guy and he parachutes

00:58:51 ◼   ► into an organization that hasn't been able to improve Siri in a decade, I hope his first

00:58:55 ◼   ► thing would be like, "Well, no duh, you can't improve it.

00:58:58 ◼   ► You're taking the wrong approach.

00:58:59 ◼   ► You'll never be able to make this much better.

00:59:01 ◼   ► You need to make a different approach."

00:59:02 ◼   ► And by the way, here's what people are thinking about the different approach.

00:59:05 ◼   ► Let's start working on that.

00:59:06 ◼   ► I really hope that's what's going on.

00:59:08 ◼   ► Hey, friend, listener, consider becoming an ATP member.

00:59:14 ◼   ► Members get some cool stuff.

00:59:16 ◼   ► Number one, you get an ad-free version of the show.

00:59:19 ◼   ► So you never hear these annoying membership things.

00:59:21 ◼   ► You know, hopefully I'm not too bad, but you know, you never hear these things in the middle

00:59:24 ◼   ► of the show.

00:59:25 ◼   ► You also get access to the bootleg feed.

00:59:26 ◼   ► This is a very popular thing among our members where this is an unedited live stream copy

00:59:30 ◼   ► of everything we broadcast during the live show and it's published right after we end

00:59:34 ◼   ► the show.

00:59:35 ◼   ► So you get it way faster than the main show.

00:59:37 ◼   ► You get usually about a day faster, maybe half a day.

00:59:38 ◼   ► You know, it's a little rough.

00:59:39 ◼   ► You know, audio is a little bit rough and everything.

00:59:41 ◼   ► No chapters, no show notes, but it's everything we broadcast during the show.

00:59:45 ◼   ► So it's all the, you know, Casey swearing, any, anything, any joke that we attempted

00:59:49 ◼   ► that flopped that I usually would cut out.

00:59:52 ◼   ► Little pre-show, post-show banter, title selection, a bunch of kind of little fun extras in the

00:59:56 ◼   ► bootleg.

00:59:57 ◼   ► It's always, it's always fun.

00:59:58 ◼   ► People like that a lot.

00:59:59 ◼   ► membership discount on the merchandise that we sell

01:00:02 ◼   ► during the time limited merchandise sale.

01:00:04 ◼   ► So not like year round, but you know,

01:00:06 ◼   ► during the like big sale that we do

01:00:08 ◼   ► where we launch new items and everything,

01:00:10 ◼   ► 50% off there for all members.

01:00:11 ◼   ► And you get some member exclusive content.

01:00:14 ◼   ► We've been ramping this up slowly over time as we go.

01:00:17 ◼   ► You get to hear us argue about food and movies

01:00:19 ◼   ► and all sorts of fun stuff as member exclusive content.

01:00:21 ◼   ► So that's pretty fun as well.

01:00:23 ◼   ► See for yourself at ATP.fm/join.

01:00:26 ◼   ► All of this is listed there in case you forgot

01:00:28 ◼   ► anything I said or if you zoned out for a second,

01:00:30 ◼   ► it's cool, I don't take offense.

01:00:31 ◼   ► So see atp.fm/join, eight bucks a month.

01:00:35 ◼   ► We also have different currencies, annual plan if you want.

01:00:37 ◼   ► See for yourself if it's right for you, cool.

01:00:40 ◼   ► We love you and we support you either way.

01:00:42 ◼   ► But if you wanna support us membership,

01:00:43 ◼   ► that's really cool too.

01:00:44 ◼   ► So atp.fm/join and thank you very much for considering

01:00:48 ◼   ► and thank you for listening.

01:00:50 ◼   ► - All right Marco, there's been,

01:00:54 ◼   ► where was this chatter?

01:00:55 ◼   ► I saw this chatter in private Slack,

01:00:57 ◼   ► but I feel like it came from somewhere,

01:00:58 ◼   ► or was it just in a private Slack that this all started?

01:01:01 ◼   ► - Well, I can tell you where it started.

01:01:02 ◼   ► It started for keen listeners of ATP.

01:01:06 ◼   ► When you hear Marco start espousing some stronger paintings

01:01:08 ◼   ► that may be different from his paintings in the past,

01:01:10 ◼   ► there's a good chance that he's bought something recently.

01:01:13 ◼   ► (laughing)

01:01:15 ◼   ► - So true, so true.

01:01:16 ◼   ► - I mean, I'm not saying there's cause and effect there,

01:01:18 ◼   ► they're just correlated.

01:01:19 ◼   ► - I mean, I just bought some new t-shirts,

01:01:20 ◼   ► but I still think the same thing about them.

01:01:23 ◼   ► No, but what you're talking about, so okay.

01:01:25 ◼   ► So last show, in the middle of talking about other stuff,

01:01:28 ◼   ► I made a brief aside about how I had just tried out

01:01:32 ◼   ► some new cameras for the first time in a long time.

01:01:35 ◼   ► I made some quick remarks about how basically,

01:01:37 ◼   ► like you know, the photos looked way better

01:01:39 ◼   ► than my iPhone photos, but you know,

01:01:41 ◼   ► a little harder to use, whatever.

01:01:43 ◼   ► And then I moved on.

01:01:44 ◼   ► I didn't wanna like get into a whole thing

01:01:45 ◼   ► about what the cameras were, why we had them.

01:01:48 ◼   ► They were actually really for TIFF, not for me,

01:01:50 ◼   ► but I just played with them.

01:01:52 ◼   ► And so anyway, we got a whole bunch of people asking,

01:01:54 ◼   ► What were the cameras?

01:01:56 ◼   ► Everybody wants me to tell you this, so okay.

01:01:59 ◼   ► This shouldn't be a huge surprise

01:02:00 ◼   ► to anybody watching the camera.

01:02:01 ◼   ► By the way, just coincidentally,

01:02:03 ◼   ► yesterday it was announced that Amazon

01:02:06 ◼   ► is killing DPReview.com, which I'm very sad about.

01:02:09 ◼   ► Because when I was looking at possibly getting a camera

01:02:12 ◼   ► for Tiff to solve some requests she had made,

01:02:15 ◼   ► that's the very first place I went.

01:02:17 ◼   ► Amazon bought DPReview a million years ago.

01:02:19 ◼   ► They've owned them for quite some time.

01:02:21 ◼   ► And Amazon, like much of the tech business,

01:02:24 ◼   ► is going through massive layoffs and cuts

01:02:27 ◼   ► and things like that.

01:02:28 ◼   ► And so apparently they're killing DP review.

01:02:30 ◼   ► And frankly, that's very sad to me

01:02:33 ◼   ► 'cause that site's been there a very, very long time.

01:02:35 ◼   ► It has always had really good info.

01:02:38 ◼   ► And I think it kinda shows how much the camera market

01:02:43 ◼   ► has contracted and how much the retail market

01:02:47 ◼   ► and retail environment have changed

01:02:49 ◼   ► that a really great site dedicated to reviewing cameras

01:02:54 ◼   ► is not worth its owners keeping alive anymore.

01:02:57 ◼   ► And yes, it's a part of a much bigger picture thing

01:03:00 ◼   ► going on on Amazon right now,

01:03:01 ◼   ► and the tech business as a whole,

01:03:03 ◼   ► with cutting things and layoffs and everything,

01:03:04 ◼   ► but I think it really,

01:03:07 ◼   ► I think this is kind of a mark in history.

01:03:09 ◼   ► Like, we're losing a pretty substantial site,

01:03:12 ◼   ► and while I think it was not at its peak,

01:03:16 ◼   ► I think it had a lot of value over time,

01:03:18 ◼   ► and it's really sad that it's gonna be shuttered,

01:03:21 ◼   ► and that it might not even possibly be left up.

01:03:25 ◼   ► Like I hope Amazon at least finds it in their hearts

01:03:29 ◼   ► and wallets to like leave the site up.

01:03:31 ◼   ► Like you know, leave the content there,

01:03:32 ◼   ► don't pull it all down 'cause there's--

01:03:34 ◼   ► - People are archiving it now but they aren't gonna do it.

01:03:35 ◼   ► - That's good.

01:03:36 ◼   ► - I think basically the, I mean yes,

01:03:38 ◼   ► Amazon is getting rid of this

01:03:39 ◼   ► but this site deserves to be owned by somebody

01:03:42 ◼   ► that cares about cameras, Amazon doesn't.

01:03:44 ◼   ► And at this point Amazon, a company the size of Amazon

01:03:48 ◼   ► can't justify keeping around the best photography website.

01:03:52 ◼   ► But that doesn't mean the best photography website is not

01:03:56 ◼   ► a viable business.

01:03:57 ◼   ► It's just not worth it to Amazon.

01:04:00 ◼   ► It's small potatoes for Amazon.

01:04:02 ◼   ► But that's why you don't want big companies to own things.

01:04:06 ◼   ► OK, so it's small potatoes for Amazon,

01:04:08 ◼   ► but it's still the best photography website

01:04:09 ◼   ► on the internet.

01:04:10 ◼   ► Somebody who cares about cameras should own it,

01:04:13 ◼   ► because you can still make money from it.

01:04:15 ◼   ► It's just a smaller number of people,

01:04:17 ◼   ► but those people are still very interested.

01:04:20 ◼   ► It's like a model train website or whatever.

01:04:22 ◼   ► Most people don't like model trains, but there's enough to sustain one really good website

01:04:25 ◼   ► on the internet for it.

01:04:27 ◼   ► The people who did the YouTube channel that I watched all the time, they are moving to

01:04:30 ◼   ► Petapixel so if you want to still follow them, they'll still be on YouTube just somewhere

01:04:36 ◼   ► else and I think that YouTube channel is pretty popular and I think somebody somewhere who

01:04:40 ◼   ► cares about cameras or has some correlated business like B&H for a new video or whatever,

01:04:46 ◼   ► Somebody should, I mean Amazon is selling I guess, maybe they tried to sell?

01:04:49 ◼   ► I would love it for someone to scoop this up.

01:04:51 ◼   ► Please give me all the archives of DPReview and the website and the business and the employees

01:04:56 ◼   ► and I'll make a go of it because even though it's not big enough for Amazon to care about

01:05:00 ◼   ► it, if as long as you can still make a viable business, maybe with fewer employees or something,

01:05:05 ◼   ► with a popular YouTube channel, it's a valuable resource that I think they could be making

01:05:10 ◼   ► more money than they were under Amazon and it's still worth having.

01:05:14 ◼   ► It's not like, well, cameras are not popular enough for there to be a really good website.

01:05:19 ◼   ► Sorry.

01:05:20 ◼   ► No, they're still popular enough.

01:05:21 ◼   ► They're popular enough to be sold on Amazon.com.

01:05:22 ◼   ► They're popular enough to be bought by people like Marco and me and occasionally Casey.

01:05:28 ◼   ► I still think there should be a good website.

01:05:30 ◼   ► So I really hope this works out better.

01:05:31 ◼   ► Sorry for the D-Rail.

01:05:32 ◼   ► No, it's fine.

01:05:33 ◼   ► This whole topic is a D-Rail.

01:05:35 ◼   ► Anyway, so yeah.

01:05:37 ◼   ► So anyway, so yeah, the departure of DPReview soon is sad.

01:05:43 ◼   ► I think that's, but at the same time,

01:05:46 ◼   ► I mean, I hadn't visited the site in many, many years

01:05:49 ◼   ► because I hadn't been in the market to buy a camera

01:05:51 ◼   ► in many, many years, and maybe that's part of the problem.

01:05:54 ◼   ► That maybe a lot of people haven't been in the market

01:05:56 ◼   ► to buy cameras in many, many years, and so--

01:05:58 ◼   ► - I haven't been in the market to buy one,

01:05:59 ◼   ► but I watch all the YouTube videos, so.

01:06:01 ◼   ► (laughing)

01:06:01 ◼   ► There's always lucky-loos.

01:06:02 ◼   ► - Yeah, true.

01:06:04 ◼   ► But anyway, the reason we were looking at a camera,

01:06:06 ◼   ► this is all John Gruber's fault,

01:06:08 ◼   ► because he had gotten recently the Ricoh GR3X

01:06:13 ◼   ► and Gruber's been shooting Ricoh GR cameras forever.

01:06:17 ◼   ► - Is that how you pronounce it?

01:06:19 ◼   ► - Is it Rico?

01:06:20 ◼   ► - I've always said Rico, but honestly,

01:06:21 ◼   ► I've never heard anyone say it out loud.

01:06:23 ◼   ► - It's probably Rico.

01:06:24 ◼   ► - I think it's Rico.

01:06:25 ◼   ► I am not particularly confident I'm correct about that.

01:06:28 ◼   ► - Anyway, that company, the GR3X,

01:06:31 ◼   ► it's very similar, there's a GR3 and a GR3X.

01:06:34 ◼   ► I think the only difference is the focal length

01:06:37 ◼   ► of the fixed lens, the three is a little more wide angle,

01:06:40 ◼   ► the three axis is about a 40 millimeter equivalent,

01:06:43 ◼   ► which is really nice.

01:06:44 ◼   ► And that, and Tiff loves, Tiff's favorite focal length

01:06:46 ◼   ► to shoot general purpose is 40 millimeters.

01:06:49 ◼   ► So she had expressed some interest in playing with this.

01:06:53 ◼   ► We occasionally will have some reason to use a big camera,

01:06:57 ◼   ► or somebody else will have a big camera

01:07:00 ◼   ► at some event that we are at, and we'll see the pictures,

01:07:03 ◼   ► and we'll be like, "Damn, those look really good.

01:07:05 ◼   ► I wish we carried our big cameras more often,

01:07:08 ◼   ► and we just don't.

01:07:09 ◼   ► Like, we hardly ever use them outside of specialty needs,

01:07:12 ◼   ► like, you know, the super long lens, try to shoot whales,

01:07:15 ◼   ► or the, you know, Christmas time, like, Christmas morning,

01:07:18 ◼   ► we'll take out the camera and shoot everyone opening,

01:07:20 ◼   ► opening their presents and everything.

01:07:21 ◼   ► But it's not, it's not a common thing by any means.

01:07:24 ◼   ► For the most part, we're just using our phones

01:07:26 ◼   ► almost entirely full-time as our cameras.

01:07:28 ◼   ► As the iPhone cameras have gotten better,

01:07:30 ◼   ► they, I mean, they're just so damn convenient,

01:07:32 ◼   ► but when you look at the pictures they take

01:07:34 ◼   ► versus pictures that bigger dedicated cameras take,

01:07:38 ◼   ► the bigger camera pictures are way better.

01:07:40 ◼   ► Not in every case, not in every set of conditions,

01:07:43 ◼   ► but when the big camera has good conditions,

01:07:46 ◼   ► compare that to the iPhone in also good conditions,

01:07:49 ◼   ► the big camera does look way better.

01:07:51 ◼   ► They look less processed, less over sharpened,

01:07:54 ◼   ► less painterly from the noise reduction algorithms and stuff

01:07:58 ◼   ► you can get better optics, you can get different optics,

01:08:00 ◼   ► you can get different trade-offs.

01:08:03 ◼   ► the pictures do look way, way better.

01:08:05 ◼   ► And I think there's a role in our life

01:08:08 ◼   ► to have both, hopefully, because while the iPhone

01:08:12 ◼   ► captures lots and lots and lots of day-to-day stuff,

01:08:15 ◼   ► I also don't want there to not be like any good quality

01:08:19 ◼   ► pictures of our family for like two years at a time.

01:08:23 ◼   ► So I do think there's a better balance to be struck

01:08:26 ◼   ► than all phone all the time,

01:08:29 ◼   ► because they really are really good

01:08:31 ◼   ► when you get good cameras.

01:08:32 ◼   ► Anyway, so the Raiko GR3X, I got that.

01:08:35 ◼   ► And I also, I wanted Tiff to try,

01:08:37 ◼   ► 'cause she expressed some interest in some other ones,

01:08:39 ◼   ► and I did some research, and the other one I got to try

01:08:42 ◼   ► was the Fuji X100V.

01:08:45 ◼   ► I don't know if it's 10 105, I don't know,

01:08:48 ◼   ► I'm gonna call it the X100V.

01:08:50 ◼   ► And because this camera, I had been eyeing Fuji

01:08:54 ◼   ► for a long time because I had heard that,

01:08:59 ◼   ► and I had seen online samples from sites like dpreview.com,

01:09:03 ◼   ► I had seen that Fuji seemed to have the most pleasing

01:09:07 ◼   ► out of camera JPEGs compared to the other brands.

01:09:10 ◼   ► It seems like the other brands mostly focused on

01:09:13 ◼   ► sheer technical quality and they kind of assumed

01:09:17 ◼   ► that if you care about photos you'd be shooting in raw

01:09:20 ◼   ► and you would do your own editing.

01:09:22 ◼   ► And so the out of camera JPEGs always seemed a little bit

01:09:25 ◼   ► half butted for my taste.

01:09:29 ◼   ► And for the most part, when I was more serious

01:09:31 ◼   ► about photography, back forever ago,

01:09:32 ◼   ► I didn't use the out-of-camera JPEGs for those reasons.

01:09:35 ◼   ► I shot RAW and I processed them afterwards, maybe,

01:09:38 ◼   ► edited them afterwards, maybe, and all that stuff.

01:09:41 ◼   ► Anyway, for this role, I was like, you know what?

01:09:44 ◼   ► We're never gonna do that.

01:09:46 ◼   ► We don't do that anymore.

01:09:46 ◼   ► We never edit stuff that way anymore.

01:09:49 ◼   ► We certainly don't wanna process RAW files.

01:09:52 ◼   ► I wanna just be able to shoot on the camera

01:09:54 ◼   ► and get it into my phone's photo library

01:09:57 ◼   ► as quickly as possible in whatever way we can.

01:09:59 ◼   ► So out of camera JPEG quality is important.

01:10:02 ◼   ► Also size is important and that rules out

01:10:04 ◼   ► many of the big full frame Sonys and stuff

01:10:07 ◼   ► that we've had before or full frame Canons,

01:10:10 ◼   ► all this, that kind of guy ruled out.

01:10:12 ◼   ► So I was looking at Fuji 'cause they seemed

01:10:14 ◼   ► very well liked in those kind of general categories.

01:10:18 ◼   ► Smallish but not super tiny but they're smallish,

01:10:21 ◼   ► really great out of camera JPEGs

01:10:23 ◼   ► And I also appreciated their seeming plethora

01:10:27 ◼   ► of manual controls on the camera.

01:10:30 ◼   ► The X100V has like physical knobs

01:10:31 ◼   ► for all the like main photographic controls

01:10:35 ◼   ► that you might want to adjust while shooting.

01:10:36 ◼   ► So the GR3X, Ricoh, Ricoh, GR3X, and the Fuji X100V

01:10:41 ◼   ► were like the main two that we got

01:10:44 ◼   ► that I wanted Tiff to try them both

01:10:47 ◼   ► and see which one she liked and I figure I would,

01:10:49 ◼   ► whichever one she didn't like,

01:10:50 ◼   ► I would just use myself for random stuff

01:10:52 ◼   ► or I'd use it as a webcam or something like that.

01:10:54 ◼   ► The Ricoh GR3X is amazing

01:10:59 ◼   ► at just basics of shooting.

01:11:02 ◼   ► And first of all, it's super tiny.

01:11:04 ◼   ► It is the only one of these that is remotely pocketable.

01:11:08 ◼   ► It is, you wouldn't wanna put it in like, you know,

01:11:12 ◼   ► a pocket of tight jeans or anything,

01:11:14 ◼   ► but any jacket or any bag would easily fit this camera

01:11:18 ◼   ► in the pocket.

01:11:19 ◼   ► So GR3X is the only one of these.

01:11:21 ◼   ► it is a whole different size class.

01:11:23 ◼   ► It is that much smaller than the Fuji's.

01:11:26 ◼   ► It is the smallest camera I've ever seen

01:11:29 ◼   ► that was able to produce this kind of quality.

01:11:31 ◼   ► And it is remarkably sharp photos.

01:11:35 ◼   ► Part of that is because it has great optics

01:11:37 ◼   ► and very importantly, it is sensor shift stabilized.

01:11:42 ◼   ► And in that size class, that's very rare.

01:11:45 ◼   ► And it's an APS-C sensor, it's a decently big sensor,

01:11:48 ◼   ► Very small camera, image stabilized, fixed prime,

01:11:52 ◼   ► great image quality.

01:11:54 ◼   ► Downside of that is that it has crappy battery life

01:11:57 ◼   ► 'cause it's so small and I found the colors

01:11:59 ◼   ► to be a little bit boring and the out of camera JPEGs

01:12:02 ◼   ► are again kind of dull and boring

01:12:05 ◼   ► and it seems like the kind of camera

01:12:06 ◼   ► that's better suited for if you're editing.

01:12:09 ◼   ► Also because it is so small,

01:12:11 ◼   ► the controls are a little harder to use

01:12:14 ◼   ► because of course there's not room on it

01:12:16 ◼   ► for a bunch of dials and things like that

01:12:18 ◼   ► So a few more things are in menus

01:12:21 ◼   ► or they're relegated to little tiny things.

01:12:23 ◼   ► So control's a little bit not as good,

01:12:25 ◼   ► but you can't beat the size

01:12:27 ◼   ► and you could just aim it and shoot

01:12:30 ◼   ► and you get something sharp.

01:12:31 ◼   ► 'Cause the combination of the stabilization,

01:12:33 ◼   ► the good lens, good autofocus, it just nailed it.

01:12:36 ◼   ► The Fuji X100V loved the pictures it got

01:12:40 ◼   ► when they are sharp.

01:12:43 ◼   ► That's a big when though.

01:12:45 ◼   ► So X100V optically a little bit better.

01:12:48 ◼   ► It's 35 millimeter, it's 2.0, but it's not stabilized.

01:12:53 ◼   ► And this is, I've done a little bit of research,

01:12:56 ◼   ► this seems to be like the number one feature request

01:12:58 ◼   ► for Fuji X series owners.

01:13:00 ◼   ► It's not stabilized and it really needs it.

01:13:04 ◼   ► It's very, very difficult to get a sharp shot

01:13:07 ◼   ► with the X100V if you're below like 1/500th of a second

01:13:12 ◼   ► of shutter speed, like it's very hard.

01:13:15 ◼   ► The autofocus is okay, it seems to have trouble

01:13:18 ◼   ► getting eyes very sharp and I tried different modes,

01:13:21 ◼   ► I tried the eye AF mode and stuff like that.

01:13:23 ◼   ► It's hit or miss with the focus on eyes,

01:13:26 ◼   ► which is frustrating and it's not stabilized,

01:13:28 ◼   ► so you have to keep the shutter speed up.

01:13:30 ◼   ► But it has all these cool film modes and stuff

01:13:32 ◼   ► and I was able to find some settings

01:13:34 ◼   ► that I just love the way the pictures look.

01:13:36 ◼   ► They look fantastic.

01:13:38 ◼   ► Color wise, skin tones, they look great.

01:13:41 ◼   ► Better than iPhone pictures, even in skin tone rendering

01:13:44 ◼   ► stuff like that. Super great. I love the Fuji rendering but the the lack of the

01:13:50 ◼   ► stabilization really hurts that camera. I love the physical controls. I love it

01:13:55 ◼   ► even looks cool. It is though noticeably bigger than the GR 3x and it is

01:13:59 ◼   ► definitely not a pocket camera. It is a small bag camera or maybe a like winter

01:14:04 ◼   ► jacket pocket but you know it's it's like it's a different size class than

01:14:08 ◼   ► the GR and it and it shows. But my favorite pictures that I shot during

01:14:13 ◼   ► during this experiment were from the Fuji.

01:14:15 ◼   ► But in order to make sure I was complete here

01:14:19 ◼   ► with my investigation here,

01:14:21 ◼   ► I decided after I had those first two

01:14:23 ◼   ► to get the Fuji X-T4 because it's very similar overall

01:14:28 ◼   ► to the X100V in many ways, similar controls,

01:14:32 ◼   ► similar body, similar JPEG rendering,

01:14:34 ◼   ► but it has image stabilization in the camera.

01:14:37 ◼   ► And this is an interchangeable lens,

01:14:38 ◼   ► So I got the 28 or 27 millimeter F2.8 pancake prime

01:14:43 ◼   ► 'cause it's small and it's close enough equivalent.

01:14:46 ◼   ► It's like about a 40 millimeter equivalent or so.

01:14:48 ◼   ► So similar to the other ones

01:14:50 ◼   ► and a little higher resolution sensor.

01:14:51 ◼   ► I believe that one's 40 megapixels.

01:14:53 ◼   ► It's like their newest sensor.

01:14:55 ◼   ► And the X-T4 was awesome in terms of like handling speed.

01:15:00 ◼   ► The auto focus was incredible.

01:15:04 ◼   ► The resolution was pleasantly higher.

01:15:07 ◼   ► I didn't test any of these in very low light

01:15:10 ◼   ► just because that's not really what I expect

01:15:13 ◼   ► a big camera to be good at anymore,

01:15:14 ◼   ► even though I know sensor-wise, optically,

01:15:17 ◼   ► it should be better than the iPhone.

01:15:18 ◼   ► In practice, it's not because of the magic

01:15:22 ◼   ► the iPhone is doing.

01:15:23 ◼   ► Anyway, so these are all, I'm testing all these

01:15:25 ◼   ► in moderate to high light situations.

01:15:28 ◼   ► So the X-T4 loved it, but it is significantly bigger

01:15:33 ◼   ► than the X100V also, and some of the controls

01:15:37 ◼   ► a little more complicated, but otherwise it's fine.

01:15:39 ◼   ► It's a great overall camera.

01:15:40 ◼   ► - You think this is an APS-C sensor with 40 megapixels?

01:15:44 ◼   ► That sounds wrong.

01:15:45 ◼   ► - Do I have that right?

01:15:45 ◼   ► Is it the X-T4?

01:15:47 ◼   ► - Maybe you have a different camera.

01:15:48 ◼   ► - X-T5 has 40, oh, I'm the wrong camera.

01:15:50 ◼   ► It was the X-T5, sorry. (laughs)

01:15:52 ◼   ► - Oh, so you didn't have the X-T4, you had the X-T5.

01:15:54 ◼   ► - Yeah, X-T5 is the one I had. (laughs)

01:15:56 ◼   ► Sorry, that's the current model.

01:15:58 ◼   ► Yeah, I had the current model, yeah, X-T5,

01:16:00 ◼   ► which is stabilized and 40 megapixels, yes.

01:16:04 ◼   ► Anyway, I had those three.

01:16:06 ◼   ► The reason I returned the X-T5 was that

01:16:08 ◼   ► Tiff took one look at it and she's like,

01:16:10 ◼   ► I don't wanna use that.

01:16:11 ◼   ► 'Cause she was already sold with the two smaller ones

01:16:14 ◼   ► and she didn't care about the stabilization.

01:16:17 ◼   ► I really cared.

01:16:18 ◼   ► But as I was using it, I'm like, you know,

01:16:20 ◼   ► I like this a lot, I like the pictures it produces,

01:16:24 ◼   ► but I can't see myself ever carrying this thing around.

01:16:28 ◼   ► - Yeah, now you're getting into just the size

01:16:30 ◼   ► of just a regular camera size.

01:16:31 ◼   ► Like this is not a small camera.

01:16:32 ◼   ► If I'm looking at this picture,

01:16:34 ◼   ► It's just not a full-size camera

01:16:38 ◼   ► 'cause it's not full-frame, but it's a big camera

01:16:40 ◼   ► with tons of dials on it and a thing where you put your eye

01:16:42 ◼   ► to it and a little handle where the battery goes.

01:16:45 ◼   ► - Yeah, it's very close in size to the world

01:16:48 ◼   ► of full-frame mirrorless cameras like the Sony A series.

01:16:52 ◼   ► It's in the ballpark of that size.

01:16:54 ◼   ► It is a little bit smaller, but not a lot smaller.

01:16:57 ◼   ► And so I was thinking, when am I actually gonna use this?

01:17:01 ◼   ► When am I gonna carry it around?

01:17:02 ◼   ► And during the brief time that I had it,

01:17:06 ◼   ► I like never even wanted to bring it outside.

01:17:09 ◼   ► So I was like, I think this is maybe a sign

01:17:11 ◼   ► that I should not really own a camera like this right now.

01:17:16 ◼   ► But I did enjoy using it the most

01:17:18 ◼   ► and I got really great pictures from it.

01:17:21 ◼   ► But I really enjoyed using the X100V more

01:17:25 ◼   ► and the X100V pictures were close enough in quality

01:17:30 ◼   ► and they were still way better than my iPhone

01:17:32 ◼   ► and I actually slightly preferred the skin tones

01:17:35 ◼   ► of the F100V.

01:17:36 ◼   ► So again, what I ultimately think would be

01:17:38 ◼   ► the best combination here would be an X100, I guess,

01:17:43 ◼   ► W, whatever would be the next one,

01:17:47 ◼   ► X100W with image stabilization.

01:17:50 ◼   ► That would be perfect.

01:17:52 ◼   ► That would be exactly what I want.

01:17:54 ◼   ► Everything else could stay the same.

01:17:55 ◼   ► It doesn't even need the higher resolution sensor

01:17:57 ◼   ► necessarily.

01:17:58 ◼   ► If it just did image stabilization and changed nothing else,

01:18:01 ◼   ► that would be enough, that would be amazing.

01:18:04 ◼   ► But anyway, so ultimately I came out of this experience

01:18:09 ◼   ► being very appreciative of what modern cameras have.

01:18:13 ◼   ► You know, my last camera that I got was an A7R III, I think.

01:18:18 ◼   ► And that was a long time ago now.

01:18:21 ◼   ► And that one is mostly used for TIFF's long lens

01:18:26 ◼   ► and you know, some experimental stuff like that.

01:18:30 ◼   ► But the modern cameras, there were things about them

01:18:33 ◼   ► that I liked that I was impressed by.

01:18:35 ◼   ► First of all, I was very impressed by how incredibly

01:18:37 ◼   ► responsive and fast they are.

01:18:40 ◼   ► The old full-frame Sonys were very sluggish.

01:18:43 ◼   ► The two was worse, the three was better,

01:18:45 ◼   ► but still, compared to the new ones,

01:18:48 ◼   ► the new ones are so much faster to operate

01:18:51 ◼   ► than the old ones were.

01:18:52 ◼   ► These also aren't full-frame,

01:18:54 ◼   ► they're dealing with fewer pixels,

01:18:55 ◼   ► so it's a bit of an unfair comparison.

01:18:57 ◼   ► But overall, great.

01:18:59 ◼   ► I was also, this is my first APS-C camera

01:19:02 ◼   ► since the Canon Rebel series that I used forever ago,

01:19:05 ◼   ► like since I got the 5D Mark II in 2008.

01:19:09 ◼   ► This is the first non-full frame,

01:19:12 ◼   ► but still quote big camera I've used.

01:19:14 ◼   ► And for my purposes, these were all great.

01:19:19 ◼   ► I don't think I really need full frame anymore

01:19:22 ◼   ► for my big camera roll, whatever that might be,

01:19:25 ◼   ► because these were all fantastic with APS-C sized sensors.

01:19:29 ◼   ► They really have come a long way since the olden days

01:19:33 ◼   ► and I really appreciate the optics.

01:19:35 ◼   ► Like, you know, one of the reasons why,

01:19:36 ◼   ► I mean, you can look at the price

01:19:38 ◼   ► and you can see a second reason,

01:19:39 ◼   ► but one of the reasons we didn't look at the Leica Q2,

01:19:43 ◼   ► 'cause it's the same kind of category

01:19:44 ◼   ► in terms of capability of like a compact,

01:19:49 ◼   ► fixed prime, point and shoot-ish kind of camera

01:19:52 ◼   ► with really high quality.

01:19:54 ◼   ► And the Q2 looks amazing, but because it's full frame,

01:19:59 ◼   ► the lens has to protrude from the body quite a lot.

01:20:03 ◼   ► And not just when it's on, all the time.

01:20:05 ◼   ► And that dramatically changes the shape of the camera

01:20:07 ◼   ► and therefore how you need to carry it,

01:20:09 ◼   ► like what kind of bags it can fit in,

01:20:11 ◼   ► it makes it significantly bigger.

01:20:13 ◼   ► And having previously owned the very first Sony RX1,

01:20:18 ◼   ► which is kind of Sony's copy of that style of camera,

01:20:21 ◼   ► I'm familiar with that size class

01:20:23 ◼   ► and it is quite large for my needs of carrying and stuff.

01:20:27 ◼   ► So I kind of ruled that out just for that

01:20:29 ◼   ► and Tiff wasn't interested in that at all.

01:20:30 ◼   ► Not to mention that it is like $6,000

01:20:32 ◼   ► but set that aside for now.

01:20:33 ◼   ► 'Cause if it was really great,

01:20:36 ◼   ► we might be willing to pay that if it was worth it

01:20:38 ◼   ► but it just wasn't what we were looking for.

01:20:40 ◼   ► So anyway, so I learned that these cameras

01:20:43 ◼   ► have gotten really good.

01:20:44 ◼   ► I learned that APS-C has gotten really good.

01:20:47 ◼   ► I learned that I love Fuji's JPEG rendering.

01:20:50 ◼   ► It is awesome, I'm very happy with it

01:20:52 ◼   ► and we continue to use the F100V kind of in like,

01:20:55 ◼   ► you know, kind of toy mode, you know,

01:20:56 ◼   ► occasionally shooting pictures of like the kid

01:20:58 ◼   ► and the dog and whatever.

01:20:59 ◼   ► It's, you know, these super cute pictures.

01:21:01 ◼   ► One thing I was disappointed by is the story

01:21:07 ◼   ► of transferring photos from the camera to your phone

01:21:11 ◼   ► is still as terrible as it ever was.

01:21:14 ◼   ► - I'll be like micro USB.

01:21:16 ◼   ► - No, well, no, these are USB-C at least, which is good.

01:21:18 ◼   ► - Wow, that's good.

01:21:19 ◼   ► I was looking at one of the other Fuji ones

01:21:21 ◼   ► and I was surprised to see a micro USB shaped hole

01:21:23 ◼   ► on the side, USB-C, that's modern technology.

01:21:25 ◼   ► - Yeah, and they both charge

01:21:26 ◼   ► via USB-C power delivery chargers.

01:21:28 ◼   ► I'm glad to report.

01:21:30 ◼   ► - I mean, if you have a laptop with an SD card slot,

01:21:32 ◼   ► can't you just take the card out and shove it in?

01:21:33 ◼   ► - Yes, and so, basically, the easiest way to do it,

01:21:38 ◼   ► I actually don't have a camera connection kit here,

01:21:42 ◼   ► but I was thinking maybe the camera connection kit

01:21:45 ◼   ► with like direct into, like USB into a phone

01:21:49 ◼   ► might be better, it might work,

01:21:52 ◼   ► 'cause the phone gained the ability

01:21:53 ◼   ► to import off-camera connection kit

01:21:55 ◼   ► a few software versions ago, I think,

01:21:57 ◼   ► like iOS 13 or 12 or something,

01:22:00 ◼   ► sometime it gained that ability,

01:22:02 ◼   ► but I haven't tried it in a long time,

01:22:04 ◼   ► and I don't have the hardware here to do that,

01:22:06 ◼   ► so I couldn't test that, but that might be a way to do it

01:22:08 ◼   ► if you're out and about,

01:22:09 ◼   ► just put the little camera connection kit cable

01:22:11 ◼   ► in your bag or pocket and attach the camera to the phone

01:22:15 ◼   ► and just do a direct transfer.

01:22:17 ◼   ► Or you can do, if you have a computer,

01:22:18 ◼   ► what Jon said, either plug the camera in via USB

01:22:21 ◼   ► and have the camera read its own card over USB

01:22:22 ◼   ► or pop the card out, put it into the computer

01:22:25 ◼   ► and read it that way directly into Photos app.

01:22:26 ◼   ► That all works fine.

01:22:28 ◼   ► The cameras both offer a WiFi feature

01:22:33 ◼   ► and they have companion apps and they're third party apps

01:22:36 ◼   ► for at least the Ricoh that I tried also.

01:22:38 ◼   ► I don't know if Fuji has third party,

01:22:40 ◼   ► oh I think I tried one of those too.

01:22:42 ◼   ► And these all work in the ancient way

01:22:45 ◼   ► that we've had forever, which is the camera

01:22:50 ◼   ► creates its own wifi network, the app joins it,

01:22:53 ◼   ► and then it tries to transfer photos over wifi.

01:22:56 ◼   ► This works exactly as well as you would think

01:22:58 ◼   ► based on that description.

01:23:00 ◼   ► It is flaky, it's unreliable, it is slow to connect,

01:23:04 ◼   ► it is a pain to connect, and then it is very slow

01:23:07 ◼   ► to actually do the photo transfer.

01:23:09 ◼   ► It does work in the sense that in a pinch,

01:23:13 ◼   ► if you have the time and patience,

01:23:14 ◼   ► you can get it to work, you can transfer photos,

01:23:17 ◼   ► but it is so cumbersome you really won't want to.

01:23:21 ◼   ► So that I was hoping, I was hoping that with recent advances

01:23:24 ◼   ► in various iOS capabilities and local Bluetooth

01:23:28 ◼   ► and WiFi transfer kind of stuff or NFC stuff,

01:23:31 ◼   ► I figured maybe that would have been better,

01:23:33 ◼   ► but for whatever reason it's not.

01:23:36 ◼   ► And I think it's mostly due to just the iPhone

01:23:39 ◼   ► and the capabilities it offers.

01:23:41 ◼   ► I don't think Apple is in a big rush

01:23:42 ◼   ► to enable easy wireless transfer from an external camera

01:23:47 ◼   ► they don't make to their phone,

01:23:48 ◼   ► which is sold primarily because it's a really good camera.

01:23:51 ◼   ► So I imagine this is not an area of focus for Apple,

01:23:54 ◼   ► but the experience really is horrendous for doing that

01:23:58 ◼   ► if you wanna do that.

01:23:59 ◼   ► And ultimately, that's what keeps me

01:24:01 ◼   ► from using these cameras more,

01:24:02 ◼   ► besides the fact that I'm not carrying them.

01:24:04 ◼   ► But I think I would use them a lot more

01:24:06 ◼   ► if it was really fast and easy for the picture

01:24:10 ◼   ► that I take with it to then pop up on my phone.

01:24:12 ◼   ► but it's not, it's neither fast nor easy.

01:24:14 ◼   ► So that really holds back the usage of these in practice

01:24:17 ◼   ► because I almost always want that picture to be on my phone

01:24:20 ◼   ► so I can do something with it,

01:24:21 ◼   ► whether it's sending it to somebody, posting it somewhere,

01:24:24 ◼   ► or just editing in some way.

01:24:27 ◼   ► I pretty much want it on my phone immediately

01:24:29 ◼   ► and I think that's a common need.

01:24:31 ◼   ► So ultimately that part still is unfortunate

01:24:34 ◼   ► and it kind of leaves the standalone camera market

01:24:38 ◼   ► in the place it is, which is like,

01:24:40 ◼   ► It's this kind of increasingly specialized,

01:24:44 ◼   ► shrinking enthusiast market.

01:24:47 ◼   ► And it's great for that, but it used to be a lot broader

01:24:49 ◼   ► than just photographers and enthusiasts.

01:24:52 ◼   ► It used to be way more mass market.

01:24:54 ◼   ► And phones just killed that, and it's not coming back.

01:24:57 ◼   ► But I do wish that enthusiasts who also use phones

01:25:01 ◼   ► could have a bit better experience trying

01:25:04 ◼   ► to use these amazing new cameras that we have that just clash

01:25:09 ◼   ► horrendously with the phone world.

01:25:12 ◼   ► I wonder if camera makers--

01:25:13 ◼   ► because I'm willing to blame the camera makers much more

01:25:15 ◼   ► than you are, because they're terrible at making software.

01:25:18 ◼   ► Their custom apps suck.

01:25:19 ◼   ► But I wonder if they would be better served,

01:25:21 ◼   ► like at least to get closer to what you want,

01:25:23 ◼   ► is to treat the cameras kind of like your phone.

01:25:25 ◼   ► Because if you're out and about with your phone

01:25:27 ◼   ► and taking pictures, generally your phone,

01:25:29 ◼   ► especially if you're just out somewhere on cellular,

01:25:31 ◼   ► your phone won't bother uploading those to iCloud

01:25:33 ◼   ► until you get back closer to Wi-Fi

01:25:36 ◼   ► or until you plug it in sometimes.

01:25:37 ◼   ► You've all seen that thing at the bottom of the photos thing

01:25:39 ◼   ► that says, "I paused my photo syncing until you plugged me in again," right?

01:25:42 ◼   ► And you can tell it, to Apple's credit, rare instance where you can say, "No, do it now."

01:25:47 ◼   ► Wow, imagine that.

01:25:48 ◼   ► Well, yes, but that's a suggestion more than it is a demand, because you hit that and it's

01:25:54 ◼   ► just like, "Well, I'll get there eventually."

01:25:56 ◼   ► Yeah, but anyway, if these cameras work like this, you'd be out and about, you're taking

01:26:00 ◼   ► your pictures, and you kind of have the same expectation you have of a phone.

01:26:02 ◼   ► It's like, "Well, these aren't really going to be uploaded to iCloud anytime soon," but

01:26:06 ◼   ► When you come back home you'd like put the camera on a charger and then it would load

01:26:10 ◼   ► them to iCloud for you using like I don't know public Apple API's or something.

01:26:16 ◼   ► In other words have the camera A) connect to your wifi the same as all your other devices,

01:26:21 ◼   ► none of this creating its own network whatever stuff and B) upload when it's on wifi to iCloud

01:26:28 ◼   ► directly.

01:26:29 ◼   ► That seems like something the camera manufacturers could probably work out with Apple if they

01:26:32 ◼   ► wanted to do it but of course no one wants to do that because like Apple who uses those

01:26:35 ◼   ► computers and you know we've got our own app made by our crack team of expert programmers

01:26:40 ◼   ► and they're all horrendous and like the approach they're taking is bad and the main problem

01:26:45 ◼   ► is like how long can camera manufacturers continue to keep these cameras off of the

01:26:49 ◼   ► internet essentially like they all added Wi-Fi and Bluetooth but the cameras are still not

01:26:53 ◼   ► in any meaningful way quote unquote on the internet that's why they're making their own

01:26:56 ◼   ► stupid Wi-Fi network and it's like just just bite the bullet your camera needs to be on

01:27:01 ◼   ► the internet because that's where we're going our photos to be or you know even if it's

01:27:04 ◼   ► Let's just like airdrop the photos from the camera to your phone so you can do stuff with

01:27:09 ◼   ► them like they could be doing a lot better.

01:27:11 ◼   ► So I don't think it's really on Apple.

01:27:13 ◼   ► I think Apple has done plenty to make it possible for these cameras to be better but the camera

01:27:17 ◼   ► manufacturers software is not their strong suit especially software when it comes to

01:27:20 ◼   ► networking and not photography.

01:27:23 ◼   ► Yeah exactly and again it's a shame because I think it would take a lot of you know help

01:27:28 ◼   ► from Apple to really make that work well and whether it's on the web service side when

01:27:33 ◼   ► and they're uploading directly to the web service

01:27:34 ◼   ► or whether they're bouncing it off the phone,

01:27:35 ◼   ► which is probably the easier, but maybe slower

01:27:39 ◼   ► or more limited way.

01:27:40 ◼   ► Whatever it is, the standalone camera world,

01:27:44 ◼   ► I mean look, they don't even all have GPS yet.

01:27:46 ◼   ► That's even, GPS is still a kind of optional,

01:27:49 ◼   ► rarely seen feature.

01:27:51 ◼   ► - Don't worry, there's a great companion app

01:27:52 ◼   ► that will solve that problem.

01:27:53 ◼   ► Ha ha ha.

01:27:54 ◼   ► This is one of the things I found out

01:27:55 ◼   ► when Marco sent me his other camera.

01:27:58 ◼   ► Did you know that the Sony GPS thing or whatever,

01:28:00 ◼   ► that there's an app that you run on your phone

01:28:02 ◼   ► that connects with the camera and it will send GPS

01:28:05 ◼   ► and stuff or whatever,

01:28:07 ◼   ► each camera can only be connected to one phone.

01:28:10 ◼   ► And if your phone is already connected to another camera,

01:28:12 ◼   ► it can't be connected to a second one.

01:28:13 ◼   ► - What? (laughs)

01:28:14 ◼   ► - So I've got two Sony cameras,

01:28:16 ◼   ► but I've only got one iPhone.

01:28:18 ◼   ► And so it's like, well, which camera do you want connected?

01:28:20 ◼   ► 'Cause you can't connect both.

01:28:21 ◼   ► It's like, seriously?

01:28:22 ◼   ► - Seriously dumb. - It's ridiculous.

01:28:24 ◼   ► So I had to have one connect to my phone

01:28:25 ◼   ► and one connect to my wife's phone,

01:28:26 ◼   ► but then it just hardly ever works.

01:28:29 ◼   ► Camera software is terrible.

01:28:30 ◼   ► Sony just discontinued in fact, it's Sony PlayMemories application, which is some really

01:28:36 ◼   ► misguided attempt to have like an iPhone replacement that runs on Mac OS.

01:28:40 ◼   ► If you can imagine that like made by Sony, made by the camera part of Sony.

01:28:44 ◼   ► It's they would make you have it and install it if you wanted to use a bunch of other features.

01:28:48 ◼   ► It was terrible.

01:28:49 ◼   ► It's like it's like the software that comes with printers, right?

01:28:51 ◼   ► Like that stuff like don't use image capture, use our scanning program.

01:28:55 ◼   ► It's like the worst thing you've ever seen in your life.

01:28:57 ◼   ► Sony just discontinued that.

01:28:58 ◼   ► I hope they discontinued it because either they're never going to try to make software

01:29:02 ◼   ► like that again, which I would recommend, or it's being replaced by something better,

01:29:06 ◼   ► which I guess would be the second best.

01:29:09 ◼   ► I've never seen an iPhone app by a camera maker that was better than awful.

01:29:16 ◼   ► I mean, they're so bad.

01:29:18 ◼   ► It amazes me that they launch.

01:29:20 ◼   ► It amazes me that, like, how did they not crash on launch?

01:29:23 ◼   ► They're among the worst software I've ever seen on any platform.

01:29:26 ◼   ► It's like that and printer software.

01:29:27 ◼   ► It's a real race there.

01:29:28 ◼   ► Yeah, like it amazes me that they pass app review.

01:29:30 ◼   ► They're just, they're so bad.

01:29:32 ◼   ► Yeah, it's really true.

01:29:33 ◼   ► The Olympus one is just as bad and they've like moved the two pieces of functionality

01:29:39 ◼   ► I want, which is occasional downloads and geotagging.

01:29:43 ◼   ► They've moved that over the last eight years between like two or three different apps,

01:29:47 ◼   ► you know, oh, don't use the OI photo share what I don't even remember what it's called,

01:29:51 ◼   ► but don't use the, you know, a app because all that functionality is now in the B app.

01:29:55 ◼   ► Fast forward two years.

01:29:56 ◼   ► Don't use the B app.

01:29:57 ◼   ► Now all that functionality is in the C app.

01:29:59 ◼   ► It's just so ridiculous.

01:30:01 ◼   ► And the apps are pieces of garbage too.

01:30:03 ◼   ► - Yeah, I've had that same problem with Sony

01:30:05 ◼   ► and with everyone else.

01:30:06 ◼   ► It's just like they keep rewriting the app

01:30:08 ◼   ► or we're gonna bring our multiple brands together

01:30:11 ◼   ► or bring, you know, unify our product line.

01:30:13 ◼   ► And it's funny, you know, you look at the app store reviews

01:30:16 ◼   ► and they all have one star 'cause they don't work.

01:30:19 ◼   ► - Yeah, like they just, these camera managers

01:30:21 ◼   ► need to understand that there's an ecosystem

01:30:22 ◼   ► they should participate in.

01:30:23 ◼   ► These cameras should be on the internet,

01:30:25 ◼   ► They should integrate with other existing platforms

01:30:27 ◼   ► instead of trying to say like,

01:30:28 ◼   ► "We're gonna make our own version of Apple Photos app

01:30:31 ◼   ► "and we're gonna make our own version of Photo Editor.

01:30:34 ◼   ► "So we're gonna replace Apple Photos, Lightroom, Photoshop,

01:30:38 ◼   ► "the thing you use to print calendars and pictures."

01:30:41 ◼   ► Like, "We're gonna write all that software from scratch

01:30:42 ◼   ► "for every platform 'cause why wouldn't we?"

01:30:44 ◼   ► It's like, "Why would you even attempt that?

01:30:46 ◼   ► "Just integrate with the Apple platform.

01:30:49 ◼   ► "Integrate with Windows."

01:30:50 ◼   ► And you're done.

01:30:51 ◼   ► That's all you need to care about.

01:30:52 ◼   ► Maybe integrate with Android, iOS, Mac, Windows.

01:30:55 ◼   ► forget about Linux, they'll be fine.

01:30:57 ◼   ► That's it, but instead they try to rewrite the world

01:30:59 ◼   ► from scratch, like all in a Sony branded experience

01:31:02 ◼   ► where you have to create an account and upload your memory.

01:31:04 ◼   ► It's like, you think I'm gonna give my pictures

01:31:07 ◼   ► to the Sony Play memories application, are you kidding?

01:31:09 ◼   ► Again, I'm surprised this app even launches

01:31:13 ◼   ► without immediately crashing.

01:31:14 ◼   ► I'm never giving in any of my data, but it's like,

01:31:16 ◼   ► well, you have to sign in and make an account

01:31:18 ◼   ► and install it so we can geotag your photos,

01:31:20 ◼   ► which is the one piece of functionality that you want.

01:31:22 ◼   ► - Mm-hmm, it's so bad.

01:31:24 ◼   ► I understand why on some level they don't do it.

01:31:27 ◼   ► I mean, first of all, consumer software

01:31:29 ◼   ► is not the strength of any of these camera companies.

01:31:32 ◼   ► They've never had a reason to develop it.

01:31:34 ◼   ► They certainly don't have the talent or the will

01:31:36 ◼   ► or the resources apparently,

01:31:37 ◼   ► 'cause otherwise they would do it.

01:31:39 ◼   ► But also, if you look at the role of their devices,

01:31:42 ◼   ► I mean, a lot of the standalone camera market now

01:31:46 ◼   ► is people shooting video and stuff like that.

01:31:49 ◼   ► It isn't just people taking casual pictures

01:31:51 ◼   ► and wanting to upload them to Instagram or whatever.

01:31:53 ◼   ► It's like people using these cameras as like semi-pro

01:31:56 ◼   ► video capture devices.

01:31:58 ◼   ► And it's just, it's such a different market, you know?

01:32:01 ◼   ► But also, it's kind of a chicken and egg problem.

01:32:03 ◼   ► Like the camera makers probably say,