ATP

00:00:00 ◼   ► I gotta go get a cup of water 'cause I didn't,

00:00:02 ◼   ► I realized I don't have one here, so hang on a second.

00:00:04 ◼   ► I'll be back.

00:00:05 ◼   ► Play some fish, I'll be back in a second.

00:00:07 ◼   ► Oh god, no, please don't.

00:00:08 ◼   ► We learned a lot more about the MacBook Pros

00:00:12 ◼   ► and everything this week, and honestly,

00:00:14 ◼   ► I feel like we spent too much time on it,

00:00:17 ◼   ► but the news is so big that we do need to talk about

00:00:19 ◼   ► a few angles of it. (laughs)

00:00:21 ◼   ► And like this, it's the kind of news that

00:00:24 ◼   ► whenever there's like a new, like brief little scandal

00:00:27 ◼   ► about our new Apple product, that news is relevant

00:00:30 ◼   ► for like a week, and then after that,

00:00:33 ◼   ► the value of it is basically zero.

00:00:35 ◼   ► - Except for the keyboard, which had like a three year life.

00:00:39 ◼   ► (laughs)

00:00:39 ◼   ► - Well, because they haven't fixed that.

00:00:41 ◼   ► Like it's one thing, like when they fix it--

00:00:42 ◼   ► - You never know. - Have they?

00:00:44 ◼   ► - Well, I guess, yeah, I guess we don't know.

00:00:45 ◼   ► But yeah, like when they fix it, then it becomes,

00:00:49 ◼   ► like then the value of like the scandal that happened

00:00:51 ◼   ► is basically zero.

00:00:53 ◼   ► So like, it's rough that we have to like spend

00:00:56 ◼   ► like pretty much three shows talking about this,

00:00:58 ◼   ► but there's a bunch of angles in this

00:01:00 ◼   ► that are kind of interesting.

00:01:01 ◼   ► I've learned a lot about this,

00:01:03 ◼   ► and so I hope that this can,

00:01:05 ◼   ► I hope we can provide some value besides,

00:01:07 ◼   ► hey, there was a problem, and now it's fixed.

00:01:09 ◼   ► (laughs)

00:01:11 ◼   ► - Yeah, you know, it's been weird,

00:01:13 ◼   ► because we caught the drama with regard to the

00:01:18 ◼   ► brand new i9, is it i9 or i7?

00:01:21 ◼   ► I'm having a brain fart.

00:01:22 ◼   ► - Well, I mean, technically it's all of them,

00:01:23 ◼   ► but the one that caused the most drama was the i9.

00:01:25 ◼   ► - Okay, I'm not crazy.

00:01:26 ◼   ► Anyway, so we caught, we recorded just after that all broke,

00:01:31 ◼   ► and in the time between we recorded and now,

00:01:34 ◼   ► when we're recording again,

00:01:35 ◼   ► it's basically all blown back over.

00:01:37 ◼   ► But we'll talk more about that.

00:01:38 ◼   ► - Our timing was perfect, I think.

00:01:39 ◼   ► - Yeah, actually that's true.

00:01:40 ◼   ► - These crazies were perfectly timed for our recording,

00:01:42 ◼   ► so thank you, Apple.

00:01:43 ◼   ► - Yeah, actually, I agree.

00:01:44 ◼   ► Like, you know, so the big fix came out,

00:01:46 ◼   ► was it yesterday, was it?

00:01:48 ◼   ► And so yeah, like we actually are perfectly timed

00:01:51 ◼   ► to talk about this in a way that is probably complete,

00:01:56 ◼   ► that has a decent amount of informing-ness to it,

00:01:59 ◼   ► informed-ness information to it.

00:02:01 ◼   ► (laughing)

00:02:03 ◼   ► And like, 'cause you know, last week we were kind of

00:02:06 ◼   ► in the middle of it.

00:02:07 ◼   ► Like we had seen the videos and Reddit posts that like,

00:02:10 ◼   ► there's a problem with the new MacBook Pro, we think.

00:02:13 ◼   ► And we could speculate a whole bunch on what that was,

00:02:15 ◼   ► and we could say it, but I spouted off about like,

00:02:17 ◼   ► what should and shouldn't be the case.

00:02:19 ◼   ► Now we have the information to actually,

00:02:21 ◼   ► like now we have way more information than last time.

00:02:24 ◼   ► We'll get to that, so I guess,

00:02:26 ◼   ► let's end this preamble for now, we'll get to that,

00:02:28 ◼   ► but suffice to say, I know this is a lot to talk about,

00:02:32 ◼   ► the new MacBook Pro and its weird throttling thing.

00:02:35 ◼   ► But I think we have value to add,

00:02:38 ◼   ► I will try to keep it brief.

00:02:40 ◼   ► - With that in mind, everyone look at your timestamps now.

00:02:43 ◼   ► All right, so to start off,

00:02:44 ◼   ► let's do some of the more mundane things.

00:02:46 ◼   ► It turns out that it is easier to replace the key caps

00:02:51 ◼   ► on the brand new keyboards.

00:02:53 ◼   ► From this small website I've never heard of

00:02:55 ◼   ► called 512Pixels, an Apple service document tells us

00:02:58 ◼   ► that the keyboard buttons is the word I'm looking for,

00:03:02 ◼   ► they're easier to replace, and the membrane is quote,

00:03:06 ◼   ► "To prevent debris from entering the butterfly mechanism.

00:03:09 ◼   ► Be careful not to tear the membrane."

00:03:11 ◼   ► So that's good news, and hopefully it works.

00:03:13 ◼   ► - You tried to combine two follow-up items there.

00:03:15 ◼   ► What are you doing?

00:03:16 ◼   ► - Dude, I'm following the show notes, man.

00:03:18 ◼   ► What do you want from me?

00:03:19 ◼   ► - That's a violation, sir.

00:03:20 ◼   ► - That's right.

00:03:21 ◼   ► Boy, you violated the indenting.

00:03:24 ◼   ► - Oh my God. - And getting this

00:03:24 ◼   ► for a reason, it's like Python.

00:03:26 ◼   ► Significant white space.

00:03:28 ◼   ► - Oh God.

00:03:29 ◼   ► (laughing)

00:03:30 ◼   ► John, why are you, I thought only Merlin got bad cop John.

00:03:33 ◼   ► Now I'm getting bad cop John.

00:03:35 ◼   ► - There's no such thing as bad cop John.

00:03:36 ◼   ► (laughing)

00:03:38 ◼   ► You messed with follow-up, Casey.

00:03:39 ◼   ► I mean, what do you expect?

00:03:40 ◼   ► - Oh, my word.

00:03:41 ◼   ► - The easier key cap replacement I think is a good move.

00:03:43 ◼   ► Like independent of the, you know,

00:03:45 ◼   ► the reliability of the keyboard

00:03:47 ◼   ► or how you like it or whatever,

00:03:48 ◼   ► being able to just fix just one key cap

00:03:51 ◼   ► is a smart thing to do.

00:03:52 ◼   ► And again, three years, we're three years

00:03:55 ◼   ► into this keyboard now, starting from the 2015 MacBook.

00:03:58 ◼   ► The fact that they went three years with a keyboard

00:04:01 ◼   ► where if a key cap like pops off or something

00:04:03 ◼   ► that you have to replace the entire keyboard,

00:04:05 ◼   ► it just seems like not a smart move

00:04:07 ◼   ► from a like cost and repairability perspective, right?

00:04:12 ◼   ► So I'm glad that they've addressed that

00:04:14 ◼   ► because it's just, you know, it's such a shame to see

00:04:17 ◼   ► there's nothing actually wrong with my keyboard,

00:04:19 ◼   ► but like I dropped something on it

00:04:20 ◼   ► and it broke this one key cap off.

00:04:21 ◼   ► Well, top case replacement.

00:04:23 ◼   ► Like that used to be the answer.

00:04:25 ◼   ► So now that you have a fighting chance,

00:04:27 ◼   ► the, you know, repair people have a fighting chance

00:04:29 ◼   ► of fixing your key cap by itself.

00:04:32 ◼   ► Although you have to be careful not to tear

00:04:33 ◼   ► the little membrane thing because the membrane,

00:04:36 ◼   ► as iFixit teardown has shown, is all one piece.

00:04:40 ◼   ► So if you tear it, it's not like you can replace

00:04:42 ◼   ► just the membrane under that key.

00:04:44 ◼   ► It's just one giant thing.

00:04:45 ◼   ► So you need a whole new top case.

00:04:46 ◼   ► And the second thing was like,

00:04:49 ◼   ► I mean, we talked about this last week,

00:04:49 ◼   ► I just wanted to put in here for confirmation,

00:04:51 ◼   ► like that Apple just said the keyboard is quieter,

00:04:54 ◼   ► but they had a patent about keeping debris out

00:04:57 ◼   ► and the patent looks like the keyboard.

00:04:58 ◼   ► So presumably that thing is not there to make it quieter,

00:05:01 ◼   ► it's to keep stuff out.

00:05:02 ◼   ► Apple's own documentation, it's internal documentation

00:05:05 ◼   ► for it's like service guide or whatever it says.

00:05:07 ◼   ► The keyboard has a membrane under the key caps

00:05:09 ◼   ► to prevent debris from entering the butterfly mechanism.

00:05:12 ◼   ► It's right from Apple's mouth, so if anyone

00:05:14 ◼   ► still had any question whether the membrane

00:05:16 ◼   ► was there to make it quieter or to keep debris out,

00:05:18 ◼   ► Apple has settled that once and for all.

00:05:20 ◼   ► - Yeah, that's like, and we have another

00:05:21 ◼   ► follow up item here that I'd like to also read.

00:05:24 ◼   ► This was from Josh Fenton, 'cause this is related to this.

00:05:27 ◼   ► When Apple issued the release of these new Apple Pros

00:05:30 ◼   ► and they said, this new keyboard design is quieter.

00:05:33 ◼   ► And everyone's like, does it fix the dust killing it

00:05:37 ◼   ► thing or not?

00:05:37 ◼   ► And it's pretty clear now that Apple was kind of BSing us

00:05:42 ◼   ► on that, like yeah, I guess it is a little quieter maybe,

00:05:47 ◼   ► but people are like straining to hear the difference.

00:05:49 ◼   ► Meanwhile, both Apple's support documents now

00:05:52 ◼   ► and iFixit tear downs and things confirm

00:05:54 ◼   ► that this is not really to be quieter,

00:05:56 ◼   ► it's really there for this ingress prevention.

00:05:58 ◼   ► So obviously that's the real reason it's here

00:06:00 ◼   ► and Apple was kind of BSing us.

00:06:01 ◼   ► One of the reasons why people speculated, including us,

00:06:05 ◼   ► why they might be not saying why this is actually here,

00:06:10 ◼   ► is that there are a number of class action lawsuits

00:06:13 ◼   ► about the keyboard failure.

00:06:14 ◼   ► And so people are speculating, maybe they didn't want to say

00:06:18 ◼   ► that it's quieter, or say that it fixes this problem

00:06:21 ◼   ► because that would admit fault that there is a problem

00:06:23 ◼   ► in the first place.

00:06:24 ◼   ► And we got a nice email from Josh Fenton, who is a lawyer,

00:06:27 ◼   ► saying that there's a legal doctrine that prohibits

00:06:30 ◼   ► introducing subsequent remedial measures as evidence

00:06:34 ◼   ► of the existence of a previous negligence or design defect.

00:06:37 ◼   ► So the idea is you couldn't make a legal,

00:06:40 ◼   ► according to this doctrine, it is not a valid legal argument

00:06:43 ◼   ► to say you definitely had a defect before

00:06:46 ◼   ► because the new product changed in this way

00:06:48 ◼   ► to make this thing better.

00:06:50 ◼   ► And Josh Fenton says, while counterintuitive,

00:06:52 ◼   ► this doctrine exists in support of a public policy

00:06:55 ◼   ► that the law should not serve as a disincentive

00:06:57 ◼   ► for people or companies to improve the safety and quality

00:06:59 ◼   ► of their products, services, or premises.

00:07:01 ◼   ► 'Cause it's kind of, that would be kind of counterintuitive

00:07:03 ◼   ► and kind of damaging over time.

00:07:05 ◼   ► So I don't know, I mean, chances are Apple is being

00:07:07 ◼   ► exceptionally conservative and not even wanting

00:07:10 ◼   ► to try to have to invoke this kind of doctrine.

00:07:12 ◼   ► Like they probably want as much immunity as possible

00:07:15 ◼   ► against these lawsuits.

00:07:16 ◼   ► So this doesn't say like, they don't have to come out

00:07:19 ◼   ► and say that this is fixed, but it does seem like they have

00:07:22 ◼   ► less legal reason than we predicted to avoid saying

00:07:27 ◼   ► that this is fixed because of ingress.

00:07:29 ◼   ► - So if that's true, what's left is what people

00:07:31 ◼   ► were thinking before, which is pride, as in corporate pride

00:07:34 ◼   ► or like why would we admit that we ever did anything wrong?

00:07:36 ◼   ► Why don't we just say we have a new keyboard, it's great,

00:07:38 ◼   ► best keyboard we ever made, and it's quieter,

00:07:40 ◼   ► and just don't answer any questions about reliability.

00:07:45 ◼   ► Because if it's not legal, keeping their mouth shut,

00:07:47 ◼   ► it's a worse look.

00:07:49 ◼   ► 'Cause if it was a legal thing, like whatever,

00:07:50 ◼   ► lawsuits are weird.

00:07:51 ◼   ► And you're right, maybe it still could be an abundance

00:07:53 ◼   ► of caution and who knows if there's some specific case law

00:07:56 ◼   ► related to this or whatever.

00:07:57 ◼   ► But if the legal reasons are taken out of the equation,

00:08:00 ◼   ► the only thing left is like, I think a miscalculation,

00:08:04 ◼   ► a PR miscalculation perhaps.

00:08:06 ◼   ► 'Cause it looks better when you say we had a problem

00:08:10 ◼   ► and we fixed it and the new ones are great.

00:08:11 ◼   ► And it's not like they're not admitting they had a problem

00:08:13 ◼   ► 'cause they do have the repair extension program

00:08:15 ◼   ► and you don't do that just for the hell of it.

00:08:17 ◼   ► You don't do a repair extension program on a keyboard

00:08:19 ◼   ► that has perfect reliability, right?

00:08:20 ◼   ► There's obviously a problem and yet still,

00:08:23 ◼   ► somehow they could not bring themselves to say

00:08:25 ◼   ► this new keyboard is more reliable.

00:08:27 ◼   ► Going back to the keyboard and these membranes,

00:08:30 ◼   ► iFixit did some really interesting work with regard

00:08:34 ◼   ► to testing whether or not the membrane does anything.

00:08:37 ◼   ► So they have this post which we'll link to

00:08:40 ◼   ► where they did a few things to kind of figure out,

00:08:42 ◼   ► does this make a difference?

00:08:43 ◼   ► And what they did was they sprinkled blue powder

00:08:47 ◼   ► on top of the keys, used them for a while

00:08:49 ◼   ► and then popped off the key caps to see, okay,

00:08:51 ◼   ► where's this blue powder?

00:08:52 ◼   ► And the conclusion they seem to come to is that,

00:08:54 ◼   ► yeah, it definitely helps.

00:08:56 ◼   ► I think the most relevant part of this test though

00:08:57 ◼   ► is so they did this glowy sand which is very, very fine.

00:09:00 ◼   ► It's kind of cool because you can track where it is

00:09:02 ◼   ► and everything and showing that it mostly keeps it.

00:09:05 ◼   ► An important part is I think they compare it

00:09:06 ◼   ► to one without the membrane.

00:09:07 ◼   ► Say without the membrane, it's all over the place

00:09:09 ◼   ► and with the membrane, it's mostly on the edges, good, right?

00:09:11 ◼   ► Well then they took a little bit of sand and put it on

00:09:13 ◼   ► and it wedged the keys instantly.

00:09:15 ◼   ► Because there are holes in the membrane.

00:09:16 ◼   ► There are big holes where the hinges come through.

00:09:18 ◼   ► So it's not like sand can't get in

00:09:20 ◼   ► and it's not like sand has to sneak in through a tiny opening.

00:09:22 ◼   ► The openings are pretty big.

00:09:23 ◼   ► So still you should really avoid getting sand

00:09:26 ◼   ► in your keyboard.

00:09:27 ◼   ► The good news, as we mentioned before,

00:09:28 ◼   ► is if you do get a piece of sand under your keyboard,

00:09:31 ◼   ► there's a chance that they can pop off the key cap,

00:09:32 ◼   ► take out the sand and put the key cap back on

00:09:34 ◼   ► and you don't have to have all top gear replacement.

00:09:36 ◼   ► But like I said, if you're gonna make some kind of barrier

00:09:39 ◼   ► to keep grit out and there are five giant holes

00:09:42 ◼   ► under every key where grit can get in,

00:09:44 ◼   ► it's only so much you can do.

00:09:45 ◼   ► So probably improved, probably better,

00:09:47 ◼   ► but maybe don't use your MacBook Pro on the beach.

00:09:51 ◼   ► - You know, it's too related to know

00:09:52 ◼   ► whether in practice this will actually reduce

00:09:55 ◼   ► the occurrence of the keyboard problems.

00:09:56 ◼   ► But it is kind of nice that they did this test.

00:09:58 ◼   ► It is kind of nice to know we aren't totally

00:10:01 ◼   ► out of the woods on this.

00:10:02 ◼   ► Like if you have one of these,

00:10:03 ◼   ► you still can't like eat a muffin on top of it

00:10:05 ◼   ► without running a pretty big risk there.

00:10:07 ◼   ► You're simply, you have a reduced risk.

00:10:11 ◼   ► But in a way I also think like,

00:10:12 ◼   ► you know Casey, you've had with your MacBook One,

00:10:14 ◼   ► you've had a lot of compressed air adventures,

00:10:17 ◼   ► as Apple recommends, like trying to get your one piece,

00:10:21 ◼   ► trying to move around the piece of dust

00:10:22 ◼   ► that gets under the keys so it goes to a different key.

00:10:24 ◼   ► I think with this, that's probably not gonna work.

00:10:26 ◼   ► With this, I think the dynamic here will be

00:10:30 ◼   ► that it is harder for things to get in,

00:10:32 ◼   ► but it's even harder for anything to ever get back out.

00:10:35 ◼   ► So if something does get in, which is, as John said,

00:10:38 ◼   ► which is still possible, it's just less likely,

00:10:40 ◼   ► but if something does get in, I bet you're done.

00:10:43 ◼   ► I bet that's it, you have no recourse.

00:10:45 ◼   ► It's nice that they're making this better.

00:10:47 ◼   ► I hope in practice it's gonna be so rare,

00:10:50 ◼   ► 'cause you know, I've fixed it,

00:10:51 ◼   ► they had to try pretty hard to finally get something in it.

00:10:54 ◼   ► Like they had to really blast it with sand, I think.

00:10:57 ◼   ► So it was, I hope that in practice

00:10:59 ◼   ► it's not gonna be a problem,

00:11:01 ◼   ► but they still have a keyboard that if a piece of dust

00:11:05 ◼   ► gets in the wrong spot, it will kill it.

00:11:07 ◼   ► They just have made it less likely for that to happen.

00:11:09 ◼   ► So progress, but not probably completely out of the woods.

00:11:13 ◼   ► - I put a thing in, the thing I put in the show

00:11:16 ◼   ► and it says that they didn't actually have to try hard

00:11:17 ◼   ► with the sand, like this is a quote from another thing.

00:11:19 ◼   ► We sprinkle a pinch of sand over the keyboard,

00:11:21 ◼   ► type on the keys for a minute,

00:11:23 ◼   ► and we don't even have to lift the key caps

00:11:24 ◼   ► to realize that something is wrong.

00:11:26 ◼   ► A few keys have seized up.

00:11:27 ◼   ► So it wasn't actually that hard with sand.

00:11:29 ◼   ► Oh, that's not good.

00:11:31 ◼   ► - I don't know what to think about this.

00:11:33 ◼   ► I do think that listening to Marco,

00:11:35 ◼   ► I've come up with the official unofficial name

00:11:37 ◼   ► for this keyboard, which is the Hotel California keyboard

00:11:40 ◼   ► where you can enter but you can never leave.

00:11:41 ◼   ► (laughing)

00:11:42 ◼   ► - It's a Roach Motel.

00:11:43 ◼   ► I was gonna make the Roach Motel joke,

00:11:44 ◼   ► but you tried to make the Hotel California.

00:11:47 ◼   ► I think the Roach Motel is better,

00:11:48 ◼   ► but that may be before your time.

00:11:49 ◼   ► - No.

00:11:50 ◼   ► First of all, it is slightly before our time.

00:11:52 ◼   ► But second of all, come on.

00:11:54 ◼   ► Hotel California is so good.

00:11:55 ◼   ► Anyway. - That's fantastic.

00:11:57 ◼   ► - All in the chat room has an interesting question,

00:12:00 ◼   ► which I think we should cover real quick before we move on.

00:12:02 ◼   ► Do you guys think this is an acceptable solution?

00:12:04 ◼   ► And taking that word specifically, acceptable,

00:12:08 ◼   ► I think this is acceptable.

00:12:10 ◼   ► I'm not in love with it.

00:12:11 ◼   ► I think that everyone is still worried

00:12:14 ◼   ► about keyboard reliability,

00:12:15 ◼   ► in part because we haven't really heard

00:12:17 ◼   ► what the reliability is like.

00:12:18 ◼   ► You know, these machines have only been in people's hands

00:12:19 ◼   ► for what, a week, two weeks?

00:12:21 ◼   ► But it's not, well, I feel like it's acceptable.

00:12:26 ◼   ► It is enough to make me think it's worth trying

00:12:29 ◼   ► these new keyboards if you've had a problem in the past.

00:12:32 ◼   ► But it certainly seems like the Achilles heel status

00:12:36 ◼   ► of these keyboards hasn't changed.

00:12:38 ◼   ► And that's a problem.

00:12:39 ◼   ► And it's also tough because I've been going back and forth

00:12:42 ◼   ► between my MacBook Adorable and my full Darth Vader,

00:12:47 ◼   ► you know, 104 whatever key wireless keyboard.

00:12:52 ◼   ► And although I do love this magic keyboard

00:12:54 ◼   ► that Apple makes, and I've said that many, many,

00:12:55 ◼   ► many times on the show,

00:12:57 ◼   ► I still like the feel of my Adorable a little bit more,

00:12:59 ◼   ► which is weird 'cause I never thought I'd like anything

00:13:02 ◼   ► more than this magic keyboard.

00:13:04 ◼   ► So it's acceptable, but I feel like time will tell

00:13:09 ◼   ► to see if it's really enough.

00:13:12 ◼   ► I don't know, Marco, obviously you have thoughts on this,

00:13:13 ◼   ► so let's start with Jon,

00:13:14 ◼   ► and then Marco, maybe you can bring us home afterwards.

00:13:17 ◼   ► - So what was the question again?

00:13:18 ◼   ► Is it acceptable?

00:13:19 ◼   ► - Yeah, so do you guys think this is an acceptable solution?

00:13:23 ◼   ► - I don't really know what acceptable means.

00:13:25 ◼   ► I think the question for me is would I recommend

00:13:28 ◼   ► these laptops to people, right?

00:13:30 ◼   ► Like is, you know, say yeah, go ahead and buy one,

00:13:33 ◼   ► they're good, like it's okay,

00:13:35 ◼   ► they fixed the keyboard things.

00:13:36 ◼   ► And the answer to that is the jury's still out

00:13:38 ◼   ► as far as I'm concerned.

00:13:39 ◼   ► Like I don't know yet, it's too early.

00:13:40 ◼   ► It would make me more comfortable, surprisingly,

00:13:44 ◼   ► like I know they're not gonna do this,

00:13:45 ◼   ► but like if these were included in the repair extension,

00:13:48 ◼   ► just because you have this keyboard

00:13:49 ◼   ► and it's supposed to be better about reliability,

00:13:51 ◼   ► and it probably is, but if it isn't,

00:13:54 ◼   ► I don't wanna pay for a new top case after a year, right?

00:13:56 ◼   ► I want the, it's almost like the 2017 and '16 and '15 models

00:14:01 ◼   ► have the assurance that like, well, you get one,

00:14:02 ◼   ► and if you do have a problem, don't worry,

00:14:04 ◼   ► they'll cover it for four years, right?

00:14:05 ◼   ► And that's probably how long your laptop will last anyway

00:14:07 ◼   ► before the battery's hosed.

00:14:09 ◼   ► Whereas this one, it's like, oh, that problem's probably gone

00:14:12 ◼   ► so you're probably fine.

00:14:13 ◼   ► But if it turns out the problem isn't gone,

00:14:14 ◼   ► and again, with the sprinkling of sand,

00:14:16 ◼   ► we see that it actually isn't entirely gone

00:14:18 ◼   ► and they're under the worst possible conditions.

00:14:20 ◼   ► Oh, someone spilled some sand on your laptop, right?

00:14:22 ◼   ► And you typed on it for two seconds.

00:14:25 ◼   ► And I don't wanna be stuck having to replace it.

00:14:27 ◼   ► So it's an improvement, it's better.

00:14:32 ◼   ► I particularly like the idea that you can replace a keycap

00:14:34 ◼   ► without replacing the whole keyboard.

00:14:35 ◼   ► That I think is a big possibility of like,

00:14:37 ◼   ► you won't have to pay for a whole new top case.

00:14:39 ◼   ► You'll just bring it in, yeah, it's annoying,

00:14:40 ◼   ► but you got some sand under your keyboard,

00:14:41 ◼   ► they'll pop it, I'll give you a new key cap,

00:14:43 ◼   ► or you break a key cap by dropping something heavy

00:14:45 ◼   ► in your keyboard and they can fix it.

00:14:47 ◼   ► But I feel like it's still not up to the bar of the 2015,

00:14:51 ◼   ► which is basically a keyboard that,

00:14:53 ◼   ► I'm sure people had problems with it

00:14:54 ◼   ► and people broke key caps off and stuff like that,

00:14:56 ◼   ► but it didn't raise to the level of,

00:14:59 ◼   ► some high percentage of people you know

00:15:01 ◼   ► with 2015 MacBook Pros had one or more keyboard problems.

00:15:05 ◼   ► Like, no one ever talked about the keyboard.

00:15:06 ◼   ► It was like, the keyboard was the keyboard and whatever.

00:15:09 ◼   ► I don't think this one has crossed that bar yet.

00:15:11 ◼   ► I feel like it's just the, as we said in the past,

00:15:13 ◼   ► Rose, it's just the accumulation of Band-Aids.

00:15:16 ◼   ► This is like the maximum number of Band-Aids

00:15:18 ◼   ► all piled on top of this little wound,

00:15:20 ◼   ► but the scab is kinda still there.

00:15:22 ◼   ► (laughing)

00:15:23 ◼   ► Well said.

00:15:25 ◼   ► - I think the jury is still out on this.

00:15:27 ◼   ► I really, I can't wait for, you know,

00:15:29 ◼   ► I did order one of these machines,

00:15:31 ◼   ► we can talk about that later if you want.

00:15:32 ◼   ► Tell you one thing I did when I ordered it,

00:15:34 ◼   ► I also ordered one of those silicone covers

00:15:35 ◼   ► for the keyboard for like 10 bucks on Amazon.

00:15:37 ◼   ► - Oh no.

00:15:39 ◼   ► - Because I just don't know.

00:15:40 ◼   ► And here's the thing, like, you know,

00:15:42 ◼   ► it's basically a $10 insurance policy.

00:15:45 ◼   ► Like, I can spend $10 on this stupid silicone cover.

00:15:48 ◼   ► This problem might have been solved, but we don't know yet.

00:15:51 ◼   ► It sure seems like it hasn't really been fully solved,

00:15:54 ◼   ► but maybe it's been improved enough that we'll stop caring.

00:15:57 ◼   ► The correct solution to this problem

00:15:58 ◼   ► is to design a keyboard that either

00:16:00 ◼   ► dust can get in and out freely,

00:16:02 ◼   ► so that dust getting in for a second isn't a big deal

00:16:05 ◼   ► 'cause it can just come right back out again,

00:16:07 ◼   ► like the old ones, or design a keyboard

00:16:10 ◼   ► that dust can't get into at all.

00:16:12 ◼   ► You know, like, if you look at the design of

00:16:14 ◼   ► like the smart keyboard for iPads,

00:16:16 ◼   ► where it's basically like a cloth membrane

00:16:18 ◼   ► over the whole thing covering the key,

00:16:20 ◼   ► like one big thing over the whole keys,

00:16:22 ◼   ► so that even like, you know,

00:16:23 ◼   ► you can basically spill water on it,

00:16:24 ◼   ► and you know, it doesn't really go in anywhere.

00:16:26 ◼   ► I feel like those two extremes are the way to go here.

00:16:28 ◼   ► Either make it so dust can come in and out freely,

00:16:30 ◼   ► or make it so dust can't get in at all.

00:16:32 ◼   ► And they haven't done that yet.

00:16:34 ◼   ► As John said, this is still a bandaid

00:16:36 ◼   ► on this keyboard design.

00:16:37 ◼   ► I still really want to know what the next design is.

00:16:41 ◼   ► Like, what's the next major keyboard here?

00:16:43 ◼   ► This keyboard, they took a lot of risks,

00:16:44 ◼   ► they did a lot of crazy things,

00:16:46 ◼   ► they optimized for factors I don't give two craps about.

00:16:49 ◼   ► Like, great, my keyboard is precise and stable.

00:16:51 ◼   ► Two things I have never once pushed the key

00:16:54 ◼   ► on any other keyboard and said,

00:16:54 ◼   ► "This is too imprecise and unstable."

00:16:58 ◼   ► That has never, like, I've never had

00:16:59 ◼   ► that opinion of anything.

00:17:00 ◼   ► So, I wanna see where they go next.

00:17:02 ◼   ► - So would you say it's acceptable?

00:17:04 ◼   ► I'm not trying to snark, I'm genuinely asking.

00:17:05 ◼   ► Like, I feel like you're kind of on both sides of that.

00:17:08 ◼   ► So, yes, no, acceptable?

00:17:10 ◼   ► - With the caveat that it's still way too early to tell,

00:17:14 ◼   ► just by the way it seems, by the way it looks,

00:17:17 ◼   ► like, based on iFixitate and stuff, and their testing,

00:17:20 ◼   ► I give it like a B.

00:17:22 ◼   ► Like, this seems like it's probably a decent solution,

00:17:25 ◼   ► but it's not the best solution.

00:17:26 ◼   ► - Fair enough.

00:17:27 ◼   ► - By the way, with your little silicone cover thing,

00:17:29 ◼   ► I'm wondering if you have just increased

00:17:31 ◼   ► the amount of finger grease

00:17:32 ◼   ► that's gonna appear on your screen.

00:17:33 ◼   ► - We're already past that point.

00:17:35 ◼   ► (laughing)

00:17:35 ◼   ► - You're all on board with it.

00:17:37 ◼   ► - I've fought that for years.

00:17:38 ◼   ► Like, when I had my very first PowerBook, my first laptop,

00:17:42 ◼   ► I was very careful with how I arranged the backpack.

00:17:44 ◼   ► I made like a whole custom padded compartment

00:17:46 ◼   ► out of black felt, so it was never squished against.

00:17:50 ◼   ► So, I told you before my theory about what causes

00:17:53 ◼   ► the keyboard imprint on the screens.

00:17:55 ◼   ► When it's closed flat, like on top of a desk,

00:17:57 ◼   ► they don't make contact, but if it's compressed

00:18:00 ◼   ► slightly in a bag, I think the screen lid flexes slightly

00:18:02 ◼   ► inwards, and that compresses the keyboard against the screen.

00:18:05 ◼   ► So, I think that's where, so it really depends on like

00:18:07 ◼   ► what kind of bag you're putting it in,

00:18:08 ◼   ► whether anything is on top of it or pushing against it.

00:18:11 ◼   ► So, I used to care a lot about that.

00:18:13 ◼   ► In the last few years, I just started losing that fight

00:18:15 ◼   ► more and more as I got like different backpacks

00:18:17 ◼   ► and different bags and everything,

00:18:18 ◼   ► and now I just don't care anymore.

00:18:20 ◼   ► - Yeah, the other thing I noticed

00:18:21 ◼   ► of looking at laptops around work is,

00:18:23 ◼   ► I think it was the 2015s and maybe the couple years before,

00:18:26 ◼   ► like the anti-glare coating would start coming off.

00:18:27 ◼   ► Have you seen that?

00:18:28 ◼   ► - Oh yeah, the delamination, yes.

00:18:30 ◼   ► - Yeah, a bunch of the ones at work have it,

00:18:32 ◼   ► but surprisingly, it's one of those problems

00:18:33 ◼   ► that if you don't, like people don't notice it

00:18:35 ◼   ► until I point it out to them and say,

00:18:37 ◼   ► "You realize your anti-glare thing is,"

00:18:39 ◼   ► I don't know if it's the anti-glare or delamination

00:18:40 ◼   ► or whatever, but you can tell.

00:18:42 ◼   ► It looks kind of like there's like grease on the screen,

00:18:44 ◼   ► but it's not grease.

00:18:45 ◼   ► It's that you can't clean it off,

00:18:46 ◼   ► and it tends to happen around the edges or whatever.

00:18:49 ◼   ► But people are like, "Eh, they shrug,"

00:18:50 ◼   ► and one of the reasons they shrug is the rest of their screen

00:18:53 ◼   ► is covered with finger grease from the keyboard

00:18:55 ◼   ► and directly from their fingers.

00:18:56 ◼   ► So, their entire screens look like a mess anyway,

00:18:58 ◼   ► and they're just used to it.

00:19:00 ◼   ► - All right, Marco, you have something to smile about though

00:19:02 ◼   ► because a little birdie has told us

00:19:04 ◼   ► something very, very interesting.

00:19:05 ◼   ► Can you tell us about this?

00:19:07 ◼   ► - Yes, so as I've ranted a lot about before,

00:19:09 ◼   ► the USB-C ecosystem is just not really panning out,

00:19:13 ◼   ► and for me, one of the biggest problems is that,

00:19:15 ◼   ► yes, you can convert everything.

00:19:17 ◼   ► If you go out and buy all new cables,

00:19:19 ◼   ► you can get all cables that have USB-C on one end

00:19:22 ◼   ► and whatever you need on the other end for all your stuff.

00:19:24 ◼   ► So, you can convert your cables pretty easily,

00:19:26 ◼   ► but you can't convert hubs 'cause there doesn't seem

00:19:28 ◼   ► to be any kind of actual USB-C hub

00:19:31 ◼   ► that just takes one USB-C port

00:19:33 ◼   ► and makes four other USB-C ports out of it.

00:19:36 ◼   ► And there's lots of complexity to the USB-C spec

00:19:40 ◼   ► with things like Thunderbolt and DisplayPort and power.

00:19:43 ◼   ► It's very, very complex, so making such a thing

00:19:45 ◼   ► is apparently very difficult,

00:19:46 ◼   ► which is why they don't really seem to exist.

00:19:48 ◼   ► There are a few products in the market,

00:19:50 ◼   ► a couple of those $300 docks

00:19:52 ◼   ► that have two output ports maybe,

00:19:55 ◼   ► but that's usually about your only option.

00:19:58 ◼   ► And I think USB-C requires hubs

00:20:01 ◼   ► that just take one port to multiple C ports

00:20:03 ◼   ► for it to really finally take over.

00:20:07 ◼   ► And we've heard from an anonymous person

00:20:09 ◼   ► that Intel is finally releasing a hub chip

00:20:12 ◼   ► for USB-C next year, so this would be a chip

00:20:14 ◼   ► that could be put into hubs.

00:20:16 ◼   ► We may not see products based on it until 2020.

00:20:19 ◼   ► I don't know enough about this kind of stuff

00:20:22 ◼   ► to say whether this is it, whether this is all we need,

00:20:24 ◼   ► whether this is even true or plausible,

00:20:27 ◼   ► but typically the chips that are in USB hubs

00:20:31 ◼   ► and Thunderbolt hubs and everything,

00:20:32 ◼   ► they tend to be made by very few manufacturers.

00:20:35 ◼   ► That's why you have a million different brand names

00:20:37 ◼   ► on Amazon for what appears to be about the same hub

00:20:41 ◼   ► or the same kind of thing.

00:20:42 ◼   ► It's 'cause there's one company

00:20:43 ◼   ► that makes those chips in there,

00:20:44 ◼   ► and usually it's somebody like Intel or Vya

00:20:47 ◼   ► or something like that, like one of the big chip makers.

00:20:50 ◼   ► And so apparently, if the chip to make USB-C hubs

00:20:54 ◼   ► hasn't really existed yet,

00:20:55 ◼   ► that would certainly explain the problem that we have.

00:20:57 ◼   ► And if Intel's making one in the next couple of years,

00:20:59 ◼   ► that's probably good news,

00:21:01 ◼   ► although I kinda wish it was out now,

00:21:02 ◼   ► and I kinda wonder, we've had USB-C ports

00:21:05 ◼   ► on MacBook since 2015, what are we waiting for?

00:21:07 ◼   ► - Yeah, yeah, yeah.

00:21:10 ◼   ► Do we wanna talk about the history lesson

00:21:12 ◼   ► that was also associated with this?

00:21:14 ◼   ► - Yeah, I thought this was a good summary,

00:21:15 ◼   ► 'cause the larger issue,

00:21:17 ◼   ► I mean, it's not really related to throttling,

00:21:19 ◼   ► but what's the deal with Intel CPUs these days?

00:21:21 ◼   ► There's the bigger picture of Moore's Law

00:21:24 ◼   ► and everything we've talked about,

00:21:25 ◼   ► but Intel CPUs haven't been getting that much faster

00:21:28 ◼   ► or that much better in recent years for a variety of reasons,

00:21:31 ◼   ► and one of the reasons is that they haven't been able

00:21:34 ◼   ► to get their new processor, 10-nanometer process online,

00:21:37 ◼   ► so they've just been making part after part after part

00:21:39 ◼   ► out of 14 nanometers, and there's only so much you can do

00:21:42 ◼   ► by rearranging the deck chairs.

00:21:44 ◼   ► I think they're adding more cores,

00:21:46 ◼   ► I think they're adding eight cores in the same CPUs

00:21:48 ◼   ► that we have six cores in now,

00:21:49 ◼   ► 'cause what else can they do,

00:21:51 ◼   ► because they don't have a smaller process?

00:21:53 ◼   ► And this sort of semi-brief history,

00:21:56 ◼   ► I think, is a good rundown of what the deal was

00:21:59 ◼   ► and why Apple might find itself in a difficult situation

00:22:03 ◼   ► designing its laptops around these chips.

00:22:04 ◼   ► So 2016, when the Skylake MacBook Rosie released,

00:22:08 ◼   ► Cannon Lake was supposed to be around the corner in 2017

00:22:11 ◼   ► with an LPDDR4 controller supporting 32 gigs, right?

00:22:14 ◼   ► So in 2016, they're like, "Oh, no problem.

00:22:16 ◼   ► "Next year, 2017, we'll be able to support 32 gigs

00:22:20 ◼   ► "with low power," that didn't happen, right?

00:22:23 ◼   ► And then Cannon Lake got pushed to 2018,

00:22:24 ◼   ► Kaby Lake was inserted,

00:22:26 ◼   ► which is just Skylake with HEVC and HDCP 2.2,

00:22:29 ◼   ► then Coffee Lake got added in 2018,

00:22:31 ◼   ► which is just Kaby Lake with more cores,

00:22:32 ◼   ► and Cannon Lake was canceled,

00:22:34 ◼   ► although this person says the same and honest person,

00:22:37 ◼   ► "Yes, really, although Intel hasn't fessed up yet.

00:22:39 ◼   ► "Now, Ice Lake is teetering on the edge of failure

00:22:41 ◼   ► "and Intel is prepping yet more

00:22:42 ◼   ► "in 14-nanometer parts for 2019,

00:22:44 ◼   ► "and here we still are, 14-nanometer

00:22:46 ◼   ► "with Skylake memory controller."

00:22:48 ◼   ► And the new 15-inch MacBook Pro uses

00:22:50 ◼   ► the built-in memory controller,

00:22:51 ◼   ► which works just fine with DDR4,

00:22:52 ◼   ► which is what we're seeing right now.

00:22:54 ◼   ► Apple has made plenty of own goals

00:22:56 ◼   ► in soccer parlance, football parlance,

00:22:59 ◼   ► in the MacBook Pro, primarily due to design hubris,

00:23:02 ◼   ► sorry, but trying to plan products

00:23:04 ◼   ► on Intel's roadmap has become masochistic.

00:23:07 ◼   ► Like, what do you do if you're gonna plan,

00:23:09 ◼   ► like, this is the new generation of MacBook Pro,

00:23:11 ◼   ► and we're gonna design three models for the next three years

00:23:13 ◼   ► and Intel tells us this is their roadmap,

00:23:15 ◼   ► and then they just miss that roadmap entirely

00:23:17 ◼   ► and just keep releasing 14-nanometer parts.

00:23:20 ◼   ► It's not a great situation to be in.

00:23:21 ◼   ► And you can argue that the correct choice

00:23:23 ◼   ► is to scrap everything and redesign a new case

00:23:27 ◼   ► to handle 14-nanometer things,

00:23:28 ◼   ► but the lead times might just not have been

00:23:31 ◼   ► good enough to do that, so.

00:23:32 ◼   ► We're all in the same boat here.

00:23:35 ◼   ► It's not like someone else has

00:23:36 ◼   ► secret better Intel CPUs than Apple.

00:23:38 ◼   ► Apple gets the same ones as everyone else,

00:23:41 ◼   ► and Intel really needs to get its process online,

00:23:43 ◼   ► and we have a topic about that we might get to later,

00:23:46 ◼   ► but it's yet another reason why people continue

00:23:50 ◼   ► to entertain fantasies of ARM-based Macs,

00:23:53 ◼   ► because one thing that has happened in recent years

00:23:55 ◼   ► is Apple's ARM CPUs are getting better all the time,

00:23:58 ◼   ► and they're really, really good.

00:24:00 ◼   ► And, like, every time I go to a Geekbench or whatever

00:24:02 ◼   ► and look at the scores for, like, the iPhone X

00:24:05 ◼   ► versus Apple's MacBooks, like, Apple's 2018 MacBooks,

00:24:09 ◼   ► they're not that far apart.

00:24:10 ◼   ► And I think about exactly how many fans there are

00:24:14 ◼   ► in an iPhone X, right?

00:24:16 ◼   ► How big it is, how big the battery is,

00:24:19 ◼   ► and then I look at the Geekbench scores,

00:24:21 ◼   ► like, ARM Macs are looking, you know,

00:24:25 ◼   ► I know it's not Apple's to Apple's,

00:24:26 ◼   ► I know it's not the same, I know it doesn't have

00:24:28 ◼   ► 32 gigs of RAM in the phone, like, I understand,

00:24:30 ◼   ► I know it's not a giant monster GPU, discrete GPU in there,

00:24:33 ◼   ► but, boy, Intel is really making ARM Macs

00:24:37 ◼   ► look more and more attractive every year.

00:24:39 ◼   ► - We are sponsored this week by the Aftershokz

00:24:42 ◼   ► bone conduction headphones.

00:24:43 ◼   ► Go to ATP.Aftershokz.com to learn more,

00:24:47 ◼   ► and check out the weightless wireless Trex Air.

00:24:51 ◼   ► I love bone conduction headphones,

00:24:53 ◼   ► especially in the summertime.

00:24:55 ◼   ► Here's why.

00:24:56 ◼   ► So any other headphone you have,

00:24:57 ◼   ► it's gonna be something that either goes on top of your ear

00:25:00 ◼   ► or has a big cup around your ear,

00:25:02 ◼   ► or something that goes into your ear.

00:25:03 ◼   ► And I have problems with both of those things.

00:25:05 ◼   ► The in-ear ones just hurt me,

00:25:07 ◼   ► I just don't have the ears for them.

00:25:09 ◼   ► The ones that go around or on your ears

00:25:11 ◼   ► are too hot for the summertime.

00:25:13 ◼   ► And when you're walking around outside,

00:25:15 ◼   ► it's hard to hear things like cars driving by,

00:25:17 ◼   ► or if someone's trying to talk to you

00:25:19 ◼   ► when you're wearing big headphones.

00:25:21 ◼   ► Bone conduction headphones solve these problems.

00:25:23 ◼   ► First of all, there's nothing in or on or around your ears.

00:25:27 ◼   ► Instead, they have these little transducers,

00:25:29 ◼   ► they're very small,

00:25:30 ◼   ► they have like a little wraparound neckband,

00:25:32 ◼   ► and these little transducers that rest

00:25:34 ◼   ► kind of in front of your ears on your cheekbones.

00:25:37 ◼   ► And they send little vibrations

00:25:38 ◼   ► that your eardrums pick up as sound,

00:25:40 ◼   ► but the world doesn't hear it that way.

00:25:42 ◼   ► And in addition, because nothing is blocking your ears,

00:25:45 ◼   ► you can still hear the world around you.

00:25:47 ◼   ► So it's very practical when you're doing outdoor activities,

00:25:49 ◼   ► walking, cycling, even just doing stuff around the house,

00:25:52 ◼   ► if you wanna hear if someone knocks on the door,

00:25:54 ◼   ► or if someone else in your house is calling your name,

00:25:56 ◼   ► they're incredibly nice.

00:25:57 ◼   ► And the Aftershokz Trekz Air are just great headphones too.

00:26:01 ◼   ► They have their Bluetooth, of course, so they're wireless.

00:26:03 ◼   ► They're pretty much weightless.

00:26:05 ◼   ► Like they say they weigh like an ounce,

00:26:06 ◼   ► but honestly, the weight is nothing,

00:26:08 ◼   ► like you feel it like nothing.

00:26:10 ◼   ► Good range, good battery life,

00:26:12 ◼   ► there's a great warranty if you need it.

00:26:13 ◼   ► They're water resistant, IP55 certified,

00:26:15 ◼   ► so if you get caught in the rain

00:26:17 ◼   ► or if you're sweating all over them, no big deal.

00:26:18 ◼   ► The Trekz Air are the only headphones I wear in the summer.

00:26:21 ◼   ► I like them so much that this summer I'm testing a lot

00:26:23 ◼   ► between my watch app and my phone app,

00:26:25 ◼   ► and I wanted two different headphones

00:26:27 ◼   ► so I could just grab one that's paired to each device

00:26:29 ◼   ► without having to re-pair them.

00:26:30 ◼   ► And I bought a second Trekz Air,

00:26:32 ◼   ► that's how much I like them.

00:26:33 ◼   ► I like them so much I bought a second one.

00:26:35 ◼   ► I highly recommend the Aftershokz Trekz Air.

00:26:37 ◼   ► Check it out today, ATP.Aftershokz.com.

00:26:42 ◼   ► And you can use code ATPBUNDLE

00:26:45 ◼   ► to get $55 off of a Trekz Air bundle.

00:26:49 ◼   ► So once again, ATP.Aftershokz.com,

00:26:52 ◼   ► use code ATPBUNDLE to get $55 off a Trekz Air bundle.

00:26:56 ◼   ► Thank you so much to Aftershokz

00:26:59 ◼   ► for just making awesome headphones

00:27:00 ◼   ► and for sponsoring our show.

00:27:02 ◼   ► (upbeat music)

00:27:06 ◼   ► In one of my big rants last episode

00:27:08 ◼   ► on this throttling thing,

00:27:09 ◼   ► you know, I've honestly, I've mostly tried to stay out

00:27:12 ◼   ► of the i9 throttling thing

00:27:14 ◼   ► because I just didn't know much about it

00:27:15 ◼   ► and it was a configuration I didn't have

00:27:17 ◼   ► and I couldn't get access to

00:27:18 ◼   ► and it just seemed like a, you know,

00:27:20 ◼   ► it was mostly playing out on YouTube and Reddit.

00:27:22 ◼   ► But one of the big, you know, lines I drew in the sand

00:27:25 ◼   ► was a CPU should never run below its base clock.

00:27:28 ◼   ► Like it should never have to do that in normal operation.

00:27:30 ◼   ► It should be able to sustain its base clock indefinitely

00:27:33 ◼   ► and if it can't, there's something seriously wrong

00:27:34 ◼   ► with the thermal design around it.

00:27:37 ◼   ► And it turns out that was not a complete picture

00:27:40 ◼   ► of how things work anymore.

00:27:41 ◼   ► So we've learned a lot since the last episode

00:27:45 ◼   ► and so I wanted to go over a few things.

00:27:46 ◼   ► First of all, we pointed out last episode

00:27:50 ◼   ► the 800 megahertz clock,

00:27:52 ◼   ► like where the CPU would drop down to 800 megahertz.

00:27:54 ◼   ► Now in the throttling problem

00:27:57 ◼   ► where it would cause like the big sawtooth graph basically

00:27:59 ◼   ► of like constant spikes up and down,

00:28:01 ◼   ► that actually was the real problem.

00:28:03 ◼   ► But sometimes the CPU will drop to 800 megahertz

00:28:05 ◼   ► because it's idle.

00:28:07 ◼   ► 800 megahertz is simply the lowest clock

00:28:09 ◼   ► the CPU will typically use in just power save mode.

00:28:13 ◼   ► It is also used in very extreme cases

00:28:15 ◼   ► for throttling for thermal reasons.

00:28:17 ◼   ► But if your CPU hits 800 megahertz,

00:28:20 ◼   ► that might just be because it wasn't doing that much

00:28:22 ◼   ► and it clocks itself down to save power.

00:28:24 ◼   ► Secondarily, so going back to a CPU must always maintain

00:28:27 ◼   ► its base clock no matter what.

00:28:29 ◼   ► Well, it turns out that when Intel's recent AVX instructions

00:28:34 ◼   ► are being used, their processors reduced their base clock

00:28:37 ◼   ► because AVX instructions are apparently so like heavy duty

00:28:41 ◼   ► and complicated to execute that the processors are designed

00:28:44 ◼   ► to actually reduce their base clock

00:28:46 ◼   ► while running AVX instructions.

00:28:48 ◼   ► And I honestly don't know that much about AVX,

00:28:50 ◼   ► but it's kind of like the continuation of like MMX,

00:28:52 ◼   ► like back in the old days, MMX,

00:28:54 ◼   ► and then what were the other ones?

00:28:56 ◼   ► There were like a bunch of SSE

00:28:57 ◼   ► and all those like extra instructions,

00:29:00 ◼   ► like big vector instructions to process lots of data at once.

00:29:03 ◼   ► So AVX is like the recent version of that kind of idea.

00:29:06 ◼   ► - Doesn't even mention Altavec, but I'll do it

00:29:08 ◼   ► because I have to.

00:29:09 ◼   ► (laughing)

00:29:10 ◼   ► - Yeah, your Altavec really took over the world there, John.

00:29:13 ◼   ► - Altavec was awesome.

00:29:14 ◼   ► So much better than MMX.

00:29:17 ◼   ► - So among the valid reasons that a CPU might operate

00:29:20 ◼   ► below base clock are either it's saving power

00:29:23 ◼   ► 'cause it's idle or it's using AVX instructions,

00:29:26 ◼   ► even if it's at full load.

00:29:28 ◼   ► And then finally, we talk a lot about the CPU's TDP.

00:29:33 ◼   ► I mean, this is a figure in watts,

00:29:34 ◼   ► which is the maximum power draw or heat generation,

00:29:37 ◼   ► which is kind of the same thing, that a CPU will issue.

00:29:40 ◼   ► And it turns out my view of TDP

00:29:42 ◼   ► and my definition of TDP was outdated.

00:29:44 ◼   ► The 15 inch is usually like a 45 watt TDP on those processes,

00:29:48 ◼   ► which I thought would mean that at peak load,

00:29:50 ◼   ► no matter what, it would not exceed 45 watts of heat output

00:29:53 ◼   ► and therefore power usage.

00:29:55 ◼   ► It turns out that's not the case.

00:29:56 ◼   ► It turns out TDP is the expected maximum average power draw

00:30:02 ◼   ► with all cores operating at the base frequency.

00:30:06 ◼   ► If turbo boost is being used,

00:30:08 ◼   ► it can exceed TDP all the time.

00:30:11 ◼   ► Like it can exceed TDP constantly.

00:30:13 ◼   ► It can do it whenever it wants.

00:30:15 ◼   ► TDP is just when turbo boost is not being used,

00:30:17 ◼   ► when it's the base frequency and all cores are maxed out,

00:30:21 ◼   ► that's roughly what you can expect

00:30:23 ◼   ► the maximum power output to be.

00:30:25 ◼   ► Now, we've been told by people who know more

00:30:27 ◼   ► about this than us, thank goodness they exist,

00:30:29 ◼   ► that before Coffee Lake, which is this generation,

00:30:33 ◼   ► and before Kaby Lake, which is the 2017 generation,

00:30:36 ◼   ► there actually used to be a lot more headroom in practice

00:30:39 ◼   ► before the CPUs would really hit TDP.

00:30:42 ◼   ► The TDP figures were apparently very conservative.

00:30:44 ◼   ► CPUs would not normally surpass them much under load,

00:30:47 ◼   ► even using turbo.

00:30:48 ◼   ► But as we mentioned a few minutes ago,

00:30:52 ◼   ► Intel not being able to get to their 10 nanometer process,

00:30:55 ◼   ► get those chips out the door,

00:30:56 ◼   ► has had to cram more and more transistors and chips

00:30:59 ◼   ► into this old 14 nanometer process,

00:31:01 ◼   ► which means that they're cramming more cores and more stuff

00:31:05 ◼   ► into the chips without shrinking their process size,

00:31:09 ◼   ► which means they're just gonna use more power.

00:31:11 ◼   ► Simple as that.

00:31:12 ◼   ► Like the design of each core hasn't changed substantially

00:31:15 ◼   ► to make them more efficient,

00:31:17 ◼   ► and there's no process shrink size,

00:31:19 ◼   ► or process size shrink to give us a cooling benefit there,

00:31:22 ◼   ► so basically these chips are gonna be more and more

00:31:24 ◼   ► complicated without increasing their efficiency much.

00:31:26 ◼   ► So they are using more power.

00:31:29 ◼   ► So we didn't just get two cores for free,

00:31:32 ◼   ► we got two more cores by basically Intel is now

00:31:36 ◼   ► approaching and surpassing the quoted TDP figures

00:31:38 ◼   ► much more often.

00:31:39 ◼   ► So when a CPU's under load,

00:31:41 ◼   ► that 45 watt TDP processor in the 15 inch

00:31:44 ◼   ► can use way more than 45 watts.

00:31:46 ◼   ► It probably won't sustain it for very long,

00:31:49 ◼   ► but it frequently and will easily pull 65 watts.

00:31:53 ◼   ► I think somebody even said it could pull 100 watts

00:31:55 ◼   ► at max load, like if max turbo is on.

00:31:58 ◼   ► This is pulling massive amounts of power,

00:32:00 ◼   ► it's generating massive amounts of heat,

00:32:02 ◼   ► and this is not just the i9,

00:32:04 ◼   ► but all of the Coffee Lake CPUs,

00:32:07 ◼   ► even down to the 13 inch ones.

00:32:08 ◼   ► So basically, everything I knew about both clock speed

00:32:11 ◼   ► and TDP was outdated and wrong.

00:32:13 ◼   ► (laughs)

00:32:15 ◼   ► I say this because it's something I really

00:32:20 ◼   ► am trying to get better at.

00:32:21 ◼   ► We wanna make sure our hot takes are well informed.

00:32:24 ◼   ► And you know, 'cause I,

00:32:26 ◼   ► it's hard, because whenever there's something like this,

00:32:29 ◼   ► whenever there is some kind of big Apple drama,

00:32:32 ◼   ► like when that annoying guy on YouTube broke his iMac Pro

00:32:36 ◼   ► and cleaned up and wouldn't fix it because it was broken,

00:32:39 ◼   ► not because he destroyed it.

00:32:40 ◼   ► (laughs)

00:32:42 ◼   ► Like everyone, whenever anything like that happens,

00:32:45 ◼   ► everyone comes to people like us on Twitter and is like,

00:32:48 ◼   ► hey, what do you think of this?

00:32:49 ◼   ► This is crazy, right?

00:32:50 ◼   ► Oh my God.

00:32:50 ◼   ► And I try to get a really fast opinion out there,

00:32:54 ◼   ► because everybody wants it,

00:32:54 ◼   ► and because I wanna give it,

00:32:55 ◼   ► because I'm excited too.

00:32:56 ◼   ► And if Apple has done something wrong,

00:32:58 ◼   ► I wanna call the Mona 2, you know?

00:33:00 ◼   ► I thought I knew enough about this stuff

00:33:02 ◼   ► to be able to say like,

00:33:04 ◼   ► this should never happen,

00:33:05 ◼   ► XYZ should never happen, and they're wrong.

00:33:08 ◼   ► And I just didn't in this case.

00:33:09 ◼   ► Like there's so much about this

00:33:10 ◼   ► that's way more complicated than I thought it was.

00:33:13 ◼   ► Some of it has changed recently,

00:33:14 ◼   ► some of it I just didn't know.

00:33:16 ◼   ► And so it's just nice to be able to step back for a minute

00:33:19 ◼   ► and try to get a little bit better

00:33:21 ◼   ► at not jumping to so many like

00:33:24 ◼   ► absolute conclusions about this stuff so quickly.

00:33:26 ◼   ► So I apologize for doing that so often.

00:33:28 ◼   ► I'm trying to get better at that.

00:33:29 ◼   ► And this is one example where I think maybe

00:33:32 ◼   ► we could all get a little bit better at that.

00:33:33 ◼   ► - So Intel is kind of in a similar situation

00:33:36 ◼   ► in that they had existing definitions

00:33:37 ◼   ► based on the way their chips used to work.

00:33:39 ◼   ► And as things have changed about chips,

00:33:44 ◼   ► as they always do, as the process shrinks,

00:33:45 ◼   ► the things that used to be steady or reliable before

00:33:50 ◼   ► now become more variable

00:33:51 ◼   ► and different things become important.

00:33:53 ◼   ► And as they've had to spend more and more time

00:33:55 ◼   ► on 14 nanometer, they've had to fudge things

00:33:57 ◼   ► just because you can't get blood from a stone.

00:33:59 ◼   ► Like you were saying, you're gonna put more cores on,

00:34:02 ◼   ► you're not getting that for free.

00:34:03 ◼   ► You gotta put more area and your things get bigger

00:34:05 ◼   ► and they dissipate more heat.

00:34:06 ◼   ► And there's more transistors

00:34:08 ◼   ► and they're the same size as they were before.

00:34:09 ◼   ► So where do you think you're getting any savings from?

00:34:11 ◼   ► And whatever headroom you had, you're eating it.

00:34:12 ◼   ► So a couple of people wrote in to pull the definitions

00:34:16 ◼   ► of like TDP and base clock and stuff

00:34:19 ◼   ► from Intel's documentation.

00:34:20 ◼   ► And what used to be a straightforward definition

00:34:23 ◼   ► is now a little bit more wishy-washy

00:34:25 ◼   ► and has more caveats and more special cases.

00:34:28 ◼   ► And in some cases it's somewhat circular.

00:34:29 ◼   ► Like the base clock is the clock that can maintain

00:34:32 ◼   ► while satisfying TDP.

00:34:34 ◼   ► And TDP is the power used when it's running a base clock.

00:34:37 ◼   ► Like you just go around in circles

00:34:39 ◼   ► with an asterisk, asterisk, except for AVX instructions.

00:34:43 ◼   ► And as we've seen from the fix,

00:34:46 ◼   ► which I think we'll get to in a little bit,

00:34:48 ◼   ► there's the whole profile of how the cooling system works

00:34:50 ◼   ► and how the hardware deals with changes in load and heat

00:34:55 ◼   ► and how that integrates with the cooling system

00:34:59 ◼   ► and what the optimal arrangement is.

00:35:01 ◼   ► And it's not just as simple as

00:35:03 ◼   ► if CPU gets hot make cool now please, right?

00:35:06 ◼   ► There's a bunch of curves and there's settings

00:35:08 ◼   ► and like it's very complicated, right?

00:35:10 ◼   ► So it was so much simpler when there was no turbo boost

00:35:14 ◼   ► and the CPU had a speed and maybe you could overclock it,

00:35:17 ◼   ► but either way just stay to that speed all the time

00:35:19 ◼   ► and there was no speed step and there was no throttling.

00:35:22 ◼   ► And it was just, and we still use those same terms

00:35:25 ◼   ► and those same sort of measurements

00:35:26 ◼   ► and the same conception of how they work.

00:35:29 ◼   ► But in this modern age, like I was saying before

00:35:31 ◼   ► with the thin and light laptop and the over provisioning,

00:35:33 ◼   ► like things are more complicated

00:35:35 ◼   ► and everything is trying to some kind of compromise

00:35:39 ◼   ► because you can't have a laptop like this

00:35:43 ◼   ► with a CPU like this that behaves like one from a decade ago

00:35:48 ◼   ► or it can run at maximum power all day

00:35:50 ◼   ► and that's how you can measure battery life, right?

00:35:53 ◼   ► Everything is a very complicated compromise

00:35:55 ◼   ► involving hardware and software management

00:35:59 ◼   ► and various curves and specific loads,

00:36:03 ◼   ► specific applications, sets of instructions,

00:36:06 ◼   ► just everything is much more complicated than it used to be.

00:36:10 ◼   ► And I can only imagine what it's like in phones.

00:36:12 ◼   ► Like the only reason we don't know

00:36:13 ◼   ► about the world's going inside there

00:36:14 ◼   ► is 'cause it's like just a sealed box to us

00:36:16 ◼   ► and we don't have a history of smartphones

00:36:18 ◼   ► where things were replaceable and we could talk about them

00:36:20 ◼   ► in this sort of a modular way.

00:36:23 ◼   ► But on laptops we do, but most of that is out the window now

00:36:26 ◼   ► so we're grappling with it and so I think is Intel

00:36:30 ◼   ► because they have to sell these things and sell a benefit.

00:36:33 ◼   ► And the last thing you wanna do when selling a benefit

00:36:35 ◼   ► is to explain all the nuances.

00:36:37 ◼   ► Like it's faster under these conditions with these caveats

00:36:41 ◼   ► assuming you did this, that and the other thing.

00:36:43 ◼   ► It's better to just say it's faster, right?

00:36:46 ◼   ► And then when people find, but what about this?

00:36:48 ◼   ► What about that?

00:36:49 ◼   ► And you look at Intel's documentation and it says,

00:36:50 ◼   ► yeah, no, that's in the docs, but it's very nuanced

00:36:52 ◼   ► and we didn't feel like explaining that

00:36:53 ◼   ► to you in the sales brochure.

00:36:55 ◼   ► - Yeah, ultimately what this boils down to is the numbers

00:36:58 ◼   ► that Intel quotes on the CPUs.

00:37:00 ◼   ► Like this is a 2.9 gigahertz processor.

00:37:03 ◼   ► It basically means nothing.

00:37:04 ◼   ► (laughing)

00:37:05 ◼   ► Basically like you can't take anything as absolute anymore.

00:37:08 ◼   ► - It means it can run faster than the 2.71 or the 2.21.

00:37:12 ◼   ► Like the numbers are still comparable for the most part,

00:37:15 ◼   ► although it was a story I just saw fly by

00:37:16 ◼   ► right before we recorded.

00:37:17 ◼   ► - But they aren't even proportional.

00:37:18 ◼   ► Like the 2.9 is not like 40% of whatever it is

00:37:22 ◼   ► faster than the 2.2.

00:37:24 ◼   ► - Yeah, well, because you have to look at the turbo things

00:37:26 ◼   ► and they might not be proportional to the base clocks.

00:37:28 ◼   ► Well, here's one that makes it even worse.

00:37:29 ◼   ► The story I just saw before we recorded is that

00:37:31 ◼   ► in the upcoming generation, which I think they're gonna go

00:37:34 ◼   ► to eight cores instead of six, again on the same process

00:37:36 ◼   ► as far as I could tell.

00:37:37 ◼   ► - Oh, geez.

00:37:38 ◼   ► - They're taking hyper-threading out of the i7s.

00:37:41 ◼   ► So now instead of it being a six core,

00:37:44 ◼   ► like 12 thread, it'll be an eight core, eight thread.

00:37:47 ◼   ► Just to confuse you even more.

00:37:49 ◼   ► And it should still end up being faster than the old ones,

00:37:53 ◼   ► but people are gonna run their CPU meters and be like,

00:37:55 ◼   ► "Hey, there's fewer lines."

00:37:57 ◼   ► This 14 nanometer forever thing is kind of like the,

00:38:03 ◼   ► what the hell was it?

00:38:04 ◼   ► A chat room could help me.

00:38:04 ◼   ► 533 megahertz front side bus on the G4

00:38:07 ◼   ► for like years and years, even when it was obscene.

00:38:11 ◼   ► So to take a half step back, there was a Reddit post

00:38:16 ◼   ► that got really, really popular a few days ago,

00:38:18 ◼   ► where somebody who strikes me as very intelligent

00:38:21 ◼   ► and very well read, and I don't say that sarcastically,

00:38:25 ◼   ► had come up with a theory as to why

00:38:28 ◼   ► this throttling was happening.

00:38:30 ◼   ► And they have since edited their posts

00:38:32 ◼   ► with some new information, which we'll get to in a minute.

00:38:34 ◼   ► But if you scroll down on the link we'll put

00:38:36 ◼   ► in the show notes, you'll see where it says,

00:38:37 ◼   ► "Original post below."

00:38:39 ◼   ► So what they said was they think the problem

00:38:42 ◼   ► is the voltage regulator module being unable

00:38:47 ◼   ► to satisfy the power desires of the i9 CPU.

00:38:49 ◼   ► And they did a bunch of crazy stuff with setting some

00:38:54 ◼   ► switches in either the CPU or the VRM in order to get it

00:38:59 ◼   ► to not limit the power of going to the CPU.

00:39:02 ◼   ► Which from what I can tell, having obviously not tried,

00:39:05 ◼   ► this seems to have worked.

00:39:07 ◼   ► If you're willing to disable SIP and install a Kext,

00:39:11 ◼   ► kernel extension, all this other crazy stuff.

00:39:13 ◼   ► So that worked reasonably well for a little bit

00:39:18 ◼   ► until, I believe it was yesterday,

00:39:21 ◼   ► as one of you said earlier, Apple released an official fix.

00:39:23 ◼   ► So I feel like this has been mostly resolved,

00:39:28 ◼   ► but do we have any thoughts on this Reddit thread

00:39:30 ◼   ► or the fix that came out of Apple just yesterday?

00:39:33 ◼   ► - I think the Reddit thread is a good example

00:39:34 ◼   ► of people trying to figure it out on their own.

00:39:37 ◼   ► They don't know what the problem is,

00:39:38 ◼   ► but they can see, we have measurements,

00:39:40 ◼   ► we have some tools that we can use,

00:39:42 ◼   ► and there's a thing about the tools

00:39:43 ◼   ► that we'll get to in a little bit too.

00:39:44 ◼   ► But we're like, what's the problem here?

00:39:46 ◼   ► What could be the problem?

00:39:47 ◼   ► Because it's not as simple as just an on/off switch

00:39:51 ◼   ► or it's always running at maximum cooling.

00:39:54 ◼   ► There's lots of decisions to be made exactly

00:39:56 ◼   ► how much power should the CPU draw for how long?

00:39:59 ◼   ► When it turbos, when should it decide to turbo

00:40:02 ◼   ► and when should it say, I've turboed enough,

00:40:03 ◼   ► I should go down, right?

00:40:05 ◼   ► And there are other chips involved in the motherboard,

00:40:07 ◼   ► not just the CPU that could be involved

00:40:09 ◼   ► in a cooling decision.

00:40:10 ◼   ► And so someone was just poking in some registers

00:40:12 ◼   ► or firmware or something and changing a bunch of value,

00:40:15 ◼   ► hex values in some big bit mask thing

00:40:18 ◼   ► and saying this controls how long it turbos to maximum speed

00:40:21 ◼   ► or how you're allowed to draw 100 watts for 30 seconds

00:40:24 ◼   ► and you're allowed to draw 150 watts for five seconds

00:40:27 ◼   ► and just playing with those numbers

00:40:29 ◼   ► with the goal of taking some workload,

00:40:31 ◼   ► I forget what they were doing as the workload,

00:40:32 ◼   ► and making that workload complete faster

00:40:34 ◼   ► because it's inefficient to bump really hard

00:40:37 ◼   ► up against your thermal ceiling and throttle way, way down

00:40:39 ◼   ► and then bump right back up against it

00:40:41 ◼   ► and throttle way, way down.

00:40:42 ◼   ► It's much better to oscillate with smaller moves

00:40:45 ◼   ► like turbo for a shorter period

00:40:47 ◼   ► so you don't hit your ceiling and don't bump way, way down

00:40:50 ◼   ► and so you can just bump from your base clock up to turbo,

00:40:52 ◼   ► base clock up to turbo or whatever.

00:40:53 ◼   ► It's better than going from turbo all the way down

00:40:56 ◼   ► to like 800 megahertz or whatever.

00:40:57 ◼   ► Again, like Mark was saying before,

00:40:58 ◼   ► if you say, oh, it's not actually the CPU,

00:41:01 ◼   ► it's the voltage regulation module

00:41:03 ◼   ► and that overheats, it responds by pulling power

00:41:06 ◼   ► from the CPU and all sorts of theories

00:41:08 ◼   ► and there's a lot more going on inside there

00:41:10 ◼   ► than anyone outside Apple's probably gonna be able to tell.

00:41:14 ◼   ► So Apple's fix, as the new updated thread says,

00:41:18 ◼   ► doesn't just operate on the CPU

00:41:20 ◼   ► but takes the entire system into account

00:41:21 ◼   ► so it can handle cases where what if your GPU

00:41:24 ◼   ► is making really hot because it's something

00:41:25 ◼   ► in your application is using the GPU.

00:41:27 ◼   ► Like it's all the same kind of cooling system inside there

00:41:30 ◼   ► and an excess of heat can make everything hot

00:41:32 ◼   ► 'cause it's all next to each other

00:41:33 ◼   ► inside the same case or whatever.

00:41:34 ◼   ► So Apple's patch takes a more holistic view of the system

00:41:37 ◼   ► as it always does and tries to balance the heat

00:41:40 ◼   ► and tries to make the components,

00:41:43 ◼   ► not slam hard against their thermal ceiling

00:41:44 ◼   ► but just creep right up to it

00:41:46 ◼   ► but have the cooling system keep them below it

00:41:48 ◼   ► and have them oscillate back and forth as you'd expect.

00:41:50 ◼   ► Basically saying if it gets too hot,

00:41:52 ◼   ► go back down to your base clock, cool off for a while

00:41:54 ◼   ► and then go, you know.

00:41:56 ◼   ► So it's some heroic armchair attempts at fixes

00:42:01 ◼   ► without access to source code or anything like that

00:42:03 ◼   ► but just access to a machine.

00:42:04 ◼   ► But it's, you know, I would never recommend

00:42:08 ◼   ► someone do something like that,

00:42:09 ◼   ► not at least because it requires a kernel extension

00:42:12 ◼   ► and disabling system integrity protection,

00:42:15 ◼   ► all sorts of other stuff.

00:42:17 ◼   ► And if it really is a real problem,

00:42:19 ◼   ► you should have some confidence

00:42:20 ◼   ► that Apple will do something about it

00:42:21 ◼   ► and in this case they did within days.

00:42:23 ◼   ► So this is a pretty fast turnaround time from Apple

00:42:26 ◼   ► and it's some interesting experiments

00:42:30 ◼   ► by people out in the field

00:42:31 ◼   ► but I would warn people, if you have one of these,

00:42:36 ◼   ► give it some time, give it a few days, give it a week,

00:42:39 ◼   ► don't rush out and install software

00:42:42 ◼   ► that some person posted on Reddit,

00:42:44 ◼   ► even if it might look cool.

00:42:45 ◼   ► - Well but you know, I think there is some value

00:42:49 ◼   ► to the side of that of like,

00:42:51 ◼   ► you know, when you look at what we mentioned earlier

00:42:53 ◼   ► about how the alleged reason that Apple PR gave

00:42:57 ◼   ► for the keyboard membrane was to make this keyboard quieter

00:43:01 ◼   ► when that appears to be pretty, you know, BSE, right?

00:43:05 ◼   ► So whatever Apple says or doesn't say

00:43:07 ◼   ► about problems like this,

00:43:09 ◼   ► it's hard to know whether it's accurate or not

00:43:11 ◼   ► without, you know, people playing around with stuff like this

00:43:13 ◼   ► One of the reasons I was watching this thread

00:43:15 ◼   ► was like, alright, so they say that the VRM throttling

00:43:18 ◼   ► is at fault here, that basically that the thermal settings

00:43:21 ◼   ► were not good for that processor

00:43:24 ◼   ► and that changing the thermal settings

00:43:25 ◼   ► would fix this problem.

00:43:27 ◼   ► So the very first thing I wanted to know

00:43:28 ◼   ► when the patch came out was, did they change those settings?

00:43:31 ◼   ► Because like with this method,

00:43:33 ◼   ► you're like using some utility to read the settings

00:43:36 ◼   ► and then you figure out what the value should be

00:43:37 ◼   ► and write new settings back.

00:43:38 ◼   ► So I was very, very curious to go back to this thread

00:43:41 ◼   ► once that patches out to see like, alright,

00:43:43 ◼   ► has this person who wrote this like gone back

00:43:46 ◼   ► and updated it with the new one?

00:43:48 ◼   ► Because if the value has changed with this new,

00:43:52 ◼   ► you know, with the patch that Apple released,

00:43:54 ◼   ► Apple's reason was, you know, BS that this is actually

00:43:57 ◼   ► the real reason here, that this value is wrong.

00:43:59 ◼   ► And it would still be great if they fixed it,

00:44:01 ◼   ► but if the value had changed and that was the real reason,

00:44:03 ◼   ► that would be like, you know, worth knowing,

00:44:05 ◼   ► worth experimenting to find out.

00:44:07 ◼   ► Fortunately, that wasn't the case.

00:44:08 ◼   ► Fortunately, that, you know, the value has not changed

00:44:11 ◼   ► of this VRM profile thing of like how much heat

00:44:14 ◼   ► it can draw for how long.

00:44:16 ◼   ► So fortunately, the PR explanation appears to be correct.

00:44:19 ◼   ► And we'll get to that, I guess, now.

00:44:22 ◼   ► - Indeed, so Apple has released a software fix

00:44:26 ◼   ► and we'll put a link to Jason Snell's coverage

00:44:29 ◼   ► on six colors.

00:44:30 ◼   ► But the quote from Apple is as follows,

00:44:33 ◼   ► "Following extensive performance testing

00:44:35 ◼   ► under numerous workloads, we've identified

00:44:37 ◼   ► that there is a," this is from them,

00:44:39 ◼   ► "missing digital key in the firmware

00:44:41 ◼   ► that impacts the thermal management system

00:44:43 ◼   ► and could drive clock speeds down under heavy thermal loads

00:44:45 ◼   ► on the new MacBook Pro.

00:44:47 ◼   ► A bug fixes including today's macOS High Sierra

00:44:50 ◼   ► 10.13.6 supplemental update and is recommended.

00:44:52 ◼   ► We apologize to any customers who experience

00:44:54 ◼   ► less than optimal performance on their new systems."

00:44:56 ◼   ► And then there's some marketing drivel after that.

00:44:58 ◼   ► So Macworld took a look at it

00:45:00 ◼   ► and they said their preliminary results

00:45:03 ◼   ► from our 4K Premiere test is that the 2017 2.9 gigahertz

00:45:07 ◼   ► Core i7, 90 minutes to do this thing.

00:45:11 ◼   ► The 2018 Core i9 before the patch, 80 minutes,

00:45:14 ◼   ► so a delta of 10 minutes.

00:45:16 ◼   ► The 2018 Core i9 after the patch is 72 minutes,

00:45:20 ◼   ► which is eight minutes quicker.

00:45:22 ◼   ► So, and additionally they also saw that the clock speeds

00:45:25 ◼   ► were pretty flat, which indicated that things

00:45:28 ◼   ► were operating as they should have been.

00:45:30 ◼   ► - Yeah, because before the patch, the symptom was basically

00:45:32 ◼   ► you have this, as I mentioned, the sawtooth graph

00:45:34 ◼   ► of the, if you look at the processor frequency

00:45:36 ◼   ► where it would peak really high

00:45:38 ◼   ► and then drop all the way down to 800 megahertz

00:45:40 ◼   ► and then peak really high again

00:45:42 ◼   ► and then drop all the way back down

00:45:43 ◼   ► and just like constant fluctuations

00:45:45 ◼   ► and something was clearly overheating or being mismanaged.

00:45:48 ◼   ► That's not normal, it's not supposed to be

00:45:50 ◼   ► that spiky and crazy.

00:45:52 ◼   ► So the patch did indeed fix the problem.

00:45:55 ◼   ► I actually talked to Apple about this.

00:45:57 ◼   ► I had a call, I asked about the VRM thing,

00:46:00 ◼   ► they said that was not it.

00:46:02 ◼   ► But ultimately I don't really have any more information

00:46:03 ◼   ► than that that isn't in the statement.

00:46:05 ◼   ► I'm really, really curious to know,

00:46:08 ◼   ► just like some more technical detail

00:46:10 ◼   ► about like a missing digital key

00:46:12 ◼   ► in the thermal management system firmware.

00:46:14 ◼   ► Like I'm curious what that means.

00:46:16 ◼   ► - Didn't ask Apple about it?

00:46:18 ◼   ► - I did, but you know, it was, you know,

00:46:20 ◼   ► I didn't get a lot of technical detail.

00:46:22 ◼   ► You know, they were very friendly

00:46:24 ◼   ► and open to whatever they knew how to answer.

00:46:27 ◼   ► But unfortunately, this statement really covers it all.

00:46:30 ◼   ► There's not, besides that I was able to confirm

00:46:32 ◼   ► that the VRM thing was not it.

00:46:34 ◼   ► I don't really have any other information

00:46:36 ◼   ► that wasn't in the statement.

00:46:37 ◼   ► But it was nice, I talked to them,

00:46:39 ◼   ► they seemed genuinely concerned.

00:46:40 ◼   ► I asked them if they bought a freezer

00:46:42 ◼   ► to rerun the YouTube test.

00:46:44 ◼   ► They did not.

00:46:45 ◼   ► They didn't seem to think that was very funny.

00:46:47 ◼   ► (laughing)

00:46:49 ◼   ► Otherwise, yeah, so it's,

00:46:51 ◼   ► I would like to hear from any listeners who know

00:46:54 ◼   ► about like just how this stuff works at this low level.

00:46:56 ◼   ► Like what is that, a missing digital key?

00:46:58 ◼   ► Does that just mean like whatever their method

00:47:02 ◼   ► of controlling this in software,

00:47:04 ◼   ► maybe like it was not running,

00:47:06 ◼   ► or if it was not being loaded by some like firmware download

00:47:10 ◼   ► at the CPU level, if the CPU was like falling back

00:47:13 ◼   ► to its like built-in profiling instead

00:47:15 ◼   ► of having Apple customize it,

00:47:16 ◼   ► which I think might be what's going on.

00:47:17 ◼   ► But I just wanna know more about this at a low level,

00:47:20 ◼   ► just out of curiosity's sake.

00:47:21 ◼   ► But the reality is that the problem is fixed

00:47:24 ◼   ► in the sense that it is, like you know,

00:47:26 ◼   ► before it had these weird thermal characteristics

00:47:28 ◼   ► and in many tests it was slower than the one

00:47:31 ◼   ► that came before it or was at least not faster by very much.

00:47:35 ◼   ► And now the performance graphs have smoothed out

00:47:38 ◼   ► and it is faster than, basically it is,

00:47:42 ◼   ► as what you'd expect, the fastest CPU option

00:47:44 ◼   ► is the fastest performing in real-world tests.

00:47:48 ◼   ► So that like, that was the real problem

00:47:50 ◼   ► and that problem seems to have been fixed.

00:47:52 ◼   ► They did tell me that this problem affected

00:47:55 ◼   ► all of the 2018 MacBook Pros.

00:47:57 ◼   ► So not just the i9 and the 15 inch,

00:47:59 ◼   ► but also the i7s and the 15 inch

00:48:01 ◼   ► and the 13 inch options as well.

00:48:03 ◼   ► This bug apparently affected all of them,

00:48:06 ◼   ► which does lend more credence to the fact

00:48:08 ◼   ► that they are most likely telling the truth on the source.

00:48:10 ◼   ► It isn't like some weird thermal thing with the i9

00:48:12 ◼   ► that it affects all of them.

00:48:14 ◼   ► But I don't think people really saw it

00:48:15 ◼   ► that much with 13 inches yet.

00:48:16 ◼   ► So I don't know if we have a lot of test results

00:48:19 ◼   ► to confirm like maybe how bad it was before

00:48:21 ◼   ► and where it is now.

00:48:23 ◼   ► But regardless, it's fixed, I'm happy,

00:48:26 ◼   ► and it seems like it's pretty much a closed book.

00:48:30 ◼   ► - I have the same academic curiosity about the explanation

00:48:32 ◼   ► because the fact that like this one paragraph thing

00:48:35 ◼   ► doesn't have technical details but does decide

00:48:37 ◼   ► to say a missing digital key,

00:48:38 ◼   ► which is just enough to intrigue you,

00:48:40 ◼   ► say really, like what does that even mean?

00:48:42 ◼   ► The theory that I've heard, like you just said,

00:48:45 ◼   ► the most prevailing one is some piece of software

00:48:47 ◼   ► that's supposed to be signed and wasn't

00:48:49 ◼   ► and so didn't get loaded or something.

00:48:50 ◼   ► But yeah, I don't know.

00:48:51 ◼   ► I mean, they're not into giving you the technical explanation

00:48:55 ◼   ► but I think the most interesting part about this,

00:48:57 ◼   ► again, it's not interesting consumers,

00:48:58 ◼   ► all you care is that the thing is fixed

00:49:00 ◼   ► and it is so great, right?

00:49:01 ◼   ► But it's not actually like,

00:49:04 ◼   ► it's one of those boring software problems

00:49:06 ◼   ► where it's basically like a build system problem

00:49:08 ◼   ► or like a CI pipeline problem or a packaging problem.

00:49:13 ◼   ► Like it's not actually a problem with the technology,

00:49:15 ◼   ► it's a problem with the boring part of the system

00:49:17 ◼   ► where you package up the bits for sale

00:49:20 ◼   ► and make sure that those are the same exact bits

00:49:23 ◼   ► that you tested but of course you have internal only things

00:49:25 ◼   ► and so it was like,

00:49:26 ◼   ► like what they were selling to consumers

00:49:30 ◼   ► was not exactly the bits they thought they were.

00:49:32 ◼   ► And it's great 'cause you can fix a problem like this.

00:49:34 ◼   ► It's not a hardware problem,

00:49:35 ◼   ► it's like purely a software thing

00:49:36 ◼   ► and they can send you a software patch and it fixes it,

00:49:38 ◼   ► right, there must be relief

00:49:39 ◼   ► to all the hardware people involved.

00:49:41 ◼   ► Like everything's fine, the cooling system is fine,

00:49:43 ◼   ► it's not a catastrophic failure of thermal paste

00:49:46 ◼   ► or something, it's just the software

00:49:48 ◼   ► we thought we were sending out to you on your laptop

00:49:50 ◼   ► is not actually the software that you got

00:49:52 ◼   ► so here's a patch and everyone's relieved.

00:49:55 ◼   ► It's kind of like how a lot of the recent leaks from Apple

00:49:59 ◼   ► have been like,

00:50:01 ◼   ► it's similarly related to like the build system

00:50:04 ◼   ► or like putting things on URLs temporarily

00:50:07 ◼   ► where they shouldn't be or like just sort of mundane

00:50:11 ◼   ► human error unrelated to the creation of the product.

00:50:14 ◼   ► 'Cause people creating their product

00:50:15 ◼   ► are working really hard and doing all the things

00:50:16 ◼   ► and doing everything right as best they can

00:50:18 ◼   ► and then someone goes and accidentally puts like

00:50:21 ◼   ► an unreleased version of operating system

00:50:22 ◼   ► up on a public URL or like the build system builds something

00:50:26 ◼   ► that they think is an exact match for what they tested

00:50:27 ◼   ► but actually isn't and they ship these laptops to people

00:50:30 ◼   ► and they have the wrong bits on them.

00:50:32 ◼   ► Stuff like that happens, like it's not a big deal,

00:50:34 ◼   ► this is a pretty fast turnaround time

00:50:35 ◼   ► but I'm still technically curious about the details

00:50:38 ◼   ► and somebody's boring tell all nerd book

00:50:42 ◼   ► about Apple 20 years from now can include this in a chapter.

00:50:44 ◼   ► - This will be the most boring chapter.

00:50:46 ◼   ► - And be like let me tell you about the time

00:50:47 ◼   ► that the build system screwed up somehow.

00:50:49 ◼   ► (upbeat music)

00:50:50 ◼   ► - We are sponsored this week by Linode.

00:50:52 ◼   ► Go to linode.com/atp to learn more and get a $20 credit.

00:50:56 ◼   ► Linode is my favorite web host.

00:50:59 ◼   ► If you need a virtual private server,

00:51:01 ◼   ► not only is it the best VPS host I've ever used

00:51:04 ◼   ► and I've used a lot of hosts,

00:51:05 ◼   ► it's the best web host I've ever used period

00:51:07 ◼   ► because now the VPSs are so fast now

00:51:09 ◼   ► they compete with dedicated in my opinion.

00:51:11 ◼   ► Like I've used a lot of dedicated servers,

00:51:12 ◼   ► VPSs I think are the way to go for almost anybody

00:51:15 ◼   ► and Linode is the place to do that

00:51:17 ◼   ► because they have 24/7 friendly support,

00:51:19 ◼   ► they have phone support if you need it,

00:51:21 ◼   ► the performance is fantastic

00:51:22 ◼   ► 'cause you're running on native enterprise grade SSDs,

00:51:25 ◼   ► you're backed by a 40 gigabit network

00:51:26 ◼   ► running on Intel Xeon E5 processors

00:51:29 ◼   ► so this is really high spec stuff here.

00:51:31 ◼   ► All that's for an amazing price.

00:51:33 ◼   ► Their one gig plan is just $5 a month.

00:51:37 ◼   ► They have all sorts of plans that go up from there

00:51:38 ◼   ► depending on what your needs are.

00:51:40 ◼   ► You can pick 10 data centers, you can put your server.

00:51:42 ◼   ► All this is backed by a seven day money back guarantee

00:51:46 ◼   ► so you can see for yourself there's nothing to lose.

00:51:48 ◼   ► The pricing is very, very fair.

00:51:49 ◼   ► As I mentioned, they're great value,

00:51:51 ◼   ► I think consistently the best value

00:51:53 ◼   ► in the hosting business by far

00:51:55 ◼   ► and they back it up to with constant upgrades.

00:51:58 ◼   ► Like they just did an upgrade now,

00:51:59 ◼   ► about every few months they will do an upgrade

00:52:01 ◼   ► where you just get more.

00:52:03 ◼   ► Whatever plan you were paying for before,

00:52:05 ◼   ► you can go click a button

00:52:06 ◼   ► and you just get upgraded to a bigger server

00:52:08 ◼   ► and you just get even more resources

00:52:10 ◼   ► for the same price you were paying before.

00:52:12 ◼   ► I love Linode and if you have a need

00:52:14 ◼   ► for your own virtual private server,

00:52:15 ◼   ► whether that is a web application

00:52:17 ◼   ► like what I'm running on mine,

00:52:18 ◼   ► you can run Docker containers,

00:52:20 ◼   ► you can use encrypted disks, you can run VPNs.

00:52:22 ◼   ► It is by far the nicest host I've ever used.

00:52:24 ◼   ► I've been there the longest

00:52:26 ◼   ► and I think it's the best value in the business.

00:52:28 ◼   ► So check it out today at linode.com/atp.

00:52:31 ◼   ► That will get you a $20 credit

00:52:33 ◼   ► and if you pick that one gig plan

00:52:35 ◼   ► for just five bucks a month,

00:52:36 ◼   ► that could be four free months.

00:52:38 ◼   ► That's pretty awesome.

00:52:39 ◼   ► And once again, there's a seven day money back guarantee

00:52:41 ◼   ► so there's nothing to lose.

00:52:42 ◼   ► Check it out once again, linode.com/atp.

00:52:45 ◼   ► Thank you so much to Linode for hosting all my stuff

00:52:48 ◼   ► and supporting our show.

00:52:49 ◼   ► - All right, and then we have one final piece of drama

00:52:54 ◼   ► over the last few days that luckily this show has avoided.

00:52:58 ◼   ► The Intel Power Gadget Utility,

00:53:01 ◼   ► which I guess is a way I've not used this.

00:53:02 ◼   ► Marco, have you used this?

00:53:04 ◼   ► - I have actually.

00:53:05 ◼   ► So it's Intel's kernel extension that you have to install.

00:53:10 ◼   ► Basically it becomes an app that you can run

00:53:13 ◼   ► that simply shows you nice little graphs and numbers

00:53:16 ◼   ► about what your processor is doing,

00:53:18 ◼   ► how fast it's going, what kind of utilization it has,

00:53:21 ◼   ► like as a percentage, and how much power it's drawing.

00:53:24 ◼   ► So that's how we're able to see things like,

00:53:27 ◼   ► oh, this CPU is spiking to this peak frequency

00:53:31 ◼   ► and then dropping and then spiking again and then dropping

00:53:32 ◼   ► and the water is just going crazy and the temperature is not

00:53:35 ◼   ► and you can see all things like that

00:53:37 ◼   ► because of this utility from Intel

00:53:39 ◼   ► called the Intel Power Gadget.

00:53:41 ◼   ► And as far as I know,

00:53:42 ◼   ► I don't think there's another way to get,

00:53:44 ◼   ► you can get heat from various system settings reading apps

00:53:49 ◼   ► but I don't think you can get clock speed and CPU wattage

00:53:52 ◼   ► that easily or that precisely from anything else.

00:53:54 ◼   ► - Right, so this thing existed and then didn't

00:53:58 ◼   ► and now does again.

00:53:59 ◼   ► So apparently it was pulled from Intel's site

00:54:03 ◼   ► right around the time everyone was trying to grab it

00:54:06 ◼   ► to figure out what the hell was going on

00:54:07 ◼   ► with these new Mac or Pros.

00:54:09 ◼   ► And then I guess was version three,

00:54:13 ◼   ► just appeared in the last 48 hours.

00:54:16 ◼   ► That sort of does the same thing

00:54:19 ◼   ► but one of you guys was kind enough

00:54:20 ◼   ► to put a quote in the show notes.

00:54:21 ◼   ► In version three, there are additional features

00:54:23 ◼   ► that include estimation of power on multi-socket systems

00:54:25 ◼   ► as well as externally callable APIs

00:54:27 ◼   ► to extract power information within sections of code.

00:54:30 ◼   ► And apparently it's important to note the latest release

00:54:32 ◼   ► also includes support for Windows 10.

00:54:34 ◼   ► - That's like the entire change log.

00:54:35 ◼   ► So the mystery again with the,

00:54:37 ◼   ► if you wanted to assume that Apple is evil,

00:54:39 ◼   ► it's like, oh, they found this problem

00:54:40 ◼   ► and their solution is to get on the phone

00:54:42 ◼   ► and tell Intel to pull the power gadget.

00:54:45 ◼   ► But then it's back a day later or whatever

00:54:47 ◼   ► and when it comes back, the changes have nothing to do

00:54:50 ◼   ► with like, 'cause it's not multi-socket system,

00:54:52 ◼   ► it's not Windows 10

00:54:53 ◼   ► and it's not about instrumenting sections of code.

00:54:55 ◼   ► Like why did it disappear?

00:54:58 ◼   ► And why did they then release version three?

00:55:01 ◼   ► Who knows?

00:55:02 ◼   ► It could have been just a coincidence

00:55:03 ◼   ► but either way, all the people screaming

00:55:05 ◼   ► that it was a conspiracy theory

00:55:07 ◼   ► about some terrible problem Apple had.

00:55:09 ◼   ► The problem Apple had was actually not terrible.

00:55:12 ◼   ► It was actually easy to fix and they fixed it immediately

00:55:14 ◼   ► and then the power gadget is back.

00:55:15 ◼   ► So don't always assume that Apple's out to get you.

00:55:18 ◼   ► - Well, but at the same time,

00:55:20 ◼   ► you could see why people got here, right?

00:55:22 ◼   ► Because Apple's solution, when the 2016 MacBook Pro came out

00:55:25 ◼   ► and had weird terrible battery life,

00:55:28 ◼   ► depending on what you were doing,

00:55:29 ◼   ► was to hide the battery life remaining indicator

00:55:32 ◼   ► in the menu bar forever.

00:55:33 ◼   ► So their solution there was,

00:55:36 ◼   ► here's this thing that's showing people

00:55:38 ◼   ► how weird our battery life is

00:55:39 ◼   ► with these new complicated laptops.

00:55:41 ◼   ► We're just gonna hide that information.

00:55:43 ◼   ► So it actually was totally plausible

00:55:46 ◼   ► that Apple's solution to people are seeing weird things

00:55:50 ◼   ► in the frequency graph might be convince your partner Intel

00:55:54 ◼   ► to pull the frequency graph utility from your website.

00:55:56 ◼   ► Like that is a totally plausible outcome

00:55:58 ◼   ► based on their past performance.

00:56:00 ◼   ► So I can see why people thought that.

00:56:02 ◼   ► - Don't you think hiding the number though

00:56:06 ◼   ► is partly because the numbers weren't accurate

00:56:08 ◼   ► and they were misleading?

00:56:09 ◼   ► Like the thing with the speed thing is like,

00:56:11 ◼   ► look, you can hide the power gadget

00:56:13 ◼   ► or not let people install the power gadget all you want.

00:56:15 ◼   ► They still have a stopwatch.

00:56:16 ◼   ► They could still run a workload and time it

00:56:18 ◼   ► versus the old version and say,

00:56:20 ◼   ► I don't need to see what the frequency is.

00:56:22 ◼   ► I don't know what the problem is.

00:56:23 ◼   ► All I know is this thing is slower than it should be.

00:56:25 ◼   ► Like you can't stop people from doing that.

00:56:27 ◼   ► And like the battery power thing is like,

00:56:30 ◼   ► we have bad estimates of battery power

00:56:32 ◼   ► and people are flipping out about,

00:56:34 ◼   ► because they're flipping out when they see the bad estimate,

00:56:36 ◼   ► they never actually say, okay,

00:56:38 ◼   ► is that actually how much time I have remaining?

00:56:40 ◼   ► Start a stopwatch now.

00:56:41 ◼   ► Like they were just flipping out on the numbers.

00:56:42 ◼   ► So like the fact that it was back like a day later,

00:56:46 ◼   ► I don't, it just doesn't add up to me.

00:56:49 ◼   ► I like, and maybe they did,

00:56:50 ◼   ► maybe some panic PR person did call and ask them to pull it.

00:56:54 ◼   ► But I don't think it was part of some grand conspiracy

00:56:57 ◼   ► to hide a big problem because there wasn't a big problem.

00:57:00 ◼   ► There just wasn't, like the problem was a small problem

00:57:03 ◼   ► and it was fixed.

00:57:04 ◼   ► And it was like, it was not unexpected.

00:57:05 ◼   ► They didn't have to research it for a month.

00:57:07 ◼   ► They didn't have to do extended testing.

00:57:08 ◼   ► I bet it was pretty quick when someone said,

00:57:10 ◼   ► I figured out the problem.

00:57:11 ◼   ► People don't have the right bits on their computer.

00:57:14 ◼   ► Like it's the best problem to have.

00:57:15 ◼   ► Like we, this is not actually a problem.

00:57:16 ◼   ► It was the build system or whatever.

00:57:18 ◼   ► I'm calling it the build system.

00:57:19 ◼   ► That's probably the wrong word.

00:57:20 ◼   ► I'm sure there's some Apple word for like

00:57:22 ◼   ► the way things get onto the computers that you buy.

00:57:27 ◼   ► So, I don't think you can,

00:57:31 ◼   ► I can see where people got there because again,

00:57:32 ◼   ► people assume Apple is always evil or whatever,

00:57:34 ◼   ► but just things happen so fast

00:57:36 ◼   ► and we're so quickly disproved in this case

00:57:38 ◼   ► that I don't think people even had enough time

00:57:40 ◼   ► to get up a full head of steam about being angry

00:57:41 ◼   ► about the conspiracy theory about Apple,

00:57:44 ◼   ► Apple making Intel pull its gadget

00:57:45 ◼   ► and then re-add it with Windows 10 support.

00:57:48 ◼   ► It could be like the sticky note on the store

00:57:49 ◼   ► where it's like, this is every time they have to

00:57:51 ◼   ► update the power gadget utility, it's always gone for a day.

00:57:54 ◼   ► - All right, so moving on,

00:57:56 ◼   ► let's talk a little bit more about Intel

00:57:58 ◼   ► because about a month ago,

00:58:00 ◼   ► friend of the show Ben Thompson wrote a really, really,

00:58:02 ◼   ► really great article at Stratechery

00:58:04 ◼   ► with regard to kind of Intel's problems

00:58:06 ◼   ► and how they got in this, I guess, mess that they're in today

00:58:10 ◼   ► it's been a while since I've read this

00:58:11 ◼   ► although I certainly did read it.

00:58:12 ◼   ► John, I think you might've added this to the show notes.

00:58:15 ◼   ► Can you come take it away and tell us about this?

00:58:17 ◼   ► - On this show, we've talked about Intel,

00:58:21 ◼   ► I think most of the time in the context of like ARM stuff,

00:58:25 ◼   ► ARM Macs or Intel chips and phones,

00:58:29 ◼   ► just kind of the comparison

00:58:34 ◼   ► between what Intel offers Apple

00:58:37 ◼   ► and what Apple is essentially offering itself

00:58:39 ◼   ► by like buying PA semi and making its own chips

00:58:42 ◼   ► and making its own GPUs and doing all of this stuff.

00:58:44 ◼   ► I think maybe a year or two ago,

00:58:47 ◼   ► one of the focuses was that

00:58:50 ◼   ► Intel had a big strategic advantage to Apple

00:58:56 ◼   ► and that their process technology,

00:58:59 ◼   ► meaning how small they can make the features on their chips

00:59:02 ◼   ► was a generation ahead of everybody else.

00:59:05 ◼   ► And so no matter how much more efficient ARM CPUs were

00:59:09 ◼   ► and no matter how much Apple could custom tailor them

00:59:11 ◼   ► to their exact needs, it's very hard to overcome

00:59:15 ◼   ► the fact that Apple had to make all of its ARM chips

00:59:18 ◼   ► using larger feature sizes, bigger transistors

00:59:21 ◼   ► than Intel did.

00:59:23 ◼   ► So even though Intel may have old and creaky chips

00:59:25 ◼   ► and x86 is old and creaky, even x86-64

00:59:29 ◼   ► and it's inefficient and all these reasons

00:59:31 ◼   ► why ARM is better than Intel,

00:59:33 ◼   ► Intel has this advantage of like they take

00:59:35 ◼   ► whatever their designs are and they can make them

00:59:37 ◼   ► much smaller than everyone else and they use less power

00:59:39 ◼   ► and they just have all these advantages.

00:59:42 ◼   ► And the discussion then was, wouldn't it be great

00:59:46 ◼   ► if Apple could get its like its ARM CPUs

00:59:50 ◼   ► or system-mounted chips or iPhone stuff

00:59:52 ◼   ► and design them, but then have Intel make them.

00:59:57 ◼   ► Say we have designs, Intel,

00:59:59 ◼   ► we don't actually want your chips,

01:00:02 ◼   ► we just want you to be a fab for us.

01:00:05 ◼   ► Here's the design, you make it for us

01:00:07 ◼   ► and you make it on your industry-leading

01:00:08 ◼   ► small than everybody else, always ahead by a year and a half,

01:00:11 ◼   ► two years ahead of everybody else

01:00:13 ◼   ► and that would be the best of both worlds.

01:00:15 ◼   ► It would be the best chip designs

01:00:18 ◼   ► and then the smallest fabrication size.

01:00:20 ◼   ► And the discussion then was like, yeah,

01:00:22 ◼   ► but why the hell would Intel do that?

01:00:23 ◼   ► Because they can't charge you as much.

01:00:25 ◼   ► They can charge you way more for their high-end Intel CPU

01:00:29 ◼   ► or Xeons or Inines or whatever,

01:00:32 ◼   ► because it's like that's our intellectual property,

01:00:34 ◼   ► we design it and we fab the chip

01:00:36 ◼   ► and so we can double dip there.

01:00:39 ◼   ► Whereas if we're just a dumb fab and we just say,

01:00:41 ◼   ► oh, we'll manufacture things for you

01:00:43 ◼   ► and you'll bring the thing yourself,

01:00:44 ◼   ► we can't get as much money from you.

01:00:46 ◼   ► So, and we're ahead of everybody else,

01:00:48 ◼   ► so I bet everybody would love for us to fab

01:00:50 ◼   ► their stupid designs, but we're only gonna fab our designs.

01:00:52 ◼   ► That's our competitive advantage,

01:00:53 ◼   ► you all suck, Intel is great.

01:00:55 ◼   ► And that seems like such a long time ago,

01:00:58 ◼   ► maybe it wasn't as long ago as I thought,

01:01:00 ◼   ► but it seems like such a long time ago.

01:01:02 ◼   ► And over the time since then,

01:01:05 ◼   ► there's been smaller stories about Intel

01:01:07 ◼   ► willing to fab chips for other people,

01:01:09 ◼   ► which I think that have appeared in our show notes

01:01:10 ◼   ► at various times and maybe we haven't actually talked about,

01:01:12 ◼   ► maybe they just got booted out

01:01:13 ◼   ► because they became irrelevant.

01:01:16 ◼   ► But the most recent story,

01:01:17 ◼   ► as we've talked about on the past few shows,

01:01:18 ◼   ► is Intel, it's supposed cadence,

01:01:23 ◼   ► it used to be the TikTok where it would be like

01:01:25 ◼   ► they do an architecture change, then a shrink,

01:01:26 ◼   ► then architecture change, then a shrink

01:01:28 ◼   ► on like a yearly cadence or an 18 month cadence

01:01:30 ◼   ► or whatever the hell it was,

01:01:32 ◼   ► had been slowing down and they made all these PR things,

01:01:34 ◼   ► it used to be TikTok and now it's like process optimized,

01:01:39 ◼   ► but like they made a three step process,

01:01:40 ◼   ► and then as Marco said a few shows ago,

01:01:42 ◼   ► now it's TikTok, tok, tok, tok, tok, tok.

01:01:44 ◼   ► They've been stuck at 14 nanometers for a really long time,

01:01:48 ◼   ► for longer than anyone thought they would be,

01:01:50 ◼   ► including Intel, Intel had grand plans

01:01:52 ◼   ► for that 10 nanometer thing.

01:01:54 ◼   ► Now about six months ago, we were hearing like TSMC,

01:01:58 ◼   ► Taiwan Semiconductor Manufacturing Corporation,

01:02:01 ◼   ► they've made it to 10 nanometers or seven nanometers

01:02:04 ◼   ► or whatever, they're ahead of Intel,

01:02:06 ◼   ► and then people were quick to say,

01:02:07 ◼   ► well, again, kind of like base clock speed and TDP,

01:02:12 ◼   ► feature size is another one of those things

01:02:14 ◼   ► that used to have perhaps a more common

01:02:17 ◼   ► and simpler definition than it does these days,

01:02:19 ◼   ► because now people are arranging,

01:02:21 ◼   ► are making transistors with these strange 3D arrangements

01:02:24 ◼   ► of elements at the microscopic level,

01:02:27 ◼   ► like building these little pyramids and structures and fins,

01:02:31 ◼   ► and it's not as easy to say, well,

01:02:33 ◼   ► what is your feature size?

01:02:34 ◼   ► What are you measuring between?

01:02:35 ◼   ► What is your, it's like dot pitch on a monitor

01:02:37 ◼   ► or versus dot pitch on a trinitron,

01:02:39 ◼   ► like it's all kind of fuzzy,

01:02:40 ◼   ► and the gist of it was that Intel's 10 nanometer

01:02:44 ◼   ► is actually smaller than TSMC's 10 nanometer,

01:02:47 ◼   ► or maybe even smaller than TSMC's seven nanometer

01:02:49 ◼   ► or whatever, and so there was, you know,

01:02:52 ◼   ► maybe six months ago still a reason to say,

01:02:54 ◼   ► yeah, it looks like Intel might've lost its fab lead,

01:02:56 ◼   ► but it's not actually true,

01:02:57 ◼   ► but at this point, with Intel still unable

01:03:00 ◼   ► to deliver 10 nanometer stuff,

01:03:02 ◼   ► and TSMC making, I think they call their seven nanometer

01:03:05 ◼   ► or whatever, like it doesn't matter

01:03:07 ◼   ► that Intel's 10 nanometer, it may be as good as

01:03:10 ◼   ► or better than a similar number from TSMC.

01:03:13 ◼   ► TSMC looks like it's gonna actually deliver its process

01:03:17 ◼   ► in shipping products for customers,

01:03:20 ◼   ► and Intel still can't, and so, I mean,

01:03:25 ◼   ► Ben Thompson's article's more about the history of Intel

01:03:27 ◼   ► and like how it got to this point,

01:03:30 ◼   ► but I feel like Intel's strategy,

01:03:33 ◼   ► aside from like selling their arm holdings and losing,

01:03:37 ◼   ► and opting out of the contract to make the CPU

01:03:40 ◼   ► for the original iPhone and various other strategic mistakes,

01:03:44 ◼   ► if Intel was still ahead of everybody on process,

01:03:47 ◼   ► their, you know, their integrated design,

01:03:50 ◼   ► as Ben calls it, where they make the CPUs and have the fab,

01:03:55 ◼   ► would still be a viable strategy,

01:03:58 ◼   ► but if you're not the best fab in the world,

01:04:01 ◼   ► and when you were, you also refused to fab things

01:04:06 ◼   ► for other people, all you're left with now is x86

01:04:10 ◼   ► and its legacy of compatibility,

01:04:12 ◼   ► and a fab that is not helping you at all.

01:04:16 ◼   ► I mean, at a certain point, it might be better for Intel

01:04:21 ◼   ► to have TSMC manufacture their chips for them

01:04:25 ◼   ► if they can never get 10 nanometers out.

01:04:26 ◼   ► Like, it's just, they're in a bad situation,

01:04:30 ◼   ► and it's one of those best situations that's explicable.

01:04:32 ◼   ► It's like, this stuff is very difficult to do.

01:04:35 ◼   ► You make some bad bets seven years ago or five years ago,

01:04:38 ◼   ► whatever the crazy lead times are,

01:04:40 ◼   ► deciding how you're gonna fab your 10 nanometer,

01:04:43 ◼   ► and sometimes it just doesn't work out.

01:04:45 ◼   ► Like, it's a bummer.

01:04:46 ◼   ► It's a bummer for Intel, but it's an honest loss.

01:04:49 ◼   ► I feel, to me, it feels honest.

01:04:53 ◼   ► It feels more honest to me than boneheaded decisions

01:04:57 ◼   ► like sticking, like, Itanium, their instruction set

01:05:01 ◼   ► that was gonna compete with x86-64,

01:05:03 ◼   ► and opting out of the iPhone and all that other stuff,

01:05:07 ◼   ► and maybe also feels like a more honest loss

01:05:10 ◼   ► than refusing to fab for other people,

01:05:14 ◼   ► because I feel like there was a way,

01:05:16 ◼   ► because they were a leader,

01:05:17 ◼   ► their fab was in the lead for so long,

01:05:19 ◼   ► there should have been a way for them to leverage that

01:05:21 ◼   ► to make money while also maintaining their top-end deals.

01:05:26 ◼   ► Like, we'll fab for certain other people,

01:05:28 ◼   ► but we won't fab things to compete with their own chips.

01:05:30 ◼   ► And I'm sure they did that to a limited degree,

01:05:32 ◼   ► and that's what those stories were about a few months ago.

01:05:34 ◼   ► But now, they're in a really rough place

01:05:36 ◼   ► where they need to get their 10-nanometer process online,

01:05:41 ◼   ► 'cause they can't just keep making new arrangements

01:05:44 ◼   ► of the same cores in 14-nanometer forever.

01:05:47 ◼   ► Eventually, if they get stuck like this for much longer,

01:05:50 ◼   ► Apple will have no choice but to leave.

01:05:53 ◼   ► It's kind of like the Power Mac G5 situation all over again,

01:05:56 ◼   ► where Apple just couldn't get better parts.

01:05:59 ◼   ► They couldn't get laptop parts, period,

01:06:01 ◼   ► and they couldn't get faster chips

01:06:03 ◼   ► for their high-end things.

01:06:04 ◼   ► So it's like, what choice do we have?

01:06:05 ◼   ► We have to go to Intel, essentially,

01:06:07 ◼   ► 'cause the option would be,

01:06:08 ◼   ► I guess we'll just keep shipping

01:06:09 ◼   ► water-cooled Power Mac G5s forever at the same exact speed,

01:06:12 ◼   ► because we have no one to buy a faster CPU from.

01:06:15 ◼   ► And Intel said, "We have faster CPUs,

01:06:17 ◼   ► "and we make them faster every year,

01:06:18 ◼   ► "and so you should come to us," and they did.

01:06:21 ◼   ► Say Intel comes out with these eight-core i9s, right?

01:06:25 ◼   ► What happens next year?

01:06:28 ◼   ► 12-core i9s?

01:06:29 ◼   ► Eventually, it becomes the size

01:06:32 ◼   ► of a Kraft American cheese slice inside your computer.

01:06:36 ◼   ► You can't just keep adding cores,

01:06:39 ◼   ► and there's not enough parallelism into workloads

01:06:41 ◼   ► to deal with that anyway.

01:06:42 ◼   ► I don't think they're in that situation.

01:06:43 ◼   ► I think they probably will get their process online.

01:06:45 ◼   ► They're highly motivated to do so.

01:06:47 ◼   ► It's probably safe to say that Intel

01:06:52 ◼   ► has lost the massive lead they used to have,

01:06:55 ◼   ► and not that we're onboard this sinking ship,

01:06:58 ◼   ► because, again, the ARM CPUs that we're imagining

01:07:01 ◼   ► are theoretical, like, "Hey, where is the ARM CPU

01:07:03 ◼   ► "to compete with the Xeon?"

01:07:04 ◼   ► Doesn't exist.

01:07:05 ◼   ► It could, maybe it does inside Apple,

01:07:07 ◼   ► but you can't say there's an obvious thing to go to.

01:07:09 ◼   ► And, of course, there's AMD, which is resurgent,

01:07:12 ◼   ► and which does not fab its own chips.

01:07:14 ◼   ► It uses GlobalFoundries, which is the fab part of AMD

01:07:17 ◼   ► that they spun off, and I think GlobalFoundries

01:07:19 ◼   ► also might be catching up to/matching Intel

01:07:23 ◼   ► in process at this point.

01:07:24 ◼   ► I haven't been keeping up with them as much as TSMC.

01:07:28 ◼   ► But, yeah, the world order in the silicon chip CPU world

01:07:33 ◼   ► has really, really turned upside down

01:07:37 ◼   ► from even just a few years ago.

01:07:40 ◼   ► And we just, you know, we collectively as consumers,

01:07:44 ◼   ► and Apple, are all kind of stuck having to make

01:07:48 ◼   ► some very difficult decisions, which is like,

01:07:50 ◼   ► our decision is, do I keep buying these new computers

01:07:53 ◼   ► every year, even though they aren't getting

01:07:55 ◼   ► that much faster or better, right?

01:07:56 ◼   ► At a certain point, it's like, what's the point

01:07:58 ◼   ► of me buying a new laptop if it's about the same

01:08:00 ◼   ► as the one I bought last year?

01:08:01 ◼   ► Or if my phone is faster?

01:08:03 ◼   ► Again, look at the Geekbench scores for the iPhone X

01:08:05 ◼   ► and compare them to Apple's entire current MacBook

01:08:08 ◼   ► and MacBook Pro line, you will be depressed, right?

01:08:11 ◼   ► (laughing)

01:08:12 ◼   ► And Apple's choice is, do we, so, like,

01:08:15 ◼   ► do we basically have to, you know,

01:08:18 ◼   ► should we have started a project to switch the Macs

01:08:20 ◼   ► to ARM three years ago?

01:08:22 ◼   ► And in retrospect, we only realize it now,

01:08:23 ◼   ► and now we're kind of screwed too,

01:08:25 ◼   ► because it's not like we have alternatives.

01:08:27 ◼   ► Did we start talking to AMD and see if they can make

01:08:29 ◼   ► some kind of, you know, Ryzen technology laptop chips for us?

01:08:32 ◼   ► Like, you know, it's so much easier when you're like,

01:08:35 ◼   ► we have the best CPUs, and every year Intel comes out

01:08:38 ◼   ► with new ones, and we're on this great gravy train

01:08:40 ◼   ► that lasted a long time, but it's, you know,

01:08:43 ◼   ► the smooth ride is over, and some hard decisions

01:08:46 ◼   ► are gonna be coming soon, and it's not really

01:08:48 ◼   ► anybody's fault, like, except for maybe you could say

01:08:54 ◼   ► Apple should have had more contingency plans,

01:08:57 ◼   ► but yeah, things are about to get real interesting,

01:09:00 ◼   ► it seems like.

01:09:01 ◼   ► We are sponsored this week by Backblaze,

01:09:04 ◼   ► unlimited cloud backup for Macs and PCs

01:09:06 ◼   ► for just $5 a month.

01:09:08 ◼   ► Go to backblaze.com/atp for a fully featured

01:09:11 ◼   ► 15-day free trial.

01:09:14 ◼   ► Backblaze is by far the best cloud backup.

01:09:17 ◼   ► I've used a bunch of these services.

01:09:18 ◼   ► Backblaze is the best one I've used.

01:09:20 ◼   ► I've been with them the longest.

01:09:21 ◼   ► I entrust them with my data.

01:09:24 ◼   ► With Backblaze, you can access your data

01:09:26 ◼   ► anywhere in the world, on the go, on your phone,

01:09:29 ◼   ► on a laptop, on the web, whatever you want.

01:09:32 ◼   ► You can even use it as not only just like disaster recovery,

01:09:35 ◼   ► but even if you just forgot a file on your home computer.

01:09:38 ◼   ► You can open up Backblaze while you're traveling

01:09:40 ◼   ► or whatever, on your phone, your laptop,

01:09:42 ◼   ► and you can go pick that file and just download

01:09:44 ◼   ► that file from Backblaze.

01:09:45 ◼   ► So it's great even just for practical uses like that,

01:09:48 ◼   ► not even including disaster recovery.

01:09:50 ◼   ► And of course, if you have a disaster,

01:09:51 ◼   ► cloud backup is what you want because it is out of your

01:09:54 ◼   ► house or your apartment or your workplace.

01:09:56 ◼   ► It is somewhere else offsite.

01:09:58 ◼   ► So the disasters that affect your stuff

01:10:00 ◼   ► can affect your backup as well.

01:10:02 ◼   ► If you already have Backblaze at home,

01:10:04 ◼   ► tell your SysAdmin at work to go to backblaze.com/atp also,

01:10:08 ◼   ► because they can hit a business backup link

01:10:10 ◼   ► over there on the top of that page.

01:10:11 ◼   ► They are for the same awesome service

01:10:13 ◼   ► with more admin tools using their Backblaze groups feature.

01:10:16 ◼   ► Backblaze, all of this is just gimmick-free, $5 a month

01:10:20 ◼   ► for unlimited, full backups.

01:10:23 ◼   ► You can plug in as many hard drives to that computer

01:10:25 ◼   ► as you want for $5 a month and they will back them up.

01:10:27 ◼   ► You can see for yourself with a 15-day,

01:10:30 ◼   ► fully featured free trial at backblaze.com/atp.

01:10:34 ◼   ► Go there, play with it, start protecting yourself today.

01:10:37 ◼   ► This is the kind of thing like,

01:10:38 ◼   ► it's like saving for retirement.

01:10:40 ◼   ► You wanna start this yesterday.

01:10:42 ◼   ► You don't wanna be caught without this if it's too late.

01:10:45 ◼   ► You want cloud backup and if you're gonna have cloud backup,

01:10:47 ◼   ► Backblaze is the best.

01:10:49 ◼   ► Go to backblaze.com/atp for that 15-day free trial.

01:10:53 ◼   ► Thank you so much to Backblaze for sponsoring our show.

01:10:55 ◼   ► (upbeat music)

01:10:59 ◼   ► - Greg Kolodiejic, he writes,

01:11:01 ◼   ► "Do you ever use higher than default scaling

01:11:03 ◼   ► "on retina screens?

01:11:04 ◼   ► "If so, when?"

01:11:06 ◼   ► So let me see if I can figure out

01:11:08 ◼   ► the right way to summarize this.

01:11:09 ◼   ► So since retina is by default

01:11:12 ◼   ► just kind of pixel doubling everything,

01:11:14 ◼   ► so you have two pixels worth of information

01:11:16 ◼   ► where one would have been before.

01:11:18 ◼   ► That's a summary, I'm sure it's somehow technically inaccurate

01:11:21 ◼   ► but you get the idea.

01:11:22 ◼   ► So what you can do is if you have good eyes,

01:11:24 ◼   ► which I do not, then you can actually go in some laptops,

01:11:29 ◼   ► you can go to the native resolution

01:11:30 ◼   ► such that you can get everything to be really, really tiny

01:11:34 ◼   ► and you can fit a whole bunch of stuff on your screen.

01:11:36 ◼   ► So I have terrible eyes and I never do that.

01:11:41 ◼   ► And in fact, I think on my,

01:11:43 ◼   ► actually maybe on my adorable, I do now that I,

01:11:45 ◼   ► no, I'll have to double check that.

01:11:46 ◼   ► But on my old work laptop, which is a 15 inch laptop,

01:11:50 ◼   ► I absolutely left it as the default

01:11:52 ◼   ► and that's because I have terrible eyes.

01:11:55 ◼   ► John, you now have a laptop, what are you doing with this?

01:11:58 ◼   ► - I do run mine at non-native res

01:12:00 ◼   ► because I never look at that screen

01:12:01 ◼   ► and remember my goal was to have my external monitor

01:12:05 ◼   ► be exactly the same resolution as my internal monitor

01:12:07 ◼   ► because I mirror them

01:12:08 ◼   ► because then every time I connect my external monitor,

01:12:10 ◼   ► none of my windows move.

01:12:11 ◼   ► So that's what I do.

01:12:12 ◼   ► My internal monitor, I run it at a non-native 1920 by 1200

01:12:16 ◼   ► and I just never look at that screen.

01:12:18 ◼   ► And that's exactly what my external monitor, 24 inch,

01:12:20 ◼   ► runs at and so yeah, I don't like non-native res

01:12:24 ◼   ► when I have to take my laptop to the meeting

01:12:26 ◼   ► and I have to look at the non-native res.

01:12:27 ◼   ► A, it's too small for me and B, it's blurry.

01:12:30 ◼   ► So it's not what I prefer,

01:12:32 ◼   ► but I do it at work for that particular reason.

01:12:35 ◼   ► - So it is worth pointing out

01:12:38 ◼   ► that the default scaling has changed.

01:12:40 ◼   ► The 2016s didn't actually change the number of pixels

01:12:43 ◼   ► or the size of the screens,

01:12:44 ◼   ► but they did change the default

01:12:46 ◼   ► to be one notch higher than native.

01:12:48 ◼   ► Rather than saying do you use higher than default scaling,

01:12:51 ◼   ► what we really should be doing

01:12:52 ◼   ► is do you use higher than native 2X scaling,

01:12:55 ◼   ► which by default with the new ones, you do.

01:12:58 ◼   ► So anyway, so on the 15 inch,

01:13:01 ◼   ► that would be if you use the one

01:13:02 ◼   ► that simulates 1680 instead of 1440,

01:13:05 ◼   ► that would be one notch up, that is the current default,

01:13:08 ◼   ► but the actual native pixels are 1440 times two.

01:13:11 ◼   ► I use the 1440 setting most of the time on my 15 inch

01:13:14 ◼   ► because even though things are kind of big

01:13:17 ◼   ► and kind of chunky, I like the scale that it's at.

01:13:21 ◼   ► When I'm doing work, like heavy work,

01:13:23 ◼   ► like Xcode or Logic, I will bump it up.

01:13:26 ◼   ► I use a utility called iFriendly,

01:13:28 ◼   ► that's E-Y-E, not I, you know the letter.

01:13:32 ◼   ► So iFriendly is, I've mentioned it before,

01:13:34 ◼   ► it's a utility where you can just assign

01:13:36 ◼   ► a certain combination of modifier keys

01:13:38 ◼   ► and then the up and down arrows

01:13:39 ◼   ► to next resolution, the up or next resolution, down,

01:13:43 ◼   ► so you can change resolutions really quickly

01:13:45 ◼   ► just with a key command.

01:13:46 ◼   ► And so I will use that to oscillate

01:13:48 ◼   ► between the super high one that Jon uses,

01:13:51 ◼   ► the 1920 simulated one if I'm doing

01:13:53 ◼   ► podcast editing in Logic, or the 1680 one,

01:13:56 ◼   ► which is the one step above default,

01:13:58 ◼   ► or one step above native, which is the new default.

01:14:00 ◼   ► I'll use that one doing coding, Xcode, stuff like that.

01:14:03 ◼   ► Or if I'm just doing like email and Twitter

01:14:05 ◼   ► and wasting time like that,

01:14:06 ◼   ► I will use the 1440, which is the native 2X mode,

01:14:09 ◼   ► where everything's a little bit bigger.

01:14:11 ◼   ► - Real time follow up, on my adorable,

01:14:13 ◼   ► I actually do have it set to more space.

01:14:15 ◼   ► So the default is 1280 by 800,

01:14:17 ◼   ► I have it set to 1440 by 900.

01:14:20 ◼   ► I guess my eyes aren't as bad as I thought, go me.

01:14:22 ◼   ► - An actual 2X pixels on that,

01:14:24 ◼   ► or something like, it's like the 1152 by whatever,

01:14:27 ◼   ► that's the actual 2X pixels on that screen.

01:14:29 ◼   ► - Yeah, 1152 by 720.

01:14:31 ◼   ► - Your eyes are bad though Casey,

01:14:32 ◼   ► because you couldn't tell it was non-native.

01:14:34 ◼   ► - Yeah, it's real blurry.

01:14:35 ◼   ► (laughing)

01:14:36 ◼   ► And again, I really, I hope so badly,

01:14:39 ◼   ► that when they do a major redesign of these laptops,

01:14:41 ◼   ► I really wanna see the 2X pixel,

01:14:45 ◼   ► actual pixel nativeness move up one step.

01:14:49 ◼   ► Like I want the 15 inch to actually have,

01:14:52 ◼   ► you know, 1680 times two.

01:14:54 ◼   ► I want, and you know, similar,

01:14:55 ◼   ► like I want the 13 inch to have 1440 times two.

01:14:57 ◼   ► Like I want them to actually move up a step in pixels

01:15:00 ◼   ► to match their current space setting defaults.

01:15:03 ◼   ► - Do they ever do that on the plus phones?

01:15:05 ◼   ► - Oh, that's a good question.

01:15:06 ◼   ► - The plus phones are still using their like 3X scaling

01:15:09 ◼   ► down to a 1080p screen,

01:15:12 ◼   ► but the iPhone 10 actually has those 3X pixels.

01:15:17 ◼   ► - Yeah, that's kind of a shame about the plus,

01:15:18 ◼   ► I was hoping that eventually there was like a stop gap

01:15:21 ◼   ► and eventually they would just do native,

01:15:22 ◼   ► but I guess they did, but just on the 10.

01:15:24 ◼   ► So maybe when they're all 10ish phones,

01:15:26 ◼   ► will we finally be done with that non-native?

01:15:27 ◼   ► Not that it's bad, it looks fine on the pluses,

01:15:29 ◼   ► but it bothers me at like a anal retentive level.

01:15:34 ◼   ► - No, honestly, I think like on the 15 inch,

01:15:37 ◼   ► you know, when you run it at the one step above

01:15:39 ◼   ► at the 1680 step, I think it looks okay.

01:15:42 ◼   ► I can usually stop noticing it after a little while,

01:15:45 ◼   ► but when I spent that, you know, day owning the 12 inch,

01:15:49 ◼   ► I never stopped noticing that the screen

01:15:51 ◼   ► just looked kind of blurry and crappy.

01:15:52 ◼   ► Like I think the smaller the screen is,

01:15:54 ◼   ► the more apparent that difference is,

01:15:56 ◼   ► and the 12 inch I think looks noticeably bad in that way.

01:16:01 ◼   ► - I am glad I cannot see it.

01:16:03 ◼   ► - Yeah.

01:16:04 ◼   ► - James asks, what tech related opinions

01:16:06 ◼   ► that you once strongly held do you now cringe about?

01:16:09 ◼   ► For example, saying something like the original iPhone size

01:16:11 ◼   ► was the optimum size, which I think I might have said.

01:16:13 ◼   ► - Me too.

01:16:14 ◼   ► - And the bigger Android phones were ridiculous,

01:16:16 ◼   ► which I think I also said.

01:16:17 ◼   ► - Me too.

01:16:18 ◼   ► - What else, or what other opinions did you have

01:16:21 ◼   ► that maybe you were wrong about?

01:16:23 ◼   ► And I was thinking about this earlier today,

01:16:26 ◼   ► and it occurred to me that not only do I have

01:16:29 ◼   ► a great opinion to share with everyone,

01:16:32 ◼   ► but I have documentation to prove it.

01:16:35 ◼   ► So when I was big into Tumblr back in like 2008,

01:16:40 ◼   ► and I at the time was running Ubuntu on a ThinkPad,

01:16:44 ◼   ► and I did a distro upgrade from like,

01:16:47 ◼   ► I got to give into Hardy Heron or something like that.

01:16:49 ◼   ► I forget exactly which one it was,

01:16:50 ◼   ► but everything crapped the bed, and I decided I'm over it.

01:16:53 ◼   ► And meanwhile, I was going back and forth