Note that the $1B figure in the headline comes from a different article [1] which references a podcast [2] where analyst Jay Goldberg estimates that the tape-out costs would have reached $1B for the M3.
This is most Apple rumor “news” articles. A sentence worth of questionable info was relayed from anonymous source A to source B to outlet C. Outlet D sees it and expands that sentence into a thousand word article with filler and old info.
My best friend’s sister’s boyfriend’s brother’s girlfriend heard from this guy who knows this kid who’s going with a girl who estimated that the tape-out costs would have reached $1B for the M3!
Off topic / this is the kind of website I immediately could not care less about the content. They immediately inundate with ads while hiding information behind 2000 words of fluff. The internet is ruined.
This is what I hope AI/LLMs will fix, having a browser or alexa/siri like experience that reads the content from the web and removes all the junk. As an added bonus it could correlate and collect informations from different sources and display it in a personal yet standardised way.
To me that is the evolution of the internet as I see it currently. No more search, point and click ad driven nonsense but a personal research bot that saves you time and makes your life richer. An evolution sort of like the move from mircofilm to digital records.
It will be interesting to see the interplay between the corruption by human greed and bias on these bots and the logical basis that builds them. Hopefully logic will thrive over time.
May I suggest uBlock origin. If you are on a platform that does not allow effective ad blocking (chrome, safari, iOS) then consider switching to Android/Firefox.
How do you do that? I recently started using a Pi-hole for network-wide ad blocking - with outstanding results - but that obviously only works on my home network.
Conversely, this site actually does have an EU compliant cookies banners with a single "refuse consent" button that's prominent as the "take all my data" button. In fact, the refuse button is perhaps more prominent then the accept button.
It's perhaps slightly annoying as it come up with a delay that kicks in after a scroll, but on the scale of usual purposefully-misleading cookie consent (mis-)handling via dark patterns, it's not bad at all.
I know a guy who once worked on Apple Silicon and then moved to Google to work on their CPUs. Something he told me really struck me. He said he liked Google because they’re really cool and he told me that he liked that they had this thing called a “blameless post mortem” (very familiar to us software folk) and he thought it was cool.
He said that when they missed a deadline for tape-out, there was a blameless post-mortem and it did a good job of identifying the problem. And it was quite effective.
I asked him why he was surprised. Was Apple different? He’d been there almost a decade “Oh, I don’t know if there was a process. The idea that we could delay tape-out was unthinkable”.
That stuck with me for the difference between the two organizations.
> The idea that we could delay tape-out was unthinkable
Apple's insistence on releasing updates/upgrades to everything once a year will inevitably bite them in the ass.
In the past few years they've already started slipping deadlines for software (updates to iOS/MacOS in early fall, some advertised software late in the fall, or next year). It's only time before this happens to hardware.
Their failure, or inability, to release tick/tock updates on Intel, was surely biting them in the ass. How much revenue was lost by not updating the trash-can Mac Pro for six years?
There's a difference between regular updates and major yearly releases with unskippable deadlines.
Also, trashcan Mac was Apple's own self-inflicted damage, and nothing to do with Intel.
A lot of the issues with lagging updates to anything that's not iPhones are also precisely because there's no time and non resources to focus on anything but the next release.
So the way I read this is that it’s not just about the performance / power efficiency improvements, but also so that Apple can be the first to actually work with 3nm technology, ensuring they’ll stay in the lead in terms of experience with these cutting edge technologies.
Seems like the latter may be even more important than the former in the long term, also in terms of attracting talent.
I'm not sure that's accurate. It seems more like Apple is allowing TMSC to be the ones that get to stay in the lead and develop experience and that that's a cost that Apple has to pay in order to be first to market with this new technology.
Perhaps Apple is lending TSMC large amounts of cash to buy equipment, plus signing supply agreements, plus giving Apple access to first tapeouts.
Apple has cash and needs to invest it, and Apple doesn't want to bring cash back to US (taxed) and the loans would stay on Apple's balance sheet as assets?
I wish this article would point out the actual cost to Apple as a line item to their overall mac business rather than total revenue for the company. Or at least reveal what 5 and 7nm tape-outs costed. Some googling suggests about $40b in 2022 for all mac revenue.
The Mx line evolution is also used in he iPad and also the iPhone, even if less directly. The overall revenu hit may not be perfect, but better than only the Mac line.
Deeply confusing comment you’ve made. I don’t know where to even begin other than to say “that’s just not really how people talk” and indicates poor style and very strange argumentation.
You could get rid of the first 4 paragraphs and still get the same amount of salient and relevant information across.
I think the article is wrong to say that Apple said using N3 over N5 gave 30% boost. The said M3 performance cores are 30% faster than M1 performance cores. Some part of this speed-up comes from the smaller node.
I see headlines like this and my first thought is how Elon Musk spent $43B to force people to read his bigoted tweets instead of literally anything else.
The gigantic improvement in performance over the previous gen M chip is so huge that it would have been worth it even if Apple had to pay every last dollar of its reserves.
> Quite true, quite true. Apple is indubitably the creme de la creme, in a league of its own, miles above all. Who can possibly put a price on being first to a nanometer? It is worth any cost to obtain pure silicon perfection. The knowledge obtained from pioneering this new nanometer will unlock infinite possibilities for years to come. The future of the spaceship campus is very bright indeed.
The luxurious silvered and brushed, bespoke titanium surfaces glinting with speckled shadows of the grand dynamo within, tantalize all Is with eyes.
As we enter into the silicon's innards, we see the sleek, chic, perfectly placed caches, with an arc of 137.5 degrees, the golden ratio. As we zoom further, we see the ultra refined dielectric gaps, for the electron's pleasure. Lastly, notice the shimmered reflection off the single drop of isotonic natrian water. This is the tear of the worker that makes 3nm go boom boom. Shit, guys, did I drop the last part?
So you're suggesting that you get to decide what goes where? But you don't! Hahahaha :)
And your comment is more like Twitter than mine. Although if you're saying that "This is not Twitter" as in "This is not a place where you can speak your mind freely" then I disagree, you most certainly can!
It may not be liked by the community, as mine wasn't above, but that's OK. My comment was a genuine expression of the beauty, the majesty and fantastic accomplishment that those chips represent.
On another day, on another thread -- on another part of this thread -- it would have been a completely different story. Lots of upvotes instead of down. That's just the internet. So you don't get to decide quality based on votes.
But, I guess, if that had happened, you wouldn't have had anywhere to stick your little nasty, fake-deciding-for-others, would you? Heh :)
I hope you'll reflect and make better comments in future! hahahaha! :)
[1]: https://wccftech.com/m3-tape-out-costs-alone-cost-apple-1-bi...
[2]: https://d2d.substack.com/podcast