cherioo

Apple may withdraw security feature, which is unclear what that achieves at all. Apple will not withdraw all sales in a country for privacy of that country’s citizens, as evidenced in China.

UK demanding to see private information for citizens in other countries though..

What hardware are you able to run this on?

They would run from Mac when nvidia offers a 128gb or 512gb laptop offering.

Until then it’s meaningless to even consider windows (laptop)

I think this is to make human /user experience better. If you use accessibility features, then user need to know how to use those features. Similar to another comment in here, the UX they shoot for is “click the red button with cancel on it”, and ship that ASAP.

I don’t have iphone 16, and this article puzzles me.

Is apple only using jxl for their “raw” camera capture, but not regular camera capture?? The non-raw use case seems to be the one that would have more impact to regular folks.

Why? Is jxl inferior to HEIC?

The azure post is very hand wavy, and it looks like azure isn’t even supporting redis 7, let alone redis minor version after the license change.

Nvidia margin is like 70%. Using google TPU is certainly going to erase some of that.

OLED uses PWM to modulate brightness, which makes some people uncomfortable due to eye sensitivity.

All phone OLED exhibits this, though iphone is one of the more egregious ones.

It’s amazing how this is now the news, not when it successfully lands!

Intel is removing SMT from their next gen mobile processor.

My guess is this will help them improve ST perf. We will see how well it works, and if amd will follow

I like your idea. But on the other hand, training an AGI, and then having a layer on top “aligning” the AGI sounds super dystopian and good plot for a movie.

This feels like rehashing much of the conversation about TikTok, and earlier Huawei.

It is a national security concern. Whether the cure is better than the poison only time will tell.

My understanding is that the low level APIs are different, CUDA vs. whatever Android provides vs. whatever Apple provides. However, higher abstraction like PyTorch may be able to target different platforms with less code changes.

Can some ELI5 how remote attestation is supposed to work? It feels like asking a remote endpoint “are you who you say you are”. What’s stopping remote endpoint always responding “yes”

Privatize the gain, corporatize the losses!

I don’t think that is anymore disincentive at the individual level. What’s lacking is high enough fine at corporate level so the corporate will more sufficiently monitor itself. Right now there’s little incentive not to cheat, much like the fight club ford quote.

The author is an engineer at AWS so he has an obvious bridge to sell.

That said, I do agree with the general point of needing more than one computer. But most people probably just need two, one for backup, far from “distributed system” in the traditional sense.

There are several way different groups are viewing right to repair

Consumer: right to repair means fixing my broken display will be a DIY job for $50? Sweet!

Repair shop: right to repair means I can source a display from lowest bidder, charge $150 for broken screen, and make $120 in profit? Let’s go!

Apple: right to repair means you must buy $275 display module to fix a broken display. So we can keep some nice nice profit.

Samsung seems to be there with Apple. Looking at the price of display module I am worried iFixit is also there with Apple. And iFixit and Samsung couldn’t find a good split on who gets to keep the profits.

It is crazy that an iPhone display module from iFixit appears to cost more than Apple 1st party repair.

It’s 3:2, which historically people said it’s “better” for coding because it allows more vertical space.

Personally I find bigger monitor available today is more than enough vertically. Horizontal space is actually quite useful to be able to put things side by side.

How is ruining and destroying the value of Sprint, that SoftBank owns, possibly good for SoftBank?

What is SoftBank to gain here from enriching TMobile?

While FPGA may prove more efficient than 3090, a primarily gaming card, I can’t see how it should be more efficient than dedicated training/inference card, as the latter are more effectively ASIC, not to mention memory and bandwidth limitations.

Is there something I am missing making FPGA potentially more viable, besides not feeding into NVIDIA’s greed?

It is not just democracy. All governance systems are susceptible to mass communication tools that can change public opinion, be it books, radio, tv, news organizations, or tiktok. It is why many non democratic countries do not have freedom of press, and a highly censored internet. This is a main contributor to Arab Spring.

It is just sad that even US is susceptible to this.

But why is that somebody spending even 40k on a barebone car?

How is it that BYD and a bunch of chinese manufacturers able to produce ~10k car but nobody else?

Pushing salary lowers help the society at large, or at least that’s the thesis of OP. While it sucks for SWE, I actually kind of agree. The skyrocketing of SWE salary in the US, and the slow progress US is making towards normalizing/reducing it does not help US competitiveness. I would not fault Meta for this though, as much as US society at large.

SWE should enjoy it while they can before salary becomes similar to other engineering trades.

Genuine question, why don’t publishers stop using doubleclick and build their own?

Wow I never made that connection.

There’s plenty of competition to netflix.

Tiktok is probably the biggest competitor to YT. But it had to come in from short form video angle, because the moat of YT in long form video is probably insurmountable. Its fate remains to be seen.

Nvidia was forced to do that because it was so early, they had no customers except PhDs. I see no evidence AMD wants to do that right now, and instead focusing on extracting value from deep pocket enterprise customers.

The way things go, I think the AMD consumer card experience will only get better once AMD manage to gimp consumer cards’ ML throughput or RAM.

Those four things have almost always been replacement in many it not most thinkpads. This feels like pure marketing spin.

The newly developed housing isn’t meant to be the one that is cheap. What gets cheaper is the old housing people left behind to get into the new housing.

And if a city cannot build as fast as people flowing into the city, then there will never be “housing left behind”.

Huawei has created the “ascend” series ai chip. Reportedly it has 80% of A100 in theoretical performance.

Rumor is that they had trouble selling it against A100 due to worse price to performance and worse software integration. But the US sanction has now created a huge market for it.