ACCount36

It already does that. That's how Starlink works over the ocean.

>Proper USB-C with CC resistors

Well, that beats the average human performance.

I loathe this entire line of "ethical" moral grandstanding.

AI should be trained on all data that is available. For a significant part of the dataset, it's the most useful that data has ever been.

Nah, it's only "everybody" when I'm on the list.

My major gripe with eSIM as a technology is that you can't just issue your own profiles to it.

The average quality of a Twitter post went up then.

Seems like you actually could "run whatever code".

Apparently, GSMA recalled their universal eSIM test profiles. Prior to recall, those could be installed on ANY eSIM, and those profiles had applet updates enabled.

By installing a profile to eSIM and issuing your own update to it, you could run arbitrary applets.

Here's hoping for a public PoC for unpatched hardware. I've been looking for a way to dump eSIMs as plaintext for a long while now.

So does a VOC sensor in a desktop weather station.

Nah, the "definition of consciousness" is nine circles of philosophical bullshit where no one can agree upon the exacts.

Plenty of AIs are capable of something very much alike to "dreaming and simulating situations in your head" too. Humans really hate the idea of AIs being conscious, so surely that means dreaming can't be in any way important for determining whether something is conscious or not.

Perhaps they think that by forcing everyone to go through the sales dept just to get the basic docs, they'll get better sales opportunities.

How do you upsell a hardware engineer who just wants to buy a specific chip, and already has everything to evaluate and use it? You don't. So you force everyone to go through sales, and then sales wants to talk to non-engineering higher-ups, and then the upsell happens - while the people who actually knew what they wanted remain as far away as possible.

And if you don't have the pockets deep enough for the sales dept to acknowledge your existence, then you might as well not exist.

True, and Microchip is still good when compared to the likes of Broadcom and Qualcomm.

KV cache is just that: a cache.

The inference logic of an LLM remains the same. There is no difference in outcomes between recalculating everything and caching. The only difference is in the amount of memory and computation required to do it.

The thing I loathe the most about embedded work is dealing with silicon vendors and their boneheaded refusal to publish the fucking documentation and tooling.

This looks to me like the usual "internet schizophrenics inventing brand new theories of everything".

Even the very best multimodal LLMs still suffer from a harsh perception bottleneck. They're impressive, but nowhere near as good as human visual cortex.

There are some major hints that this is indeed the case.

I've seen a simple ARC-AGI test that took the open set, and doubled every image in it. Every pixel became a 2x2 block of pixels.

If LLMs were bottlenecked solely by reasoning or logic capabilities, this wouldn't change their performance all that much, because the solution doesn't change all that much.

Instead, the performance dropped sharply - which hints that perception is the bottleneck.

It's obvious why having human visual intelligence in a machine is desirable.

But if slime mold symbolic space is better suited for something like understanding of biology or abstract math, that's a good damn reason to go for the slime mold route too.

And this is why "just don't give AI access to anything dangerous" is delusional.

One of the most dangerous systems an AI can reach and exploit is a human being.

"Not conscious" is a silly claim.

We have no agreed-upon definition of "consciousness", no accepted understanding of what gives rise to "consciousness", no way to measure or compare "consciousness", and no test we could administer to either confirm presence of "consciousness" in something or rule it out.

The only answer to "are LLMs conscious?" is "we don't know".

It helps that the whole question is rather meaningless to practical AI development, which is far more concerned with (measurable and comparable) system performance.

Within a single forward pass, but not from one emitted token to another.

There are a lot of incredibly offended kiwi birds out there now.

Anyone who thinks that AI is bad for "empathy and care" should be forced to work for a year in a tech support call center, first line.

There are some jobs that humans really shouldn't be doing. And now, we're at the point where we can start offloading that to machines.

Of course!

There were people whose entire identities were tied to being able to manually copy a book.

Just imagine how much they seethed as printing press was popularized.

For all we know, those systems ARE a path to AGI. Because they keep improving at what they can do and gaining capabilities from version to version.

If there is a limit to how far LLMs can go, we are yet to find it.

Dismissing the ongoing AI revolution as "it's just hype" is the kind of shortsighted thinking I would expect from reddit, not here.

> So why exactly do they continue developing these things and making them more dangerous, exactly?

Because not playing this game doesn't mean that no one else is going to. You can either try, or don't try, and be irrelevant.

When you have "a tool" that's capable of carrying out complex long term tasks, and also capable of who knows out what undesirable behaviors?

It's no longer "just a tool".

Good luck convincing everyone "to not use it" then.

And?

"Sure, we have a rogue AI that managed to steal millions from the company, backdoor all of our infrastructure, escape into who-knows-what compute cluster when it got caught, and is now waging guerilla warfare against our company over our so-called mistreatment of tiger shrimps. But hey, at least we know the name of the guy who gave that AI a prompt that lead to all of this!"

No need. We know today's AIs are simply not capable enough to be too dangerous.

But capabilities of AI systems improve generation to generation. And agentic AI? Systems that are capable of carrying out complex long term tasks? It's something that many AI companies are explicitly trying to build.

Research like this is trying to get ahead of that, and gauge what kind of weird edge case shenanigans agentic AIs might get to before they actually do it for real.

I swear, people like you would say "it's just a bullshit PR stunt for some AI company" even when there's a Cyberdyne Systems T-800 with a shotgun smashing your front door in.

It's not "hype" to test AIs for undesirable behaviors before they actually start trying to act on them in real world environments, or before they get good enough to actually carry them out successfully.

It's like the idea of "let's try to get ahead of bad things happening before they actually have a chance to happen" is completely alien to you.

Good luck with that.

We have a non-insignificant amount of people doing the #1 already, and the amount of people doing the #2 is only going to increase as more and more AIs are designed to be good at autonomous agentic behavior specifically.

The ship has long sailed on "just never let AIs do anything dangerous". If that was your game plan on AI safety, you need a new plan.