It would be very easy to write a script that burns tokens when your leaderboard position changes. The problem is competition, which if scripted, would result in tokens burned going through the roof at an accelerating pace.
Naturally, even with people, that will still happen. It’ll just be a little bit slower instead.
Eventually they will run out of money. The real question is whether the planet survives while they drain their bank accounts.
That codebase must be abysmal.
Well I use AI every day in Photoshop and Lightroom. AI tools are common and extremely useful in all sorts of media production already.
Photoshop used AI long before generative AI took off. Specialized models have existed for various domains for decades. This is unrelated to the current bubble.
Science is using it on modeling of protein folding, and large dataset analysis. I personally know one person using AI tools to analyze fMRI data in a study.
Science used AI long before generative AI took off. Specialized models have existed for various domains for decades. This is unrelated to the current bubble.
News media uses it in formulaic articles in finance and sports. They’ve been doing that with specialized software for a decade or more already.
News media is also dying. It’s saturated with low quality clickbait, and most major news sites are barely worth a mention anymore. Not only are the writers losing their jobs, but the businesses themselves are being bought out by larger investment companies and being turned into tabloid clickbait, propaganda tools, and listicles. I wouldn’t expect most of them to survive past the bubble, and that even has very little to do with generative AI anyway and more to do with a cultural shift in how people receive and consume news.
AGI wouldn’t be about a model that knows everything about language and every advancement in every field, but rather a model that is better than humans at finding solutions to problems
A LLM (or any other kind of model) that cannot adapt to changes in a field cannot perform better than humans in that field after that field experiences significant changes. Any such model would eventually degrade in output quality over time.
Also, AGI (artificial general intelligence) usually refers to an AI capable of performing all cognitive tasks at least as well as a human. It’s as much of a buzzterm as “AI” is, of course, so there’s an endless number of definitions for it. Such an AI should be capable of, at minimum, adaptation over time.
The point of a “smarter” model is not that it knows all the facts, that would be wasteful as it is trivial to look up facts at inference time.
An omniscient model would be impossible, but that’s not what I was referring to at all. LLMs these days fill their context windows with relevant information through careful prompting, tool calls, and so on. This is generally how a model is supposed to adapt. Context windows are bounded in size, though. It would have an increasing amount of information to include in that window over time, meaning the amount of data it needs to fit in the context window is unbounded.
Unless someone creates a LLM with infinite context (which would require infinite VRAM), such a LLM can never exist. Therefore, a LLM trained today will never be equivalent to (or better than) humans at all cognitive tasks for the entire future of humanity. There will always come a point where such a LLM’s output quality degrades, and it can do nothing to resolve that.
Edit: Here’s a simple example: a new written language emerges with all the complexities of a language like English. Humans can learn that language and communicate in it. A LLM cannot.
While I’m interested to see the proof, it’s more of a formality. It doesn’t take a PhD to ask what happens when the “AGI” LLM is trained on out of date information. They don’t learn over time, and they have a limited context buffer. At the very minimum, it would run out of context just keeping up with changes to spoken language over 30 years, let alone advancements in fields, new fields, and so on.
What you’re referencing is distillation. Anthropic even has an article on distillation “attacks” (as if they have some divine right to the data behind their models) that goes over it a bit.
Are any of the reported polling numbers reliable? I’m not too sure who VTsIOM is, but if they’re state-owned, then I’m inclined to believe that anything they say is bullshit.
recovery email which they did not hash
How do you recover an account on the other providers? Do you have to provide the same recovery email you set before during account recovery? If you hash the email, you have no way of reading it anymore, so someone has to provide it to you again.
I’ve had laptops where this is a BIOS setting. A simple switch on the side of the keyboard to flip which keys are on the fn layer solves the problem too. Same with key remapping, for keyboards that support that (and really any nice keyboard should or it’s not worth the cost).
In good news, we got a summary of his 9950x3d2 review, which was basically that it’s a ripoff at $900. Unironically, if you’re somehow in the market for a CPU like that, consider either the 9950x3d for productivity and core count, or the 9800x3d for gaming. The 9950x3d2 brings nothing to the table for anyone outside of maybe some niche applications which need both core count and cache size and can afford the latency for data transfers between CCDs.
Or, I guess, don’t buy anything because all the companies suck and everything is unbelievably expensive. Who needs a computer anyway?
Silicon Valley owes a moral debt to the country that made its rise possible. The engineering elite of Silicon Valley has an affirmative obligation to participate in the defense of the nation.
I agree with this. They have a duty to permanently rid the world of these sick fucks, especially for the sake of defending their country from them and people like them.
Here’s hoping John Apple does what it takes to make me actually consider an Apple device! Like, I don’t know, making user-friendly decisions without the EU getting on their ass about it first.
My phone’s keyboard lets me compose multiple dashes into an emdash. I believe I can also bind a compose key on my desktop, though I haven’t needed it there yet.
This (the audio, which is from Portal 2) is what I’m referencing by the way. I like that one too though!
The Hill is reporting that Iran is saying the Strait “will close again”.
Edit: other sources are saying essentially “… if the US blockade continues”.
Edit 2: And now it’s closed again.
Before the crisis, more than 130 ships a day travelled through the strait, but this has reduced to a trickle under threats from Iran’s Revolutionary Guards.
Feels like one other party might also be responsible for blockading the Strait. I can’t quite put my finger on which, though.
The benchmark European gas contract fell by about fell by about…
Lol.
We’re only seeing the start of this transition, and so far it seems to be going in a good direction. Here’s hoping his government does what he’s saying and actually builds a media to serve the country rather than just himself. It seems like that is most likely the case, which is awesome!
Life gave him lemons. Seems like he’s ready to make life take the lemons back.
They’re able to hit the ground running.
In fact, it often has nothing to do with what’s on the canvas, and entirely to do with the context the art was made in.