Well, yes. But they didn't actually know that that was what the design implied.

Tests have shown a lot of AI actions that weren't expected ahead of time, but in retrospect should have been obvious.

This requires that you believe they're being honest, with no backing evidence. I can see possible evidence of consistency, but that's a much weaker claim.

If they were lies, then I agree with down-ranking them.
If they were just opinions, then I only agree with avoiding them.

OTOH, I don't use X/Twitter. Slashdot is as close as I come to social media.

I think I heard this story last year, or perhaps a few years ago. I've no reason to think it was the same plant, though.

We're also discovering unceasingly dim objects that are closer. They just don't get the same news coverage. But I'm more interested in wandering planets and planetoids (and even Oort belt objects).

Perhaps the answer is that the "Big Bang" wasn't the first, and these black holes are left over from a prior "universe". In that case, though, why are they so close together.

Yeah, my question is "What isn't prior art?". The summary sure doesn't list anything.

OTOH, it does indicate the direction he wants to push things. He wants these integrated into the lives of people whether they, themselves, use them or not.

Investors are chasing a company that achieves some kind of AI singularity. Let's set aside the fact that there's no reason to believe there is anything but diminishing marginal returns by making marginal refinements to current frontier models. Let's imagine someone hits the jackpot and gets, not even AGI, but a system that's as far ahead of today's frontier model are ahead of 2020's GPT 2.0 in performance.

Globally AI revenues are 150 billion, against a cumulative burn rate of 450 billion. A model that is a generation ahead of others would almost certainly capture the lion's share of that revenue.

If AGI magically appears as a Sam Altman has promised investors it will, a hundred million is way too low. Add, maybe, another zero to the revenues.

Conservatively, a safer assumption is that frontier models will get marginally better based on refinements in training and reinforcement and the other bits and bobs that go into these systems. The nobody is winning the lion's share of anything, at least overnight. But you have to define "safe". By "safe" I mean unlikely to lose money. But some investors are clearly defining "safe" as "having the greatest chance of owning a piece of the biggest thing ever."

I'm not following this super-closely, but if Anthropic is pursuing adding multi-step model based reasoning to their system, that could be the basis of a generational leap in capability. But if that is an approach that looks like it has a chance of working, then their competitors are no doubt pursuing the same thing. In that case you'd expect the revenue pie to grow as the scope of model utilty increases, but that growth to be split among several competitors. This could credibly result in a revenue stream for some of them that is as big as the entire industry's revenue stream today. But there's going to be hell to pay on the data center impacts end of things.

But that *does* require that you believe that they are honest to you about their future actions...and that they won't change their mind later.

Pirates were often hanged by the same countries that issued them letters of marque.

Why do you believe what they say when they continually misrepresent things. I'd say "lie", but that implies they know what they're doing.

If I read that summary correctly, this will only ever be usable to clean small amounts of stuff in a closed container. It's not going to work for, say, the Great Pacific Garbage Patch.

That's what it meant ever since RUR. I'll grant that PR people keep spinning different meanings for it though.

Please modify this to:
"""Whenever a "discovery" is announced it's best to wait for verification before popping champagne corks."""

If there isn't independent verification then it's poor policy to believe. It may also be poor policy to disbelieve.

Actually, I had measles twice before there was a vaccine. There are two different kinds of measles, but presumably the vaccine covers both.

Not necessarily. If they're designed to achieve maximum penetration, then you're probably correct, but if they're designed as networks of friends, and you can't invite whole networks, then it's a lot more manageable. Say that no group can have more than, O, 254 members, and you don't have the problem you posit. Of course, everyone would have trouble finding the groups they wanted to be a member of that also had room for them, but there are problems with every approach.