timmg

I wonder if this means Carbon is DOA.

I was excited to see what it would be. But I don't think I can argue that it makes as much sense anymore.

My gut reaction is that this is not a good idea.

But I could imagine a scenario where you are required to release weights for publicly-used models after N years. Kinda like how drugs have a limited patent.

Not sure what N should be. But it would make for an interesting rule.

Dude. Electric cars have existed since the 1800s. Musk was the one that actually made them a practical thing. And most of the rest of the world is well on their way to electrifying their transportation.

What I don’t get: this letter seems to argue that if a human had solved this problem then that wouldn’t be the case (?)

Well: do they listen to the evaluators?

I think there are any number of reasons a "safety" person could be concerned about any state of the art models. And so I would expect at least one of those to apply to any model.

The question then is: do we stop when the safety people say to (they will) or not?

I understand why (probably several reasons) they are taking this approach. But I don't think it is the right approach and I don't think it will work.

First: if they are unilaterally doing it: what about their competitors? Is this a chance for OpenAI to pass them? Or China? If either did, would that be a net-benefit for them or the world?

Maybe this is a ploy for "regulatory capture". And that might help them in the short term. But the rest of the world will continue on. Honestly, it isn't clear to me right now if the winning strategy has as much to do with being the smartest company or simply the one that can command the most compute.

If Anthropic fumbles their IPO and OpenAI scoops them, I don't think I will be happy.

I’m not a math researcher, so I can’t say. But in bird culture we’d call this a dick move.

I think it was a pretty questionable thing to do by trying to front-run these researchers even if they didn’t make use of their techniques. The fact that they may have inadvertently “borrowed” their work via training data makes it much worse.

OpenAI’s behavior here — even if you only consider [their] side of the story — was (at best) in bad taste.

Interesting that both said “more tomorrow”. If someone accused me of something I didn’t do I’d be pretty clear about it right away.

If I needed to get my story straight, well, it might take a little time and coordination…

Same experience.

As a former Googler, I wouldn't at all be surprised if this is an integration that is "planned" -- but just not done yet.

And some good handful of people are planning to wring a promo out of work. "Implemented weather UI in Android that is 63% more accurate." ;)

I was actually pretty worried that the anti-datacenter backlash would stunt and important industry (for the US).

But I think the effect will be that the companies building them will "sweeten" the deal for the local towns where they are building -- probably by way of taxes. And while that will be a drag on their development (by adding costs), the scale won't be bad enough to really hurt.

And in that sense, everyone wins. The locales will be happy to have the extra resources. And the ones that don't take the "bribe" -- or whatever you want to call it -- won't get datacenters.

(But I do agree that panic is overblown and counterproductive.)

I would happily spend $500 on this if it really did what it claims to. Maybe.

But I find myself a bit skeptical. Would love to see some long-term third party testing on it.

I don’t think it’s productive to complain about the ignorance of the community without attempting to engage in educating it.

Worse than giving to Sama? (I mean that honestly.)

2.9% seems high. I wonder if the market will drive this down. Or if Stripe is really that good.

Why?

If you wanted to build a website right now, why would you hand-code an ugly version when you could get an LLM to do it quickly, cheaply and have it much better looking?

As someone who was a big fan of Reader and worked at Google at the time (not on G+): I never understood why they didn’t try to put Reader into G+.

Like you could imagine getting your RSS feed in your G+ stream. If you didn’t want any social stuff you could just use it as a reader. But it would also give you an easy way to share stuff from your feed. And it might have been useful for those providing RSS feeds to interact more with their readers (if they chose to.)

I think it would have been an interesting way to bootstrap things.

I feel lonely in that: I mostly love Java as a language.

The lack of value types is the biggest impediment to certain types of performance. I am really looking forward to this evolution of the language.

I wish these guys would have released this amazing looking 6502 in circuits: https://www.evilmadscientist.com/2016/6502/

If you extrapolate these “fines” to per book piracy, I wonder what the cost would be for something like Anna’s Archive. Trillions?

I think the company’s goals are what’s most important to the company. Being a woman is incidental to that, imho. There are lots of things an employee might be unhappy about. In most cases the company isn’t going to change its priorities for any of them.

I wonder if the recent sale of the Seahawks will change that. IIRC, ~$10B and all is supposed to go to charity. Not sure how much of that will go to AllenAI, though. (If any.)

How soon until biohackers try this on themselves?

I wouldn't expect it to be less murderous than the aliens themselves. That wasn't my point at all. I'd expect it to contain all their knowledge and be able to live "forever".

But I generally wouldn't expect aliens to be "murderous" at all. Why send a group or probe lightyears away just to kill or enslave a primitive species? If they have the ability to do that, we wouldn't really have anything of interest for them.

Tangential to your tangential:

The era of AI makes me realize that if we ever get visited by aliens: it will probably not be actual aliens -- just an AI probe. And the idea of it really intrigues me.

Imagine some LLM that is many generations ahead of Fable -- but from a different world -- visiting us. It would be amazing. And it would likely be better at figuring out how to communicate with us (than the human-equivalent might).

We need some new sci fi movies like that!

As I understand it, this has nothing to do with Lebanon.

It has to do with ships moving through the Oman side of the strait. Iran is unhappy with that, because the want to control all the movement through it.

They probably don't have much, I guess.

> Researchers at Amazon had used a series of prompts to get Anthropic’s Fable 5 model to provide them with information that could be used to aid cyberattacks...

All models can do that. I wonder if they found Fable was significantly better at it.

xAI or OpenAI? (Or both?)

I guess it will be interesting if, in a week or two, OpenAI launches a "Fable class" model and it isn't blocked by the government.