From the article it sounds like what they're actually doing is reviewing LLM-generated code, for that you do need good software engineers.

Although it goes without saying that good software engineers won't enjoy doing this very much

There’s a meme in finance circles that by the time some trend appears on the economist front cover it’s a top signal and time to get out of the trade and this is why.

It’s a very normie magazine written by and for normies, and by the time an idea has got that far down the easy money has already been made

By the time it appears on People Magazine that must be the third or fourth dead cat bounce!

Acid is always orally active and tabs are tiny, it being hard to administer was never the issue

The DOM and JS are joined at the hip - the DOM APIs are designed assuming JS is used to access them, and the design of JS and some of its more “unique” features is partly because it was designed for use with the DOM

At the end of the day it's not really a decision at all, the government has a lot of men with guns, Anthropic has (presumably) low numbers to zero of men with guns.

You don't really have a choice if the government decides to play hardball

There are other things that can happen. For example, Anthropic could secretly transfer a copy of the model to a Chinese company in exchange for a large sum of yuan at a Chinese bank. Both parts of the transaction would be invisible to US authorities. I don't think this is a likely occurrence but if you think it's impossible then you need to think outside the box more.

Fair point, but doing that from inside the US under the direct view of multiple three letter agencies would be extremely risky to put it mildy

If it's lucrative but risky, it will happen.

Was going to say, I don't think Opus has really got much better in the last 6mo.

It just goes in cycles of being better and then being worse again, presumably based on how much Anthropic are having to optimise inference

Whatever their rhetoric they desperately crave respectability, making Israel a pariah state à la 1980s South Africa would hurt them badly.

Why else are they even trying to be in Eurovision and UEFA in the first place

They’re in UEFA because half of the countries in Asia (AFC) and Africa (CAF) wouldn’t play them. Same reason Russia considered leaving UEFA recently.

>They’re in UEFA because half of the countries in Asia (AFC) and Africa (CAF) wouldn’t play them.

And how's that a problem of UEFA?

Everything is israel is and always will be related to palestinians in some sense because it's being done on their land

Yeah is funny anthropic going overboard with "omg this model is so dangerous guys!!!" and then the US government going "okay... well, that sounds bad, let's ban it".

Serves them right

Should they lie and say the model is not dangerous?

If you're actually worried then do the right thing and don't make the model, or do make the model and admit you are doing it to make money.

You don't get to have your cake and eat it by making the (supposedly) world ending model but also getting on a moral high horse.

It's the hypocrisy and obvious mendacity that's obnoxious to me.

Having said that, OpenAI is just as bad (probably worse) and they're friendly with the admin, so this isn't really a case of any kind of justice unfortunately

False dichotomy. You can make a safe, aligned model that blocks dangerous actions, even at the expense of overblocking benign ones.

You can release the powerful model to companies that will use it to fix instead of break.

Once a technology exists it will be used for good and for ill, you can't create it and put it in a box forever.

If you really think a technology could be dangerous then the only morally correct thing to do is not help to create it.

They're not doing that, so either they're lying about it being dangerous, or they're lying about the fact they care (maybe a bit of both)

This is just a terrible view.

Nuclear can be used for great ill or great good. Today, most actual use is for great good.

By your logic we should have stopped at the discovery of fire.

Sam Altman doesn't really know all that much about LLMs, he's a sales/marketing guy, not technical.

So it doesn't really matter what he thinks

Those are the folks who run the industry

Except for the uncomfortable fact that he controls the salary and job status of the people who do know much about LLMs.

OK. So? Would you say Harry Truman was a nuclear scientist?

Franklin D. Roosevelt is a better fit for an administrative nuclear program "founder" analogy.

Truman was totally in the dark until April 1945 by which time the bulk of the PoC and weapons prep work was done and the project was running fully independently w/o POTUS involvement.

Yet the guys in the lab coats worked for him.

Not for the bulk of the Manhattan project and not all the people in lab coats .. the intellectual founders that repeatedly pushed for the project and demonstrated feasibility weren't even US citizens.

I picked the guy whose contribution was smaller on purpose to highlight the hollowness of the claim about “controlling the careers” of people who understand what’s going on.