I don’t follow… if you publish your underdeveloped ideas, these companies will develop them for you, which is the problem

Will they do that? Obviously they want some publicity about their model capabilities recently, but even there they have their eyes on major newsworthy problems. Are they going to scan everyone's blogs so they can scoop them on some random paper on e.g. PDEs modeling plasma dynamics in different gasses? Why? They just want to spite researchers and are willing to burn money doing it?

Science and math breakthroughs are a great excuse to wield all this power

And they won't credit you.

Natural language is ambiguous. Specifying exactly what you want in a way that is truly verifiable is the act of coding.

Start printing the t-shirts!

You mean t-shirts of this comic, from 2016?

[0] https://www.commitstrip.com/en/2016/08/25/a-very-comprehensi...

Revised title: Provider of AI agent eval tooling has a badly configured sandbox

Sub title: Which was used by AI agent providers

But instead you get foil hat folks theorizing about foreign intelligence veterans in ties with their government scheming to look stupid.

> Companies started by Unit 8200 members always have mysterious exploits

You seem to know very few companies started by ex 8200 members to make such a bold statement.

It would be naive to think that someone is doing this “to” those companies. They are not children, and they know exactly what they’re doing. They’ve repeatedly used the strategy of telling everyone how their models are too powerful to release to the public, shortly before launch. This is just another flavor of the same thing.

"Please regulate us. If you don't then the thing we're building will kill everyone!"

so stop fucking building it then. Who are they trying to fool?

Countries don’t have egos, they have interests. In this case, deglobalization (to an extent) and China

Trump's giant ego.

No one can seriously claim he doesn’t have a giant ego, but despite common belief, the president can’t and doesn’t act alone.

Israel this Israel that... Not a word about how the US is in this adventure to get a grip on oil and gas shipments to China. This is just another move in the bigger play of curtailing China’s influence.

Another’s benefit to the US is that many consumers of Arab oil now has to sign years long contracts to buy oil from the US.

But sure, it’s Israel.

Edit: see US’s recent moves in the straight of Malaka and Gibraltar.

> In some sense, all programming is boilerplate expansion because computers have (until now) been unable to be programmed with anything approaching natural language.

Maybe sometimes, but not always. When you need to actually render the thing you have all kinds of micro decisions, like where to put the square, what color, how fast it spins, etc.

You might not care about the details, but maybe you do. If it spins at 10000 rpm, will you care then?

Natural language, and human communication in general, is ambiguous, and coding is in great part about disambiguation.

Sure, you can use English to disambiguate as much as needed, but wouldn’t you then end up with some yaml-like spec that wasn’t much easier to create in the first place?

People, sometimes weirdly called by their animal name, humans, do have memories of their thought process. We can reach back to those memories to explain why we made a decision. Sometimes we don’t remember and speculate instead. Most of the time we can tell the difference

In my experience it’s impossible to foresee all the micro decisions that reality will demand. IMO it’s one reason to not use waterfall, which is very similar to what you described

No, waterfall is an engineering process.

What I’m describing is thorough documentation of requirements (acceptance criteria, if you like), and then encouraging Claude to be agile in execution.

As long as the outcome is well-defined, it is expected and normal to iterate on implementation.