devnonymous

I agree, this is why I still use vi :-) Sorry couldn't resist

Coincidentally, I am reading 'Scale' by Geoffrey West where this concept of impedence matching is discussed (amongst many other things). It is fascinating how impedence matching was arrived at in biological systems viq evolutionary pressure to optimize energy usage.

Who was it ? The ^US government^ .. or for that matter, any government or corporation doesn't pursue litigation. It is humans who make this decision (at least for now ...).

Who in the government bureaucracy made the decision and followed through? It's sad that we've become accustomed to accept governments/corporations as entities by themselves when they clearly aren't.

tl;dr NP-hard isn't that hard if you relax constraints.

While not novel its a pity warrants a legitimate HN front page.

> For example, I worry that biology will have a strong attacker-defender asymmetry, where sufficiently capable models may be able to quickly weaponize pandemic-level viruses with widely available materials,

If he had just left that bit out it wouldn't be so obvious that he's just clutching at straws at this point. In some twisted sense it's almost sad to see.

Not with LLMs, it won't. (The probabilistic prediction property prevents proper persistent programmable predictability)

You're applying the doorman fallacy here. This has been discussed a lot. Honestly, the sooner the industry accepts that we aren't there yet nor does the current architecture enable us do get there, the sooner we can make progress towards that ideal state.

> We simply have better magic wands and more powerful spells now.

Wouldn't it be nice though if the incantation of the same spell would always do the same thing every time ? You see that's how my old wand and spells worked.

> In the new system prompt we say: Write code that reads like the surrounding code: match its comment density, naming, and idiom.

Hmm, so what happens in greenfield projects ? In any case, at least all the slop will be consistent.

tl;dr Since nobody really understands why or how any of this shit works, last time around we just threw a bunch of crap at it to see what stuck. Seems like we could just have have thrown less crap.

> So then it’s seems it’s either that this was intentional(ish) or bad security

My take - don't attribute to intention that which can be sufficienty explained by inexplainability or incompetence (although I doubt the latter).

> rocket engineers/scientists are in high demand with lots of likely sponsors for immigration

I wouldn't be too sure about the immigration part, when even a software engineer hire for a defence/defence adjacent job requires jumping through a number of bureaucratic hoops of security clearances. Even the companies with deep pockets don't always get the best people in the world -- they just get the best people that HR can actually hire.

The defaults seem sensible. This isn't keep-alive-forever by default solution.

From the README:

> Detect the danger window. When the main lineage goes idle and a subagent is actively running, the main prefix is at risk of expiring.

So, this is not demanding to be at the front of the queue, it's just paying someone to take the place you already had in the queue, when you want to take a leak.