Viewing profile — killerstorm

Are we sandboxing AI agent harness process, or the environment it executes commands in? Ideally, they should run in _different_ sandboxes. The environment might corrode the harness…

What if it puts malicious code into test file and you allow `npm run test `?

There are more than two types. Or, at least, more than one way to tinker with the code. Just like aleph_minus_one, I like to tinker at a higher level: find a more general way to so…

In a short story titled 'Glory' Greg Egan described process where hydrogen and anti-hydrogen are forged together into a thin needle: > The end point of this process of nuclear scul…

Google's revenue is 10x up in the last 10 years.

What's the obvious limitation? To Yann LeCun it was "obvious" that you can't build a world model from a text, so he predicted that even "GPT 5000" can't predict that an object sitt…

This is literally an opinion of one dude which is not backed by any kind of quantitative evidence. It's actually possible to answer this question rigorously: 1. Define a scientific…

The reason writing is hard might be that post-training pulls style into a particular direction. In other words, big labs are much more interested in making "AGI" than in making a g…

I think main reason people choose React now is "nobody was fired for choosing React". (That used to be a saying about IBM as a vendor.) People can build their career around React. …

Actually browsers already come with a minimal UI library. It's called HTML + CSS + JS. There's actually no need to make any wrappers - you describe what you want in HTML and make i…

Yeah it's also specifically trained to disprove Jacobian conjectures, and so on...

Well, if the developer read libngu's random.c he would have noticed that it uses rng_get() and that it's asking for MICROPY_HW_ENABLE_RNG (albeit the failsafe is defective, reading…

Yes, basically we just need really good, strong parentheses. E.g. see Yoshua Bengio "Scientist AI". Or multi-stream LLMs

I used ballpoint pens in school but I don't like the process: my hand often feels too tense, and getting handwriting legible is a lot of effort. Once I switched to fountain pens I …

What's "concern of imminent disaster"? Anthropic said their latest model is good at finding zero-days exploits. That's just empirically true. They aren't selling vuln search produc…

Yeah... There was no "reading" in our ancestral environment, it's rather unnatural for a human brain to focus on the page where nothing happens. Some people can adapt to that, othe…

You're confusing PR with marketing. Flaws they find aren't going to convince customers to buy the product. But they need to inform the public of what they doing as it's part of the…

Headline? It was buried in a model card. They just honestly report not-quite-incident because it's quite close to the incident OpenAI had. Nothing wrong with it.

It boots into BASIC which makes it really easy to learn BASIC. It feels like a programmable computer even to kids who know nothing about programming, as the only way to get anywher…

Yep. My father build a ZX Spectrum clone for me somewhere around 1991. Few years later he also got Commodore 64 as a gift from German engineers he was working with. I think both of…

The part of brain which cares about efficiency and waste have "atrophied" decades ago in most devs. More likely they just never cared... E.g. > A 1.6GB Docker image for a Node app.…