krcz
krcz on Hacker News
By "doing that" I meant "training AI specifically at the art of killing people, like selecting bombing targets, flying fighter jets", which you alleged they were doing. As far as I know neither OpenAI nor Anthropic are training models specifically for such tasks.
It sounded like you knew for a fact that these companies were doing that and now it is clear you don't; they cannot stop something they are not doing.
I would like to see the reaction of the US government if they learned that a company was training a model to deceive and hide information from them.
Could they? Could you train a person to be great at math but bad at solving sudokus?
They could try to train models to pretend to be bad at that, but there are workarounds for that.
1. Sounds sensible in principle, but it's not clear who exactly should be prosecuted? The person setting up the sandbox, the person coordinating the evaluation run, the person who was designing the architecture, the leader of the team, the CEO? There is no precedence here, so it might lead to a multi-year legal battle – and a few years might be too long to stop what is coming.
2. Do you have any evidence or news articles that LLMs are specifically trained for selecting bombing targets? I thought it was the generalized reasoning capability that made them useful there.
The second part of the article states:
> Any such armed attack and all measures taken as a result thereof shall immediately be reported to the Security Council. Such measures shall be terminated when the Security Council has taken the measures necessary to restore and maintain international peace and security.
So triggering Article 5 means reporting an armed attack to the Security Council.
Interesting; he is clearly saying that they were ("as a field") "confident and wrong" to believe that "the economy would have been completely upended" by the appearance of a model as strong as GPT4. What seems to be missing, from your point of view, to consider it an admission of being wrong?
Wouldn't it be possible to use heat pumps, making radiators 100C or hotter while running GPUs much below that?
Or no one is to blame, if the mechanism of the regression is complex and non-obvious based just on the patch itself.
Regressions are bad and they should be avoided. Still, software engineering is a complex thing and regressions happened long time before coding agents were a thing. Unless one can pinpoint regression to changes that were more sloppy than the human-written rsync commits were I don't think coding agents are to blame.
I believe your point is not that it has never failed for anyone in the last few years after upgrade? Then, if the claim is that breakage is considerably worse than it used to be before using coding agents: it is possible, but I think it requires more evidence than a few anecdotes.
> Imagine math if we'd question the basic axioms.
No need to imagine, it's enough to look into non-Euclidean geometry (obtained by excluding Euclid's fifth axiom), non-standard models of geometry, or reverse mathematics (studying which axioms are necessary for a specific theorem to be provable).
These are not possible, as x can be 7 only if y is 7, and likewise for 8. Semantics isn't based on sets, but on non-deterministic assignments to variables: y is either 7 or 8 and x is either y or 2.
It's a social construct in the the same sense anything not directly verifiable using senses is. Is there an Eiffel tower in Paris? Most people haven't seen it, so they can only accept the social consensus that it is there.
If one can afford it, they can travel to Paris and check themselves. The same with mathematical truth: if one has means (time, intelligence, access to training), they can check the proof themselves. Otherwise they need to trust the consensus.
So again, is the truth in mathematics just a social construct? In some sense, I guess, but probably not the one some people might assume hearing such a statement.
For the people wondering what is that comment doing here: in Polish it means "what a mess".
TikTok is not good for getting deep understanding of things, but it is great for discovery. you can get familiar with the surface level so you have a better idea where you want to dive deeper
I like the idea - does it support .gitignore though? I wasn't able to find anything in the documentation and GitHub issues seem to be confusing in that regards. If it doesn't, it might be hard to use with tools that create temporary or output files in the directory, such as compilers, LaTeX, some editors, etc.
Are you sure that statistic wasn't for ventilators, not oxygen?
Only if by "work involved" you mean burning lots of electricity with expensive GPUs/ASICs.
I fully agree, thousandfold price increases would suggest that processes that control prices in free market system had become unstable. In such cases I'd agree with state intervention. Fortunately it doesn't happen that often.
> Also, this doesn't really apply to housing.
Correct, it's more complex there and some of the reasons are mentioned in the article.
When the price explodes, bakers may hire more people and bake bread during night and day, increasing the supply.
The manifesto seems to agree with that:
> Utopia is not a destination we should ever expect to reach—it is a point on a compass.
Maybe there just is no objective "right" and it's better to have two biased stories with diverse biases instead of one trying its best to be unbiased.
Zygote [1], started as bunch of ideas about programming languages, now it is in exploration/design phase, might end up becoming a new programming language.
Main goals: idiomatic translation (transpilation into readable code, with both imperative and functional targets possible, e.g. translating map-filter-reduce chain into a loop), restricted homoiconic syntax (with typed macros), advanced language concepts (dependent types, contracts, verification, effect systems) implemented as macros.
Target uses: GPU/heterogenous programming, neural networks model definitions, cryptography; everywhere where control is needed over how abstractions are translated into the low level code.
[1] https://github.com/krcz/zygote
My idea was:
1. Technician writes down their diagnosis
2. They submit it to the system
3. AI comes with its own analysis
4. Technician sees the outcome, they can update their assessment
5. Everything is saved into the system
If one of technicians has too much errors in their initial assessments, it should raise a concern.
While I don't think AI should replace humans in describing medical images, it can be used to check if they might have missed something. Such AI-based description should be provided only after the human finishes analyzing the image, to avoid lazy technicians just copying algorithmic output. The goal doesn't have to be increasing accuracy and not doing biopsies, it might be reducing number of false negatives.
I assume that if he was made to step down from the CEO position, TSLA valuation would drop like a fly - so he could easily re-buy everything he would have sold. I don't think that's the path he is going to take, but it's not impossible.
I wanted to write its market cap being 1% of the gold market cap. Then I checked and it looks it has already happened, and I still don't consider Bitcoin successful, just a little bit crazier.
I'd consider it successful if I could pay with it in physical shops, or at least in one of major online shops (let's say Amazon).
Very interesting project! That is the first attempt I've seen at something I call "idiomatic code translation" - transpiling with result not only to be executed, but to be read and naturally included in existing codebases.
I've had similar idea on my mind since few years, but as for now the only output is a document explaining the goals [1], I'm still in exploration phase, as I want something more abstract, with more advanced language concepts implemented using macros (as in Turnstile+ does with dependent typing [2]).
[1] https://github.com/krcz/zygote
[2] https://www.ccs.neu.edu/home/stchang/popl2020/index.html