Every time I bring this phenomenon up, I hear from HN commenters that it used to be like that, but today it is solved by modern powerful models, harnesses.and workflows. I don't know. At work I still direct and review everything LLMs generate, because if HN are wrong and I am right, I am in huge huge trouble.
Not sure about that. Writing detailed specs you deal just with the inherent domain complexity. Writing code you also get accidental complexity that has nothing to do with the business, whuch not only piles on top, but tend to interact vexingly with the detais of the domain.
Where you would have a point is if you'd say we have excellent tooling for wrangling code, but less tooling and tradition to write and manage specs.
AI is a computer program. It calculates numbers from other numbers. By itself it does not "want" to do anything and "cannot" do anything. Before it becomes an agent in the universe (in the classical meaning), it requires being supplied by an execution environment, energy, initiative (agentic loop, specific instructions), and modality (readonly and mutating connections to real world). It is like a game of chess - it does not exist just by itself: someone must play it, having the board and the energy to do so. With the huggingface incident the AI was supplied with all of these components by humans before it broke out. So unless humans are actively involved, I so far cannot see how AI can become truly autonomously agentic and start doing anything on its own, thus posing danger. I could be wrong of course, but I do not see it for now.
You can say "yes and you have to fear the humans weilding the AI" - that I agree with.
Humans are bioreactors. They only turn one organic matter into another. By itself they do not "want" and "cannot" do anything.
They do not exist just by themselves. Some bacteria in the gut must provide them with the energy to do so. So unless bacteria are actively involved, I cannot see how humans become truly autonomously agentic and start to do anything on their own.
> how LLMs can develop new previously unseen biases for demographic groups, even when there are no differences between groups in the first place!
In real life there ARE differences between groups, and models are trained on real life data, so I do not find surprising that models anticipate differences in this synthetic situation as well, and fail to see the significance of this result.
> X. Dev/prod parity Keep development, staging, and production as similar as possible
Notably, there is no requirement or recommendation that the dev environment be a single, shared environment. Development processes where this environment is single is shared is as terrible as it is ubiqitous.
Person outside the USA here (reasonable, hopefully).
First off, it is too simplistic to try and categorize a political regime as "ok" or "not ok". A political regime, the problematics of a large country, and and a huge voting base of people like the US population, are an extremely complex phenomenon. One will not reach useful insights by labeling a regime as "not OK" and its voters as unintelligent or evil (I am not saying you are doing so, but lots of people do).
Talking about reasonable people: for reasonable people, the Trump regime, and any regime for that matter, is a mixed bag of things to like and dislike, to weigh and trade off. For sure, for each reasonable person the scales will tip in different directions. But for a reasonable voter this not a sports game where one roots for their team just because it's their team.
For many reasonable voters Trump is not "ok" or maybe even "good", but he still responds to some of their pressing priorities and concerns enough that he merits voting for. I personally voted for politicians in my country that I disliked deeply, but I'd done the calculus and concluded that the alternatives are worse. Similarly, you shouldn't conclude that the electorate of Trump predominantly loves him. You hear mostly from the loud MAGA crowd, but they do not decide.
What are the concerns and priorities that may force a reasonable person to hold their nose and vote for someone like Trump? I wouldn't like to engage in a political debate here, so I will keep things meta. But as an outsider I can see political and social trends of the last decade that are quite prominent, rather sudden, pretty radical from my vantage point, and concerning to conceivably reasonable voters enough to tip the scale, even if you or I may not really align with them specifically.
> Being poor has zero correlation with being “lazy”.
This doesn't sound right. If that were literally true, and the level of industriousness didn't affect prosperity at all, why would anyone bother to work?
Health improvements are not measurable any more if you go below 5 mcg/m2 average level of PM2.5 particulates. For 50x reduction to be beneficial, you need to live in an area where the pollution is 250 mcg/m2, which is an extreme level found only in a few unfortunate places on Earth. 5-10x reduction is reasonable almost everywhere.
I know precisely the subjective difference between a 5x reduction and a 50x reduction. Only the latter feels sufficiently good to me, and my brain operates better with it. You will know it when you experience it. Maybe it's my middle-aged brain that needs the extra support. I live in an urban area near traffic. Lately there intermittently also has been wildfire smoke coming down from Canada.
I believe that any simulation of consciousness is consciousness, so ethical implications apply. But I also don't believe fruit flies are conscious for any definition of that word.
reply