urbsgpw

Probably requires incredible amounts of energy to maintain an equilibrium like that over any prolonged periods of time. Maybe a feature of human societies that makes it improbable for any one society/state to dominate over longer stretches of time. Or it could be I just have the US, the UK and the roman empire in mind because I'd need more attention and I'm hallucinating the most probable answer based on my training data and your prompt. Very Cartesian, I know.

Yup, it seems unfathomable to some commentators that the likes of AfD or Le Pen or Bolsonaro would get popular support were it not for US big tech (interference). Which truly boggles the mind, it seems to be the same narrative that couldn't believe Hillary lost the election and it was the russians fault. Even if that were true, you should still review your policies and how you ran the campaign and learn something from it. Here it's like we're in a world with no reinforcement learning feedback loop.

I think if you are a partisan you're always going to find differences in the two scenarios which are meaningful enough to explain to yourself and others why your outrage (or lack of it) is warranted.

And here's the thing, I'm sure there are meaningful differences. But then again there's also loads of similarities. A similar thing on the geopolitical stage is the status of Kosovo. The moment it was accepted as a separate entity you opened a floodgate. President of Croatia had a very nice clip about it a while ago, basically accepting both things to be true at once: Kosovo is a state that we need to accept but we also need to admit it was illegally taken from Serbia. This sort of duality (what Nader called being able to walk and chew gum at the same time, although for a completely different situation, I forget what the topic was) is very rare.

On HN i made the mistake of defending, for lack of a better word, a certain tech oligarch who shall not be named where I eventually made a similar point to yours and got downvoted to oblivion.

I could understand the framing being: all censorship is bad (even though I suppose this view would be called childish by some). I could also understand the framing that there is too much concentration of power in the hands of these corporations (and then the inevitable cries to break them up or nationalize them). But, for example, just pointing out that nobody was against these same practices when they were aimed at the other end of the political spectrum (SF billionaires were the darlings of the Democrat and not Republican establishment) causes such outrage from what I would guess are generally well meaning and intelligent people, that it boggles the mind.

It really seems to be something in our species that turns us into crusaders for certain topics. I can very well imagine that this same biological(?) impulses, where you are incredibly convinced in your own morality, must have led to the spread of early christianity (and then to the same or even larger degree, Islam). And it has to come from a place of conviction because otherwise u cant explain the zeal. And anybody who isn't as convinced (maybe partially agreeing) gets dealt with viciously. I'm making Tom Hollands point (in a very erratic manner), that what we consider modern universalism (aka wokeness) is a natural continuation of christianity. I would just go further and say that I think there is something in us all (to varying degrees) that we want to spread a certain gospel of what it is to be a moral person. And if someone disagrees they are not moral.

Wait, so like, reinforcement learning for humans? I think you might have stumbled on to something here!

No but seriously, this. And a few comments above a commentator also mentioned on changing the training (again reeinforcing the LLMs to not seek behaviour like this) and obviously continuous work on harnesses (which I suppose, ought to be more paranoid).

It did look friggin cool tho:)

But I always preferred the F22. That thing is beautiful and the metrics are insane. But for a while it seemed like there was no peer contender and they stopped the programme.

Yup, while I'm sure that part of it is that the big players got used to this modus operandi, part of it it just your client not knowing what they want. I work for a financial institution, I see this everyday. Our IT partners just have to wait for our ill-conceived and ill-defined tickets and changes of priorities, keep track of the tickets and the money keeps flowing by itself. I know the analogy isn't perfect, but still.

Hats off to Anduril for shaking up the space. Wonder if the Swedes will be able to match them.

Huh, never thought about that 2nd point - I'm also transitioning to pi, but didn't think to change to neovim as well. I guess your logic holds for claude code users as well though.

But isn't it the case that we can't reach this safeguard with the current architecture? I remember Karpathy making an interesting point 2 years (cca) back, that I would summarize somehow like this: the mechanics behind every LLM answer are the same, what you then call hallucination is more or less a consequence of whether or not the answer was factually correct/useful.

Which would mean, as is so often the case, that the "killer feature" of the LLMs is also its biggest weakness and the two can't be disentangled. Now, we are inventive creatures and we might come up with a remedy for these issues, but what you basically see so far is more guardrails, the use of harnesses and building a whole bunch of infrastructure around the LLMs to get useful work out of them.

Which, btw is not a critique, I do it as well and it's a fun engineering challenge.

Maybe less shocking, but still, so is google. I know they are one of the labs, but that being said they rely far less on gemini for revenue as opposed to the big 2.

It seems like they're sticking to a static pricing plan that was made when the only relevant competition were the US labs (im not counting deepseek 2025 as serious competition -> glm and then kimi on the other hand, now that's a different story).

This was exactly my thought process. Coupled with a less idealistic one: mid 2025 if you looked arena ELO scores google seemed to be dominating and I was sure the trend would continue. But they really dropped the ball on coding and tool use.

That being said, controlling android and apple mobile devices is kind of a big deal. And their video models are still top notch.

Exactly my experience. I'm building an AI document-extraction platform, so I had to benchmark a bunch of models — on the cost:quality:latency:adherence picture, flash wins hands down for structured extraction. (Caveat: I've only tested the three US labs and Mistral.). So like u said, totally viable in prod for a relatively static tool. Didn't build the tool suite with gemini, but if you use service mode it currently mainly runs on flash.

Haven't done any serious coding work with the flash models though — but I'm seeing more and more HN comments from people who seem to have picked it up for that in the last couple of months.

Haven't been following the articles and snippets we get from these labs about training their models for a while. But I'm guessing the latest chinese models are way less based on distilling? If not, then your speed of progress is still limited by the two labs (which we are collectively, in various forms subsidizing).

I use hermes only ever saw their repo. Atrocious. I was sure you were exaggerating.

In other words, theory does not conform to reality, i.e. ugly data. Being an academically trained economist, not surprised.