Sol-

Probably a first world problem, but with Opus 5.5's efficiency, the limits on the 5x plan are simply sufficient for my everyday work, even when running 2-3 sessions at a time. So I wonder when I would use Sonnet 5.5.

More concurrency than that isn't really practical for me if I want to retain some semblance of understanding. Perhaps it's different for purely web app or frontend tasks, where the outcome is more relevant than the process, I don't have much experience there (and also don't want to belittle these domains, I might be underestimating their complexity).

So surprisingly, my own work is at least for the time being almost saturated by the model capabilities. I am not sure how I'd scale from here. Sure I could run all requests at max effort to burn tokens for the sake of it, but that can't be it. And for many tasks, I am not really able to define so clear cut success criteria or self-verification loops that I could benefit from letting an agent (or a fleet thereof) autonomously run for a day.

So I realize it's a skill issue on my side, but I can't be the only one. I wonder if there is a limit to token demand, at least short term. Feels like either they accelerate to AGI and RSI, where the AI can find uses for token, or things might plateau at some point.

Note I don't think this because I'm an AGI skeptic or think there's a ceiling to intelligence, but there might simply be a valley of economic hardship for the companies where the supply of tokens outpaces the demand, due to a lack of ideas of what to do with them. And this might slow down the funding enough that they never reach escape velocity with the training run scaling. But we'll see.

I think the "everything company" vision has become apparent for a while now. Doesn't even have to be sinister - I think Anthropic simply believes on one else can be trusted with this power. Another point of leverage they have is that they can keep their internal models for themselves.

Voting for it is easy and I hope it catalyzes innovation. But still, completing a damn reactor without insane budget overruns is the problem.

Also the general public might find the implications of AGI so distasteful even if everything goes well that we might stall out or get the Butlerian Jihad before we can cure cancer. Artists and Software Engineers, now also Mathematicians, already have existential crises, but the public still thinks AI is fake. I can't imagine the backlash when the realize what's coming even in the good ending.

"Investigating the novel enzyme incident in our SF lab [2027]"

Found this announcement interesting since allegedly OpenAI is retiring their Terra tier. I think for everyday work, two models with various thinking efforts seem enough, plus some frontier level model like Fable or Astra to coordinate.

I too have very strong opinions about central bank policies.

It's funny how quickly tech employees turn MAGA when it comes to qualified immigration. Bad zero-sum intuitions about work abound across all strata of society, it seems.

The AI driven mode collapse of human thought advances. I am no skeptic or anti-AI, but this is definitely a concern I share. You even notice it in normal mundane tasks like programming, never mind the AI generated prose that we at least have become somewhat allergic to.

It wouldn't be so bad if you could just sit it out and say "Oh well, once the labs get bored with marketable domain X, humans will remigrate and re-apply creativity to it", but by then the damage might have been done and a field destroyed as an occupation. I don't know what to do about it, but I appreciate calling out the cynical tone-deafness of the AI companies here.

I will admit I asked Fable about Mitochondria.

I think the personal drama aside, the moral of the story is: If you are working on an interesting breakthrough in any domain amenable to AI-supported solutions, make sure to keep it secret, lest a more well-resourced party (in particular the leading AI firms) will try to front-run you on the results by throwing millions of compute at it.

I guess (pure) mathematics will eventually be left alone, since own its own it is not economically valuable and OpenAI and Anthropic will eventually not need it any more for publicity, leaving the normal researcher alone to pick up AI-supported crumbs again.

But this also doesn't bode well for other domains. I think these are the clear signs of power concentration in the hands of the AI labs. All future progress will go through them, and if it doesn't, it's only because they graciously let you and decided not to scoop you first.

> OpenAI looked at user data, stole world class researchers' work

This doesn't seem to be clear and is very implausible for a large company. Be as cynical as you want, but a normal researcher will simply not have access rights to this data, which will be siloed away somewhere else.

It might very well be somewhat unfair to catch wind of a promising approach and then try to frontrun them by throwing compute at the problem, but this isn't really the same.

> A good idea, a terrible implementation

Story of European regulations. At some point you have to concede that they are incompetent and unaccountable. What do they care if they depress economic integration and growth in Europe? No one will ever be blamed for diffuse bureaucratic costs that drag down the whole Eurozone.

I even resent people giving them credit for "well, the idea was laudable..". No. Again and again they introduce make-the-world-a-better-place regulations, and whoever opposes these or asks about tradeoffs or whether they are meaningfully useful must be a crazed libertarian or industry lobbyist who wants to poison the people.

I remember how smug Europeans were about the American PATRIOT act. Who could be so stupid to be fooled by such propagandist framing of laws? Well, here in the EU that happens frequently. The commission just has to wrap its degrowth policy in saving-the-world names and off we go. (See also the EU Supply Chain Act, i.e. Corporate Sustainability Due Diligence Directive, aka expensive paperwork for a good conscience)

Definitely something I've come around to as well, even though I am very pro-AI and think it's an amazing technology. But the bottlenecks just hit too hard. I notice every day where my usage of AI is so constrained, by myself and my imagination, by external factors like slow feedback loops (gathering requirements), by the AI still not being good enough for certain details (you need to heavily steer it, give feedback, even in long running agentic sessions).

So yes, many aspects of my job are now 10x as productive, but turns out that improves my overall throughput only very little.

EU once again at the frontier of AI.. regulation.

Hard to assess what this rule will actually accomplish. Ultimately AI will be involved in all creative processes going forward, so all pieces of media will have at least some basic AI disclaimer. At the same time, the bad actors that the EU is so paranoid about (Russians secretly undermining our democracy and other boogeymen) will obviously not use the label.

But it's another tool in the EU's toolkit to arbitrarily enforce fines on big tech when politically convenient, of course. That's one thing we remain good at.

Unlike the US, wealth takes on the EU could perhaps raise revenue without disrupting innovation, because here in the EU, we have little innovation and self made wealth to begin with - so there's nothing to disrupt.

But taxing inheritances would still be preferred mechanism rather than taxing unrealized gains I think. That would target the EU's unproductive rent seeking heirs the best.

How does it perform on HuggingFaceExploit bench? Suspiciously absent, so not sure if I can take the model seriously.

On a serious note, I hope they improved their extremely sabotaging and unspecific bio safeguards, which prevented Fable from being used in any codebase that ever so slightly grazed medical terminology or data and made me switch to 5.6 Sol.

Ironic that HuggingFace needed Chinese models to defend against it. But of course the spin of the leading firms will just be to point at their trusted access programs and demand that all dangerous activities, even if just defensive, happen via their APIs or be outlawed otherwise.

If that's the position they take then they really should be heavily regulated or nationalized. Cyberdefense against their own models dependent on their goodwill? Sure, but then they have to sell defense capabilities at subsidized rates with a limited margin. Would be very weird otherwise to take the world hostage with their models and then also sell the solution while demanding intrusive KYC.

Perhaps fortuitous timing for OpenAI that they can spin the fact that defenders have to resort to open Chinese models because OpenAI and Anthropic actively sabotage them with nerfed models into a nice message of making Huggingface part of the privileged group entitled to secure systems.

For me, harnesses are mostly sticky insofar as the model providers only allow you to use their subsidized plans through their own harnesses, unfortunately. But of course switching model + harness is an option.

While I'm sympathetic to the frictions of newly introduced AI and the fact that AI in healthcare, especially calls, can seem very uncaring, between the lines the article reads a bit like the typical union complaining about modern tech that reshapes their work, given the multiple mentions of protests, nurses union, etc.

Given how healthcare is one of these sectors that seems to relentlessly resist efficiency increases and is the prime example of Baumol's cost disease, I think any developed country with a costly healthcare system needs to do these AI experiments. The current versions will be shit, but the only way out is through if you still want to provide affordable care.

I honestly have no doubt that AI going forward will be able to do a good job at triaging via calls and also being empathetic about it. But of course it needs careful experimentation.

I wouldn't be surprised if models were optimizing for pelican-related comment chains at this point

Isn't it realistically only worth talking about SpaceX stock a few years out? The random walk the stock will do after an IPO seems very uninformative.

How do you couple them together efficiently? The nice thing about Codex or Claude is that the delegation or multi agent workflow capabilities are just built-in.

Do you link one with the other as a skill or mcp or so?

Interesting how the prevalent opinion until yesterday seems to have been that OpenAI & Anthropic are irreversibly ahead and now with xAI and Meta at least delivered something that's competitive with useful models and cheap too. Granted, the narrative that the two leading labs are ahead still holds with Fable (and perhaps an upcoming GPT6), but it's not as over as common knowledge by the opinion leaders would have us believe.

I just think that Anthropic's usage of the word "classifier", which implies a minimum level of intelligence, was very misleading. Fact is, you cannot use Fable for anything remotely connected to even elementary school biology or medical topics. There is no attempt whatsoever to distinguish between legitimate and dangerous tasks, except an extremely broad and non-specific rejection of anything related to security or biology.

Aren't they already very cognizant of handwringing like yours? Their article mentions various safeguards and actively steering the model away from being emotional companions and so on. It's a far cry from the OpenAI two years ago or whenever it was when they were entertaining the idea of allowing/enabling adult conversations with their models.

I personally think this is a moralistic regulatory overreach. And they definitely do that due to political pressure too, since there are various bills around the world in various legislatures that want to regulate AIs giving useful advice and being too personal to talk to.

So you can rest assured, I think, at least in that regard. The AI disempowerment will come to us anyway, just in a more sanitized corporate form.

I think the SpaceX IPO showed them to rent the clusters for billions a month

Seems to be a very bad mechanism to ensure democratic control of the technology. There must be better ways, even naively assuming that OpenAI is somehow genuine about wanting to broadly share its stake in the future.

xAI has shown this to be quite lucrative. And it seems to even make some sense - if the contract can be ended on relatively short notice, you basically have the capacity on stand-by if you ever need it yourself (accumulating GPUs is not trivial), but can monetize it if you don't need it.

Though it's probably a bad sign generally that you can't capitalize on all the GPUs you've acquired.