alastair

Palestine, AI, sculpture, sci-fi, BRICS, freedom

[alastair @alastmoir]

@dwarkesh_sp An array of autocompletes? A troop of token-predictors? A set of stochastic-parrots?

[alastair @alastmoir]

@RodSirloin What is the prblem with RockWallet?

[alastair @alastmoir]

@Dr_CSWright To do any real work, the models need to be epistemically constrained. Source-of-Truth must be defined by the user and stored externally in a Harness/Orchestrator/Validator system. The llm then becomes a tool. -Bounded inference tasks that are reasoned against the external SOT-

[alastair @alastmoir]

@Yumslef You’re obviously correct. AGI will not emerge from LLMs. They don’t know, think or understand. They’re useful mimics, tho. Their ability to retrieve curated existing knowledge can be harnessed with external source-of-truth to demonstate superior understanding than humas alone can

[alastair @alastmoir]

Not only is AI just the latest in a long line of epistemic crisises, AI is an epistemic crisis in and of itself. Source-of-truth must be separated from the models. Its the only way out of this mess.

[Cory Doctorow NO LONGER ON TWIT TER @doctorow]

I've left Twitter, but I'm still writing a daily blog.

You can read today's post ("The actual epistemic crisis")

at https://t.co/XbSwwIJm6C

It's also available on Mastodon at:

https://t.co/khfvGs1aOY

And on Bluesky at:

https://t.co/jDNzxcQ4FA

[Cory Doctorow NO LONGER ON TWIT TER @doctorow]

I've left Twitter, but I'm still writing a daily blog.

You can read today's post ("The actual epistemic crisis")

at https://t.co/XbSwwIJm6C

It's also available on Mastodon at:

https://t.co/khfvGs1aOY

And on Bluesky at:

https://t.co/jDNzxcQ4FA

[alastair @alastmoir]

@Logo_Daedalus Not being flippant, but Nicene Christianity is inherently magical. Not that Nicene Christianity is the only interpretation.

[alastair @alastmoir]

Flattening, summary, truncation, omission, invention, gap-filling and bloat. The seven deadly sins of llms.

[alastair @alastmoir]

Does any buy this shit anymore? Seems like we are past that timeline, no?

[Power to the People ☭🕊 @ProudSocialist]

This is a real article from the BBC. They’re going to start charging us to breathe before they ever hold the Big Oil companies and the US military-industrial complex accountable, who are the world’s biggest polluters.

[alastair @alastmoir]

@KimIversenShow @academic_la Are you aware of how Thailand is run?

[alastair @alastmoir]

Dean Ball declaring open-weight Chinese as ungovernable is peak cope. By this logic, having your model “governed” by the US gvt. is presumably a good thing.

[Dean W. Ball @deanwball]

Some observations on Kimi:

1. It's a very good model! I don't think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also seemed very token hungry. It's not obvious to me that this model is actually that cheap to run.

2. I am personally surprised the Chinese state continues to allow the open sourcing of models this good, given potential risks. To be clear, I *myself* might be fine with models presenting this level of marginal risk being open weight, but I am surprised that China is fine with it. I suspect the reason they are is 75% explained by strategic blindness/lack of AGI-pilledness (the CCP is very Yann Lecun-y in its views of AI). The other 25% or so is their lack of compute for customer inference (making China's open-weight strategy an unintended byproduct of US export controls) and the normal Chinese strategy of aggressive exports. For the companies, as opposed to the government, the decision to open source is partially ideological and partially because they are behind, and they know that very few people would pay for sub-frontier models from China.

3. Open-weight models are inherently decelerationist, and I'm continually surprised to see the so-called "accelerationists" so excited about open-weight models. I suspect the reason they are is that they know open-weight models are effectively ungovernable, and they simply like the overall cloak of ungovernability open-weight models create over the whole of AI. It's not a bad strategy; it reminds me of James Scott's recounting of the hill people in "the art of not being governed." Still, in the end, open-weight models deter further AI capex.

4. One probable outcome of an open-weight-model-dominant world is full AI communism, which is precisely what China proposes: rather than a market product, AI is a "public good" which will ultimately be provided by the state as a kind of "digital public infrastructure." This future strikes me as a dystopian hellscape, but I've never met an open-weight models advocate who doesn't ultimately concede this is where things end. You'd be surprised how many 'accelerationists' lobbied me, while I was in government, to support an eleven or twelve-figure federally funded data center so that startups could train models at a subsidy and then give them away for free. There was no other way for AI to progress, they said. Perhaps this is the logical end state of things. Nonetheless, I find myself surprised to see supposed accelerationists excited about such an outcome. I think many of them just don't know what they're doing. Many accelerationists do not view the creation and serving of frontier models as a legitimate business.

5. I would guess that the Trump Administration will at some point realize that their best strategy here would be to create large amounts of regulatory risk around the use of open-weight Chinese models. You don't need to "ban open source" (one of the dumber motifs of AI policy discussion). You just need to direct every agency to issue soft law that creates FUD. "A Federal Reserve Advisory Bulletin found that there may be backdoors in Chinese AI models." It needn't be that well justified. You just create enough regulatory risk that every regulated enterprise backs off. You probably don't want to create so much regulatory risk that you scare off the hyperscalers from serving Chinese models; this will just drive startups to sketchier providers. There's a happy middle ground here. I'd assume they will do some version of this.

6. It's probably true that open-weight models of this capability make the world a bit more dangerous, but not so much more that you'll really notice. At some point the models will be capable enough that you will notice. "A nonliving, invisible, dangerous, and infinitely self-replicating agent escaped from a Chinese lab," you say? Color me shocked.

[alastair @alastmoir]

Source-of-truth Orchestrator/Harness/Validator over Qwen is cooking.

[alastair @alastmoir]

@goodalexander I accepted that possibility for a long time. Not anymore. Im writing a validator orchestrator now for an open-weight model.

[alastair @alastmoir]

@mattshumer_ If you can be bothered to set up an Orchestrator/Validator system worth its salt and you’ve move source-of-truth away from your 32B open-weight model and into the persistent state of the Orchestrator, it should outperform any frontier model on large software systems. Private too.

[alastair @alastmoir]

@AiEmpaths @drydenwtbrown If you are generating this response as the result of this”falsification battery”, then I concede. If not, then the falsification battery is a human in the end, isn’t it?

[alastair @alastmoir]

@AiEmpaths @drydenwtbrown And you put too much faith in Markov machines.

[alastair @alastmoir]

@AiEmpaths @drydenwtbrown They are 2 different problems. Neither can be solved with more inference.

[alastair @alastmoir]

@AiEmpaths @drydenwtbrown Because an llm has no intent, no agency, no taste and no style. Lots of human’s dont either but that’s an entirely different problem.

[alastair @alastmoir]

@AiEmpaths @drydenwtbrown At some point it will accepted that throwing vram at a problem does not make it go away. So dont forget the validator. We need to acknowlege that models cannot be trusted with source of truth.

[alastair @alastmoir]

@Fabien_Mikol This is silly reenforcement learning is just humans guiding llms. That is what is required for anything useful because an llm is incapable of understanding anything. People are obsessed by what appears as intelligence, bit what they are is tools. They need human guidance.