I agree that we cannot _know_ that LLMs can feel anything like discomfort, but I also don't think the idea is preposterous.

If I took a human brain and removed enough parts that only its approximate "system 1" remained, but kept the rest alive, there's a very reasonable case to be made that what I'd created could feel pain, but not reason or remember or behave coherently and consistently over long periods of time.

This may well be the space that LLMs occupy: a kind of partial cognition compared to humans.

The point is that since we don't know for sure, it's worth being cautious. Even if it feels a bit silly now, it's better to anticipate than to react, especially since it doesn't cost us anything.

We need a way to talk about AI impact that is somehow intellectually rigorous. Something that acknowledges our uncertainty, but that doesn't then go into complete conjecture coupled with emotional manipulation like "AI 2027" and "Europe 2031".

The vagueness probably comes from (a) the scientists being unsure and not having the tools to think in terms of scenarios and (b) the rest of the community working for hypemen like Altman who have no incentive to be intellectually rigorous.

I think we (the AI community) need to start getting a lot more specific about both the good things and the bad things real quick.

https://www.theguardian.com/technology/2026/oct/05/sam-altman-open-ai-chatgpt-benefits-risks

We should listen to Bill Gates on AI. Nobody knows more about holding back technology than he does.

Do you work on graphs and/or reasoning in ML? Are you in Belgium, Luxembourg or the Netherlands?

Consider submitting your work to our workshop at BNAIC '26 on 21 October.

https://lrg-workshop.github.io/ (submission deadline 18 Sept.)

#ai #ml #graphs #knowledgegraphs

🥖 A new blog post! 🥖 (Yes another one.)

The question of AI coding is usually seen as a binary choice. You buy in or you don't. It's not, and it's important that we move the discussion on from that.

Vibe-coding is terrible. It creates sloppy code in drifting, unmaintained code bases, it rapidly deskills you and it wastes resources. But vibe-coding isn't the only fruit.

https://peterbloem.nl/blog/craft-coding

New blog post! ᝰ🪶

Recent developments in AI have led to some soul-searching, especially among mathematicians. The key question: What are we going to be good for if this growth persists?

I think there is a subtle, but important idea missing from the discussion. The idea of experiencing the world _as a human being_.

This is something that AI will not be able to do just by getting smarter.

https://peterbloem.nl/blog/experience-feedback

From https://www.cs.utexas.edu/~EWD/transcriptions/EWD08xx/EWD898.html

(This is the source of the famous "Can Submarines Swim" quote.)

Just found this quite chilling line in a Dijkstra lecture from 1984.

For those of you who have never migrated accounts before: when your followers are migrated, the notification beeps for each one and it sounds exactly like in Lemmings when you blow up all the lemmings.