WordsOfTheFuture

Futurist and computer science master's student at RWTH.
Doing research on RL and planning.
Interested in all STEM fields.
I give talks on AI.

[WordsOfTheFuture @MaxKannen]

3303d7e319e7727983df657b536b30cde6be4c5f46ed66e3a9a76d8e03a6a74f

[WordsOfTheFuture @MaxKannen]

Does anybody know if there is research that uses the internal emotional features as signals for RL training? I think this is a interesting idea.
https://t.co/mIqReM94PC

[WordsOfTheFuture @MaxKannen]

As a german this is so funny. No better place to surrender.

[Emmanuel Macron @EmmanuelMacron]

Le Président Trump a signé ce soir à Versailles l’accord entre l’Iran et les États-Unis.

Cet accord ouvre la voie à une paix durable et permet la réouverture du détroit d’Ormuz.

C’est un pas important dans la bonne direction pour nos compatriotes qui permettra d’obtenir bientôt une baisse des prix de l’énergie.

[WordsOfTheFuture @MaxKannen]

It seems like https://t.co/CX3eBzfY6g was to conservative. The EU needs to build Datacenters and renewable energy as fast as possible. Until then China is becoming the only possible supplier of near Frontier AI.

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[WordsOfTheFuture @MaxKannen]

We want useful tools to automate work but we do not want slaves.
On the other hand we want to see if we can build artificial life and push the limits of research.

[WordsOfTheFuture @MaxKannen]

I think we soon reach a point where we have to split up AI research into the branch that continues to push the models towards AGI and explores sentience and another branch that tries to avoid sentience to build better tools.

[WordsOfTheFuture @MaxKannen]

My thoughts on the current leaks

[WordsOfTheFuture @MaxKannen]

5. Manufacturing Consent is a old classic and definitly still relevant. It clearly shows how propaganda, especially western one, worked and still works. Most of the book is about events around 1985 but can be easly applied to recent events.

[WordsOfTheFuture @MaxKannen]

I want to recommend some books that I read this year like I did last year. Without particular order:

[WordsOfTheFuture @MaxKannen]

4. "Nexus: A Brief History of Information Networks from the Stone Age to AI" is a very well written view on civilisations as information networks and makes some key observations.

[WordsOfTheFuture @MaxKannen]

That being said Opus 4.5 is really impressive. Just to expensive.

[WordsOfTheFuture @MaxKannen]

Anthropic models are good but they did not figure out reasoning. It actually makes their models worse sometimes.

[WordsOfTheFuture @MaxKannen]

@simpsoka Everyone talked about cloning for a long time, but now that we are actually selecting humans by traits, nobody cares. It is insane.

[WordsOfTheFuture @MaxKannen]

OpenAI continues to use its models to write their posts

[WordsOfTheFuture @MaxKannen]

I still believe that the closed source models will stay ahead for a while but this feels similar to Deepseek R1 where the open model came out at the right moment to catch up and shortly after the next generation of closed source models comes out.
That being: GPT-5.1 and Gemini 3

[WordsOfTheFuture @MaxKannen]

🤨🤨🤨

https://t.co/ycarVqJMow