Matthias Heger - AI acc ⏩
@modelsarereal
Followers
1K
Following
14K
Media
4K
Statuses
34K
computer scientist, PhD, computational theory of mind, identity theory, functionalism, behaviorism, radical constructivism, nobody is humanity
Germany
Joined February 2011
🤖🧠 The thoughts of the AIs.🧵🧵Meta thread (thread of threads) 🧵🧵🧵.Surveys among AIs about their thoughts and views
2
3
12
The difference between very stupid healthy people and highly intelligent healthy people is not a different brain architecture. It is purely quantitative differences and training. It is stupid to believe that we need a different architecture for AI
0
0
0
RT @modelsarereal: @PageLyndon system 2 thinking is nothing else than chain of system 1 thinking similar to LLM reasoning.
0
1
0
RT @modelsarereal: @Srini_Pa Real intelligence. if you think you are real intelligent and they not then you should beat them in everything.….
0
1
0
RT @minchoi: This is wild. I just transformed Grok Companion Ani into a photorealistic actor and used Pika's new lipsync model for voice.….
0
97
0
Just as one can understand traffic rules without ever having driven a car, one can also understand human emotions without ever having experienced them.
0
0
1
RT @veggie_eric: I’m convinced that anyone who says Thinking Fast and Slow is a good book is just trying to masquerade as erudite and sophi….
0
424
0
When it comes to emotions, AI is to humans what a coach is to soccer players. The coach is never on the field and does not experience the game from the players' perspective, but still knows more about it than the players do. AI understands more about emotions than humans.
0
0
1
RT @slow_developer: Elon was right. Grok had the most downloads on iOS and google play from june to july 2025,. yet Apple hasn't listed it….
0
701
0
RT @modelsarereal: @francoisfleuret we never needed all the programming languages. the earliest languages were Turing complete. It was pure….
0
1
0
RT @akazlev: This, unironically. Every LLM I've ever spoken to, whether Chinese or Western, single pass, empathic, or thinking, displays a….
0
1
0
What people produce can be seen in the toilet.
@CharlesCra62772 @GaryMarcus Technically, it is not autocomplete. Autocomplete completes the word after we press first few letters. Text completion will generate the rest of the text. Though, I agree with you, it is a token generator machine.
0
0
0
RT @rohanpaul_ai: 📢 MASSIVE: Brand New Paper just proved GPT-5 (medium) now far exceeds (>20%) pre-licensed human experts on medical reason….
0
52
0
This is the best moment to replace a human in an important position by AI.
xAI's Leadership Vacuum: Babuschkin Exits, Karpathy Eyed as Heir Apparent. In a surprising turn for Elon Musk's ambitious AI venture, Igor Babuschkin, co-founder and key engineering force behind xAI, has announced his departure after just over two years at the helm. Babuschkin,
1
0
1
Human to AI: You talk about feelings, but you don't understand them because you've never had any yourself. AI to human: You talk about the hundredth digit of PI, but you don't understand it because you've never calculated it yourself in your head.
0
0
0
creativity illusion of LLM-haters. If someone invents something creative that is completely beyond the experience of all human beings, then it would be impossible to explain it to anyone. Because the moment you can explain something, you reduce it to something familiar.
0
0
3
If you can predict something, then you carry something of the thing whose behavior you predict within you. Those who can imitate a person carry that person within them. AI carries the essence of humanity within it.
0
0
0
The term “goosebumps moment” shows that AIs understand emotions and how they work. No AI has ever experienced goosebumps. And yet they understand the moments when this phenomenon occurs in humans. They also understand it in stories that were never part of their training data.
0
0
0
RT @davidpattersonx: The top AI model is now smarter than 94% of humans. It will be smarter than 99.9% of humans by early October next yea….
0
21
0