No single individual will ever be "in charge" of superintelligence.
@ylecun
-
Stop the AI hype: Just give us models, not societal surgery
By
–
Please please please. I'm on my knees begging every AI exec on the planet. Just stop with this stuff. Stop. Just give us models. Let the collective distributed intelligence of people figure things out in real time like we always do. Let people adapt. It's what we do. It's all just so tiresome. We just want models. We'll figure it out. We promise. We don't need societal level surgery and UBI and robot taxes and ham-fisted legislation and populists politicians passing dumb law after dumb law and lobbying groups and all this craziness. We are not giving birth to magic super miracle machines that suddenly invalidate every single pattern of the entirety of human history and technological development. We're not. Really. AI is amazing. It's wonderful. But it's not magic. Can we please just let AI be cool and useful and problematic in realistic ways instead of all this crazy talk? We are hallucinating at a collective scale. It's a madness really. A societal meme level madness. Just give us a products please and leave all the politics in the garage. Stop proposing societal level surgery with drastic measures for things that have not happened and may not happen and probably won't happen. Just stop. Mike Allen (@mikeallen) 🚨🚨@sama tells me he feels such URGENCY about the power of coming AI models that @OpenAI is unveiling a New Deal for superintelligence – ideas to wake up DC He says AI will soon be so mindbending that we need a new social contract 👇Altman's top 6 ideas axios.com/2026/04/06/behind-… — https://nitter.net/mikeallen/status/2041099089031356468#m
-
Language Is a Roof, Not Foundations for Thinking
By
–
The vast majority of our thinking is not language based.
That doesn't mean language is useless for thinking. A roof is a useful thing to have over your head.
But a roof without foundations and walls to support it is considerably less useful. -
Language Thinking Limitations and Abstract Mental Models in AI
By
–
Thinking in language has limited applications, largely in coding and mathematics where the language itself can help reasoning. But, as I've been saying for years, thinking manipulates mental models in abstract (continuous) representation space. Soooo, xAI gonna use JEPA now?
-
Adversarial Training: A Key Use Case in AI Development
By
–
A perfect use case of adversarial training.
-
Reinforcement Learning Limitations on Fine-tuned Model Prompts
By
–
No, RL doesn't fix it. It merely makes e smaller for prompts present in the fine-tuning set.
-
LLMs and Code Generation Systems: Clarifying Autoregressive Architecture
By
–
1. I never said LLMs were not useful 2. Code generation systems are not strictly auto-regressive LLMs. They produce multiple outputs and pick the best ones. 3. Your argument is as if I said "perpetual motion is impossible" and you responded "meanwhile, it's been 300km since I
-
Autoregressive Models Error Propagation in Discrete Sequences
By
–
That's a ridiculous argument. – all auto-regressive models diverge, whether they are generative (in input space) or not. – for discrete symbol sequences, the probability of correctness decreases exponentially with the sequence length, assuming independence of errors. – THAT
-
Clarifying the utility debate around large language models
By
–
I never said LLMs were not useful. We're discussing a different question here.
-
Error Recovery Impossibility in Autoregressive Models
By
–
You didn't understand either. Yes, the independence of errors is an assumption, which may or may not be reasonable. No, errors are NOT RECOVERABLE in an auto-regressive setting because the set of correct answers form a subtree in the tree of all possible sequences. Once you get