MTAN

@ropirito Real

given one long weekend free of all distractions, i could vibecode a top-10 app and retire my entire bloodline

@yifever Spontiferaneousity

what would you say in 2025 to prove you're a human?

Enjoyed chatting w @AndrewYNg * Bleeding Edge of Agentic AI * Will Models Bootstrap Themselves? * Vibe Coding vs. AI Assisted Coding * Successful Founder Profiles * Next Industry Transformations youtube.com/watch?v=SYisFb…

The most interesting part for me is where @karpathy describes why LLMs aren't able to learn like humans. As you would expect, he comes up with a wonderfully evocative phrase to describe RL: “sucking supervision bits through a straw.” A single end reward gets broadcast across every token in a successful trajectory, upweighting even wrong or irrelevant turns that lead to the right answer. > “Humans don't use reinforcement learning, as I've said before. I think they do something different. Reinforcement learning is a lot worse than the average person thinks. Reinforcement learning is terrible. It just so happens that everything that we had before is much worse.” So what do humans do instead? > “The book I’m reading is a set of prompts for me to do synthetic data generation. It's by manipulating that information that you actually gain that knowledge. We have no equivalent of that with LLMs; they don't really do that.” > “I'd love to see during pretraining some kind of a stage where the model thinks through the material and tries to reconcile it with what it already knows. There's no equivalent of any of this. This is all research.” Why can’t we just add this training to LLMs today? > “There are very subtle, hard to understand reasons why it's not trivial. If I just give synthetic generation of the model thinking about a book, you look at it and you're like, 'This looks great. Why can't I train on it?' You could try, but the model will actually get much worse if you continue trying.” > “Say we have a chapter of a book and I ask an LLM to think about it. It will give you something that looks very reasonable. But if I ask it 10 times, you'll notice that all of them are the same.” > “You're not getting the richness and the diversity and the entropy from these models as you would get from humans. How do you get synthetic data generation to work despite the collapse and while maintaining the entropy? It is a research problem.” How do humans get around model collapse? > “These analogies are surprisingly good. Humans collapse during the course of their lives. Children haven't overfit yet. They will say stuff that will shock you. Because they're not yet collapsed. But we [adults] are collapsed. We end up revisiting the same thoughts, we end up saying more and more of the same stuff, the learning rates go down, the collapse continues to get worse, and then everything deteriorates.” In fact, there’s an interesting paper arguing that dreaming evolved to assist generalization, and resist overfitting to daily learning - look up The Overfitted Brain by @erikphoel. I asked Karpathy: Isn’t it interesting that humans learn best at a part of their lives (childhood) whose actual details they completely forget, adults still learn really well but have terrible memory about the particulars of the things they read or watch, and LLMs can memorize arbitrary details about text that no human could but are currently pretty bad at generalization? > “[Fallible human memory] is a feature, not a bug, because it forces you to only learn the generalizable components. LLMs are distracted by all the memory that they have of the pre-trained documents. That's why when I talk about the cognitive core, I actually want to remove the memory. I'd love to have them have less memory so that they have to look things up and they only maintain the algorithms for thought, and the idea of an experiment, and all this cognitive glue for acting.”

@vedantdotrpm Incident response teams with overworked engineers

So, if everything depends on AWS, what does AWS depend on?

@fireandvision @repligate Would love to read the rest!

I love words

@ropirito Real

@jsuarez @grok It may be worthwhile to try the Ai2 Paper Finder, or you could even try building an agent which searches through arxiv. My guess is that making your own agent would be more expensive but more effective.

Quick check in on LLM-land: can any models do this? "Go find me all RL papers that use a fast-slow RNN and pick out the ablation numbers with/without" @grok fails terribly

@ropirito Real

The team I used to work for previously has now made themselves obsolete by focusing on shitty RAG pipelines for a year instead of getting the internal safety team to approve long context models for company use. When I told them this would inevitably happen, the tech PM acted like he knew better (hasn’t coded in 5 years). I can assure you that the average corp is unable to keep up to date with AI advancements and will end up dependent on random B2B SaaS instead of just keeping a skunkworks team at hand to implement the latest tech.