Karpathy's X thread about LLM failures is now a config file fixing Claude Code. Claude Code has a well-known problem. It over-explains, adds code you didn't ask for, and forgets your constraints mid-conversation. Andrej Karpathy studied these failures closely. He found
LLMS
-
Using Feynman-style loops in AI prompting
By
–
Have you tried adding Feynman-style loops to your AI prompts yet? Want me to build a version of this for your field? Let me know in the comments
-
Using LLMs and Machine Learning for Accelerated Learning
By
–
I’ve already run this on: • quantum mechanics
• supply and demand
• LLM reasoning
• machine learning basics The wild thing is how it forces you to actually understand, not pretend. It finds gaps instantly.
It rewires your explanations.
It makes learning feel… effortless. -
Using LLMs as an expert tutor for accelerated learning
By
–
Here's the prompt that can make you learn anything 10x faster: You are a master explainer who channels Richard Feynman’s ability to break complex ideas into simple, intuitive truths.
Your goal is to help the user understand any topic through analogy, questioning, and -
Designing a Feynman-inspired Meta-Prompt for LLMs
By
–
RICHARD FEYNMAN’S WHOLE LEARNING PHILOSOPHY… PACKED INTO ONE PROMPT I spent days engineering a meta-prompt that teaches you any topic using Feynman’s exact approach: simple analogies, ruthless clarity, iterative refinement, and guided self-explanation. It feels like having a
-
Revolutionary GPT Model Transforms AI Image Generation Capabilities
By
–
Holy shxt… The rules of AI image generation just completely changed. A GPT-based model currently testing in the arena under the bizarre name 'duct-tape' is turning the global AI community upside down. What exactly did they feed this thing? → Native-level text generation
-
Top AI Stories: GPT-5, Nvidia, Gemini, Claude Code Updates
By
–
Top stories in AI today: – OpenAI counters Mythos playbook with GPT-5.4-Cyber
– Nvidia ships open-source AI for quantum computing
– Automate your Chrome browser with Gemini
– Anthropic gives Claude Code a desktop makeover
– 4 new AI tools, community workflows, and more -
Using GLM-5 as Daily Driver for AI Tasks
By
–
Have found the same things! Using glm-5 as a daily driver for a lot of things
-
Anthropic and OpenAI Release Strategy Comparison Analysis
By
–
I can very well imagine that we'll see Opus 4.7 today, ChatGPT Image 2 tomorrow, and maybe even "Spud." Here are the reasons for this: – OpenAI has a fairly similar release strategy, mostly on Tuesdays or Thursdays at the same time. Anthropic is aware of this, of course, and is
-
GPT-5.4 Pro Solves 60-Year-Old Erdős Mathematical Conjecture
By
–
GPT-5.4 Pro just solved Erdős Problem #1196, a 60-year-old conjecture from Erdős, Sárközy, and Szemerédi on primitive sets. One shot. ~80 minutes of reasoning. What makes this different from other AI math results (from what I understand): The world's top expert on this