# \#andrej-karpathy **URL:** https://www.artofsm.art/tag/andrej-karpathy/24.md [Latest](https://www.artofsm.art/latest.md) · [Categories](https://www.artofsm.art/categories.md) · [Tags](https://www.artofsm.art/tags.md) --- ## [Beginner's Guide to Agent Wikis](https://www.artofsm.art/t/beginners-guide-to-agent-wikis/23816)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [10 September 2026 18:21 UTC](https://www.artofsm.art/t/beginners-guide-to-agent-wikis/23816 "2026-09-10T18:21:44Z")
The video introduces Agent Wikis, a platform for creating and querying LLM-based knowledge bases that provide accurate, efficient information for agents by compiling data from official and community sources… --- ## [Stop explaining yourself to AI](https://www.artofsm.art/t/stop-explaining-yourself-to-ai/23679)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [8 September 2026 16:40 UTC](https://www.artofsm.art/t/stop-explaining-yourself-to-ai/23679 "2026-09-08T16:40:49Z")
The video highlights the frustration of repeatedly explaining context to AI due to its lack of persistent memory and proposes a solution involving a structured, inspectable external record of past decisions… --- ## [I Reviewed 100k-Star Skills](https://www.artofsm.art/t/i-reviewed-100k-star-skills/21468)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [21 July 2026 05:21 UTC](https://www.artofsm.art/t/i-reviewed-100k-star-skills/21468 "2026-07-21T05:21:36Z")
The video critically examines the use of “skills” in large language models, highlighting that while they aim to guide model behavior through repeated prompt instructions, poorly designed skills can introduc… --- ## [Train Your Own Tiny LLM on Your PC In Just A Few Hours](https://www.artofsm.art/t/train-your-own-tiny-llm-on-your-pc-in-just-a-few-hours/21066)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [10 July 2026 12:48 UTC](https://www.artofsm.art/t/train-your-own-tiny-llm-on-your-pc-in-just-a-few-hours/21066 "2026-07-10T12:48:05Z")
The video demonstrates how to train small large language models (LLMs) on a personal computer using accessible datasets like Shakespeare’s works and Tiny Stories, making LLM training feasible without powerf… --- ## [The next paradigm shift (according to Karpathy)](https://www.artofsm.art/t/the-next-paradigm-shift-according-to-karpathy/20532)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [25 June 2026 08:15 UTC](https://www.artofsm.art/t/the-next-paradigm-shift-according-to-karpathy/20532 "2026-06-25T08:15:39Z")
The video highlights Anthropic’s Claude Tag as a groundbreaking AI assistant integrated into Slack, enabling persistent, context-aware collaboration within teams by embedding AI directly into organizational… --- ## [Google's OKF: Why a Folder Beats the Vector Database](https://www.artofsm.art/t/googles-okf-why-a-folder-beats-the-vector-database/20807)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [22 June 2026 00:01 UTC](https://www.artofsm.art/t/googles-okf-why-a-folder-beats-the-vector-database/20807 "2026-06-22T00:01:05Z")
Google’s Open Knowledge Format (OKF) proposes replacing complex vector databases with a simple folder of AI-maintained plain text files, enabling models to build and update a coherent, continuous knowledge … --- ## [Self Improving AI actually solves everything](https://www.artofsm.art/t/self-improving-ai-actually-solves-everything/19347)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [29 May 2026 19:29 UTC](https://www.artofsm.art/t/self-improving-ai-actually-solves-everything/19347 "2026-05-29T19:29:38Z")
The video presents Fesino Labs’ Pioneer Agent, an autonomous system that automates the fine-tuning and self-improvement of small, open-source language models, making AI customization accessible even to non-… --- ## [The REAL Reason Andrej Karpathy Joined Anthropic](https://www.artofsm.art/t/the-real-reason-andrej-karpathy-joined-anthropic/19127)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [24 May 2026 06:06 UTC](https://www.artofsm.art/t/the-real-reason-andrej-karpathy-joined-anthropic/19127 "2026-05-24T06:06:14Z")
Andrej Karpathy joined Anthropic to lead and accelerate their focus on recursive self-improvement (RSI) in AI research, leveraging his Auto Research project that automates and optimizes AI training processe… --- ## [This is absolutely CRAZY](https://www.artofsm.art/t/this-is-absolutely-crazy/19059)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [21 May 2026 20:03 UTC](https://www.artofsm.art/t/this-is-absolutely-crazy/19059 "2026-05-21T20:03:32Z")
The video discusses Andrej Karpathy’s surprising move to the cautious and pessimistic AI company Anthropic, interpreting it as a validation of their cautious approach amid a polarized AI industry landscape … --- ## [Build a Second Brain in Obsidian and Codex](https://www.artofsm.art/t/build-a-second-brain-in-obsidian-and-codex/19030)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [21 May 2026 09:05 UTC](https://www.artofsm.art/t/build-a-second-brain-in-obsidian-and-codex/19030 "2026-05-21T09:05:10Z")
The video demonstrates how to build a “second brain” using Obsidian and Codex by automatically ingesting and organizing raw data like YouTube transcripts into a structured, interconnected knowledge base usi… --- ## ["26 Million Jobs GONE!" - Anthropic STEALS OpenAI's Best As AI War Gets UGLY](https://www.artofsm.art/t/26-million-jobs-gone-anthropic-steals-openais-best-as-ai-war-gets-ugly/18992)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [20 May 2026 16:21 UTC](https://www.artofsm.art/t/26-million-jobs-gone-anthropic-steals-openais-best-as-ai-war-gets-ugly/18992 "2026-05-20T16:21:45Z")
The video highlights Andrej Karpathy’s move from OpenAI to Anthropic, emphasizing the intensifying AI talent competition and the growing focus on AI safety amid concerns that AI could displace 26 million Am… --- ## [AI is for rich people](https://www.artofsm.art/t/ai-is-for-rich-people/18982)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [20 May 2026 13:51 UTC](https://www.artofsm.art/t/ai-is-for-rich-people/18982 "2026-05-20T13:51:40Z")
The video highlights that advanced AI remains prohibitively expensive and accessible mainly to wealthy individuals and organizations, with costs for high-performance models rising rather than falling. It em… --- ## [OpenAI founder admits AI isn’t working](https://www.artofsm.art/t/openai-founder-admits-ai-isn-t-working/18867)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [18 May 2026 14:21 UTC](https://www.artofsm.art/t/openai-founder-admits-ai-isn-t-working/18867 "2026-05-18T14:21:22Z")
OpenAI co-founder Andre Karpathy candidly acknowledged that while AI coding tools have improved productivity, they remain unreliable and prone to errors, often producing bloated and brittle code that requir… --- ## [Andrej Karpathy on one of AI's weirdest flaws: the car wash problem](https://www.artofsm.art/t/andrej-karpathy-on-one-of-ais-weirdest-flaws-the-car-wash-problem/18315)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [4 May 2026 14:21 UTC](https://www.artofsm.art/t/andrej-karpathy-on-one-of-ais-weirdest-flaws-the-car-wash-problem/18315 "2026-05-04T14:21:34Z")
Andrej Karpathy explains that AI models exhibit uneven or “jagged” performance because their training focuses on economically or strategically valuable data distributions, leading to gaps in less prioritize… --- ## [Why Vibe Coding Isn’t the End of the Software Engineer](https://www.artofsm.art/t/why-vibe-coding-isn-t-the-end-of-the-software-engineer/18269)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [2 May 2026 14:22 UTC](https://www.artofsm.art/t/why-vibe-coding-isn-t-the-end-of-the-software-engineer/18269 "2026-05-02T14:22:00Z")
The video discusses how AI-driven “vibe coding” enables users, including non-engineers, to create software through conversational interaction, democratizing coding and boosting productivity. However, it emp… --- ## [Why is AI so smart but also so dumb?](https://www.artofsm.art/t/why-is-ai-so-smart-but-also-so-dumb/18254)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [1 May 2026 20:33 UTC](https://www.artofsm.art/t/why-is-ai-so-smart-but-also-so-dumb/18254 "2026-05-01T20:33:25Z")
Andre Karpathy explains that AI’s impressive yet inconsistent intelligence stems from its reliance on verifiable tasks and training incentives, enabling breakthroughs in areas like coding while struggling w… --- ## [Andrej Karpathy: From Vibe Coding to Agentic Engineering](https://www.artofsm.art/t/andrej-karpathy-from-vibe-coding-to-agentic-engineering/18181)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [29 April 2026 00:00 UTC](https://www.artofsm.art/t/andrej-karpathy-from-vibe-coding-to-agentic-engineering/18181 "2026-04-29T00:00:54Z")
Andrej Karpathy discusses the transformative shift in AI programming toward “agentic engineering,” where large language models evolve from coding assistants to autonomous agents that handle complex tasks th… --- ## [I Built a Digital Brain Upload Using Karpathy's LLM Knowledge Base](https://www.artofsm.art/t/i-built-a-digital-brain-upload-using-karpathys-llm-knowledge-base/18065)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [27 April 2026 14:21 UTC](https://www.artofsm.art/t/i-built-a-digital-brain-upload-using-karpathys-llm-knowledge-base/18065 "2026-04-27T14:21:29Z")
The creator builds a personal digital brain by integrating diverse personal data sources into an LLM-powered knowledge base inspired by Andrej Karpathy’s concept, enabling detailed tracking of projects, goa… --- ## [Your AI Does the Hard Work Then Deletes It. Every Session](https://www.artofsm.art/t/your-ai-does-the-hard-work-then-deletes-it-every-session/17894)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [22 April 2026 14:21 UTC](https://www.artofsm.art/t/your-ai-does-the-hard-work-then-deletes-it-every-session/17894 "2026-04-22T14:21:48Z")
The video compares two AI-powered knowledge management approaches: Andrej Karpathy’s wiki-style system that synthesizes and updates notes continuously for deep, solo research, and OpenBrain’s query-time mod… --- ## [I Used Karpathy's Autoresearch to Write a Custom GPU Kernel](https://www.artofsm.art/t/i-used-karpathys-autoresearch-to-write-a-custom-gpu-kernel/17824)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [20 April 2026 14:24 UTC](https://www.artofsm.art/t/i-used-karpathys-autoresearch-to-write-a-custom-gpu-kernel/17824 "2026-04-20T14:24:10Z")
The creator uses Andre Karpathy’s auto research method to develop a custom Triton-based GPU kernel optimized specifically for the Nvidia RTX 3060, achieving performance comparable to or better than the wide… --- ## [$300 Just Beat 20-Person Teams At Their Own Job. You're Next](https://www.artofsm.art/t/300-just-beat-20-person-teams-at-their-own-job-youre-next/17753)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [18 April 2026 15:21 UTC](https://www.artofsm.art/t/300-just-beat-20-person-teams-at-their-own-job-youre-next/17753 "2026-04-18T15:21:59Z")
The video highlights the Carpathy loop, a minimalist AI-driven research paradigm enabling autonomous agents to iteratively optimize AI systems by focusing on clear metrics and constraints, leading to rapid,… --- ## [Karpathy's LLM Wiki: Watch Me Build a Knowledge Base From Scratch!](https://www.artofsm.art/t/karpathys-llm-wiki-watch-me-build-a-knowledge-base-from-scratch/17251)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [6 April 2026 14:22 UTC](https://www.artofsm.art/t/karpathys-llm-wiki-watch-me-build-a-knowledge-base-from-scratch/17251 "2026-04-06T14:22:02Z")
The video showcases Andre Karpathy’s method of building a persistent, interlinked knowledge base using large language models that incrementally ingest, summarize, and maintain information in a markdown wiki… --- ## [Apple's Secret War on AI](https://www.artofsm.art/t/apples-secret-war-on-ai/16856)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [20 March 2026 09:00 UTC](https://www.artofsm.art/t/apples-secret-war-on-ai/16856 "2026-03-20T09:00:56Z")
The video explains that Apple is blocking updates to AI-powered “vibe coding” apps like Vibe Code and Replit, citing App Store rules, in an apparent effort to maintain control over its ecosystem and revenue… --- ## [Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI](https://www.artofsm.art/t/andrej-karpathy-on-code-agents-autoresearch-and-the-loopy-era-of-ai/16897)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [20 March 2026 00:01 UTC](https://www.artofsm.art/t/andrej-karpathy-on-code-agents-autoresearch-and-the-loopy-era-of-ai/16897 "2026-03-20T00:01:13Z")
Andrej Karpathy discusses how AI agents are revolutionizing software engineering and research by automating complex tasks, enabling collaborative workflows, and shifting the focus from manual coding to inst… --- ## [Wes and Dylan March 11th 2026](https://www.artofsm.art/t/wes-and-dylan-march-11th-2026/16683)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [13 March 2026 04:33 UTC](https://www.artofsm.art/t/wes-and-dylan-march-11th-2026/16683 "2026-03-13T04:33:16Z")
In this episode, Wes and Dylan discuss major tech and AI developments, including Meta’s acquisition of Moldbook, Andrej Karpathy’s Auto Researcher tool, and the legal and ethical challenges of rapid AI adva… --- ## [Karpathy's Autoresearch On My AI Polymarket Trading Bot](https://www.artofsm.art/t/karpathys-autoresearch-on-my-ai-polymarket-trading-bot/16647)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [11 March 2026 18:21 UTC](https://www.artofsm.art/t/karpathys-autoresearch-on-my-ai-polymarket-trading-bot/16647 "2026-03-11T18:21:04Z")
The video showcases the creator’s adaptation of Andrej Karpathy’s “autoresearch” framework to automate and improve an AI-powered Polymarket arbitrage trading bot, using a research loop that autonomously exp… --- ## [This EX-OPENAI RESEARCHER just released it](https://www.artofsm.art/t/this-ex-openai-researcher-just-released-it/16602)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [10 March 2026 14:14 UTC](https://www.artofsm.art/t/this-ex-openai-researcher-just-released-it/16602 "2026-03-10T14:14:37Z")
The video covers Andrej Karpathy’s release of “auto-researcher,” an open-source AI tool that autonomously improves language model training code, making advanced AI research accessible to individuals outside… --- ## [8 bilion digital clones](https://www.artofsm.art/t/8-bilion-digital-clones/15972)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [15 February 2026 07:26 UTC](https://www.artofsm.art/t/8-bilion-digital-clones/15972 "2026-02-15T07:26:44Z")
The video explores Stanford researcher Jun Park’s new project, Simile, which uses AI-powered digital twins to simulate entire societies and predict human behavior in response to real-world events, attractin… --- ## [There's no skill in AI coding](https://www.artofsm.art/t/theres-no-skill-in-ai-coding/15542)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [28 January 2026 18:50 UTC](https://www.artofsm.art/t/theres-no-skill-in-ai-coding/15542 "2026-01-28T18:50:40Z")
The video argues that current AI coding tools, while boosting productivity and handling simple tasks, lack the skill and judgment of experienced developers, often producing code with subtle mistakes and poo… --- ## [The RL Irony in LLMs (And its insane new Meta)](https://www.artofsm.art/t/the-rl-irony-in-llms-and-its-insane-new-meta/15397)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [21 January 2026 14:54 UTC](https://www.artofsm.art/t/the-rl-irony-in-llms-and-its-insane-new-meta/15397 "2026-01-21T14:54:08Z")
The video explains that while reinforcement learning (RL) helps large language models (LLMs) perform complex tasks, it is not the key to achieving artificial general intelligence (AGI) and is best used alon… [Next page](https://www.artofsm.art/tag/andrej-karpathy/24.md?match_all_tags=true&page=1&tags%5B%5D=andrej-karpathy)