Upon popular request, you can now also get this as a physical poster via Zazzle zazzle.com/llm_architecture_… This is based on a 56 MB PNG file with 182 megapixels. I just ordered one myself but please be aware that I haven't been able to verify the quality, yet.
GENERATIVE AI
-
Rethinking AI Cost Evaluation Beyond Traditional Software Pricing
By
–
I started evaluating AI cost very differently the second ChatGPT Plus launched, but I understand people/companies are still evaluating it as only software.
-
EverMind publishes high-quality AI research paper this week
By
–
Just a slight spoiler, @EverMind will publish another high-quality paper this week.
-
Product scaling beyond expectations by orders of magnitude
By
–
selling more than spectacles by a few orders of magnitude
-
Claude Usage Patterns: Weekends vs Weekdays Workflow
By
–
I use Claude more on weekends than weekdays at this point haha. I also experiment with cowork a lot more during weekends running while I'm doing other things vs. during the week focusing on urgent tasks TBH.
-
RAG Deployment Gap: Model Capability Versus Production Reality
By
–
Benchmarks go up every month but most companies still can't get a basic RAG pipeline to work reliably. There's really a big gap between model capability and actual deployment.
-
Constraints Drive AI Engineering Creativity Beyond Hardware
By
–
Love this challenge. Constraints force creativity. The best AI engineering often comes from "how do I make this work with what I have" not "give me more GPUs."
-
OpenSquirrel: Rust IDE with AI agents as central unit
By
–
Karpathy asked. I delivered. Introducing OpenSquirrel! Written in pure rust with GPUI (same as zed) but with agents as central unit rather than files. Supports Claude Code, Codex, Opencode, and Cursor (cli). This really forced me to think up the UI/UX from first principles instead of relying on common electron slop. github.com/Infatoshi/OpenSqu… Andrej Karpathy (@karpathy) Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level – the basic unit of interest is not one file but one agent. It’s still programming. — https://nitter.net/karpathy/status/2031767720933634100#m
→ View original post on X — @thom_wolf, 2026-03-15 22:03 UTC
-
LLM Job Exposure Scoring and Employment Impact Complexity
By
–
The "exposure" was scored by an LLM based on how digital the job is. This has no baring on what actually happens to these occupations, which has to do with demand elasticity and a lot more. People are sensationalizing the visualization tool and putting words in my mouth.