Karpathy’s Pelican
twitter.com
August 2, 2026
2 min read
🔥🔥🔥🔥🔥
70/100
Summary
We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's...
Coverage & discussions
Hacker News · twitter.com
Community discussions
Community Sentiment
Positives
- It's incredible that we can generate entire games with just a few words — that's a huge leap in what AI can do, even if the results are lackluster.
- Using LLMs for one-off tasks is fantastic; they can quickly produce software tailored to specific needs, which opens up new possibilities.
Concerns
- These AI-generated games are painfully boring and lack the creativity that makes games fun — it's like watching paint dry.
- LLMs currently have zero imagination and seem stuck in a loop of producing the same predictable content over and over.
- We're entering an era of throwaway software, but that might mean sacrificing quality for convenience — is this really progress?
