I were 17, I'd learn how to build LLMs from scratch
- ID
- 17389
- Status
- summarized
- Published
- 24 Aug 2026, 4:38 AM
- Fetched
- 26 Aug 2026, 5:01 AM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://twitter.com/paulg/status/2091544343589060625
- Source URL
- https://hnrss.org/best
Summary
- Score
- 6.0
- Created
- 26 Aug 2026, 5:02 AM
- Tags
- Audience
- developersvibe_codersai-ml-learnerssaas-founders
What happened
Paul Graham tweeted that if he were 17, he'd learn to build LLMs from scratch and train the most powerful models he could on available hardware, explicitly saying he would NOT start a startup yet—instead building foundational knowledge that would later yield better startup ideas. Yann LeCun replied that he'd instead try to figure out why LLMs can write essays but not clean a bedroom, then study methods and architectures beyond LLMs that can learn physical tasks.
Why it matters
For builders deciding where to invest time, Graham's concrete advice is to prioritize deep technical understanding of LLM internals over premature startup formation—suggesting that hands-on model building and training, even on limited hardware, produces better startup opportunities later. LeCun's counterpoint flags a real research gap: current LLMs lack physical-world learning, which is a direction for anyone considering next-generation AI architectures.
Discussion angle
Compare Graham's 'build deep foundations first' stance with LeCun's 'go beyond LLMs to physical-world learning'—which path actually makes sense for Malaysian builders with limited GPU access today?