What Tobias Lütke thinks about reinforcement learning
Co-founder and CEO of Shopify; long-time Linux user and Omarchy contributor.
Tobias Lütke did not write this page.
We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.
Filter reinforcement learning · Newest first
Topic
All 63reinforcement learning 1management 11design 6innovation 4LLMs 4coding agents 3motivation 3AI alignment 2AI slop 2
37 more topics
behavioural science 2energy 2infrastructure 2politics 2product management 2simplicity 2software quality 2startups 2taste 2writing 2AGI 1AI agents 1AI and jobs 1ambition 1America 1Anthropic 1chat interfaces 1code generation 1courage 1developer tools 1discipline 1Europe 1government 1ideas vs execution 1measuring intelligence 1Microsoft 1nuclear power 1Omarchy 1OpenAI 1recursive self-improvement 1semiconductors 1space 1SQLite 1team size 1TypeScript 1vibe coding 1video games 1
Only Tobias Lütke on reinforcement learning. See everyone on reinforcement learning
1 dated position, 2026, in their own words. Our reading of what Tobias Lütke has said — not written or endorsed by them.
-
Their wordsLLMs understand the concept of forward evolving a system to solve a need very well. This comes from the agentic RL. They understand how to change the state of a codebase or acquire data from databases and then operate on it similarly to humans. But without an execution environment, they actually cannot do this properly.