LLM outputs are 'ultra processed information', quips blogger comparing them to engineered desserts
jwt0625 · x · 2026-09-06
Blogger jwt0625 shares a witty analogy: after marveling at how many parameters go into optimizing an ultra-processed dessert, it struck them that LLM outputs are literally "ultra processed information" — heavily engineered, optimized products of computation, much like engineered junk food.
More from Fun
- User Shares GPT-6's Portrait of Them Drawn in Classic MS Paint Style — iScienceLuvr · 2026-09-06
- Meme: say 'AI' three times in the mirror and Dario appears to say your job has 6-12 months — _jaydeepkarale · 2026-09-06
- AI snail tells user to go outside after long Astra sessions in amusing demo — aaraalto · 2026-09-06
- Heracles' ninth labor: factor this 300-digit number — NathanpmYoung · 2026-09-06
- GPT-6 Astra builds browser games with three.js, even modeling cars in Blender — gaganghotra_ · 2026-09-06
- Shopify CEO Tobi Lütke ships headless NetHack via GPT6, usable as MCP — DeryaTR_ · 2026-09-06