As Hand-Coding Fades, the Edge Shifts to Research Insights LLMs Can't Supply
janusch_patas · x · 2026-11-11
A developer argues that as LLMs take over hand-written code, the center of gravity shifts toward research: solving problems an LLM can't just code up because no solution is known yet or the needed knowledge isn't in its weights.
Key points:
- Great ideas can't be outsourced; you still need careful analysis and to actually understand what's going on.
- Open source sees more contributions, but the juicy bits driving progress are rarely among them, and solution quality varies wildly.
- Software experiments can be run and discarded fast, accelerating overall progress, yet individual developers are more compute/GPU-constrained than ever, and tokens aren't free.
- The drudgery of the workflow is unsolved: producing something beyond useless demos, and designing real user workflows, still takes patience.
More from AGI Musings
- The Carlsen vs Stockfish lesson: nobody pays to watch the engine — teodorio · 2026-11-11
- Ross Girshick posts CVPR slides recounting 1.5 decades of object detection and its link to the LLM era — CatAstro_Piyush · 2026-11-11
- Ben Todd wrestles with a 50:50 chance only 18 months remain of 'the human era' — ben_j_todd · 2026-11-11
- Mathematicians face identity crisis as AI solves problems in hours, Corriere interview warns — ValerioCapraro · 2026-11-11
- Why inference speed will be AI products' biggest competitive edge in 1–2 years — itsOmSarraf_ · 2026-11-11
- AI Defense Shifts to Scale Economics: Attacks Buried in Legitimate Traffic — DavidLinthicum · 2026-11-11