Coding doesn't take that much intelligence — LLMs brute-force it like Deep Blue
ivan_bezdomny · x · 2026-09-27
- The author muses that LLMs are "so stupid, yet so productive on code," suggesting coding itself may not require much intelligence — humans find it hard mainly due to memory constraints.
- Coding agents largely brute-force programming, analogous to Deep Blue in chess: machines can win without deep understanding.
- Irony: we designed programming languages for humans (machine-interpretable as a side effect), yet it turns out machines are the far better programmers.
More from AGI Musings
- Are frontier models' inner optimizers obsolete? CoT as goal-directed search sparks alignment debate — xuanalogue · 2026-09-27
- DHH roasts frontier-lab doomer class at Rails World 2026: more P-Bloom, less P-Doom — rohanpaul_ai · 2026-09-27
- Full video: DHH's Rails World 2026 talk roasting AI doomers — rohanpaul_ai · 2026-09-27
- Why mesa-optimization fell out of favor: goal misgeneralization is more precise — xuanalogue · 2026-09-27
- AI safety folks debate whether mesa-optimization is still the right risk-communication concept — xuanalogue · 2026-09-27
- Grinding LeetCode today is like cramming for the 1910 imperial exam, says ex-Tesla AI lead — henloitsjoyce · 2026-09-27