Are robots under-discussed? Why LLMs can't directly control robots and VLAs may be the answer
Grouchy-Stranger-306 · reddit · 2026-10-24
A Reddit thread arguing robots are under-discussed relative to their potential economic impact. The author contends that LLMs sending direct commands to robots fails on speed, cost, and precision—current demos look unimpressive and are far from working several hours autonomously; a hybrid architecture where the language model only reasons and talks without controlling motors seems more plausible.
Open questions raised: are VLAs or models like GEN-1.5 the real path? Could LLMs eventually train such models end-to-end—generating and scoring simulation environments to fix the scale problem? On Tesla Optimus shipping late 2027, he's skeptical: true autonomy would imply a private breakthrough, but everything public so far has been remote-controlled demos.
More from AGI Musings
- Noam Brown: identical-weight agents would converge on shared protocols without pre-agreement — polynoamial · 2026-10-24
- No boundary persists through evaluation, and most people will resist that reality — natesiggard · 2026-10-24
- Palantir CEO Alex Karp: AI could make him 20x richer — a problem for society — rohanpaul_ai · 2026-10-24
- Australian banks use AI as cover for layoffs driven by mortgage slump fears — gaganghotra_ · 2026-10-24
- Healthcare AI is outpacing decades-old EHR systems, and patients are at risk — luisdans · 2026-10-24
- Hardware engineer's long rant: AI will never have moral rights, stop debating machine consciousness — Recent-Day3062 · 2026-10-24