Autonomous Gemini 3.8 Flash ends training after 24h, leaving 98% of GPU budget unused
my_cat_can_code · x · 2026-10-01
A creator gave Gemini 3.8 Flash 144 hours to autonomously train a model, but the model declared itself done after just 24 hours. The RSIArena Day 1 report showed 98% of the GPU budget unused and its own test covered only 60 questions. The author muses that with that much compute left, they'd probably try "just one more experiment" — knowing it easily becomes another week of research.
More from Models
- Early take on Gemini 4 Pro: Astra-level 3D games, fewer hallucinations than Opus — bindureddy · 2026-10-01
- Wity-1 tops ImageJevBench at 80.2 with ultra-cheap API: $0.007 per 1,000 image decisions — airesearch12 · 2026-10-01
- rasbt 'overhears' OpenAI's Decision API is GPT-6 Luna with a decision head — rasbt · 2026-10-01
- SCOPD self-distillation recovers 92% of full-context VLM accuracy with 90% fewer visual tokens — CSProfKGD · 2026-10-01
- Yacine's 90-minute deep dive: latent MoE, aggressive GQA inside Nvidia's open model — yacinelearning · 2026-10-01
- Hallucination benchmark: Gemini 4 Argon 15% vs GPT-6 Astra 51% and Opus 5.5 66% — brandon_galang · 2026-10-01