Karpathy: Jev is an under-invested point on the LLM Pareto curve — single-token, low-latency frontier-ish intelligence
karpathy · x · 2026-09-22
Karpathy says the surprise hit Jev sits on the LLM Pareto-optimal curve in a regime of no thinking, single-token, low-latency acceptable intelligence — a regime with large revealed latent demand that was under-invested as the industry raced toward higher intelligence. He calls it a zero-shot classifier with frontier-ish intelligence and wonders which other old ML ideas are worth rescuing with modern models.
Related event: Jev's viral zero-shot classifier sparks talk of reviving old ML ideas(5 posts)→
More from Models
- LangChain hosts open-source decision model SemIf free for a week via LangSmith Gateway — hwchase17 · 2026-09-22
- Xiaomi's MiMo-V2.6 debuts with 1M context, benchmarked just behind GPT-5.6 Sol — MaziyarPanahi · 2026-09-22
- Xiaomi's MiMo-V2.6 lands: 1.02T-param MoE takes top open model spot on Artificial Analysis — _AndrewZhao · 2026-09-22
- Merge Gateway lists Grok 4.7 from xAI, citing coding gains on CursorBench and DeepSWE — shensi · 2026-09-22
- SemiAnalysis says open source is dying, yet 20+ open models shipped in the past month — _lewtun · 2026-09-22
- Grok 4.7 posts 59% recall on defensive cyber bench at half the cost of rivals — andreamichi · 2026-09-22