The Rise and Fall of Jev
Web page · https://blog.alexewerlof.com/p/rise-and-fall-of-jev-eval-hype-vs-llm
First seen by Oriel on .
AI summary
The article uses Jev’s rapid hype cycle to argue that AI progress has mostly been incremental, while marketing, industry drama, and unresolved technical and economic limits fuel inflated expectations.
- Jev is described as a fast, low-cost general-purpose classifier that produces schema-conforming outputs, but it cannot reason and can still be wrong.
- The author says similar decision models predate Jev, while Jev’s architecture and RLCD training method were not disclosed; open-weight alternatives may offer better performance per cost for some tasks.
- The article traces progress since ChatGPT through instruction tuning, function calls, coding harnesses, agents, larger contexts, multimodality, tool protocols, orchestration, sandboxes, and memory.
- The author identifies abstract reasoning, continuous learning, real-time embodied control, and the economics of running AI as major unsolved challenges.
- The article argues current AI can still cause harm through tools and automation without reaching AGI, while hype and predictions about its broader impact remain unreliable.
#artificial intelligence #llm agents #ai safety #robotics
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 84, latest observation ; +70 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 63, latest observation ; +57 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50206361 | https://blog.alexewerlof.com/p/rise-and-fall-of-jev-eval-hype-vs-llm Read discussion |