analiza aiqbench
The proposal advocates for Verified Compute Allocation Disclosures (VCAD) to enforce a 15% cap on frontier model training starting in 2027, aiming to delay an intelligence explosion by approximately one year through reallocation of resources toward safety research and customer service. While the authors acknowledge this specific measure may be insufficient as a standalone solution against rapid recursive self-improvement, they argue it is a necessary foundational step for implementing more durable policy regimes like Plan A or Plan S. The strategy relies on reducing inputs to AI development or enforcing inefficiencies via safeguards rather than relying solely on voluntary compliance.
Pacing AI via Compute Reallocation

De la AI Futures Project
FragmentThis is a linkpost for pacing-ai.com. I encourage folks to read it via the site (which has interactive graphs and expandables), but we’ve copied the text here as well.
We offer a concrete proposal for pacing the frontier.
1. Transparency: Verified Compute Allocation Disclosures (VCAD)
2. Reallocation: Shift compute away from capabilities R&D
The first component is Verified Compute Allocation Disclosures (VCAD): AI companies regularly disclose their compute usage, and an independent auditor verifies the disclosures. The second component is to reallocate compute away from capabilities R&D and toward other use cases, such as serving customers and safety research (non-capabilities R&D).
We roughly estimate that pacing via compute reallocation starting January 2027, with a 15% cap, would delay the intelligence explosion by a year. This is most likely insufficient to prevent AI takeover (or mitigate other problems created by RSI), and so this year should be used to implement a more durable policy regime such as Plan A or Plan S.
Extras afișat conform termenilor RSS ai publicației · 161 cuvinte · textul integral și figurile sunt pe site-ul publicației
Subiecte conexe

Pacing AI via Compute Reallocation
LessWrong · 18 nov. 2026
Learning to Solve Hard Problems in RL for LLMs by Never Giving Up
arXiv · 16 sept. 2026
Unifying Data, Memory, and Compute Efficiency in LLM training: A Survey
arXiv · 10 iun. 2026
LatentLoop: Scaling Looped MoE Models with Asymmetric Compute Allocation
arXiv · 15 oct. 2026