読書リサーチレーダー投資フレームワーク
ログイン / 新規登録
← 読解に戻る
ニューストピック · 可纠正性研究基金クイックテイク1 分で読めるDisconfirmAI 編集部

Corrigibility fund adds $48,000 in prizes for recent papers

A second retroactive prize round lands, with over $60,000 planned for December; whether it becomes routine is what to watch.

2026-10-01 00:08 GMT+8

The Corrigibility Research Fund announced an additional $48,000 on September 30, rewarding roughly two dozen researchers across about a dozen teams working on corrigibility — making AI systems accept human correction.

Fund manager Max Harms handed out $27,000 in a first round in July and plans to disburse more than $60,000 in December. The largest single award, $14,000, went to Rubi Hudson's paper on a corrigibility transformation.

The selection is one person's judgment; Harms himself calls the awards ad hoc and says the purse sizes should not be taken too seriously, and he deliberately passed over established figures such as Yudkowsky and Christiano.

ソース

  • lesswrong.com↗
トピック

可纠正性研究基金

可纠正性研究基金(Corrigibility Research Fund)是 Lightcone Infrastructure 旗下、由 Max Harms 管理的小型 AI 安全资助项目,2026 年至少发放 20 万美元,资助让 AI …

7 ステップ1 件更新日 2026-10-04 14:52

経緯

ログイン / 新規登録
読書リサーチレーダー投資フレームワーク
  1. 2026-07-17基金正式宣布设立,说明其定位、可纠正性研究的界定及资助与奖励双轨机制。corrigibilityresearch.org ↗
  2. 2026-08-04Granted AI 发文详解该基金:2026 年至少发放 20 万美元,资助单笔 5千–3.5万美元,首轮申请截止 8 月 23 日,奖励分 9 月底约 4 万与 12 月中 6 万+两波。grantedai.com ↗
  3. 2026-08-23第一轮资助申请截止。grantedai.com ↗

未解決の問題

  • 追溯奖励能否常态化:10 月 1 日已是连续第二轮发放,12 月中旬超 6 万美元的第二波是否如期落地是直接检验。
  • 第二轮资助(10 月 31 日截止)的获资助名单与金额。
  • 这种小额、低摩擦模式在 AI 安全资助格局中是否会被更大机构复制,材料未给出判断。
トピック全文を読む →
リサーチを続ける →