# Noam Brown：测试时计算无上限，或解百万美元难题

- 来源：Chubby♨️ (@kimmonismus)
- 发布时间：2026-08-01 18:25
- AIHOT 分数：38
- AIHOT 链接：https://aihot.virxact.com/items/cmsa8agxq00anrojd8umexzjm
- 原文链接：https://x.com/kimmonismus/status/2083499230765744537

## AI 摘要

OpenAI 推理模型关键架构师 Noam Brown 称，尽管未发布模型已找到多个难题解法，但进一步增加测试时计算（TTC）或可解决百万美元级问题。他承认尝试其他重大难题尚未成功，也未攻克千禧年大奖难题，但每个问题投入不高，测试时计算仍有巨大提升空间。

## 正文

Noam Brown is one of the key architects behind OpenAI's reasoning-model push and a foundational contributor to o1， the model family that established test-time compute as a powerful new scaling axis.

Therefore， one should take very seriously the statement that， despite the fact that the unpublished model has already found numerous solutions， with additional TTC （test-time compute）， perhaps even million-dollar problems could be solved.

In short： there's no end in sight. The sky's the limit.

### 引用推文

> Noam Brown：And yes we did try other major problems without success. Sadly no Millennium Prize problems (yet). But also, we didn't spend a lot on each problem. It's possibl...