Updated · 1 episodes · 1 show · 1 source notes

entity Topics: Technology

Gu Yu / 谷雨

Overview

Gu Yu is presented as a co-founder of New York Cognition working on cognitive learning, continual-learning evaluation, and AI research-company development.

Current Profile

In the supplied interview, Gu Yu is most useful as a reflective operator who revises prior beliefs about RSI startups, computer use, informal mathematical reasoning, memory notes, academic evaluation, and product release. His recurring distinction is between technical capability and the surrounding system that makes it useful: workplace cognition beyond interface operation, sequential learning beyond single long tasks, and market closure beyond internally coherent research.

Key Characteristics

  • AI researcher and startup co-founder focused on cognitive and continual learning.
  • Willing to make prior judgments explicit and revise them after capability or market evidence changes.
  • Advocate for evaluating agents through accumulated production learning rather than isolated tasks alone.
  • Skeptical of paper counts and low-selectivity review as reliable research-quality signals.
  • Founder who emphasizes autonomy, technical taste, market feedback, and tolerance for uncertainty.

Evidence

Research and benchmark work

Revised technical judgments

Research and company evaluation

Qualifications

  • The profile relies on one structured podcast summary rather than a transcript, CV, publication record, benchmark documentation, or company filing.
  • Statements about teams, models, companies, academic systems, financing, and future capability remain attributed judgments.
  • The supplied source gives limited biographical detail beyond his company role and reported operating preferences.

What Changed

  • Established Gu Yu’s profile around continual-learning evaluation, revisable technical judgment, and research-to-market discipline.

Relationships

Sources

1 source notes across 1 show
  1. 【年终复盘②】RSI可能是NeoLab的机会、PA真的有用、创业要造灭火器还是砖头?|对话谷雨 十字路口Crossing