header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP
Partnered With:

Sarah Guo Counters AI Investment Despair: Quantifiability Will Eventually Be Cheap, Private Correctness Is the Moat

According to DataBeat monitoring, Forbes Midas List venture capitalist and Conviction founder Sarah Guo responded to the AI Investment Despair Syndrome spreading throughout the 2026 venture capital community with a detailed article.

This type of emotional concern revolves around the fear that the evolution of large models' capabilities will eventually engulf everything, rendering all applications as valueless thin wrappers. However, she emphasizes that any work that can be quantified through benchmark testing is rapidly becoming commoditized, with the true premium shifting to the untrainable corners – those highly proprietary, validation-costly implicit tasks.

Taking software engineering as an example, in 2024, Devin could only address 13% of evaluation tasks. Currently, top AI agents are scoring above 80% on the SWE-bench. Yet, research by Mert Demirer from MIT and others shows that the latest programming agents have led to a 180% surge in code written, but the code successfully merged and deployed increased by only 30%. Code composition has become a cheap commodity token, while ensuring the smooth operation of decades-old complex legacy systems in the real world remains a slow moat that capital cannot forcibly breach. As OpenAI scientist Noam Brown puts it, the only reliable way to assess an agent's performance over a year is to run it for a year.

Becoming smarter through large models does not make private reality public; access permission and accountability are the bottlenecks. Models cannot sign agreements or assume responsibility, which was the core logic behind her early-stage lead investment in Baseten and investments in OpenEvidence and Harvey. The three have built barriers by locking in the environmental interfaces and user habits. Currently, OpenEvidence has already covered over half of American doctors, and her first client for Harvey was also a referral.

Sarah Guo's Conviction Fund has precisely invested in 27 companies, with a total valuation of over $62 billion, including six star AI unicorns such as Baseten.

举报 Correction/Report
Correction/Report
Submit
Add Library
Visible to myself only
Public
Save
Choose Library
Add Library
Cancel
Finish