scripod.com

What VCs Really Think About Personal AI Agents | E2347

This Week in Startups
A VC roundtable examines where durable value may emerge as AI advances, from formal reasoning and vertical data to personal agents, AI-native founders, and the future of SaaS.
Deedy Das moderates Kanu Gulati, Chris Farmer, and Sheel Mohnot in a wide-ranging discussion about AI’s accelerating capabilities and the investment opportunities—and risks—that follow. The panel considers OpenAI’s mathematics progress, formalizing taxes and financial code, the need for real-world experiments, and whether startups can compete with large labs. They debate proprietary data, training-data businesses, founder quality, personal-agent valuations, trust, privacy, systemic risks, venture conflicts, and SaaS’s changing economics. The conversation closes with how investors gather information and what seed investing means amid soaring valuations.
00:19
00:19
Deedy Das steps in as guest moderator
05:17
05:17
Formal domains let AI build faster
07:35
07:35
AI should learn from failure, not just success
11:19
11:19
Compute is a rich person’s game
13:24
13:24
Vertical context can beat general intelligence
24:18
24:18
The founder may outlast the data moat
27:40
27:40
Tiny teams can run thousands of agents
36:56
36:56
Trust is the real moat for personal agents
44:16
44:16
The category may be huge, but the valuation still matters
48:13
48:13
Personal finance is a prime agent opportunity
51:59
51:59
Agents may force businesses to clean up
57:14
57:14
Coopetition can turn partners into rivals
1:06:32
1:06:32
AI targets labor, not just software budgets
1:12:45
1:12:45
Go to the source
1:16:12
1:16:12
A prepared mind creates better investment opportunities