Interview, Fireside Chat
Former Chief Scientist at Salesforce, Richard Socher | You.com, LLMs, AI Agents, Complex Work
- The company plans to leverage a $50 million Series B fundraise over the coming years to transition from search to a "productivity engine" for complex knowledge work in sales, service, marketing, research, and VC firms, aiming to be 10x more accurate than Google.
- Revenue is currently doubling every quarter, with a strong pipeline of enterprise customers including hedge funds, biotech, media, and insurance sectors expected to utilize private RAG on massive internal data sets.
- The firm is actively hiring engineers, VPs of engineering, and sales, marketing, and product leaders to support scaling efforts and the development of capabilities allowing agents to surf the web, interact with specific databases, and execute web-based clicks.
- Predictions indicate that in the next few years, AI agents will increasingly surf the web for users, potentially shifting the internet's operational model and impacting the advertising economy.
- Over a five-to-10-year horizon, agents are expected to automate repetitive workflows performed a dozen to two dozen times, leading to a labor model where humans manage teams of AI agents while the marginal cost of intelligence declines.
- Long-term forecasts suggest that countries embracing AI will achieve significantly higher success metrics, while biology is expected to transform into an engineering discipline where AI designs proteins for medicine.
- Two distinct AI agent company types are expected to emerge: horizontal platforms facilitating entire corporate transformations and vertical-specific agents fully automating roles such as recruiting.
- Technical risks include the high failure rate of multi-step agent workflows (e.g., 20 steps with 95% accuracy per step failing half the time) unless implemented with robust error handling.
- The labor market may see reduced demand for average programmers as AI handles simpler tasks, increasing reliance on elite coders to guide agents, though a "unicorn with a single employee" is unlikely within the next few years.
- AI funding is projected to remain positive for companies demonstrating consistent revenue growth and concrete milestones, while investment will focus on founders combining deep industry expertise with AI-native knowledge.
- Automation of physical work and robotics is anticipated to progress more slowly than software due to biological and hardware constraints compared to pure intelligence scaling.