Stanford MS&E435 Economics of the AI Supercycle | Spring 2026 | Enterprise Internal Knowledge

Why It Matters

Specialized enterprise models unlock the commercial value sitting in proprietary corporate data, offering firms a way to convert frontier AI advances into operational advantage and new business-critical applications. This shift could reshape competitive dynamics by privileging companies that can effectively integrate and deploy domain-tuned models.

Summary

Yash Bottel, founder and CEO of Applied Compute and an OpenAI alum, described his rapid path from Stanford to OpenAI research and then to founding a startup that helps enterprises build specialized AI models. Drawing on work in post-training, evals, and agentic coding (e.g., CodeEx and Long Horizon Tasks), he argued that frontier reasoning models are powerful but lack domain-specific business knowledge. Applied Compute’s approach is to combine these frontier techniques with proprietary enterprise data to create tailored models that improve real-world workflows. The talk also placed recent AI advances in historical context—highlighting how scale in compute and data since AlexNet has driven dramatic gains in capabilities.

Original Description

For more information about Stanford’s graduate programs, visit: https://online.stanford.edu/graduate-education

To follow along with the seminar schedule, visit: https://hci.stanford.edu/

This seminar covers intelligence and unlocking enterprise internal knowledge.

Guest speaker: Yash Patil, CEO of Applied Compute

Instructor: Apoorv Agrawal is a Stanford Adjunct Lecturer in Management Science and Engineering and a partner at Altimeter Capital.