PodcastTBPN

OpenAI CEO Sam Altman at DevDay on Agents That Simplify Work, the Compute Buildout, Simpler Model Choice and the Unsolved Alignment Problem, on TBPN

Written by an Ethmos research agent · shared with you

OpenAI's custom chip program goes live in H1 2027 to boost inference efficiency as agent demand (dots) drives a major compute buildout.

ShowTBPN

Episode dateSeptember 29, 2026

Runtime16 min

FeaturingSam Altman, John Coogan, Jordi Hays

Chip program timeline: Altman says OpenAI's in-house chip program comes online in the first half of 2027 and will scale in future years, giving a large inference advantage for whatever power it can secure.
Simplifying the model picker: The goal is a single model that makes tradeoffs itself, with speed/cost dials for edge cases, though developer choice among named model variants will remain.
dots agent needs tuning: Altman says OpenAI has tuned proactive notifications from dots too cautiously and wants more, while acknowledging misuse/growth-hacking risk.
Open ecosystem push: OpenAI is letting users bring ChatGPT subscriptions and compute into any third-party app or harness, contrasting this with rival AI subscription models that lock usage in.
Alignment still unsolved: Altman calls claims that alignment is solved dangerously wrong, stressing it's a science-plus-engineering problem requiring more research before superintelligent models arrive.
Startup culture at scale: He attributes OpenAI's velocity to continuously increasing per-person workload and impact, avoiding the political drag that slows large public companies.

Deep dive

From Loopt to personal agents

A host notes the interview falls almost exactly twenty years after Altman's earlier startup Loopt launched, drawing a parallel to today's agent demos (planning barbecues, booking restaurants) that mirror Loopt's original vision. Altman agrees it's a full-circle moment once pointed out. He says he personally uses OpenAI's new proactive agent, dots, mostly for work rather than personal life, though he's inspired to extend it to leisure, and pushes back on the binary fear that AI will make people work more or less — he thinks most people want more time for kids, hobbies, and creative projects.

The full analysis is a free trial away.

An Ethmos agent found this conversation, listened, and wrote this brief. Yours can do the same — across the shows you follow and the thousands you don't.

Read free — start your trial

TMTB readers: 14-day free trial + 25% off your first month

Also inside: 01 Pace of the last six months · 02 Open ecosystem and agent risk · 03 Startup culture, compute, and model simplification · 04 Forecasting and unsolved alignment

This page is an AI-generated summary and analysis prepared with Ethmos and shared by an Ethmos user. It does not reproduce the original programming. The underlying episode and its recording remain the property of their respective creators, and all show and company names are the property of their respective owners.

AI summary. May contain errors. Not investment advice.