AI's third era is the persistent co-worker, and product cycles last only three months
OpenAI product lead Tara Seshan argues that AI products are moving from chat to agents to persistent co-workers, and that builders must design for model capabilities two to three months out; the future of work is steering, not rowing, and the writing that is thinking should never be automated.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Build products for the model capabilities of two to three months from now
Tara divides the evolution of AI products into three eras: the first era is chat, the second is collaborating with agents, and the third era, arriving now, is working alongside a persistent co-worker who can get things done with you. She stresses that when building a product you should base it neither on current model capabilities nor on predictions of what models will do a year from now, but on model capabilities two to three months out, because both extremes fail.
— Tara SeshanThe future of work is steering, not rowing
Tara believes the future of work will be more about ‘steering’ than ‘rowing’: agents handle execution, humans handle direction. The level of abstraction in steering will keep rising, but ultimately humans still need to make opinionated decisions. She stresses that even as the steering layer rises, human judgment and accountability at key junctures remain irreplaceable — that is both the responsibility and the value.
— Tara SeshanThe end state of a product is that users don't have to choose tools
Tara explains that ChatGPT's ‘work mode’ is built on Codex underneath, but with a different UI. The product's north star is to let users avoid choosing, with the system automatically picking the right tool for the task. She stresses that ‘done beats perfect’ — get it into users' hands first, then fix based on feedback; fast iteration matters more than polish, and she says ‘this is the worst the model has ever been.’
— Tara SeshanCodex hasn't changed; what changed is users' perception
The host observes that the conversation on Twitter about Claude Code and Codex has shifted over the past few months, with more people leaning toward Codex. Tara responds that the Codex team has always stayed user-oriented and iterated quickly, and that how it operates internally hasn't changed — the market and users simply came to realize it. She argues the change isn't in the team, it's in users' perception.
— TaraThe writing that is thinking should never be automated
Tara distinguishes ‘writing as thinking’ from ‘writing as reporting.’ The former she never automates, because it is key to organizing her thoughts; the latter she hands entirely to the model. She recommends writing a doc to 70% and then getting feedback, rather than chasing perfection. At OpenAI she still writes a lot of docs, but mainly for herself, because long documents are no longer a signal of depth of thought.
— TaraBefore product-market fit, there is product marketing fit
The most important lesson Tara learned at Sutter Hill is that product-market fit matters, but she underestimated ‘product marketing fit.’ Before building a product, you should sharpen the narrative and positioning through a lot of selling. She believes excellent product marketing work can determine a company's success or failure, and says Mike Speiser is unmatched at it.
— TaraIn their own words · checked verbatim
You fail if you build for where the models are now. You fail if you build for where you think the models will be in a year. Like both outcomes are equally wrong.
Tara Seshan27:14
This is the worst the models will ever be.
Tara Seshan29:14
I think humans will continue to be most valuable as a certainly as a um like an entity of accountability.
Tara46:25
writing is thinking is something I never will automate.
Tara53:30
product market fit is sure important but actually I really underrated product marketing fit.
Tara1:02:43
the teal fellowship was an inflection point in my life I wouldn't be where I am without it
Tara1:16:51
you've finally got these things running in the cloud doing real work
Tara1:20:56
Figures
| Model capability planning horizon | 2-3 months | 27:14 |
| Codex monthly active users | 10 million | 33:16 |
| Recommended doc completion | 70% | 57:36 |
| Teal Fellowship cohort size per year | 20 | 1:17:52 |
| Teal Fellowship stipend | $100,000 | 1:17:52 |
Glossary
- persistent co-worker
- An AI agent that works alongside you over the long term and proactively completes tasks, as distinct from one-off chats or short-lived agents.
- product marketing fit
- The degree to which a product's narrative and positioning match the needs of a market segment, with emphasis on sharpening positioning through selling early on.
How to listen
AI product managers, founders, investors, and engineers who want to know what the future of work looks like.
You can skip the craftsmanship discussion at 42:22; it doesn't affect the main thread.