Lists / People

David Luan

Research-and-engineering leader (OpenAI VP Eng, Google Brain LLM lead) who crossed into building applied AI agents — co-founding Adept and running Amazon's AGI Lab.

AI research-and-engineering leader turned agent-company founder. An early VP of Engineering at OpenAI (helped ship GPT-2, scaled the team from ~30 to ~120), then led large language model efforts at Google Brain, before co-founding Adept (2022) as CEO to build agents that can "do anything a human does on a computer." In June 2024 Amazon hired Luan and senior Adept staff and licensed Adept's technology; he became VP of Autonomy and head of Amazon's AGI Lab (formalized December 2024), shipping Nova Act in March 2025, before departing in February 2026 to start something new.

Website
#102Technical Founders
39/ 100
Authority scoreEmerging
1lists#102peak1primary1category
Frameworks

Attributed to David Luan

Mental models, principles, and operating frameworks extracted from sources where David is the credited author or speaker.

6frameworks
01
Behavioral-Cloning Shortcut to AGI
You don't have to re-run evolution's compute to reach AGI — you can behaviorally clone everything humans already know.
Innovation·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
02
Drive by Vision, Not APIs
Itemize the workflows where every step has an API and the count is near zero — so the agent must see and act on the screen like a human.
Innovation·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
03
Multimodal for Knowledge Work, Not Cats and Dogs
Economic value is in knowledge work, so train multimodal models on charts, tables, invoices, PDFs and UIs — not COCO.
Innovation·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
04
Reliability Over Generality
Demos work 60% of the time; enterprises need the nines. Prioritize reliability over everything, then climb the abstraction ladder.
Strategy·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
05
Research Leader as Portfolio Allocator
A research org wins by concentrating compute on a few big bets and disbanding the rest — not by letting a bottom-up credit market allocate it.
Leadership·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
06
The Augmentation Company
Always solve tasks just a little too hard for the model — keeping a human in the loop is what builds the data flywheel.
Strategy·from Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept
Domains

Framework distribution

6frameworks
Innovation3 frameworks · 50%
Strategy2 frameworks · 33.3%
Leadership1 framework · 0%
Appears alongside

Top neighbors of David Luan

8people
Last updated
29 May 2026
Suggest an edit →