Ben Thompson argues in Stratechery that AI competition reflects competing tech philosophies

Ben Thompson argues in Stratechery that AI competition reflects competing tech philosophies

Ben Thompson’s Stratechery article, published July 8, 2025, argues that AI’s promise of abundance is already producing scarcity in chips and sought-after talent. He connects that competition to different ideas about what computers should do: act for people, help people work better, or improve enterprise results by replacing some work. In Thompson’s account, Apple’s view of AI as a tool helps explain its restraint, while Meta’s ambition to have computers do things for people helps explain its aggressive hiring. He sees Google as pursuing AI despite a tension between direct answers and an advertising business built around search results. Microsoft’s tools depend on employees choosing to use them; agents, by contrast, can do work in place of people. Thompson says Anthropic’s opportunity is to build agents that deliver business results, while warning that the boundaries between these approaches are fuzzy and companies can pursue more than one. The stakes are not only which companies attract talent, but whether AI augments employees or replaces them. Thompson ends by suggesting that incumbent tool-makers may need to partner with those who believe in computers doing everything. Whether Anthropic can build the agent product he describes remains unknown.

Publication: 49f42cd3-81b3-0f37-f012-156086b25943

attribution — pass: The reading consistently attributes the framework and arguments to Ben Thompson in Stratechery. This matches the single-author analytical article in the evidence, and quoted material from Bloomberg and Christensen is not misattributed to him.

It was AI, at least the pre-LLM version of it, that inspired me in 2018 to write about Tech’s Two Philosophies

summary — pass: The summary faithfully conveys Thompson's abundance/scarcity paradox and the tool-versus-doer philosophies. It also covers the placements of Apple, Meta, Google, Microsoft and Anthropic, his fuzzy-boundaries caveat and his closing partnership suggestion. Agents are explained plainly as doing work in place of people. A minor omission is his suspicion that mission-driven researchers favor computers doing everything, but this does not distort the meaning.

my suspicion is that to the extent mission is a motivator the more likely an AI researcher is to be enticed by the idea of computers doing everything

date — pass: The evidence gives originalPublished as 2025-07-08T15:06:51Z, and the source dateline also reads Tuesday, July 8, 2025, so the stated July 8, 2025 date matches. The evidence kind is publisher-page, so no archive disclosure is required. The reading does not use the later footer link titles and contains no hindsight framing.

Tuesday, July 8, 2025

support — pass: Each claim is supported by a source passage. The abundance/scarcity point, Meta's philosophy, agents as replacements, Anthropic's agent goal with the one-framework caveat, and the incumbent partnership suggestion all match the source. The augment-versus-replace stake reflects Thompson's own discussion. The reading keeps open as unknown whether Anthropic succeeds.

Anthropic’s task is to build the best agent product in the world, including not just state-of-the-art models but all of the deterministic computing scaffolding that actually makes them replacement-level workers.

media — pass: No media is selected, and no verified candidates were supplied, so there is no identity or relevance problem.