OpenAI's 'Code Red,' Anthropic's 81,000-user study, and the new AI org chart

OpenAI refocuses on coding and 'Work AGI.' Anthropic publishes the largest user study of an AI product to date. And major companies start restructuring around AI.

A ‘great convergence’ week. The major labs are all narrowing focus to similar core bets, and the human-side conversation is finally getting more nuanced.

OpenAI shelves Sora, doubles down on coding and 'Work AGI'

My take: Video generation is being sunset. Compute is being shifted to a new model called 'Spud.' The Codex/ChatGPT/Atlas combo will be unified into a desktop super app. The product division is being renamed 'AGI Deployment.' Read between the lines: even OpenAI is making hard choices about where to focus. If they're choosing coding and knowledge work over consumer video, that should tell you something about where the durable value is.

Microsoft consolidates consumer and commercial Copilot teams

My take: Mustafa Suleiman is now focused entirely on proprietary model training. The unified Copilot product should resolve a lot of the user confusion that's plagued Microsoft's AI offering. If your business runs on Microsoft 365, the next 6-12 months should bring noticeably better AI integration. Plan to revisit your training and adoption strategy when it lands.

Anthropic ships 'Claudification of Everything'

My take: Computer control, mobile-controlled desktop sessions, scheduled cloud tasks, channels for external events to push into Claude Code sessions, 1M context generally available, memory and connectors on the free plan. The pace is staggering. Practically: if you've been waiting for Claude to be 'enterprise-ready,' that gate fell over in March.

Jack Dorsey writes up the new AI org chart

My take: Block's plan: replace traditional management hierarchy with an AI 'intelligence layer.' Three roles: Individual Contributors, DRIs (Directly Responsible Individuals), and Player Coaches. No middle managers as information routers. Whether or not Block makes it work, this is the kind of thinking that's going to spread. Worth reading and pressure-testing against your own org structure.

Anthropic's 81,000-user study: hopes and fears coexist

My take: Top hopes: professional excellence, personal transformation, winning back time. Top concerns: unreliability, job/economic impact, loss of autonomy. Existential risk was a very low concern. The interesting bit: early economic gains are heavily skewed toward entrepreneurs, small business owners, and people with side projects. If you have one, lean in.

FedEx training all 400,000 employees on AI

My take: HSBC, by contrast, is considering laying off up to 20,000 (10% of workforce) over 3-5 years. Two very different models. Worth deciding now which side of that line your organization sits on, because the messaging difference matters enormously for retention.