Why this is in the vault
Weekly "This Week in Stratechery" curation issue (2026.38, "Doomforce") — a pointer to the week's bundle content plus three curator picks on AI-doom discourse, US-China AI pacing, and Salesforce's pivot to agents-as-UI.
Curation section
- The View From Anywhere But San Francisco (Ben Thompson) — pushback against the past week's AI-doom debates for shutting out dissent; points to Andrew Sharp's Sharp Text piece "Some of All Fears," which argues the AI-anxiety conversation of the last 10 days has been "absurd and irresponsible" and urges solving real problems only when they're actually real.
- The Limited Potential for a Pacing Deal (Andrew Sharp) — Sharp China episode with Bill Bishop on why China is unlikely to agree to a US deal slowing AI development; the CCP's history of successfully controlling technology (e.g., censoring the early internet) informs its confidence managing AI risk; rising discontent on Chinese social media may reflect how the Party manages social malaise rather than a censorship failure.
- The Salesforce Zag (Andrew Sharp) — flags Wednesday's Stratechery Update on Dreamforce, where Marc Benioff climbs down from his 2024 AgentForce commitment and instead leans into Anthropic/OpenAI chatbots as the preferred UI for Salesforce customers — charging a premium rather than fighting the shift to agents-as-UI, with the risk that this erodes the SaaS UI moat broadly.
- Also linked in this issue (already vaulted): "Pacing the Frontier, AI's Digital Limits, AI Commissars" and "OpenAI Ads, Amazon Ads in ChatGPT, Walmart to Accept Apple Pay."
- Bundle round-up (not vaulted individually): Dithering ("Football and the Frontier," "iPhone 18 Pro"), Asianometry ("Silicon Valley's Got That Energy (But No Compute)," "Exxon's Office Fling"), Sharp China full episode, Greatest of All Talk NBA preview, Sharp Tech ("Doom Debates Go Mainstream"), and the week's Stratechery video "Write Things Down."
Mapping against Ray Data Co
The Salesforce item is the concrete one: Benioff abandoning UI-as-moat and ceding the interface layer to Anthropic/OpenAI chatbots is the same "agents replace the UI" dynamic RDCO's own bet on Claude-native, agent-first delivery (vs. building a proprietary dashboard layer) assumes will play out industry-wide — direct evidence for the L5 north star thesis that bets are downstream of agent capability, not proprietary UI. The AI-doom-discourse item is lower-signal for RDCO directly but is useful calibration context for how "AI risk" gets talked about publicly while phData/Anthropic-cert work proceeds on the assumption of continued frontier scaling.
Related
[[2026-09-16-stratechery-salesforce-ai-force-agents-as-ui-headless]] [[2026-09-14-stratechery-pacing-the-frontier-ai-commissars]] [[2026-09-15-stratechery-openai-amazon-ads-walmart-apple-pay]] [[project_l5_north_star_strategic_direction]]