Big Technology Podcast Why OpenAI Killed Sora, Did Apple Just Save Siri?, Meta’s Big Loss
160 snips
Mar 28, 2026 
Ranjan Roy, writer of the Margins newsletter and sharp tech industry analyst, joins a lively chat on why Sora fizzled and who could seize the AI video opening. They dig into OpenAI and Anthropic’s assistant rivalry, Apple’s shaky Siri repair, Meta and YouTube’s courtroom setback, platform liability, and why sliding tech stocks may signal deeper AI nerves.
AI Snips
Chapters
Transcript
Episode notes
Why OpenAI Shut Down Sora
- OpenAI killed Sora mainly to focus compute and talent on GPT-style reasoning, not because video AI lacked all demand.
- Alex Kantrowitz says Sora sat on a different tech tree built around world models, which Greg Brockman framed as too costly to pursue alongside core reasoning models.
The AI Race Is Converging On Agents
- OpenAI and Anthropic are converging on the same prize: AI assistants that control devices, access data, and complete multi-step work.
- Alex Kantrowitz ties this to the OpenClaw style use case, while Ranjan Roy notes Sierra, Notion, and Writer are chasing the same autonomous knowledge work market.
AI Trust Builds Through Small Delegations
- Trust in agentic AI will grow gradually as models prove reliable on bounded tasks, not through one breakthrough launch.
- Ranjan Roy describes a 7 p.m. workflow that drafts replies from his Gmail history; he still reviews before sending, but sees quality slowly earning trust.