Monday, September 29, 2025

Start of archive·May 16

Top story

The Agent RaceOpenAI Blog

Sora 2 System Card

OpenAI's next-gen video model introduces capabilities (physics accuracy, synchronized audio, enhanced steerability) that redefine the competitive landscape for multimodal AI. This is a direct capability leap that will force competitors to recalibrate their roadmaps.

Sora 2 announced as 'state of the art' video and audio generation model

The briefs

As AI agents proliferate, organizations must distinguish between automating meaningful work versus creating elaborate presentations, with the future favoring companies that deploy agents for actual productive tasks rather than administrative theater.

Race between human-centered work and infinite PowerPoints

Vercel is converting Claude Sonnet 4.5 into a production developer tool, not just a model. This moves coding AI from chat to automated workflows—repos, CI/CD, and shipping.

Claude Sonnet 4.5 now available on Vercel AI Gateway with full SDK support

OpenAI's next-gen video model raises the bar on physical accuracy and real-time control, signaling intensified competition in the generative video space and potential implications for creative workflows and synthetic media production.

Sora 2 released by OpenAI

OpenAI is launching Sora 2 and a consumer-facing Sora app with embedded safety protections as core architecture, signaling a shift from research model to productized video AI platform. This matters because it's the first major test of whether video generation can reach mass users without policy/brand backlash.

Sora 2 released with safety-first architecture

OpenAI is dogfooding an AI research assistant for internal knowledge work at scale—a signal that agentic workflows for document analysis and insight generation are moving from concept to operational reality in enterprise settings.

OpenAI research assistant processes millions of support tickets

OpenAI is dogfooding its own agent capabilities for enterprise workflows—a real-world validation that AI agents can handle complex document processing at scale, with direct implications for how companies should think about deploying agents internally.

OpenAI built internal contract data extraction agent

OpenAI is dogfooding its own AI to solve go-to-market efficiency. This is a real-world deployment case study showing how AI agents can scale sales productivity — and a signal of what OpenAI believes works operationally.

OpenAI built internal GTM assistant for sales productivity

OpenAI is positioning itself as a reference customer for its own technology, using a public case-study series to demonstrate real-world deployment patterns and build credibility with enterprise buyers considering AI adoption.

OpenAI launches 'OpenAI on OpenAI' content series

OpenAI is using its own models to automate inbound sales workflows, demonstrating real-world deployment of AI agents in revenue-critical processes—a case study in dogfooding that shows practical business impact beyond research.

OpenAI built internal AI sales assistant for lead conversion