OpenAI acquires presentation startup NextSlide
Two moves in the last day quietly extended what AI assistants can do at work: OpenAI is pulling presentation-making talent into ChatGPT, and Anthropic's Claude Code sessions just learned to talk to each other.
OpenAI has acquired NextSlide, a year-old startup whose AI turns prompts, notes, documents and research into polished, editable presentation decks. Founder Ahmed Beshry announced the team has joined OpenAI and is now working on ChatGPT; financial terms were not disclosed. Beshry said the public announcement was coming "a few months late," since the deal actually closed earlier this year, and described the product's goal as making "visual communication more accessible." It's his second startup sale — he previously co-founded Caper AI, the smart-cart checkout company Instacart acquired in 2021.
Why it matters: OpenAI has been pushing ChatGPT deeper into workplace productivity, and presentations are one of the last formats where a text answer isn't the deliverable — users need structure, visual hierarchy, and a file they can keep editing. Folding in NextSlide's team is a bet on ChatGPT becoming the place where work gets turned into communication, not just generated. It also raises the pressure on standalone AI-deck tools: when the biggest assistant ships presentations natively, specialists need to be clearly better to survive.
Anthropic's Claude Code now lets your separate sessions message each other directly. Shipped in version 2.1.224 for macOS and Linux, the feature lets one session hand a finding to another, ask a question, or warn a parallel session before a breaking change lands — no more copy-pasting context between terminals. Only text summaries travel, not conversation history, files or permissions, and same-machine messages stay local, never routed through Anthropic's servers; replies across your other machines go through Remote Control. Windows support isn't here yet.
The genuinely useful part isn't agents chatting for its own sake — it's one session warning another before a shared rename breaks its queries, the kind of coordination humans are slow at. It's quiet infrastructure for multi-agent workflows where the human supervises instead of shuttling context between windows.
What to watch: whether ChatGPT's slide features go from "can generate a deck" to a real editing surface, and whether Anthropic extends cross-session messaging beyond text into shared state.
Would you let two Claude sessions coordinate a codebase without you in the middle — and where's the line before you want approval controls back? Tell us in the comments.
