Howdy wizards,
Here’s what’s brewing in AI.
The big thing
ChatGPT can now answer with little interfaces and explainers.
Source: OpenAI
Answers in ChatGPT can now be little apps or interfaces that help you understand better: a chart, a map, a form, a calculator, a checklist, a small game.
It decides when a question gets turned into a UI.
Or you can just ask: make me a savings calculator, etc. And ChatGPT just makes it and runs it, right in the chat.
It’s available with GPT-6 on all plans, although the paid plans have access to better versions of the model and will make better explainers.
It’s not really something new. Claude has had charts and diagrams inside its answers for half a year already. The difference is distribution: OpenAI has 1.2 billion weekly users.
The next day, Anthropic released a version for work: Claude Dashboards. Claude can now build a dashboard on your Salesforce data right in the chat.
Why it matters
Visualisations are becoming a more common way for AI to communicate with you, across tools (OAI is actually late to the party, Claude and Gemini have both had it for months).
And with good reason. A while back I wrote about how I was using visualisation as a technique to help me understand what I’m building with Claude Code.
It’s a great way to break down complexity and understand concepts fast.
Just remember: the information fuelling what you’re seeing is still processed by AI. You already know AI hides its inability to judge what’s useful and what isn’t in eloquence and em dashes.
It will increasingly hide it in pretty apps and charts. So keep your own judgement sharp.
IN PARTNERSHIP WITH GRANOLA
Most AI notetakers just transcribe.
Granola is the AI notepad that helps you stay focused in meetings, then turns your conversations into real progress.
Engineers use Granola to:
Draft Linear/Jira tickets from standup notes
Paste code into Granola Chat and instantly receive a pass/fail verdict against the requirements that were discussed in your meeting
Search across weeks of conversations in project folders
Create custom prompts for recurring meetings (TL;DR, decision log, action items) with Recipes
Granola works with your device audio, so no annoying bots join your meetings. It quietly works with any meeting app, like Zoom, Meet or Teams, and in-person with Granola for iPhone.
NEWS NEWS NEWS ❦ NEWS NEWS NEWS
All the small things
Industry moves
Three OpenAI safety researchers who were recently fired wrote an open letter. OpenAI says they mishandled sensitive information, and they say working with outside safety people was part of their job. The irony in this story is strong: Sam Altman promising outside auditors employee-like access just recently, then this. Their first demand is that the firing doesn’t become an excuse to drop the outside evaluators. David Robinson, who led OpenAI’s safety reports, quit in the middle of this, saying the culture is broken.
Anthropic subscriptions offer 5x the value of OpenAI, according to an independent test. SemiAnalysis checked what each plan’s usage would cost at API rates, and Anthropic comes out about 5x ahead on both labs’ everyday models. If you were considering switching to Claude — now is a good time.
Anthropic’s usage policy bans being cruel to Claude from November 12. “Sustained and needless” cruelty, specifically. Frustration, pushback and dark creative themes are (still) allowed. Claude just ends the conversation if you persist.
New tools & product features
OpenAI launched Dots, an always-on personal agent. It’s their answer to Meta’s Muse and Grok Bot. You know the drill: A dot has its own computer in the cloud. It can take action. You set the rules for when. It comes with ChatGPT’s subscription tiers starting from $100 a month. Now that everyone is technical enough to run an overnight agent, the actual challenge is finding actually worthwhile problems to do overnight.
Models
Google released its new frontier model Gemini 4 Argon. On the independent tests it’s level with GPT-6 Astra, but with a hallucination rate of 15% compared to Astra’s 51%. For now only a hand-picked group of cyber defenders have it; then rolling out to API customers and Google’s most expensive plan (AI Ultra).
CONTEXT WINDOWS ❦ CONTEXT WINDOWS
🟦 CASE STUDY OF THE WEEK
FIFA let AI screen its World Cup press applications
A five-person team used to check every journalist's application by hand, and 5,000 cases piled up. Now AI reads the documents, checks them against FIFA's criteria and drafts a reply for the team's final decision, and the backlog was cleared in three weeks.
This case study is part of the Agentic Customer Service use case, which has a strong proven impact relative to the 35 other AI use cases we track.
I’ve built Context Windows just so you can see what’s working with AI.
Get a free personalized overview for your industry and your role. Take the 60‑second quiz →
What's your verdict on today's email?
Subscribe to participate
Disclosure: To cover the cost of my email software and the time I spend writing this newsletter, I sometimes work with sponsors and may earn a commission if you buy something through a link in here. If you choose to click, subscribe, or buy through any of them, THANK YOU – it will make it possible for me to continue to do this.
