In partnership with

SPD-BEEHIIV:8d13b716-9d4e-4019-bd3b-7b1b50ed38d9:R44:59beba27fc00779460c0b98eU.S. Pro subscribers get first access, with a choice: approve each purchase or authorize agents in advance.
| Monday digest | Monday, November 16, 2026 |
| | Our toolsSuperpower ChatGPT/WFH.team/Snipman/Superpower Daily New Tab | Today's briefing What matters todayChatGPT agents now have a way to buy, not just find. Claude’s work on Gödel’s notebooks also brings a question of mathematical credit into focus. | Inside today's briefing | | 01 | ChatGPT’s new wallet leaves an important choice about who approves purchases. | | 02 | Federal officials are investigating a possible AI link to grounded flights. | | 03 | OpenAI’s security agents can investigate flaws, then check the proposed fixes. | | 04 | Claude’s work on Gödel’s notebooks raises a tricky question about credit. |
|  | Lead story / launch OpenAI launches ChatGPT Wallet so agents can submit purchases for usersOpenAI launched ChatGPT Wallet on November 13, letting its “dots” agents submit purchases on users’ behalf. The initial rollout is for U.S. Pro subscribers aged 18 or older, using U.S.-issued cards with a U.S. billing address. Developed with American Express, Mastercard and Visa, Wallet charges an existing card rather than holding a balance. The terms allow purchase-by-purchase approval or advance authorization, rather than requiring a fresh checkout click every time. Before submission, Wallet shows available order details, including the merchant, items, total or maximum price, known taxes and shipping, selected card, and recurring charges or deposits. If a material detail changes before submission, Wallet asks for renewed approval. OpenAI warns that AI can make mistakes and tells users to check the details. OpenAI says its agents use transaction-specific payment tokens instead of accessing full card details. Those substitute credentials can reach the merchant alongside order information. Saving a card still permits OpenAI and its providers to collect, verify and store necessary information; users are told to use the secure card form rather than paste card details into a conversation. Approval is not a purchase guarantee: the card issuer can decline payment, and the merchant decides whether to accept and fulfill the order. Merchants handle shipping, returns and warranties; billing errors and unauthorized charges go to the card issuer. OpenAI says it is not taking a transaction cut at launch, and availability can vary by account, device, merchant and checkout. Read full story ↗ |
|
|
The voice-to-text app that built its own speech model
Most dictation apps are a wrapper around someone else's speech model. Wispr Flow trained its own.
It's called Canto, and it was built for the way people actually talk to AI: thinking out loud, changing direction mid-sentence, half-whispering in an open office.
Wispr tested it on 10 hours of real dictations from more than 2,300 speakers. Canto had the lowest word error rate of every model tested, including models from Google, OpenAI, AssemblyAI and Deepgram. That means your prompts to ChatGPT, Claude and Cursor come through the way you said them.
Talk naturally in any app and Wispr Flow types clean, formatted text, 4x faster than typing.
 | government action Federal officials investigate whether AI contributed to nationwide flight haltFederal agencies are investigating whether the FAA’s SMART traffic-management tool contributed to the November 13 communications failure that halted departures nationwide. Officials have not established a link, and contractor Air Space Intelligence denies that its products or FAA integration caused the outage. The tool uses real-time weather, congestion and flight-path data to adjust routes. FAA specialists are expected to retain oversight during the rollout. Continue reading ↗ |
|  | A tool for your workflowSuperpower Daily New Tab | | Open a new tab to AI news, daily games, and useful discoveries. Make it your own with backgrounds and components. | | See the new tab |
|  | launch OpenAI gives teams early access to AI agents that verify security flawsOpenAI’s Daybreak Red Team uses coordinated agents to examine web-app code and investigate whether suspected vulnerabilities are exploitable. Teams can bring findings from scanners, bug-bounty programs or disclosure reports, then test whether a patch blocks the attack. Assessments run in isolated environments with Guardian review and defined application and endpoint limits. No pricing or broader rollout timeline was provided. Continue reading ↗ |
|  | research Researcher says Claude surfaced 663 theorems and proof ideas in Gödel’s notesLogician Jan von Plato says Claude produced a 650-page account after receiving transcriptions of Gödel’s notebooks and related scholarship. Of the 663 theorems and proof ideas it surfaced, he says about 400 checked as stated, about 200 needed repairs and about 50 were false. One reconstructed proof reportedly passed computer checks, but its historical authorship—how much belongs to Gödel versus Claude—remains unsettled. Continue reading ↗ |
|
|
Your buyers search in ChatGPT. Are you there?
Your buyers still search. They just do it in ChatGPT, Claude and Perplexity now. ChatFeatured shows where you rank in AI answers against your competitors, then writes and publishes the content that gets you named.
| | Beyond the feed / AI events | | | This hands-on AI workshop will explore useful tools and prompting techniques to help businesses save time and boost productivity. Participants will leave with practical strategies they can apply to their work. | | See the event → Explore more events |
| Daily tool drop 6 AI tools worth knowing today | |
| | | The Internet Had a Point | Password Change Sheet Leaks the Passwords |  |
| |
| Your daily challenge Cross / Moderate / 3–5 min First clue: Part of a circle. Can you fill the grid? Take the challenge →Explore all games | | Need a new website? Try BayBold, a Bay Area web design agency. | | AI PULSE / YOUR PERSPECTIVE Should applicants disclose using AI on a take-home hiring assignment? Take part → | Reader check-in Help shape tomorrow's briefingOne click tells us what to keep, improve, or tighten. | |
| Prefer one email a week? Get the Monday briefing, including Friday and weekend news. |
|