In partnership with
News of the day
1. Anthropic launches Personal Finance for Claude, letting US Pro, Max and Ultra users connect bank accounts via Plaid for read-only spending analysis. → Read more
2. Meta's president says midterm backlash against data centers will make building AI infrastructure harder, as local opposition grows. → Read more
3. Airbnb will launch its AI travel agent in H1 2027, using a custom travel interface rather than a chatbot, with Q3 revenue up 19%. → Read more
4. Deloitte finds most firms use AI for speed but few have transformed operations, limiting real returns on AI spending. → Read more
Our take
Hi Dotikers!
Claude wants to see your bank account. This week, Anthropic rolled out Personal Finance, a beta feature that lets US users on Pro, Max, and Ultra plans connect their bank and credit card accounts via Plaid and ask Claude questions like "Where did my money go this month?" The chatbot breaks down spending by category, flags duplicate charges and creeping bills, and explains financial jargon in plain language. Access is strictly read-only, Claude never sees your banking password, and conversations in the finance hub are not used to train models.
Let's be honest: this is Anthropic playing catch-up, since ChatGPT shipped similar account-linking tools back in May. But the execution here is what makes it interesting. No ads, no sponsored recommendations, no house-brand credit cards being pushed at you. That alone puts Claude ahead of most budgeting apps, whose business model often boils down to selling you a loan while pretending to help you save. Turns out the most radical feature in fintech is simply not trying to upsell you.
The bigger picture matters too. After Claude for Legal and Claude for Financial Services, Anthropic is methodically turning its chatbot into a vertical-by-vertical assistant, and personal finance is one of the most common things people already ask AI about. Giving the model real numbers instead of hypotheticals changes the quality of the answer entirely.
The real question is trust. Handing two years of transaction history to an AI company will feel uncomfortable for many, and rightly so. The privacy guarantees are solid on paper, but this is a market where paper promises have a short shelf life. Still, if an AI assistant is going to be useful for anything beyond drafting emails, your money is exactly where that usefulness gets tested. Watch this space, and maybe your subscriptions too.
Alex.
Frontier models perform better in Attio’s agent harness.
We gave GPT-6 Astra the same job twice: analyze 100 sales calls and draft the follow-ups, once in Ask Attio and once in ChatGPT.
Ask Attio finished in 4 minutes 49 seconds. ChatGPT took 26 minutes 11 seconds, about 5.4 times as long.
The most ambitious GTM work requires more than frontier intelligence. It requires architecture that can put that intelligence to work at scale.
Meme of the day
