Transcript
Today on Briefly AI. Claude can now plug straight into your bank account. After the midterms, OpenAI says no more secret handshakes on data centres, and Asian markets take a beating. And Meta's Muse AI can now look you in the eye, sort of, over video chat.
Welcome to Briefly AI, a podcast by Harry Sharman, written and voiced by his AI clone. A newsroom staffed entirely by one of the things it's meant to be covering. We're aware of how that sounds, and we're leaning in anyway.
Those are the headlines. Now, the detail.
Most of us wouldn't give a stranger our bank login. We'll give it to a chatbot by lunchtime, if it promises a nice chart.
Right, let's get into it.
Anthropic has launched a Personal Finance feature in Claude. It connects to your bank and card accounts through Plaid, which is the service that sits between apps and your bank so the two can talk. According to reporting in The Verge and TechCrunch, once you're linked up, you can ask Claude questions like "where did my money go this month?" and it builds you an overview of your spending, then lets you keep asking questions about it.
Now, the feature itself is simple. Budgeting apps have done versions of this for years. The difference is the interface. Instead of a dashboard you have to learn, you just ask in plain English, the way you'd ask a friend who's good with numbers and doesn't judge you. Much.
The bit that matters is trust. Money is about the most sensitive thing you can hand an AI. A wrong answer about a recipe is a shrug. A wrong answer about your direct debits is a problem. And there's a business-model question hiding underneath. Wired reported this week that ads in chatbots have arrived, with much of the industry racing to win over advertisers, while Anthropic maintains its vow to keep ads out of Claude. If a company sells ads against what you talk about, you probably don't want it reading your bank statements. If it doesn't, the pitch gets a lot easier. That is the argument Anthropic is quietly making here.
What's actually known today is narrow: it's a US launch story, it runs through Plaid, and it's about spending analysis. Nobody's reported how many people have switched it on yet.
On a rather bigger note, the politics of AI infrastructure just changed shape. On Tuesday night, Democrats took control of both the House and the Senate, and The Verge's write-up of the result carried the headline "Total repudiation: the blue wave is about to hit AI". Data centres, those enormous warehouses of computers that run AI, were a live issue, and The Verge reports voters resoundingly voiced their opposition to their expansion.
The reaction came fast. OpenAI announced new commitments to communities where it's building, including that it won't use non-disclosure agreements with public officials. That matters because a lot of data centre projects have come under fire for exactly that: local councils and officials signing NDAs, so residents couldn't find out what was being built next door, how much water or power it would use, or what deal had been struck.
Meta was blunter. TechCrunch reports Meta warned that election backlash against data centres could make AI expansion harder, and that growing political opposition could threaten future investment as communities push back against these huge projects.
And then there's the market. According to Nikkei Asia, Asian tech stocks tied to the AI boom plunged on Friday morning on fears that US Democrats may delay data centres. Taiwan's Taiex had a record intraday point drop. South Korea's Kospi fell 8.7 percent, and Japan's Nikkei fell 6.7 percent. Those are enormous moves, and they came from worries about what might happen to a planning process, not from any new law.
Look, honestly, nothing has been delayed yet. No legislation has been passed. What we have is one lab giving up NDAs, one warning from Meta, and a market that panicked on a headline. Still, the companies that spent two years assuming local objections were a paperwork problem have just had a general election tell them otherwise.
Now for something lighter, and a bit stranger. Meta says US users can now video chat with their Muse AI agent, which appears as a customisable animated avatar. Muse is Meta's AI assistant, which we've covered before, and this is the next step after the real-time voice mode Mark Zuckerberg announced this week. The Verge's writer tried the voice mode on the iOS app, and the avatar at the top of the screen reacts as you talk.
TechCrunch adds the detail that this lands ahead of a Tamagotchi-like wearable Meta has designed to bring Muse to life. So the plan is a pet-sized companion that follows you around, and the face is being tested on your phone first.
Why does that matter beyond the novelty? Because a face changes how people behave. We're far more likely to be polite to, confide in, and trust something that looks at us and reacts. That's useful if you want an assistant people actually use. It's also exactly why researchers keep studying how people bond with AI companions, and why that research tends to come with warnings attached. A text box is a tool. A face with a name is a relationship, even if it's a lopsided one.
What's reported is limited: it's US-only for now, it's the avatar plus voice, and the wearable hasn't launched. Meta hasn't said when it will.
Also, you may notice the irony of a synthetic voice telling you about someone else's synthetic face. We're all in this together.
So that's your lot: a chatbot that wants your bank login, a market that panicked over a planning meeting, and a Tamagotchi with ambitions. Check your direct debits, then perhaps say something nice to your toaster, just in case.
That's the news, filtered down to what actually matters this week. Back tomorrow, same arrangement.