Good morning from the Financial Times. Today is Thursday, October 24th and this is your FT News Briefing. There’s more bad news for Boeing. And US lenders are trying to close the door on open banking. Plus, over in AI, a new virtual agent could run your life for you.
Boeing workers voted to reject a deal with the company yesterday. That means the nearly six-week strike will continue. The deal would have increased pay by 35 per cent over four years and improved retirement benefits. But workers wanted a 40 per cent raise and for the company to restore a defined benefit pension plan they previously lost. It’s a blow to Boeing’s new CEO and his plan to stabilize the business. The company reported a $6bn loss last quarter and said it would continue to bleed cash.
The US has announced new rules that will make sharing bank information a little bit simpler. They’re called open banking regulations and they’ll help customers connect their accounts to apps like Venmo or PayPal. But big banks are not a fan. Here to tell us more is the FT’s Josh Franklin. Hey, Josh!
Deutsche Bank delivered some mixed news in its earnings report yesterday. The German lender saw record profits last quarter. That’s thanks to revenues from its trading and investment advisory businesses. But it also warned investors to brace for higher-than-expected loan losses next quarter. Deutsche Bank blamed that on the commercial real estate crisis and some, quote, larger corporate defaults. Its chief financial officer did say those issues should disappear next year, though.
Keeping up with day-to-day tasks can be a bit of a time suck. But it’s a fact of life, right? Well, not according to AI start-up Anthropic. On Tuesday, it unveiled a new computer use feature for its model, Claude. It says this tool will enable Claude to basically act like a personal assistant and control your computer. The FT’s Cristina Criddle has been following this new era of virtual agents. She joins me now. Hi, Cristina.
So most tasks that you can do on your computer, Claude can do for you. And that includes like quite mundane tasks like filing expenses or booking travel. So you would put an instruction into Claude — much like you would if you were interacting with the chatbot interface — and ask it to do a task for you. But the types of tasks it can do now are controlling your mouse and keyboard. So it can browse the internet and do research. It can schedule calendar appointments for you. It can fill in forms and it can transfer data from one document to another. These kind of tedious things that take us quite a long time to do, but wouldn’t take your computer that long to do if you instructed it in the right way.
This is the first time I’ve seen a feature like this. It is quite unusual and I would say a big step. It feels at the moment like every tech company is announcing their version of AI agents, though. Microsoft announced some recent autonomous agents in its 365 product suite and then OpenAI recently launched developer access to some of its voice models.
So one example demoed by OpenAI was where you would ask the chatbot to find you locations that would sell strawberries near you. And then you could select the business and ask the chatbot to order you a certain amount of strawberries for a certain price. And it would call this business and put in the order for you using its voice. And so this is kind of the next frontier. Many of the AI companies are calling it for this technology, but we’ve yet to see widespread development and implementation into our lives yet.
It can make mistakes and even that was one of the things that was highlighted by Anthropic when they were demoing this feature, was that it might still click on a website in its search that isn’t that reliable. So it is in its early stages. The company warned that and they say it’s kind of like self-driving cars at the moment where consumers don’t really trust it to drive on its own. And it’s very much still got to have your hands on the wheel making sure that those kind of mistakes aren’t happening or if they do happen, you’re able to step in.
It is human-like. But we have interacted with technology to make these experiences easier for us already. I think there is a wider concern about if these models reach artificial general intelligence — which is intelligence at the same level as most humans — and that’s what a lot of these companies hope to achieve. If that happens, then maybe there’s cause for concern because the level of control that you have over those systems is still unclear and those safeguards haven’t been put in place yet.
You can read more on all these stories for free when you click the links in our show notes. This has been your daily FT News Briefing. Check back tomorrow for the latest business news.