Jeffrey Epstein’s web of influence

Today's markets move fast. Get the insights you need in 10 minutes with the Barclays Brief, a new podcast from Barclays Investment Bank. Through sharp dialogue and scenario-based analysis, our leading experts analyze key market themes each week. So, whether you're managing a portfolio or leading a business, the Barclays Brief Podcast can help you make smarter decisions today. Stay sharp, stay briefed. Find Barclays Brief wherever you get your podcasts. Good morning from the Financial Times.

Today is Thursday, February 5th, and this is your FTE's briefing. Alphabet is still putting a lot of its weight behind artificial intelligence, and we'll talk about the challenges of unpacking millions of documents on Jeffrey Epstein. Plus, why OpenAI Senior Staff are leaving the company. I'm Mark Filipino and here's the news you need to start your day.

Google said it plans to spend at least$55 billion more on capital expenditure this year than Wall Street had expected. The focus is on, you guessed it, AI. That brings its CapEx to as much as$185 billion. Alphabet CEO Sundar Pachai said, quote, We're seeing our AI investments and infrastructure drive revenue and growth across the board.

The spending spree overshadowed a second straight quarter with revenue of more than$100 billion. That was driven by strong earnings from the company's advertising and cloud computing divisions. Shares were volatile in after hours trading. And speaking of U.S. tech stocks, they had another bad day yesterday. The tech heavy Nasdaq composite fell 1.5%, and the SP slipped a half a percent.

Disappointing results from the chipmaker AMD added fuel to the AI related fire. AMD seemed to do everything right. It reported strong first quarter sales guidance and higher than expected revenue in the fourth quarter of last year, but investor expectations were just too high in the run up to its results. The sell off spread to companies like Memory Manufacturer Sandisk, Software Group Oracle, and the chipmakers Broadcom and NVIDIA.

Same kind of wide-ranging sell-off happened on Tuesday after Anthropic released AI tools that can help automate legal work, which freaked everyone out.

The U.S. Department of Justice has released millions of documents in recent days related to the late convicted pedophile Jeffrey Epstein. They show his broad ties to the world's most powerful people. But the release has been rocky, and so is the task of trying to parse these files. Chris Cook has been overseeing a team of FT journalists pouring over the documents. He's our enterprise editor and head of the computational journalism team, and he joins us now. Hi, Chris.

Hiya. So tell us how this information has been released because I you know it seems like it's kind of a unique case and frankly kind of hard to follow. So this story is like a particular technical challenge for newsrooms. First of all, just because of the sheer size of the document drop, which is far larger than anything else we've really ever had to deal with at speed.

And the second thing is that the Department of Justice's IT systems have really struggled to deliver stuff. You've probably read that they've had trouble redacting documents. Some victims' names have appeared in things, some explicit photographs have been published too. And as they've been Spotting those things, they've been pulling down thousands of documents from their website, so they're no longer available, you know, fixing them up and then pulling them back.

And what that means is that basically at any given moment when you look at their website, you'll get a slightly different selection of documents that you can download. And it's meant that as we have been trying to gather these things so we can sift them, what is available has changed from hour to hour. Okay. So this sounds like an incredibly complicated thing.

For you to parse. How are you going about it? So the first thing we want to do is actually make it searchable by our reporters. So we've been putting into tools that make it searchable. We've also got this challenge that Because the DOJ has said that some of the things that we've downloaded should be redacted and are not appropriate, we've been having to set up processes to basically check. those documents for that so that we can destroy things that we frankly do not want to end up holding.

The central challenge is that humans on their own going through millions and millions of documents, it would take months to handle this scene for a newsroom of our size. But we have to do it, you know, for the dignity and respect of these victims and because we don't want our colleagues to come across anything that they shouldn't see. Chris, because this has become a bit of a mess. Aaron Powell How should people think about some of the names that have been released in the Epstein files? Because

From my understanding, just because you're named in the Epstein files doesn't necessarily mean that you had direct contact or did any of the wrongdoing that is associated with Jeffrey Epstein. Aaron Powell So one of the things that sort of emerges when you look across such a wide range of his correspondences. You see it's not just the photos of him and the Pope and the Clintons and the very famous people and the bank CEOs.

There's clearly a big spectrum of people who had these sort of long, deep relationships with them, like Jess Daley, a former CEO of Barclays. And there are people who have more ambiguous relationships with him. So Kathy Rummler, who is the general counsel at Goldman Sachs, talks to him a lot. He does a lot of favours for her. You know, he sends

Beauticians around her house and things like that. There are sort of nice little gifts he sends her all the time. Then there are people who have much more passing relationships with him. His sort of gift. is working out what favors he can do for you to help you and keep you in his network. Now both Staley and Rumler say they were unaware of Epstein's activities.

Chris, the Department of Justice recently held a press conference about the document release. A reporter asked Deputy Attorney General Todd Blanche about victims' frustration over the release process, and uh I want to play part of his response. I hope that the work that that the men and women within this department have done over the past two months. Um st hopefully is able to bring closure. Um I think that um what we what we told our reviewers is that that was the goal. You know, there was

There's this mantra out there that oh, you know, the Department of Justice is supposed to protect Donald J. Trump. And that's what we were telling. That's not true. That was never the case. Um, we are always concerned about the victims. Chris, what have you found? We are quite time bound. So this is particularly from around twenty ten. This is what this document really relates to. It's after the period where we think that their friendship was at its peak.

So to some extent it's understandable there's not a lot of Trump in there. There's a lot of him complaining about Trump and a lot of him spreading rumors about him and gossiping about him. And he's obviously like a character in the Epstein universe. But he's clearly a departed character from the Epstein universe in the documents.

The extent to which it's like a deliberate thing and the extent to which they've partially released information so that it won't touch the devil cut Trump stuff is unclear. But they've produced, if you like, quite a natural feeling body of work. That does manage to exclude the president and his family pretty successfully.

So this is still an incomplete release. We know there are hundreds of thousands of pages more out there. So the short answer is we don't know. And we don't know whether that stuff is ever going to be released. We've got quite a partial view of Epstein's life. So to give you an example of what I mean by that. We see a lot of emails he sends himself which appear to be him drafting up letters to send to people, and we never see the thing that's actually sent to the recipient.

And we think that basically we're missing a sort of chunk, if you like, of office correspondence. There's a lot of stuff that's not really about Jeffrey Epstein. So he invested in a lot of companies and we've got a lot of internal correspondence from some of his investments, which we would w otherwise not have. Just from what we've got now, there's a lot of information in there. There's probably going to be quite a long tail.

That's the FT's Chris Cook. Thanks so much, Chris. Thanks for having me.

OpenAI is prioritizing improving Chat GPT, and now some senior staff members are leaving the company. The changes mark an important shift for OpenAI as the group makes the jump from Research Lab to one of Silicon Valley's biggest companies. Here to tell us more is the FT's Christina Crittle who broke the story. Hi, Christina. Hi. So tell us about the people who left OpenAI and why they did it. So there have been a few exits over the past couple of months.

In January, Jerry Torik is one of the biggest figures. He's been there for seven years and led its efforts on reasoning of AI models, which you can see helped to process complex problems more easily. He left OpenAI saying that he wanted to explore types of research that are hard to do at OpenAI.

And basically I looked into this a bit further and Last month, Andrea Vallone, who led model policy research at OpenAI, she was basically tasked with trying to solve the mental health of users who are becoming attached to ChatGBT. She left to join Anthropic last month and basically my sources were telling me it's because she was given this impossible mission and really wasn't supported or given the resources to be able to try and fix that issue.

And then finally, um, there was a researcher called Tom Cunningham who was on the economic research team and he was basically suggesting that open AI was turning away kind of more impartial research to focus on research that really promoted the company and cast the company in a good light. Give some background here. What does this shift in focus look like across OpenAI and why is it happening?

What you're really seeing here is a flattening of the research org. It has been restructured recently to make it much flatter. And so access to computing power, resources, is very tightly controlled at the top.

They approve different projects and my sources have been telling me that they've been rejecting projects that aren't very significant or aren't very aligned with this core research direction to advance understanding of transformers and large language models, all of the technology that underpins the core product, ChatGBT. Um and what you're also seeing on the product side is OpenAI is hiring very rapidly, bolstering its teams and releasing new products a lot more frequently.

And what does open AI have to say about this shift in focus and these departures? OpenAI rejected this characterization that it's not possible to do long horizon or blue sky research at OpenAI. Mark Chen, the chief research officer, told me that it's still focused on long-term foundational research. And that continues to account for the majority of its compute and investment. But obviously the story that the more than ten people I was speaking to is a very different one.

What do these moves mean for investors, Christina? Are they necessarily in interested in the direction open AI is going in? Well, OpenAI is on course for an IPO, potentially even this year, and it is looking to create better returns for its investors. And so you can really see that coming out in the product. Chat GPT is testing ads at the moment. There are lots of e commerce features.

And one venture capitalist who used to work at OpenAI told me that basically obsessing over who has the best model is no longer an issue. And actually if open AI focuses on what it already has, which is a consumer product which has a huge audience, one of the most popular ones in the world on the market, if it's able to retain those users

and find a way to monetize them, then that will be a successful thing. However, there is a bigger question here for investors, which is as ChatGPT started as a research preview, it was one of these long horizon bets. Does this focus on just pursuing the playbook that has worked for open AI so far and not looking towards those longer term horizons gonna mean that open AI potentially misses out on the next big thing, the next big breakthrough?

Christina Crittle covers artificial intelligence for the FT from San Francisco. Thanks, Christina. Thank you. 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. I'm Rob Armstrong. And I'm Katie Martin, and together we make the Unhedged podcast. Ooh. Since you said in September last year that you were short gold, it has doubled in price.

So listeners, when we say this show is not investment advice, we mean it is not investment advice. There are we oughta put up A plaque in the FT offices to how wrong I have been about gold. Unhedged from the Financial Times, get it wherever it is you get podcasts.