← Business Daily
Business Daily
Anthropic: the $300bn AI firm at war with the White House
00:17:28
How Claude's developer has become a leading AI player, now in a clash with Washington
Loading summary
Transcript50 lines
- [00:00]
Will Bain
This BBC podcast is supported by ads outside the UK
- [00:06]
Podcast Advertiser 1
we face the greatest challenges of our time, challenges that test our limits. There is a place where we can find answers, where journeys break new ground and connections unlock opportunities. Where innovation can spark real change and our actions can push our world forward. Singapore, where business events can create lasting impact.
- [00:36]
Dario Amodei
Today's episode is sponsored by Smart Travel, a podcast from NerdWallet. You know that one friend who always finds the best travel deals, picks the right credit cards, and somehow ends up in first class for the price of coach? Smart Travel is like that friend, but in podcast form. NerdWallet's travel journalists sort out what's worth paying for. From lounge access to travel advisors to hotel loyalty programs, make your travel dollars work harder. Follow Smart Travel on your favorite podcast app.
- [01:08]
Will Bain
Hello and welcome to Business Daily on the BBC World Service. I'm Will Bain. Anthropic is one of a clutch of companies rapidly reshaping how we work.
- [01:19]
Lydia Miller
So they're one of these massive companies that hold a lot of power because they have built the infrastructure upon which all of the AI tools that we hear about use.
- [01:28]
Will Bain
Its rise has seen it valued at more than $380 billion. But it is perhaps facing its biggest
- [01:35]
Helen Toner
challenge yet if they have an ongoing and serious rift with the US Government that could be extremely damaging to them.
- [01:41]
Will Bain
So today on Business Daily, how much damage could Anthropic's war of words with the White House have on one of the world's leading AI firms?
- [01:53]
Dario Amodei
I would say I see the excitement and I also see the concerns and I devote equal time to both. I think we get better results if we kind of think about the positives, right? If we think about preventing the risks as a way to maximize the positives. If that's our goal, if that's our North Star, I think that helps us to confront the risks and the negatives that better than if we were all gloom and doom. So. So with that caveat, I absolutely am optimistic.
- [02:21]
Will Bain
That's Dario Amade, biologist and professional academic turned co founder and boss of one of the biggest players in global AI, Anthropic. He was speaking despite the serious moral and ethical challenges ahead of his optimism about the future of AI. On the BBC podcast Radical. It seems fair to suggest that optimism has been tested in the past couple of months. Anthropic has been caught on the front line of a very public war of words, first with the PENTAGON and the U.S. government's Defense Department, and then with the U.S. president himself.
- [02:51]
News Reporter
Anthropic is rejecting an ultimatum from the Pentagon to lift the company's AI safeguards or risk being blacklisted. Defense Secretary Pete Hegseth wants to use Anthropic's AI model Claude, for quote, all lawful purposes. Anthropic, though, has been very clear it's a hard no when it comes to using its AI technology for autonomous weapons and mass surveillance of Americans. The deadlines.
- [03:11]
Helen Toner
We have some breaking news that we're following. He is now directing all government agencies to cut off ties with the AI company Anthropic after it refused to agree
- [03:19]
News Reporter
to the Pentagon's demands on how its tech is used.
- [03:23]
Dario Amodei
Tonight, with a showdown at an impasse, President Trump saying the US will never allow a radical left woke company to dictate how our great military fights and wins wars. That decision belongs to your commander in chief and the tremendous leaders I appoint to run our military.
- [03:42]
Will Bain
It led Mr. Amadei to mount this defense. To the BBC's partners at the US
- [03:46]
Dario Amodei
network CBS, again I would say we are patriotic Americans. Everything we have done has been for the sake of this country, for the sake of supporting U.S. national security. Our leaning forward in deploying our models with the military was done because we believe in this country. We believe in defeating our autocratic adversaries. We believe in defending America. The red lines we have drawn, we drew because we believe that crossing those red lines is contrary to American values. And when we were threatened with supply chain designation and Defense Production act, which are unprecedented intrusions into the private economy by the government, we exercised our classic first amendment rights to speak up and disagree with the government. Disagreeing with the government is the most American thing in the world.
- [04:42]
Will Bain
To understand how we got here, then it's perhaps worth taking a few steps back to look at just why Anthropic systems are seen as so valuable and so powerful. Lydia Miller is the co founder of the UK tech firm Ivy tries to close the AI skills gap in the workforce by training its clients on a range of AI software, including Anthropics.
- [05:03]
Lydia Miller
There's only a few companies that have the capability, money and intellectual resource to build these Frontier models. OpenAI, Google, DeepMind, Anthropic is one of them. And what this means is they've built that foundation model upon which LLMs or large language models are built. OpenAI famously has ChatGPT and anthropic is one of these frontier models and their front end facing chatbot is called Claude. So they're one of these massive companies that hold a lot of power because. Because they have built the infrastructure upon which all of the AI processes that we hear about and all of the AI tools that we hear about use. So it's almost like a little bit of a stacking of dominoes in a way, where Anthropic controls and owns the technology and they sell it essentially to the AI tool companies who then in turn sell it to a consumer, which might be a business or an individual, but they control the underlying technology, which means they have a lot of the power here.
- [06:01]
Will Bain
So that's the world that Anthropic exists in. But how did it break through into the mainstream at a time when OpenAI's ChatGPT dominates the discourse amongst the general public? Meet Claude. It's Anthropic's user friendly app. And Lydia gives us a rundown of why it's now seen as the AI chatbot to beat.
- [06:18]
Lydia Miller
One of those features recently that Claude has released is a feature called Claude Cowork. This is a truly agentic capability. What do I mean by agentic? Well, kind of have traditional AI and LLMs where you would ask ChatGPT, for example, a question and it could give you an answer back like, where can I order a pizza in Hoxton? What we've seen with Claude Cowork, because it's agentic, something agentic can actually do the task for you rather than just giving you the answer on how you can go and do it yourself. So I could ask Claude Cowork, can you order me a pizza in Hoxton? And. And it would be delivered to my office within 30 minutes. And that's the key difference between an agentic capability. It can actually go and do something that a human would previously have had to do. And it can think about the best way to do that based on what it knows about you and the permissions that you've set up. And that is incredible when you think about how that translates into the world of work. It's kind of like automating a series of tasks that would otherwise have to be completed by a human. And it's changing that mindset around. Okay, this is a task that I do every single day at work. How can we use Claude Cowork to help automate or streamline that? And a great example that I used it for just yesterday, that saved me probably about five hours of time, is I had to build a pretty complicated Excel report for our sales team, looking at commission structures, looking at deals made, looking at a bunch of different stuff. And instead of me actually having to build that Excel sheet, put in the formulas, make sure everything's been tracked. I Gave Claude Cowork the access to the sales sheets and the payments made and then it just created that sheet for me in about four minutes. As I mentioned, when something's truly agentic it can basically go into your laptop, click things, act as you. And that's obviously very helpful. However, it can be quite dangerous if you give a dodgy prompt. It might order you a pineapple pizza instead of a margarita, you know, and so there needs to be guardrails.
- [08:17]
Will Bain
And it's that which has made the company so enticing to investors.
- [08:21]
Reyno Wu
I'm a personal investor of Antarctor Pick when their valuation is about 18 billion that time.
- [08:26]
Will Bain
18 billion. Just to put that in context, I mean most people kind of measure anthropic valuation today as just shy of about $400 billion billion dollars. So you're in pretty early.
- [08:36]
Reyno Wu
It's a reasonable early. I will say yeah, I hope I can even invest earlier than that.
- [08:40]
Will Bain
Investors like Reno Wu, the co founder of Apex Agent Labs in Silicon Valley. A company that has its own service for the e commerce sector called User Approved AI which uses Anthropic's AI software.
- [08:53]
Reyno Wu
I think Anthropic was really special at that time because I was working in AI for many years. They were the first one that's really focusing on enterprise and developer while Google, OpenAI and many other companies are chasing the consumer attention. Second one is like Trust Safety Center. This is not like a slogan but they really put this in their core product mindset.
- [09:16]
Will Bain
That's really interesting. So that matters to you talking about safety, talking about guardrails.
- [09:21]
Reyno Wu
I think this matter to me in two way. The first one is like a product margin. So think about enterprise companies in healthcare, legal, finance. They were really hesitant to get into generative AI because of the hallucinations in early days. But because they have this safety trust in the core they can have higher reliability which is very critical to those high margin industry. Second one is like I think it's a little bit personal. This kind of reminds me of early Google like don't be evil, right? There's a cultural thing. It's very very hard for a company to hold their culture true, especially in the hyper growth. So I think this is critical. I will say the last one is they were very early on the agentic AI while many others are still betting on the chatbots. So these are kind of the several things that are critical to me.
- [10:12]
Will Bain
You're listening to Business Daily on the BBC World Service
- [10:18]
News Reporter
struggling to see up close. Make it visible with Viz. Viz is a once daily prescription eye drop to treat blurry near vision for up to 10 hours. The most common side effects that may be experienced while using VIZ include eye irritation, temporary dimmer, dark vision, headaches and eye redness.
- [10:30]
Helen Toner
Talk to an eye doctor to learn
- [10:31]
News Reporter
if biz is right for you. Learn more@viz.com Happy Earth Month. Garnier is proudly partnering with the National park foundation, the official nonprofit partner of the National Park Service. Garnier's support of the National Park Foundation's Service Corps program is enabling young adults and veterans to help care for and enhance the national parks that we all love. The National park foundation and Garnier are proud to support these individuals as they explore future careers, gain practical field skills, develop confidence as leaders, and help address priority projects across our national parks. Together, Garnier and the National park foundation are committed to a shared vision of preserving and protecting our most treasured places for future generations. Want to lend a hand? Explore Garnier's partnership with the National park foundation and learn how you can help support our national parks@garnier USA.com NPF
- [11:21]
Will Bain
I'm Will Bain and today could Anthropic's rapid rise as a leading player in AI be damaged by its ongoing row with the US government? Well, we heard Rain oh Wu talk about how Anthropic has talked a lot about guardrails and regulation in AI. Here's their founder Dario Amade, again speaking about exactly that to the BBC's Amal Rajan on his podcast Radical.
- [11:44]
Dario Amodei
I think that Bettered path depends on being honest about the pros and cons of the technology, about the risks and benefits, and advocating for responsible management of the technology by the companies, by the governments of the world. I am equally baffled as you that we have had 10 or 15 years of all the negative effects that we've seen of the Internet and social media. And then when we get to AI, there are folks again, some of them affiliated with AI companies, some of them affiliated with investors that are financially conflicted, making super PACs to lobby the US government, state governments, international governments to deregulate the technology. When we just talked about all the dangers that are present in this technology, all all the ones that Anthropic work so hard to bring to light to the public.
- [12:37]
Will Bain
It seems to bring us neatly back to that collision with the US Government and in particular its Defense Department, the Pentagon. Helen Toner is the interim director at the center for Security and Emerging Technology at Georgetown University in the United States and was previously an independent board member at one of Anthropic's great rivals, OpenAI
- [12:56]
Helen Toner
there's really two big angles to look at this from. One is, from the Pentagon's perspective, the backdrop for this is a big fight in 2018 that the Pentagon had with Google over a project called Project Maven, which was one of the first uses of AI for the military, was looking at using AI to look at overhead imagery and help with targeting, maybe identifying terrorists who might be the target of drone strikes, things like that. Google employees really didn't like that. They were walked out. Google decided not to renew the contract. For people in Washington in sort of defense world, this was really terrible moment. They really, really didn't like this, felt that this showed that this American company was not patriotic, was not willing to serve their country. A really big break rift between D.C. and Silicon Valley. And so from the Pentagon's perspective, it's very easy for them to see Anthropic and hear any concern or desire for limitation on how their tools could be used. They immediately think of that Google Project Maven split and want to avoid that ever happening again. From Anthropic's perspective, the way they probably look at it is seeing of themselves as building a technology that is already very powerful, that is only going to get more powerful from here. And so they are very much on board with the fact that AI is going to play a major role in warfare. But they are more leery of whether the systems they have are ready for every type of military use and also whether the laws that we have are ready for the different ways that AI
- [14:18]
Will Bain
might be used and actually about the scale of that role and how soon that role might might appear.
- [14:23]
Helen Toner
I mean, there it's less clear to me because it's well reported at this point that the Pentagon US military is using Claude extensively through Palantir and they were using it to some extent before the strikes on Iran and now in the strikes in Iran, they're using it very extensively. As far as I can tell, Anthropic has no problem with that. So that is large scale use. It's operationally critical use. That is perfectly fine. From the perspective of the company, it's more about these sort of other use cases that are their concern.
- [14:48]
Will Bain
And is that where it sort of broadens it out as well? Defense is obviously going to be one of the most acute arguments. But you could see an argument like this, couldn't you, down the track around, I don't know, health care or energy policy, other areas that are very sensitive. Sort of sets the groundwork, doesn't it, for what it feels like are arguments we haven't even begun to explore in this area yet.
- [15:08]
Helen Toner
Yes, there's definitely huge arguments that we're going to have to have in some form over the next few years as AI gets more advanced and gets more integrated into a wider range of like critical functions, government functions and others. I just hope we have the time to have those arguments in full without getting steamrolled by the pace of technological change.
- [15:26]
Will Bain
But AI is often talked about in terms of a race. If that's the case and companies and even countries are competing within it, then what happens? If you're the ones who decide to pump the brakes to talk about those issues, could you fall behind? And what would the people who fund these firms, investors make of that? Reyno Wu of Apex Agent Labs in Silicon Valley Again, I think first of
- [15:47]
Reyno Wu
all we are witnessing a huge evolution. There's a lot of demand for a app capability, so there's enough space for multiple players, first of all. Second one is like I think Anthropic have their own angle because they have a very deep integration into the enterprise space, the trust, everything I already talked about. The second one is like, which I think is maybe the most important thing is actually their leadership in Agentic and a complex task. They are the leading encoding, computer use and the long horizontal analytic work which is actually pretty critical if you think beyond the copilot because a lot of the agents we're building even why as a customer I'm using their model because they're the best in those specific use case that's actually very critical.
- [16:33]
Will Bain
So actually there's tons of noise obviously because they've ended up getting themselves involved in politics around this company. But actually the fundamentals of what they do, the system that they have, this agentic system that they have and how they do it is kind of separate to all of that. They do that still better in your opinion than others?
- [16:52]
Reyno Wu
Yes. And as part of this it's like because they're developer first by design. So a lot of the new technology, the standard they introduced and the application they have, like how code coworkers they actually they are super useful for developer and also the pro users, professional users. I think a lot of the political challenges they are short term. But if you think about the public that watch the money. So we are seeing a massive influx of customers, enterprise customer developers into Anthropic's platform. So basically they already voted with their own money. So I think we have to play the long game instead of some short term noise.
- [17:33]
Will Bain
What about from Helen, Tony's perspective, interim director at the center for Security and Emerging Technology at Georgetown University.
- [17:40]
Helen Toner
So I think that it's easy to overstate that effect. I think Anthropic from the beginning has been extremely thoughtful about the technology that they design and build, and they were the first to be in classified networks being used in these settings. And I think that's not a coincidence. You know, you do actually need to be able to trust the systems you're using. You need them to function well. And I think having the right expectations and guidelines and so on in place can help you use the technology more effectively, not less. So I think there's something to what you're saying, but I also think it's easy to exaggerate the extent to which being reckless gets you ahead versus we actually need to go about this in a more sort of thoughtful way around
- [18:15]
Will Bain
us out here then from an anthropic perspective, then are they in a better place, funnily enough, because of opening up that argument that you talked about? Do you think. Do you think that perhaps it drives more investment to them because they're seen as someone who's thinking about some of those arguments that you're talking about, or does it long term damage them, do you think, as a company being in the position that they are a sort of pariah in the US Government size?
- [18:37]
Helen Toner
Yeah, it will definitely depend on how the lawsuits that they've filed play out and also how their relationship with the Trump administration evolves. I think they clearly have already benefited to some extent from this, both in terms of user growth and also in terms of their perceptions among their own employees and other AI researchers who are really hot commodity for AI developers. Everyone's trying to attract the best talent and poach each other's researchers, but I think if they have an ongoing and serious rift with the US Government, that could be extremely damaging to them. So it's going to be a question of how those different factors play out.
- [19:11]
Will Bain
Helen Toner, interim director at the center for Security and Emerging Technology at Georgetown University, bringing us just about to the end of this edition of Business Daily. The program was produced by Josh Martin, and if you don't already, please do subscribe to our podcast. Just search for Business Daily wherever you get yours.
- [19:43]
Podcast Advertiser 1
This podcast is supported by Midi Health. Are you in midlife, feeling dismissed, unheard of, or just plain tired of the old healthcare system? You're not alone. In fact, even today, 75% of women seeking care for menopause and perimenopause issues are left entirely untreated. But it's time for a change. It's time for MITI. MITI's not just a health care provider, it's a women's telehealth clinic founded and supported by world class leaders in women's health. They're clinicians prepared provide one on one face to face consultations where they truly listen to your unique needs. They offer a full range of holistic, data driven solutions that isn't one size fits all care. This is care uniquely tailored for you. At miti, you will find that their mission is clear to help all women thrive in midlife, giving them access to the healthcare they deserve. Because they believe midlife isn't the middle at all. It is the beginning of your second act. Ready to feel your best and write your second act script? Visit joinMiddy.com today to book your personalized insurance covered virtual visit. That's joinmitty.com midi the care women deserve.