WSJ What’s News - OpenAI Rival Anthropic Sets Sights on Enterprise Market

Episode Date: September 5, 2024

A.M. Edition for Sep. 5. Artificial-intelligence startup Anthropic launches an enterprise-grade product, aimed at helping it scale inside companies. The WSJ’s Isabelle Bousquette says it’s a bid to compete in an increasingly crowded–and lucrative–field. Plus, details emerge on the 14-year-old suspect in yesterday’s school shooting in Georgia. And, the U.S. accuses Russia of spending millions on a covert campaign to influence the presidential election. Luke Vargas hosts. Sign up for the WSJ’s free What’s News newsletter. Learn more about your ad choices. Visit megaphone.fm/adchoices

Ecolab Water for Climate. Less water, more growth. Results will vary. Learn more at Ecolab.com slash EWC. Ecolab Water for Climate. Transforming the way the world thinks about water. The U.S. accuses Russia of spending millions to influence American votes. Plus how North Korean hackers found a way into U.S. companies by getting hired by them and OpenAI rival Anthropic tries to break into the valuable enterprise market. Anthropic was founded by a group of OpenAI employees who decided they wanted to do something similar, but with a bigger focus on models that were safer, more transparent about how they're coming out with a certain answer. And Anthropic is saying, yeah, that is something that opens the door for us with companies.

It's Thursday, September 5th. I'm Luke Vargas for The Wall Street Journal, and here is the AM edition of What's News, the top headlines and business stories moving your world today. We begin in Georgia, where new details are emerging about the 14-year-old suspected of killing four people in a shooting at Appalachee High School outside Atlanta yesterday. Authorities have identified the suspect as Colt Gray, a student at the school, with the FBI's Atlanta Division saying that law enforcement officials had interviewed Gray last year following online threats to commit a school shooting.

However, according to officials, Gray denied making the threats at the time and there was no probable cause for arrest or further action. The Biden administration has accused Russian President Vladimir Putin and his inner circle of orchestrating a covert influence campaign in the hopes of helping Republican-nominee Donald Trump to reclaim the White House and eroding international support for Ukraine. American spy agencies say the effort included a scheme to bankroll an American media startup and to use Russian PR companies to promote state-sponsored narratives.

In a bid to counter the effort, the Justice Department has seized more than 30 Internet domains that it says were designed to look like leading U.S. news sites but were instead filled with Russian propaganda. The DOJ also announced charges against a pair of Russian nationals, while the Treasury Department levied sanctions on 10 people. In comments posted to Telegram, Russia's U.S. ambassador rejected the accusations against Moscow. Meanwhile, hundreds and potentially thousands of North Korean spies have infiltrated American

companies by getting hired as remote IT workers using the stolen identities of foreigners, according to U.S. officials and security researchers. Our Korea Bureau Chief Timothy Martin has more on a scheme that the Justice Department says has netted Kim Jong-un's regime hundreds of millions of dollars annually, allowing it to evade international sanctions and keep bankrolling its nuclear weapons and ballistic missile programs. They're using things like generative AI to spoof potential employers into thinking that they're hiring an American sitting there in the US not someone at a desk in Pyongyang. And what does the regime need? The regime

needs cash and that's because the North Korean won doesn't really buy you much outside of North Korea. And this type of cyber activity has really been a skeleton key for them to do this. And whether it's hacking a crypto exchange, whether it's posing as fake IT workers in the U.S., this gets them to U.S. or other foreign currency. North Korea's diplomatic mission at the United Nations didn't respond to a request for comment. And former Wyoming Congresswoman Liz Cheney has broken with the Republican Party by saying she plans to vote for Democratic nominee Kamala Harris, citing the quote, danger that Donald

Trump poses, end quote. Cheney was a fierce critic of Trump during his presidency and was one of ten Republicans who voted to impeach him for his actions related to the January 6th capital riots. Competition is heating up in the market for Enterprise Artificial Intelligence. The journal's Isabelle Busquets reports that Anthropic's Claude Enterprise plan is part of the company's efforts to go toe-to-toe with the likes of OpenAI, and Isabelle joins me now to discuss this new product and how realistic of a goal that is.

Isabelle, let's talk about Claude Enterprise here. Claude, we should say, is the name of Anthropic's LLM, or Large Language Model, just like OpenAI has chat GPT. To date, Claude has mostly been available to people for free, though there is also a $20 a month pro version, but now Claude is basically going to work, I guess, being pushed out to businesses at scale. Yeah, that's right. This is the first time that Anthropic is making this big sell directly to large enterprises.

We've seen them in the business space so far, but this is the first time that Anthropic is embedding all the enterprise-grade controls and security that you would need to scale Claude out. And that includes just sort of basic logistical things like single sign-on and domain capture, identity management, being able to log into Claude with your work email, along with some additional features that are aimed at helping the enterprise work. One of their big things is the ability to collaborate within Claude, which is new-ish for AI model engagement. Normally you would just sort of put in a prompt, get out of prompt, but

Claude is pushing more into this idea that you can share prompts and you can have a group conversation with Claude, multiple people asking the same model questions and sharing answers. Right, let's talk dollars and cents here. We don't know the actual pricing that Claude Enterprise is going to be offered at, but more broadly, what is the importance of the Enterprise segment? Why might Anthropic be heading into it now? The Enterprise is a hugely valuable market here,

and enterprises are ready, willing, and able to spend a huge amount of money on generative AI, mostly in this testing and exploratory phase they're in. A lot of the pricing models we've seen so far when it comes to generative AI are on a per-seat basis, so companies are basically paying per month per person for the amount of licenses they have, which can really, really add up. Pete Slauson Who doesn't love recurring revenue, right? That's kind of the pipe dream.

Anna Voelker Yeah, there's also sort of a usage-based consumption model that's sort of based around the idea that the bigger a prompt is, the more information you're calling on, the more expensive that's gonna be. And some companies are finding that they end up using really, really big capable models for a job that they probably could have done with a smaller, cheaper model. And as a result, we've seen that OpenAI is in Anthropics and Google's of the world coming out with smaller models

that are maybe less capable of doing these really big broad analytical tasks, but maybe really good at one specific thing and can do that thing a lot more cost efficiently. As you said, Anthropic sees a big opportunity here. Companies willing to spend a lot to get their employees hooked up with AI tools. But there's kind of a financial imperative here for Anthropic and other businesses who've created these large language models to get paying customers, right? Yeah, these models are hugely expensive to build and the

enterprise market is such a valuable market. It's really critical for them to get a foothold here. And they are up against an 800 pound gorilla in Open AI. Open AI says that 92% of companies in the Fortune 500 are using some iteration of their products. That's not to say those companies are only using OpenAI products or using them all at scale, but Anthropic is still competing against, you know, a company that had a big head start and has a lot of name brand value and a

lot of penetration in the enterprise market. Do they think they have an edge? Anthropic was founded by a group of OpenAI employees who sort of broke off and decided they wanted to do something similar, but with a bigger focus on models that were safer, more transparent about how they're coming out with a certain answer. And Anthropic is saying, yeah, that is something that opens the door for us with companies. You know, they'll take calls with CIOs and corporate tech leaders and the CIOs will say to them, I took this call because I really like what you're doing with

responsible AI. Anthropic feels like that maybe is enough to start the conversation, maybe not totally enough to make the final sale, but Anthropic is confident that they can still get an in here because it is super early in this adoption cycle. And they told me that companies haven't necessarily made final decisions about what vendors they're working with yet. And they're still exploring and new models are still coming out all the time. And every time there's a new model, it's a chance to sort of reevaluate, get in front of companies. So I think, yeah, we'll see if they're able to catch up.

Yeah, big stakes here, to say the least, for how competitive this enterprise AI landscape will be in the years to come. Isabelle Busquette, thank you so much. Thank you. Coming up, a quick spin through the day's top business stories and sports reporter Andrew Beaton breaks down what's got him fired up for the NFL season ahead of tonight's kickoff. That's after the break. This episode is brought to you by CIBC.

From closing that first sale to opening a second store, as a business owner, you've hustled to accomplish a lot. But the rewards don't stop there. When you earn two times more points on things that matter to you and your business, easily track those business expenses and experience flexible Aventura rewards, you'll realize how much more rewarding your hustle can be. Get up to $1,800 in value

when you apply for the CIBC Aventura Visa for Business at cibc.com slash Aventura Business. Terms and conditions apply. OpenAI co-founder Ilya Suscovars new startup, conditions apply. OpenAI co-founder Ilya Suskova's new startup, which aims to build what it calls Safe AI Models, has raised a billion dollars in a fundraising round, with Safe Superintelligence saying on X that its investors include Andreessen Horowitz and Sequoia Capital, among others. The firm was founded in June, not long after Suttskover left OpenAI following his involvement

in the short-lived ouster of CEO Sam Altman. We are exclusively reporting that Jeep parent company Stellantis has paused production of two of its top-selling U.S. models. According to people familiar with the matter, factories temporarily stopped making the Jeep Wrangler and Jeep Grand Cherokee SUVs in the past week, with Stellantis confirming that production had been adjusted at two Detroit plants plus another in Ohio. However, how many days the facilities have been idled couldn't be determined.

The carmaker has been grappling with high inventory levels and complaints from dealers as sales have sagged. And in other news, moving markets, shares of U.S. Steel fell 17 percent yesterday on reports that President Biden is planning to scuttle the company's planned sale to Japan's Nippon Steel. According to our reporting, Biden's final sign-off hasn't happened yet, and the two companies say they've yet to receive an update or executive order related to the national security review of the sale. And coming up today, weekly jobless claims data is due at 8.30am Eastern, with earnings from the likes of wholesaler Costco and chip and software maker Broadcom expected after

the market close. And finally, football is back. The NFL season kicks off tonight with the Baltimore Ravens facing off against the back-to-back Super Bowl-winning Kansas City Chiefs, whose pursuit of a first-ever Super Bowl 3-peat is one of the big storylines that Journal sports reporter Andrew Beaton told me that he's watching over the months to come. This is not a sport designed to have dynasties. There are injuries. There's a salary cap that prevents teams from keeping the best players. But the Chiefs are trying to defy that. We're also looking at a season when NFL players may be wearing guardian caps for the first time and for the uninitiated.

Those might be these funny looking gel foam things that you see on top of helmets, and it looks definitely strange for fans. But the idea is that players have the option to wear them because it helps reduce concussions, and they're really looking at anything that might make the game safer. Including Androvia, a notable rule change, right? Yeah, the very first play of this game will look extraordinarily different because the NFL has changed its kickoff. And so they've introduced a radical change where instead of the teams sprinting all the

way down the field and then bashing into each other, they now line up way closer to each other and there's a designed kickoff zone so that more kicks are actually returned. And I think everybody's just sort of curious to see how it works. And that's it for What's News for Thursday morning. Today's show was produced by Daniel Bach and Kate Bullivant with supervising producer Christina Rocca and I'm Luke Vargas for the Wall Street Journal. We will be back tonight with a new show. Until then, thanks for listening.