Technology

➚ MOVE FAST: Indonesia. AI data center growth in Southeast Asia is fueling competition among investors, boosting valuations. The sale of a minority stake in Indonesia’s NeutraDC could value it at around $1 billion, more than 20 times its after-tax earnings.

➘ BREAK THINGS: China. Beijing responded to the latest US chip curbs with an export ban of critical minerals goods used to make semiconductors. It shows the tech Cold War isn’t slowing down in the remaining weeks of the Biden administration and could get worse under Donald Trump.

Going once, going twice, sold.

A Swiss auction house has become the first to sell a piece of artwork authenticated solely by artificial intelligence, ARTnews reported. Russian artist Marianne von Werefkin’s unnamed watercolor sold for $17,000 last week, almost double its high estimate.

Germann Auction House partnered with Swiss AI company Art Recognition for the authentication. The AI looks at minute details of each piece, including the brushstrokes, proportions, and color variations. Art Recognition claims its algorithm can authenticate pieces that it has never seen before based on these features in other works by the same artist.

Estimates vary for how widespread counterfeit art is, though experts agree fakes represent a significant market share. In 2018, a French art museum discovered that 82 of its 140 paintings from landscape artist Étienne Terrus were fake. Within the last year, an Italian art theft group has rounded up 2,100 forgeries claiming to be from artists including Pablo Picasso, Andy Warhol, and Banksy — many of which were sold through auction houses.

Sanjit Biswas is the CEO and co-founder of Samsara, which helps businesses track and monitor assets like vehicle fleets and equipment. It’s using large language models to simplify the experience for customers and improve productivity. Biswas also co-founded the enterprise Wi-Fi firm Meraki, acquired by Cisco for $1.2 billion in 2012.

Reed Albergotti: You’ve always been good at figuring out the bottlenecks in new technology trends, like with Meraki and wireless connectivity. It seems like we’re at a similar moment now with AI where it’s very cool, but there’s a lot of infrastructure that needs to be built to make it really useful. What are the bottlenecks now with AI?

Sanjit Biswas: It does remind me a lot about Wi-Fi 20 years ago, when John [Bicket] and I first started working on it. You use it for the first time and you’re like, “This is amazing.” The first time you opened a laptop and you could just sit at your couch and surf the internet, that was a really big deal and it became obvious that everyone is going to want this. But how do you make it happen?

At the time, if you wanted to build a big Wi-Fi network, you needed a Ph.D. in computer science. We’re seeing the same thing with sensors and AI. We have all gone through our ChatGPT moment. We know this is going to be world changing. But then, if I want to improve my risk mitigation at my construction job site or for my trucking fleet or something like that, there’s a lot of “how do we make this happen?” The data needs to be really clean. It needs to be trained on data that knows the answer, you need on-the-road data. And then you need to provide the driver with some real time feedback: “Please put down your mobile phone,” or provide them with some kind of coaching and scoring. That’s another bottleneck we’ve tried to break. We have the data. We have the insight. How do you take the action?

You put all of this hardware in the world. Do you see opportunities for edge compute?

We actually do a lot with edge compute and AI today. We have millions of these cameras that are deployed by fleets. They run AI inference at the edge in real time. Instead of having to wait hours for footage to go to the cloud and get analyzed, it tells you within a second, “Please put down your mobile phone,” or “You’re tailgating,” and it gives you real time coaching and feedback that’s all done with inference sitting at the edge. We use Qualcomm chips to do that. They’re very powerful compared to computers 10 years ago or 20 years ago. These are like little super computers.

Are you able to run some of these newer transformer architectures?

We started with the kind of original convolutional neural network models that were state-of-the-art a couple of years ago. We’ve moved to a transformer-based architecture that helps us do a ton of things. We can detect all different kinds of risks. We can detect very complex cases. And then it also lets us do more and more sophisticated detection. A recent one we rolled out was drowsiness detection. It turns out drowsiness can’t be detected with an image-based model. What you really want is a historic record, and a transformer model that is trained on what happened five, ten, 30 seconds before accidents. There’s amazing insights around how people move as they’re getting tired. We trained a whole model based on that and rolled it out to those cameras.

Here’s what legendary game developer John Carmack had to say about the changes at Intel: “I’m concerned to see Pat Gelsinger ousted as Intel CEO. He wasn’t a firebrand visionary, and it wasn’t exactly going great, but he was deeply technical, and I don’t expect his replacement to equal him there. ‘Business harder’ isn’t going to return Intel to greatness, only technical achievement will.”

It’s a good point. There is no other solution. Splitting up Intel’s fab and design businesses doesn’t solve anything. It probably just kills the fab business. Everything is riding on the success of the new node and Intel’s switch to a high numerical aperture extreme ultraviolet (EUV) lithography system as the cutting-edge tool for chip manufacturing. But these chip bets are so long-term we won’t really know whether Intel made the right move for some time, and well after it got rid of its general at the start of such a major campaign.

The “who cares” in all of this is that Intel’s fate is wrapped up in the US cold war with China. And it’s a microcosm of a bigger reality that can’t be solved by McKinsey or through private equity. It’s going to take mega investments in science and research — and the biggest breakthroughs will win the war.

Start your day smarter with AI Breakfast. This thrice-weekly newsletter delivers key insights into the world of artificial intelligence, from startups and disruptors to blue-chip leaders. Covering the latest tools, products, and projects, AI Breakfast gives you analysis and insight into the latest in AI. It’s the perfect addition to your media diet — subscribe for free here.

The rough percentage of the S&P 500’s gain this year that came from Nvidia’s stock surge. Vanguard’s chief economist Joe Davis warned that Wall Street’s AI-heavy portfolio risks a “correction.”

Amazon and Anthropic announced yesterday that they are teaming up to build a massive compute cluster containing hundreds of thousands of powerful AI chips designed to train the next generation of models.

Amazon was caught off guard by the “ChatGPT moment” a couple of years ago, but this phase of the AI race may actually favor them over rivals.

After dominating the cloud computing market for nearly two decades, the company has built up the muscles to solve the hard logistical hurdles standing in the way of the next AI breakthroughs.

I recently interviewed Dave Brown, the head of the company’s chip efforts, at Amazon’s headquarters in Seattle. I asked Brown about the challenge of spreading model training out over multiple locations, and it was clear Amazon was pushing for a quantum leap there.

“There’s a whole new area of science that’s going to have to come up,” he said. “It’s just a physics problem. You can only fit so much land and power into a certain given area. Let’s say — and I’m just making up numbers — a model is going to require millions of accelerators in the fullness of time. You’re not going to get that in a few city blocks.”

BlackRock’s $12 billion acquisition of HPS Investment Partners shows that what was once set up as an existential war between public and private lending is anything but, Semafor’s Liz Hoffman wrote.

For most of its decade-long rise, private credit has been cast in opposition, and competition, to loans and bonds underwritten by banks. That’s a satisfying but outdated lens, and as more money flows into private credit, the lines will start to blur.