Technology

➚ MOVE FAST: Boring. While Alphabet and other tech stocks in the Magnificent 7 took a hit this week, shares of not-so-household names like SAP, Salesforce, and Workday were up. There’s still a lot of revenue potential in AI, especially in enterprise customers, which is less showy but probably more dependable.

➘ BREAK THINGS: Blocking. Reddit has gotten more strict about protecting its data from AI companies who don’t pay to license its content. It has updated its website to prevent bots from accessing its web pages. Microsoft said its Bing search engine now cannot show results from Reddit, though there are ways to get around these types of website blocks.

OpenAI teased its long-awaited search engine on Thursday, and The Atlantic immediately found an error in the search results … in OpenAI’s own promotional demo posted on its website. Brutal. This has now happened over and over again to generative AI companies.

But this kind of criticism is ultimately meaningless. I’ve said this many times, but it’s really a PSA at this point: Don’t trust generative AI products to do anything important. Yet. What generative AI is, at this point, is a proof of concept. Think of it like the first time universities in California were networked together in the 1960s and could send messages back and forth. It was kind of cool. It eventually became the internet.

We won’t know if generative AI will spawn into something truly world-changing until models get much, much bigger, and companies get much, much better at corralling them to eke out the best performance. (My personal view is that these models will get there and they will be world-changing, but plenty of smart people would disagree with me.)

The other notable thing about SearchGPT is that OpenAI made nice with the media. The announcement came with gushing praise from big industry executives. My old boss, News Corp. CEO Robert Thomson,m said in the announcement that “For the heavens to be in equilibrium, the relationship between technology and content must be symbiotic and provenance must be protected.”

It shows how companies like The New York Times, which is currently suing OpenAI for alleged copyright infringement, are really outliers.

But it appears that OpenAI will not be sharing revenue with publishers in exchange for their blessing to be scraped by SearchGPT. If that’s the case, it might be a missed opportunity. Publishers will never have more leverage than now over companies like OpenAI.

San Francisco arrives on the “late freight” to surveillance tech

THE NEWS

The technology hub of San Francisco, one of the last holdouts on the use of policing tools from drones to facial recognition, is beginning to get comfortable with the controversial methods.

The city’s Board of Supervisors voted to ban facial recognition technology in 2019. But residents have been angered by what they see as a post-pandemic increase in crime, homelessness, and drugs, spurring lawmakers and voters to slowly back away from the anti-surveillance viewpoint. That has allowed law enforcement to experiment with the technology as the city deals with a significant shortage of police officers.

San Francisco will soon complete a rollout of 400 license plate readers made by Flock Safety that have the capability of notifying police immediately when a stolen or suspicious car enters the city limits, allowing them to apprehend suspects before they commit additional crimes.

“We’re on the late freight in using this technology,” said District Attorney Brooke Jenkins in an interview with Semafor. “This is a real game changer for a department that has not had any use of helicopters, drones, or these kinds of cameras. They’ve been old school, just cops chasing robbers, and we can’t continue to move forward that way.”

The shifts come as the city’s place in the American imagination — as the home of the counterculture, the birthplace of the tech boom, and the poster child for downtown anarchy, all at once — reaches one of its periodic peaks.

“In my hometown of San Francisco, Democrat rule has turned the streets of our beautiful city into a cesspool of open encampments and open drug use,” investor David Sacks told Republicans gathered in Milwaukee last week.

And as Vice President Kamala Harris moves closer to becoming her party’s presidential nominee, her background as a former district attorney of the city will likely bring it even more into the conversation.

But San Francisco isn’t immune from the national backlash against progressive policing and prosecutors.

On Thursday, California Gov. Gavin Newsom, himself a former San Francisco mayor, ordered state and local agencies to begin removing homeless encampments.

Even as San Francisco gets comfortable with surveillance technology, the DA’s office and police are still deliberating on what is acceptable.

For instance, Jenkins said she has so far resisted the use of AI to write first drafts of police reports using bodycam footage. “I’m more of a proponent that we, as humans, need to be the ones sifting through what technology gathers for us,” she said. “I can cross examine a human witness or a human police officer to find out what sort of bias and other things may have played a role in an investigation. I can’t cross examine software.”

Jenkins has also banned the use of AI in the DA’s office to draft legal documents.

But the office is planning to use the technology to redact documents for facts that may indicate a suspects’ race, so that prosecutors can avoid using that identifier in a decision on whether to charge a suspect.

On this week’s Mixed Signals from Semafor Media, Ben Smith and Nayeema Raza debate whether Kamala Harris can bring brat to the ballot box and how memes and the “childless cat lady” stereotype are expanding the K-hive. Then, they call up Veep executive producer Frank Rich to get his notes on an election month that’s stranger than fiction and ask whether an idealist show like “The West Wing” could be made today. Finally, Max Tani joins for Blindspots on RFK’s phone-free farming program.

Catch up with the latest episode of Mixed Signals.

AI reaches milestone in global math contest

THE NEWS

Researchers at Google DeepMind have trained a new type of AI system capable of solving complex math problems from this year’s International Mathematical Olympiad, reaching a score equivalent to a silver medalist for the first time, the company said on Thursday.

The IMO is the world’s hardest math competition for high school students. Teams of six of the brightest young mathematicians around the world are given six problems, which they each have to try and crack over two days. They can earn a maximum of 7 points for each problem, with a perfect score totaling 42 points. The USA Math team won first place this year with five out of six teenagers bagging at least 29 points, the minimum required to win a gold medal.

Google DeepMind’s latest system, dubbed AlphaProof, wasn’t too far behind. Although it didn’t solve the six IMO problems in the competition’s time limit, it managed to achieve 28 points – the highest score that AI has reached so far. Companies like Google DeepMind believe that solving difficult abstract math problems is necessary for developing the logic and reasoning skills required for artificial general intelligence, where the technology is better than humans at most tasks.

“No AI system has ever achieved a high success rate in these types of problems,” Pushmeet Kohli, VP of Research, AI for Science at Google DeepMind, said in a briefing. “A lot of reasoning is required to prove results.”

Its latest system is an improvement over the company’s previous AlphaGeometry model, and has solved 83% of all IMO geometry problems over the past 25 years, compared to the previous rate of 53% – including the hardest geometry problem ever set in the competition so far. AlphaGeometry could only handle geometry, whereas AlphaProof can tackle other areas of math like number theory, algebra, and combinatorics.

We’ve written in the past about PsiQuantum, which aims to be the first company to build a massive, industrial-scale quantum computer. The company earlier announced a $620 million deal with the Australian and Queensland governments to build a computer there by 2027.

On Thursday, it touted a similar deal, worth $500 million, with Illinois and Chicago to build one in a former US Steel site on the South Side of the city. Gov. J.B. Pritzker showed up for the announcement, calling it the “anchor tenant” in what will become a “global hub for quantum computing.”

What makes PsiQuantum’s plans interesting is that the company is essentially saying it’s figured out how to make a working quantum computer at scale. It just needs to build it. We won’t know for a few years whether the approach will work, and there are other methods that could prove better.

But quantum computing could be so game-changing that it’s almost impossible to be hyperbolic about it. And it seems like it’s getting closer to reality every year.

Toast co-founder Aman Narang took over as the company’s CEO last year, two years into its debut as a public company. I spoke to him about what it’s like going from a scrappy startup to a public company, and what the AI-enabled restaurant of the future might look like.

As generative AI gets more powerful, is Toast eventually going to become the automation component of a restaurant? Will you be handling everything from AI chatbots to robotics?

I’ll share a bit on the evolution of the company and maybe that’ll give you some perspective on that question. The reason platforms like Toast took off is a lot of this technology, believe it or not, in 2012 and 2013 was actually in servers in the restaurant. So you can imagine restaurateurs. They’re not technologists. They’re not CIOs. It was like oil and water trying to manage all this infrastructure. Just moving to the cloud made it operationally simpler, made it more cost efficient.

As we grew, we added capabilities. A lot of this was customer driven. We added fintech capability like payments and lending. We added the capability to manage the guest experience. This is the website, the online ordering, the loyalty programs.

As we think about the next decade, that’s a big part of our ethos. We believe we have to innovate on behalf of restaurants, so we’re certainly tracking all the trends very closely. There are things that will happen in the next decade. There might be voice AI at the table that augments the server. There might be times when you just want to leave or have your water refilled.

How it plays out in the restaurant ecosystem, whether it’s robotics or AI, I think a lot of it is figuring out what people want.

What will be automated in the future and what will remain human?

Some of the work that is more manual [will be automated]. When you go out to eat, you don’t realize how much manual work happens in the background. Scheduling staff, paying staff, tracking inventory, there’s so many workflows. There’s a lot of opportunities to automate what is more manual.

If you’re a restaurant, it could take a lot of time to manage your staffing, scheduling and paying employees. That’s a problem that’s ripe for disruption.

On the kitchen side, I don’t think robotics could do the last mile, in terms of what makes a chef a chef. But in terms of some of the prep, I do see opportunity. Devices that can make eggs or chop food or do some basic things. That’s likely where you see innovation in the mid term.

In terms of the front-of-house experience, you see opportunity for technology to create automation in workflows but also create personalization. One of the things restaurants struggle with is yield optimization. You see that in hotels, flights. Restaurants might do happy hour, but you could do more data-driven strategies to drive more demand.

The technology exists to recognize that Reed is here, and Reed might have or a family member might have an allergy, or what is Reed going to like on the menu. Checking out could be a matter of just walking out.

It’s like what a fine dining restaurant might do or a nicer hotel might do. Bringing that to life at scale with data, I think, is something that’s very much possible.

It started with an off-handed political tweet from venture capitalist David Sacks. It then became a full-fledged feud on X between Sacks and startup founder Parker Conrad.

Sacks, a staunch Donald Trump supporter, accused the Democrats of staging a “coup” by making Kamala Harris the party’s presumptive presidential nominee when Conrad chimed in, calling Sacks an expert on the topic.

Conrad, you might remember, was ousted from his company, Zenefits, after regulators began investigating the company for malfeasance. Sacks, of PayPal Mafia fame, was chief operating officer at the time and stepped in to take the helm.

Conrad faded from the public eye for a bit, then started Rippling, a competitor that has become wildly successful. But he has always felt like the fall guy in the scandal.

By Thursday afternoon, more investors and founders had chimed in on the debate, taking one side or the other, with national politics raising the temperature of the conversation.

The exchange made for great drama on X and offered a rare glimpse into Silicon Valley infighting that largely remains private.

• Africa’s food crisis needs innovative funding.
• China’s relationship with the US is in for a reckoning.
• Young voters are moving toward Harris and feeling less ‘meh’ about voting.