“You can get caught up in the moment of the melodrama of it all,” he told Fortune’s editor-at-large Michal Lev-Ram at the Fortune Brainstorm Tech dinner in San Francisco this week. “Zoom out, and we’re going to be in this period where we’re investing in infrastructure at a pace that no one’s seen before. The smartest minds in the world are all working on the same problem, and every company is trying to figure out their strategy at the same time.”

That cafeteria drama Taylor referenced included Microsoft shocking the tech world this week by hiring Inflection cofounder Mustafa Suleyman as chief of Microsoft AI (Inflection’s future is uncertain). Around the same time, news leaked that Apple, widely considered to be behind in the AI race, is considering partnering with frenemy Google by integrating its Gemini artificial intelligence engine into iPhones.

The chatter sounds a lot like adolescents gossiping about who’s dating who.  

Oh, by the way, Taylor was involved in one of the biggest tech dramas in the last year. He was recruited to be OpenAI’s chairman after a certain nasty breakup between the company’s board and CEO Sam Altman that threatened to implode the business. At the time, Taylor said his tenure at OpenAI would be temporary. On stage, he confirmed it, but said there’s no end date other than when OpenAI is in a “stable state.” 

In any case, Taylor says he spends most of his time these days focused on Sierra, which he cofounded with ex-Googler Clay Bavor, who was also at the Brainstorm Tech dinner. The two met nearly two decades ago while working as associate product managers at Google and reconnected last year after Taylor left his position as co-CEO of Salesforce. A year ago today, they created Sierra, whose customers now include WeightWatchers, SiriusXM, and audio equipment maker Sonos. 

Taylor sees conversational AI as the next step in a long evolution of technology benefiting corporate customers. Thanks to the dotcom bubble and widespread adoption of the internet, “every company has a website,” he said. “With the birth of social media, most companies had a profile page. Now that computers can actually understand what we’re saying, every company is going to need their own AI agent.” 

As with all startups and all AI, there are kinks to work out. At one point in building the technology, the cofounders created a test in which their bot messaged with another AI bot that simulated a customer. Taylor and Bavor left for lunch and returned to find the two bots profusely thanking the other over a string of 2,000 messages because they didn’t know how to say goodbye, the cofounders told Fortune. 

“It turns out you need to give AI a goal to stop,” Taylor said.

And so you know I’m not a bot, I’ll sign off with just one “thank you for reading.”

Also, if you’re interested in attending the Fortune Brainstorm Tech conference in Park City, Utah, July 15-17, you can find information here. 

Here’s the biggest news in tech today. 

Rachyl Jones

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

The rest of today’s Data Sheet was written by David Meyer.

NEWSWORTHY

UN AI resolution. The world has unanimously agreed that governments must safeguard human rights and protect personal data in the AI age. The entire United Nations General Assembly yesterday passed the resolution, which also says AI development must be monitored for risks. However, as Reuters reports, it’s a nonbinding resolution that lacks any real teeth.

SpaceX vs NLRB. The National Labor Relations Board, which SpaceX claims is unconstitutional, has filed new allegations against Elon Musk’s rocket firm. The watchdog claims SpaceX “interfered with, restrained and coerced its employees” from discussing “wages, hours, or conditions of employment,” and also discouraged workers from contacting the NLRB, Quartz reports.

Threads enters fediverse. Starting with users in three countries—the U.S., Canada, and Japan—Meta’s Threads app now allows interoperability with other fediverse social networks like Mastodon. That means a Mastodon user can read and share a Threads user’s post if they choose to turn on the feature. However, as TechCrunch explains, there are currently limits, such as the inability for Threads users to see replies and likes from other fediverse servers.

ON OUR FEED

“With each technological progress, there are hesitations and concerns. You just have to take the plunge and learn how to use it.”

—Celebrity photographer Annie Leibovitz tells AFP she’s not at all worried about generative AI’s effects on her medium, as “photography itself is not really real.”

IN CASE YOU MISSED IT

Reddit closes nearly 50% higher on 1st trading day in latest sign IPO market heating up, by María Soledad Davila Calero

Can AI hype fuel an IPO resurgence?, by Sage Lazzaro

The DOJ’s blockbuster lawsuit against Apple boils down to long-running argument within tech, by Kylie Robison and Rachyl Jones

Tesla challenger Fisker, flirting with bankruptcy, gets both a design award and another scathing review in the same week, by Steve Mollman

This little-known Japanese company is a key player in the leading-edge chip industry—and that’s setting it up for success amid the AI boom, by Lionel Lim

Facebook is bringing back the poke feature, by Chris Morris

Ex-French minister sparks controversy after calling for 3GB per week limit on data to make the internet less toxic, by Ryan Hogg

BEFORE YOU GO

RIP Vernor Vinge. The science fiction author Vernor Vinge passed away on Wednesday, at the age of 79. The concept of a “technological singularity” in which superhuman artificial intelligence would end the “human era”? That was his. Cyberspace? Well, William Gibson coined the term, but Vinge beat him to the concept—which he called the “Other Plane”—by five years. Read his obituary in The Register.