Eric Schmidt says the quiet part out loud

It’s increasingly rare to hear what the power players of Silicon Valley really think these days. Over the last decade, armies of PR professionals and lawyers have cocooned tech executives with talking points they’re trained to never deviate from.

Sometimes, though, a slipup still happens and the quiet part is said out loud. Such was the case during ex-Google CEO Eric Schmidt’s recent talk at Stanford University, which has mostly garnered headlines for his now-walked-back comments about Google’s slowness to compete with OpenAI. The more eye-popping thing he said during the talk, however, was his advice for AI startups: that it’s okay to steal content if you’re successful because you can just hire lawyers to “clean up the mess.” As he told the room full of students: “If nobody uses your product, it doesn’t matter that you stole all the content.”

Having run Google from 2001 to 2011, Schmidt knows a thing or two about having lawyers clean up messes. YouTube grew in its earliest days off the backs of videos it didn’t have the rights to. One could argue that the business of Google Search itself was initially built by speedrunning the legal system.

Google isn’t alone here, however. The same strategy built Silicon Valley’s most influential companies of today. With few exceptions, they decided early on that it was better to ask for forgiveness than permission. By the time the fines needed to be paid, the money had been made.

History repeats itself, and the hottest AI companies are following this same playbook with support from old-guard leaders like Schmidt, who called Sam Altman a “close friend” in the same Stanford talk.

Talk privately to many of the leaders in AI today, and they’ll tell you a similar version of what Schmidt said. They see the money and power on the horizon as being more than enough to paper over the corners that are being cut to get there. It worked for Google, so why shouldn’t it work for them?

Schmidt has naturally gotten a lot of heat for his comment that Google was caught off guard by ChatGPT because employees decided that “working from home was more important than winning.” He had Stanford take the talk offline after it starting getting attention and has said he regrets what he said. Still, there is perhaps no better rage bait than a billionaire executive blaming work-from-home culture for a massive company’s executional failings.

The reality of why Google didn’t see this wave of generative AI coming is, of course, more nuanced. Below I’ve collected some pushback to Schmidt’s view from current and former Googlers:

“In my experience, Google is stacked with people willing to work their guts out, but Google’s decentralized, federated model creates high inefficiencies, and most Googlers aren’t given the opportunity to work on the sorts of problems that would move the needle.” – a current Googler

“Google’s main problem is NOTORIOUSLY our decentralized org structure and try-many-things culture, which have pros and cons, but are ultimately a management choice (going right back to Schmidt’s time).” – a current Googler

“Weird, I remember putting in 16 hour days *from home* at the beginning of the pandemic to work on keeping Meet up and scaling. But it’s cool to see execs without an original thought in their tiny brains call me lazy.” – a current Googler

“I don’t think that is as big of a factor as the culture and process that was put into place 2016-2022ish timeframe that made everyone complacent and not take risks.” – a current Googler

“I don’t know but it is clear given the evidence that it was nothing to do with WFH.

A guess of what it was? It could be as simple as the fear to unleash another Google Buzz or [Tay chatbot].” – an ex-Googler

“🤡” – a current Googler working on Gemini

My last issue was about Google’s acquihire of the leaders of Character.AI. Plenty of other reporting on the unusual deal has come out since then, including that Google paid about $3 billion to hire roughly 30 people from the startup, including CEO Noam Shazeer, who I’m told now reports to chief scientist Jeff Dean — not DeepMind CEO Demis Hassabis.

Though Google framed this payment as a “licensing fee” to the press and regulators, I’ve confirmed the company has no plan to use any of Character’s models, which apparently are not that impressive compared to what the Gemini team is building. This may be the most expensive per-head acquihire in the history of the tech industry. Given Shazeer’s double-digit ownership of Character’s equity that Google is essentially buying him out of, he stands to personally gain hundreds of millions of dollars for quitting the startup he left Google to build for less than three years.

Even with Shazeer’s legendary status in the AI world, I’m still trying to figure out why Google’s upper brass was compelled to pay this much for a small team and no IP. (If you know more, get in touch.) I’ve heard this has also been a puzzling question to the rank and file inside DeepMind, which Google paid a mere $500 million to acquire 10 years ago.

Some notable career moves:

If you aren’t already subscribed to Command Line, don’t forget to sign up and get future issues delivered directly to your inbox. You’ll also get full access to the archive featuring scoops about companies like Meta, Google, OpenAI, and more.

As always, I want to hear from you, especially if you have a tip or feedback. Respond here, and I’ll get back to you, or ping me securely on Signal. I’m happy to keep you anonymous.

Thanks for subscribing.