Google and OpenAI are racing to rewire the internet

“Google will do the Googling for you.”

Out of everything said onstage at Google I/O this year, I’ve been thinking the most about that line from Search executive Liz Reid. It summarizes not only how Google is fundamentally changing Search but also how the company is increasingly on a collision course with OpenAI.

At this point, it’s no secret that OpenAI is working on a search engine. Shivakumar Venkataraman, a Google Search veteran VP, just joined OpenAI to lead search efforts under engineering chief Srinivas Narayanan. My understanding is that the project has been messy with a lot of turnover and that OpenAI is realizing how difficult it is to get through the messy behind-the-scenes details of building a search engine. Still, it’s easy to see why OpenAI feels the pressure to go in this direction; just look at how Google is changing Search:

After talking with Google employees at I/O, it’s clear the company has known for some time that the historical experience of surfing endless links feels antiquated and exhausting. What OpenAI and search-focused startups like Perplexity have shown is that there is a more intuitive, conversational way to navigate the web — a way that can get you to what you’re looking for more directly. That experience looks a lot more like ChatGPT than the Google Search interface of today.

Google and OpenAI are working toward the same goal here: collapsing the traditional notion of a search engine into a conversational AI interface. If this is successfully achieved, both believe it can be the interface that fully absorbs how people navigate the internet. In essence, it’s a race to see if Google Search becomes more like ChatGPT before ChatGPT can become more like Google Search.

It’s hard to say which company will have a harder time getting there. Google is aggressively rolling out AI search results now because it is seeing positive trends in user engagement. Not only do these AI results you’re going to start seeing everywhere lead to more engagement for Google, but they are also widening the kinds of queries that Google is seeing come in. The logic internally is that, in the long run, this overall growing of the pie chart will be good for Search and the sources of information it relies on. (Yes, the proliferation of these AI summaries will still cause short-term carnage to the click-through traffic of many websites, but that’s a deeper topic for another time.)

There’s also plenty of historical data showing that, if Google can’t quickly provide an answer to a question, people will look elsewhere. There hasn’t been much competition elsewhere for a long time. That is, of course, increasingly no longer the case.

There were clearly a couple of motivations behind OpenAI’s pre-I/O event this week: to upstage Google’s Astra demo the next day (that’s the thinking inside Google anyway), and to keep developer momentum going with a new, cheaper model. A few dozen outsiders (no journalists) were invited to OpenAI’s HQ to watch the presentation with Sam Altman and crew. Afterwards, there was an NDA’d talk for builders of custom GPTs, where I’m told the OpenAI team asked for feedback and shared a high-level look at the roadmap.

Watching the public livestream, I was struck by the amount of time spent showing the conversational, Her-like upgrades to ChatGPT’s voice modality. I came away with the (perhaps intended) impression that it would be a much better version of Siri, which is still about as useful as it was 10 years ago.

There have now been multiple reports that Apple is trying to find an AI partner for iOS 18 ahead of WWDC in a few weeks. I’m hearing through the grapevine that OpenAI feels like it has the upper hand over Google in winning the deal, though anything can happen until the paperwork is signed. I don’t know much about the specifics of what Apple wants product-wise and how the deal is structured, other than it’s separate from Google’s increasingly fraught search default deal in Safari. It also won’t be a global exclusive since Apple will need to use a local model provider in China due to the regulatory constraints there.

For OpenAI, a deal with Apple would be an obvious win on the distribution front and a blow to Google. Whatever happens next, I have a feeling that the rivalry between OpenAI and Google is only going to get more intense in the weeks ahead.

My notes on what else is happening right now:

Some interesting career moves I’ve noticed recently:

If you aren’t already subscribed to Command Line, don’t forget to sign up and get future issues delivered directly to your inbox.

As always, I appreciate your feedback and tips. Thanks for subscribing.