Real Alpha
Try Pro free

Decoder · 3 min · May 26, 2026

Google expects agent teams to spread beyond engineering

Sundar Pichai · CEO, Google

Sundar Pichai says some engineers are already directing teams of agents and expects the shift to reach the rest of Google.

In this episode

Full analysis
  1. 8:31Google unified its AI teams to make decisions faster4 min
  2. 26:07AI anxiety reflects social change, not just product quality7 min
  3. 33:40Publishers are adapting as search traffic and user behavior shift5 min
  4. 20:55Pichai says Search should be less opinionated on some queries5 min
  5. 41:15Pichai says AI still cannot make scientific discoveries alone5 min
  6. 17:04Google expects agent tasks to work across products4 min
  7. 12:53Google expects agent teams to spread beyond engineering3 min
Up next · Unsupervised LearningA world model predicts consequences, which LLMs cannot do3:12
“We use anti-gravity internally. That's what we are providing outside.”
Sundar Pichaiat 14:54 · Decoder
14:54

Takeaways

  1. Some Google engineers are directing teams of agents.
  2. Pichai expects the shift to extend beyond engineering.
  3. Google uses internally what it plans to offer users.

About this segment

Sundar Pichai says some engineers are already directing teams of agents and expects the shift to reach the rest of Google. He focuses on deploying those tools for users rather than reshaping the org chart.

Sundar Pichai

Sundar Pichai is CEO of Google and Alphabet.

Helped build Google Toolbar and Chrome, then led Google from 2015 and Alphabet from 2019.

  • Nilay Patel, Host · Editor-in-chief, The Verge

    Helped start The Verge in 2011 and has hosted Decoder since 2020.

From “How Sundar Pichai is rethinking Google for the AI era | Decoder”, May 26, 2026. Watch the full episode

Answers “Which jobs will AI replace, and which are safe?” See all 132 answers

Answers “Will AI replace software engineers?” See all 40 answers

Answers “Are AI agents ready for real work?” See all 104 answers

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

  1. 13:05

    course of action you need to take. Yeah. Looking around the industry, your peers in big tech have some of the wildest org chart ideas I've ever heard in my entire life. I think Meta wants to have 50 engineers report to a single manager with the power of agents. Jack Dorsey at Block wants all 6,000 people to report to him. Are you having similar thoughts that you should invent some of the craziest org charts with AI ever? Look, I think leaders and people are incredibly important. And I think our companies, it depends on what the company, some companies are, you know, have a much more narrower suite of products. And so, you know, different structure may work. When you're running something at the scale of, take the scale of Google Cloud, right? I think it's important that there is a CEO in charge and we are serving all the top enterprises in the

  2. 13:52

    world at a scale. So how do you set up for that? So I think great leaders end up mattering a lot, like we have Thomas there. So I do think about it. But what I do think about it is how are we using AI more effectively? And we have seen the transition internally, particularly amongst our developers, where we have transitioned from using AI tools to assist coding to them, a portion of the engineers, working, directing teams of agents effectively more and more.

  3. 14:27

    And so I think those are transitions underway. And I think that will flow beyond just engineering into the rest of the organization. It's already happening. Like even the work we are doing in Gemini Spark is to give that superpower to the hands of consumers what you can do with these agentic workflows, etc. So I'm more focused on making sure we are actually deploying that capability in a native way and that it's working well because for us, it's more than just me.

  4. 14:55

    the company efficient because it's the products we provide to others, right? So I look at it with a very different lens. How we do it internally is what we are giving to users outside. We use anti-gravity internally. That's what we are providing outside. So the agents in anti-gravity is what our developers are using. And so that's what we are trying to put outside. So it has that