Episode summary
Aaron Levie and Steven Sinofsky on the AI-Worker Future
The a16z Show
· · Hosts: Erik Torenberg, Martin Casado · Guests: Aaron Levie, Steven Sinofsky
Summary
The conversation argues that AI agents are less like a single autonomous supermind and more like a growing network of specialized, long-running workers. The panelists define agency as systems that can keep working in the background, feed outputs back into themselves, and do real work with minimal intervention, but they stress that current systems still need human checkpoints and clear task boundaries.
A central theme is that the most practical agent systems will likely be deeply specialized. The panelists describe code agents mapped to individual microservices, each with its own instructions and boundaries, as a way to avoid context rot and improve reliability. That same logic, they argue, will likely spread across enterprise work: agents will not simply slot into existing workflows, but will help redefine the workflow itself.
They also push back on the most dramatic AGI narratives. Hallucinations are improving, experts are learning how to verify outputs, and the economic value of AI still depends on who is using it and how. The broader claim is that AI will accelerate division of labor, create new specialties, and generate many new vertical products rather than collapsing work into one universal system.
Key takeaways
- The panelists argue that agents become more useful as they do more work in the background with less human intervention, but they still need human review.
- The conversation leans toward a specialized-agent model: one agent per task, codebase section, or microservice, rather than one all-purpose super-agent.
- The host and guests say hallucinations are improving and enterprises are learning to verify outputs, which is making AI usable in more critical workflows.
- One argument is that AI will increase specialization across the economy, creating new roles, tools, and vertical businesses rather than eliminating whole job categories.
- The panelists think the biggest near-term gains will come from experts using AI as a productivity multiplier, not from fully autonomous systems.
Notable quotes
“these are autonomous things that run in the background on your behalf and executing real work for you”
“we're seeing what what these are becoming”
“the more that you subdivide the tasks out then actually the more that they they can go pretty far on on a single task”
“the experts are now becoming the the productivity of an expert is is outpacing everything else”
How the episode opens
We thought that we were looking at the form factor of AI, which is you're talking back and forth to something. The real ultimate end state of AI and thus AI agents is these are autonomous things that run in the background on your behalf and executing real work for you. The more work that it's doing without you having to intervene, the more agentic it's becoming.
Somehow it produces output that it feeds back into itself. It's literally just the amperand in Linux, which is it's a background test. Okay.
And it's like the worst assistant in the world. and agentification is just hiring a lot of these really bad interns. I thought we I'd start this wide-ranging podcast by asking the very simple but very provocative question, what is an agent?
Oh boy. To who? Stephen.
Okay. Exactly. Go for it.
So, so I I actually have a very old person view of what an agent is, which is it's literally just the amperand in Linux, which is it's a background test. Okay. Because like you type something into 03 and then it's like, "Hey, I'm trying this out.
Oh, wait. I need a password. Can't do that.
And it's like the worst assistant in the world. And really, it's just cuz they need to entertain you while it's taking a long time to answer your prompt. And so that's my old person view of what an agent and agentification is just hiring a lot of these really bad interns.
The intern, they're getting better. They are getting better, but they still don't remember if I have a password to nature, you know, like it's just Is it possible you guys just had bad interns in like the ' 80s and '90s? We were We had terrible interns.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai agents
- ai agent definitions
- long-running self-improving systems
- specialized sub-agents
- future of work
Companies and tickers mentioned
- BOX
- MSFT
- Box
- Microsoft
- a16z
Watch the episode
Watch “Aaron Levie and Steven Sinofsky on the AI-Worker Future” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from The a16z Show
-
Cyberattacks Used to Take Months. AI Can Do It in Minutes. -
How AI Could Bring Halo and Other Classic Games Back to Life -
Mike Volpi on Why AI Infrastructure Will Outlast the Hype Cycle -
Why HubSpot’s Co-Founder Is Betting Against the SaaSpocalypse -
Inside Salesforce’s Acquisition Strategy for the AI Era -
How AI Is Turning Figma’s Finance Team Into Builders
The a16z Show, this episode and its artwork belong to Andreessen Horowitz. Octagon is not affiliated with the show; this page is an independent summary.
