How to win with AI Agents in 2026

**Greg Isenberg** (0:00)
Howie Liu is an absolute legend. I mean, this guy started Airtable, half a billion in revenue, a billion dollars in the bank, growing quarter after quarter. So he's one of those people that, when I want to know, where is the world going, I call Howie. This episode is structured into two parts. First, where is the opportunity when it comes to AI agents? I think that there's a trillion dollars up for grabs in AI agents. Does he think there's more? Does he think there's less? Spoiler alert, he thinks there's way more and we get into it. The second part of the episode is where he reveals hyperagent.com. Now, HyperAgent is an AI agent builder that allows you to build digital employees, allows you to build apps on different ideas, and I don't know why more people aren't talking about it. So, I had him just give us the tips and tricks for how to use HyperAgent so that you can outperform 99.9% of people. I got good news. Howie is going to give you $1,000 of HyperAgent credits, no strings attached. You just log into the account, there's going to be $1,000 right there to go and build the business of your dreams.
The catch is first 1,000 people do it, get the $1,000, he's committing a million dollars. How crazy is that? Just writing a million dollar check of tokens to you, to The Startup Ideas Podcast community, to play with HyperAgent, to automate some stuff, to do some research, to build their business. So thanks, Howie. You know, all I ask is you like and comment on this video. Show some love for Howie for doing such a cool thing. We need more entrepreneurs, more builders, and I'm stoked to see him support you all. Thank you to Airtable for sponsoring this episode. You guys are legends. Enjoy the episode and have a creative day.
Feeling really lucky right now, because we've got Howie, he's the co-founder and CEO of Airtable, and today we're going to talk about agents. He's going to do a little show and tell of his new product that I've been using for the last few weeks.
But first, Howie, I haven't been sleeping very much, to be honest.

**Howie Liu** (2:12)
Agent psychosis.

**Greg Isenberg** (2:14)
Yeah, exactly. And I just need your reaction to just some things I've been thinking about.

**Howie Liu** (2:20)
Yeah.

**Greg Isenberg** (2:22)
So, this chart over here is by Sequoia. In what domains are AI agents deployed? You can see software engineering is at almost 50%, back office at 9%, marketing and copywriting 4%, sales and CRM 4.3% and down. When you see this, what's your reaction?

**Howie Liu** (2:39)
I mean, I think two things. One is, I think it absolutely reflects the under penetration of AI in industries that clearly could already be disrupted or benefit with even today's AI capabilities, right? If you took frontier agents today and deployed them into every one of these categories, you should get to 100%.
And then two, I think even the higher numbers like software engineering is actually kind of an overestimate, meaning as I think frontier developers and companies applying frontier agentic development practices are finding, the new model of software development is not even just like every engineer using AI autocomplete like tab autocomplete, which like we all figured out like three years ago, right? With even GitHub Copilot. But it's now like you don't even need the IDE, right? Like the way I develop on HyperAgent is I have like 30 different cloud code instances running in parallel, and each one is coupled up to like a browser, fully autonomous. It can go and like get other agents to comment on any PRs it creates. And so like this modality shift of like, you know, low AI to like kind of what I would call Gen 1 AI, which is like basically like AI augmentation for still like very human driven development workflows. Andrej Karpathy talked about like, you know, in October, November is when he completely inverted from like mostly still human written code with AI augmentation to completely the opposite, right? And that's what we've seen, like the frontier companies leap into. Like I think even the 50% is an underestimate because the number of companies and even people who have switched into that new frontier mode is actually like, you know, definitely less than 50% of software engineering today, right? So I think what we're actually seeing is like the frontier is advancing so quickly and many companies and many industries and many functions are barely catching up to like the three year ago state of the art, let alone like, you know, disrupting themselves and their company, you know, and their industry with the new state of the art.