5 Lessons from TechCrunch All Stage '25

If anyone can build with AI, anyone can build a company. That was the message of nearly every talk at yesterday's TechCrunch All Stage.

I cut my teeth as the founding PeopleOps member in early-stage startups, so this was a kind of homecoming for me. An opportunity to see a cross-section of what the next generation of founders are thinking about in the AI generation. 

But I was also attending as a founder myself. AI has lowered the barrier to entry for software development. Now that any bozo can play in the sandbox, I've been building a product of my own.

Now that anyone can call up the basic skills of any specialization, generalists who know what to call and when will win the market. But now that AI has leveled the playing field, range is what separates those who will make it from those who'll be left behind. Here are my takeaways about what's changed and what hasn't:

1️⃣ Only spend time on what moves the business forward.

This hasn't changed. People and operational responsibilities are still the biggest timesucks that pull a founder away from the critical tasks that help them build to product-market fit. If this is you, DM me. 

2️⃣ When you're lost, ask founders who serve the same customers.

When you're testing an idea or going to market, insights from your peers serving the same customers will be the most useful, not advice from the pop stars in the Founders Hall of Fame. This advice hasn't changed, but the context you're working in has. 

No one knows what will win in this new landscape. Chart your own path. Your guess is as good as anyone else's.

3️⃣ Don't hire anyone you need to manage.

It took me a while to see the wisdom of this one, but it was true back then and it's still true now. Your first 20 hires need to contribute to velocity, and that means they need to work autonomously rather than taking up a teammate's time. 

4️⃣ Most teams aren't using AI yet. 

Despite companies investing millions into AI tooling, most teams aren't actually using it yet. If you want to leverage the benefits, don't just implement. Focus on adoption. 

5️⃣ Build micro teams with a strong leadership backbone.

The larger the team, the bigger the communication lag. Now that one person can achieve orders of magnitude more, micro teams can move at break-neck speed. But as collaboration is atomized, strong leadership at the center is more important than ever. 

6️⃣ Context translation is now a core skill.

Teams have always been limited by their ability to translate decisions into action. The most successful teams (and products) will be able to reliably build prompts that bridge real-world context to action.

Humans need to learn to better collaborate with AI, but AI still hasn't mastered collaborating with more than one human at once. This is the next frontier in AI-enabled work.