Episode #114: Building AI-Native Developer Infrastructure Erik @bernhardsson is the Co-founder and CEO of @modal, building high-performance AI infrastructure. We talk about building what is essentially a new cloud provider, created from the ground up, optimized for AI. We also

I talked to @bernhardsson about being bullish on AI while also questioning if it's a giant bubble: "On one hand, the revenue growth is insane. And in markets you would have never predicted, like vibe coding. And these people don't care that there's an AI bubble. They're paying

I asked Erik @bernhardsson what he learned joining Spotify as the 40th employee: "For most startups, if you just have smart people who are pretty autonomous and have a mission, you need a lot less structure than you think. People should just come in every morning and just ask

I talked to @bernhardsson about creating an AI-native infrastructure provider from scratch: "When I was doing machine learning at Spotify, 90% of the work was just building the infrastructure, not the application. And traditional infra is not built for fast feedback loops. When

@bernhardsson @modal Links to stream YouTube: https://t.co/bVxtmAkqLD Spotify: https://t.co/8bkBkyNVQG Apple: https://t.co/GrjRdpR5ZG Subscribe here to get each episode + the transcript in your inbox every week:

Upcoming guests on @ThePeelPod: - Erik Bernhardsson (@bernhardsson), Co-founder of @modal - Adit Abraham (@aditabrm), Founder of @reductoai - Steven Fabre (@stevenfabre), Founder of @liveblocks - Daksh Gupta (@dakshgup), Founder of @greptileai - Will Gaybrick (@gaybrick),

@awwstn @howie_ai Full episode YouTube: https://t.co/ulDyGlNjYB Spotify: https://t.co/IpEZCeBpr6 Apple:

The pitch deck strategy @awwstn used to raise $6m for @howie_ai: "Docsend analytics prove that the majority of people click all the way through the end of a deck, no matter the length. But the vast majority never spend more than 3 seconds per slide. Whether it's 10 or 100

I talked to @awwstn about the mistake of optimizing for growth too soon: "There's many proven ways to growth hack your product and make metrics look the way you want. But if you're trying to build a durable product, none of those matter. You should live in the ground truth of

I asked @awwstn how they came up with the name @howie_ai: "We wanted it to be a name. And we had a bunch of parameters, and were building this big list. And then Dave sent me a meme of a guy named Howie Dewitt. And its a screenshot of his Facebook profile, with the Montell

I asked @awwstn what its like working for @jason

I asked @awwstn why email-based AI assistants are so hard to build: "The architecture itself is not hard to build. The hard part is all the context and intuition that goes into each decision. None of them are hard 100% of the time. But they're very hard for an AI model to get

@awwstn @howie_ai @hanover_hq h/t @adaugelli, @sophiaamoruso, @anothercohen for their help brainstorm topics for this!

@awwstn @howie_ai @hanover_hq Links to stream YouTube: https://t.co/ulDyGlNjYB Spotify: https://t.co/IpEZCeBpr6 Apple: https://t.co/GSbEQZxUml Subscribe here to get each episode + the transcript in your inbox every week:

Episode #113: Howie: The AI Secretary I talked to @awwstn at @howie_ai about why AI assistants are so hard to build, deciding to make Howie super narrow, and everything that went into build the product, brand, naming, and viral launch video. Thanks to @Hanover_HQ for supporting

@frantzfries Listen to our full conversation on YouTube:

I asked @frantzfries why sending emails is still a huge problem for software companies: "Software companies tend to use 2 to 5 different tools to email their users. It's everything from notifications, to product updates, newsletters, investor updates, customer or new hire

Links to stream YouTube: https://t.co/lNoYnY283C Spotify: https://t.co/mVWQPcHvFM Apple: https://t.co/3okMYweUpG Subscribe here to get each episode + the transcript in your inbox every week:

I talked to @frantzfries why it's so hard to improve your CAC: "The best way to improve your CAC is not to have better ads. It's to have a better product. Everything's built on the product. That's what you're ultimately selling. But if the product sucks, no amount of marketing

From @frantzfries on the biggest lesson from selling his first company: "Never let anyone take your own compensation out of your control"