Today we’re opening applications for the sixth cohort of @firstround’s PMF Method, happening in Sonoma from 9/27 - 9/30. This intensive, four-day retreat has helped over 100 founders take a straighter path to PMF. We take 0% equity in your company. No accelerator fine print. PMF Method gave @jgreze the yardstick to see how his company was doing against both the different levels of PMF and his peers. His business was working but he threw it all away to pivot @TownAI into a bigger market. The company has now raised over $70M from a16z and First Round. @annembrandes had the idea for @conwayinc but learned new ways to pressure test it, along with enterprise sales frameworks she still uses today. She raised a seed from First Round and Kleiner Perkins, and now has millions in revenue. And @pk_iv founded @browserbase and had early traction, but got the benchmark data he couldn’t find elsewhere to make sure the company was on the right track. The company has raised $67M from CRV, Kleiner Perkins and Notable Capital. These three founders join PMF Method alumni — senior leaders from Anthropic, SpaceX, Stripe, Rippling and more — who’ve raised over $750M in the last 18 months. There are many paths to PMF. Few of them are a straight line. But one program can help increase your odds of finding PMF faster. Apply below:

Wow...@endurancegeo put a power plant on an undersea volcano nearly a mile below the Pacific and generated net-positive electricity. It’s the first end-to-end demo of subsea geothermal power. Andrew Redd has built a killer team with enormous ambition, and they’re backing it up with real (and incredibly fast) technical progress. We’ve been lucky @firstround to partner with this team since the pre-seed and less than 2 years later, they’re generating power on the ocean floor. This awesome video gives you a sense of the work it took to get here.

I first got access to the alpha version of @TownAI at the start of the year, and it’s almost hard to count all of the new AI assistants that have launched since then. Most of them are free and built for your personal life. Town is intentionally built for work. The team isn’t focused on helping you plan a trip or book a restaurant reservation (although you can definitely do those things with Town). Instead they’re obsessed with the everyday work that truly makes you and your team faster. Several interesting product decisions have flowed from that: - You need multiplayer mode to truly be effective. AI has made each of us faster on our own, but the work we do together hasn’t kept up. Town is doing some of the most interesting work on multi-player collaboration (with humans as well as other agents) IMO. You can CC your Townie on an email or add them to a group text to let other people work with it directly. Your Townie can also reach out to your teammates’ Townie to help get you unblocked. - You are not the product. Town has been a paid product from Day 1, and none of your context is used to train a model. The team believes your business’ own data is your asset, not something for a 3rd party to monetize. - Your life is easier when AI figures you out, not the other way around. You don’t need to hack anything together or figure out what to ask Town. In addition to making proactive suggestions and picking up on your style, projects, etc., Town also picks the right model for each job behind the scenes, so you’re never locked into one provider. And the team keeps shipping. You can now talk to your Townie live, have it build decks, and make purchases and manage passwords through new partnerships with Stripe Link and 1Password, with tons more to come. I’m a bit biased after working with @jgreze and @tonydevincenzi at Dropbox and backing them from the very beginning, but I love their relentless focus on making you more effective at work, and I’m excited to see them put their $108M Series B (led by @ForerunnerVC and @MenloVentures) to work as the agent wars continue to heat up. More from the Town team here: town.com/operate-at-the-spee…

.@AranyaInc just made @fastcompany's 2026 Next Big Things. Aranya cuts GPU setup from weeks to <48 hours. Already running on 3,700+ GPUs. Congrats to the whole team! fastcompany.com/91612620/sma…

Rare to see an internal tool go from 4 uses to 158 the week it launches. @useactively built AI FDE, a Slack agent with access to their whole codebase and every database. It fields 70%+ of customer escalations and cut first-response time in half. When it gets something wrong, they ask it to write a code change to itself. More here 👇