Miami's problem was never capital or founders. It's engineers. Proud to call this place home.

How I write product specs in 2026: 1. handwrite 10 bullets on paper like a caveman 2. hand-draw a mock that looks like a 5 year old did it 3. ramble at my cofounder for 30 min while Granola transcribes 4. feed the messy transcript to Claude and argue with it until it stops agreeing with me 5. cofounder reads the doc, leaves 7 comments. all correct. Claude agrees and contradicts itself. 6. Claude cuts it in half 7. ship the hand-drawn mock anyway because it was better than the doc and actually what eng needs what’s your version of this

Already on Episode III: Return of the Jocks 🍿

Chag sameach to my Jews. Passover is about breaking free of what’s holding us back. We call it yetziat mitzrayim. The vibe shift is to just take action. @pmarca gets it.

Saw this at every big org. Never building one like that.

OH at the office today: "gonna get some sun while agents are running"

Just bought a printer. Two modes these days: @Claude or pencil/paper.

The type of event you actually want to watch the panels. Nice work @CGarrett_15 and team Hill & Valley '26

How many times have you copied something out of Claude into a Google Doc, only to lose all your comments when you regenerate? That workflow is broken. @moment_dev fixes this 👇

Best founders aren't following anyone's playbook @rabois

Accurate 👇

startup idea: streaming inference for LLMs. i just want to push updated transcripts and get live feedback without every version filling the context window. chat interfaces weren't built for this. Anyone doing this?

AI founders these next two weeks 👇

officially declaring eye in the sky by alan parsons project as the GPT-5.2 theme song 👇 open.spotify.com/track/2sIbH…

Congrats to @technosophos, @matabortosei and the whole @fermyontech team on joining @Akamai. Honored to have been on this ride with you.

Early stage companies have to get a few critical decisions right despite making many mistakes along the way. Big companies that move fast face the same reality - lots of poor decisions, but with longer feedback loops. Existing processes and captured customers delay the 'no' that startups hear immediately. By the time big companies realize they've failed, there's post-hoc rationalization and distributed accountability to soften the blow. Founders can't hide. Decisions are exposed. Failures are immediate. That brutal transparency fires me up every day.