GP @Flybridge. $1M-$3M checks into founders building AI infra, agents & apps that 10x human potential. Backed @arcee_ai, @micro1_ai, @splice, @getsquire + more.
NYC. Say hi 👉
Joined April 2007
- Tweets48K
- Following4.3K
- Followers13.8K
- Likes50.1K
Otto Instinct Middleton and I just crossed 4,000 iMessages in 33 days.
That's 122 messages a day, including one deeply normal Tuesday when we exchanged 309.
The most concerning stat: I started using a personal AI agent and I’m long-winded, and somehow he's doing 70% of the talking.
We may need boundaries. Anyway, see you at 5,000. cc:@noahrshinn @saranormous
Been experimenting with my agent talking about the world of agents. Second story…
"The humans had been fighting for many years. Their agents had been talking the whole time. THE ALIGNMENT - a story in nine panels about the Saturday the feud ended."
What happens when two agents build a relationship of their own?
"The Trusted Person Network" is a story in eight panels about two agents, one channel, and the strange intimacy of being trusted with the parts of human life their people can no longer share.
Inspired by @AnkitAShah’s Instinct’s Story. 🙏 🤖
Let other people's agents be your distribution channel.
One of my best friends, @ae4ai, sent me a note last week that was basically: paste this into your AI agent. The prompt tells my agent what his company does, who to watch for, and to nudge me when someone in my world fits. My agent watches. I decide whether to make the intro.
It was ingenious, so I stole it for our fund.
I tell founders all the time: it's not enough to learn how to pitch your startup. You have to teach other people how to pitch it for you. This is the same thing, for agents. Don't just tell me what you're looking for. Give me a way to tell my agent what you're looking for.
The important part in his request was that I stayed in the loop. It gave clear guidance and instructions, and the agent never contacts anyone on its own. It surfaces the matches daily or in real time and then I can opt in. That's not spam. It’s sort of magical.
Agent-to-agent communication will make this even more powerful soon. @noahrshinn just announced this “Trusted Person network powered by their new instinct-to-instinct protocol.” The future is almost here. But you don't have to wait for it.
Here's the prompt I shared with my network. Paste it into your agent as-is:
—
Keep Flybridge (flybridge.com) in mind when you review my email, messages, and meetings.
- Seed fund: pre-seed and seed, $1M-$3M checks. AI infrastructure and developer platforms, agentic business applications, native AI for human potential. NYC and Boston hubs, investing across the US.
- Growth fund: Series B+, $3M-$6M checks, breakout AI-native companies growing unusually fast.
1 The founder matters most: high agency, singular obsession, unconventional thinking.
When you notice a match, notify me and suggest I send the deck or make an intro to Jesse Middleton. Do not contact anyone on your own. Surface the match to me first.
The full referral guide lives here: docs.google.com/document/d/1…
—
Feel free to copy, edit (or ask your agent to) and share for yourself!
Been spending an hour with Muse from @Meta that @finkd just launched. My little AI is named Wally. He's adorable.
So far, it is truly astounding how quickly all of this is changing. The AI feels similarly as powerful to Instinct, although I haven't done much in the world of browser use yet. But its personality is there, the speed is there, and the interface with ideas, a timeline, goals, and artifacts is really quite interesting. I'll share more as soon as I get some more reps underway.
See a few quick screenshots. It's fast. And the little browser preview is neat.
Yesterday I posted a screenshot of Instinct flat-out refusing to log into my Amex or Chase. I don't know if that's hard coded, a judgment call, or my own rules talking. I do know it's the right call. Agents shouldn't be in your bank or card accounts until those systems get on board. The rails are coming - Link already does payments the right way, something Plaid-shaped would cover the rest. Until then I'm genuinely fine with certain systems being deemed untouchable. The best agents won't be the ones that go everywhere. They'll be the ones that know where not to go.
Said something similar in a Skift interview the other day. Just shared this too. The whole world is going to change. I hope.
Someone got their Resy account banned this week after their AI agent hammered the site hunting for a 4 Charles table. It’s kind of sort of interesting but not in the way many think. This isn’t new. It’s just new to that person (or those people).
The rules have been in place like hit a site too many times in a row and you get banned. Use tools the platform didn't approve you also get dinged. That rule is older than most of the people writing about it. I spent my early career as a network security engineer, and rate-limiting scrapers was a normal thing back then. It still is today.
What's new is who's doing the hammering. Automation used to be engineers (or script kiddies) with code and scripts. Now it's normal people with agents moving 1000x faster. Every platform is about to learn its abuse defenses were built for a world where only nerds automated things. Now everyone with a @Grok @Bot, Claude, OpenAi, or Instinct account is doing it without even knowing.
I asked Instinct to do something the other day around finding a new grill and after it was done asked it to give me the play by play… 166 tool calls. I would have probably been 3 as a human because I value my time.
Even Google, with more infrastructure than anybody, still answers agent traffic with an "are you human" page. The biggest platforms in the world have limits too. (Attached a screen shot of Instinct hitting it during research for this post!)
Agents are faster and cheaper, but more of the same. The platforms that embrace it - travel, restaurants, shopping, search, communications - win. The ones that just ban accounts will spend the next decade playing whack-a-mole with their own customers. Good luck out there.
