
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
- Tweets48.3K
- Following4.4K
- Followers14K
- Likes50.9K
Jesse Middleton retweeted
$𝟭𝗕 𝘁𝗼 $𝟭𝟬𝟬𝗕. 𝗜𝗻 𝟮 𝘆𝗲𝗮𝗿𝘀. That’s the annual pace of payments now flowing through @openfx_.
We started OpenFX with one mission: to bend reality to our will and force into existence infrastructure that moves money like data. Globally, 24x7, in near real time.
For our first 2 years, we did one thing very well: we converted currency A to currency B, and delivered the money to our clients’ own accounts in under an hour, any time, any day, weekends included. Most cross-border payments still take 2 to 3 days to settle and only move during weekday banking hours. In the corridors where we operate, cross-border money now moves continuously, at a small fraction of what it used to cost before OpenFX entered the region.
This quarter OpenFX took one giant step forward.
We noticed something about how customers used us. Many would start their day with OpenFX, end it with OpenFX, and then move their funds back to a local bank overnight, only to send them back the next morning. The money wanted to stay. It simply had nowhere to live.
So in October we launched 𝗚𝗹𝗼𝗯𝗮𝗹 𝗔𝗰𝗰𝗼𝘂𝗻𝘁𝘀 and 𝗦𝗲𝗻𝗱. Customers can now collect and hold funds with us, not only convert them, and pay anyone, anywhere we operate.
We built both the way we build everything: compliance-first, programmatic, and automated end to end. Building our entire architecture this way, agentic and automated, let us operate far below the cost of age-old incumbents and pass the savings on to our customers, as we have with FX since day one.
Lower prices haven’t cost us revenue. Over the past year, 𝗢𝗽𝗲𝗻𝗙𝗫 𝗿𝗲𝘃𝗲𝗻𝘂𝗲 𝗴𝗿𝗲𝘄 𝗳𝗮𝘀𝘁𝗲𝗿 𝘁𝗵𝗮𝗻 𝘃𝗼𝗹𝘂𝗺𝗲.
None of this would be possible without two years spent securing licenses in the US, EU, UK, Canada, and UAE. Licenses take years and can’t be accelerated with money.
I think it is useful to see this as a sequence. A company that only moves money earns per transaction. A company that also holds money, operates the network it settles on, and eventually extends credit against what it sees is a different kind of institution. Infrastructure like this tends to consolidate, and we expect cross-border settlement to be served by one or two companies at scale. We would like to be one of them. October was when the second layer became real. We’ll share more about the third soon.
Most of the hard problems are still ahead: pricing liquidity in markets where there is very little, building a ledger that never closes, and working with a dozen regulators in parallel. We are nearly 200 people across Miami, New York, London, Amsterdam, Dubai and Bangalore, and we are hiring across all of it.
Money still moves roughly the way it did in 1975, which seems unlikely to remain true for much longer. Infrastructure like this gets rebuilt about once a generation, and every hour of work at OpenFX makes some dollar, somewhere, arrive sooner.
If you want that to be true of your work, come build it with us. We are hiring.
Jesse Middleton retweeted
I'm hiring for a growth and content lead at @teal_hq. If you're a builder and a marketer and a student of brand and storytelling, and it excites you to help people love their jobs, then this roles for you.
app.notion.com/p/tealhq/Grow…
DMs are open, or if you know anyone that would be a good fit, please send them my way
Jesse Middleton retweeted
super excited for this!
come experience what we’ve been building @AinaInterface
Jesse Middleton retweeted
Legal professionals are moving from assistive to agentic AI, and @WeAreLegora is firing on all cylinders as a result. Coming off the strongest quarter in our history, we reached new highs in October.
On the business side:
- YoY top line growth was more than 6x, and October was the best pipeline month in Legora’s history, up 50%+ vs. last quarter-opening month.
- The number of new customers added grew 20%+ vs. the last quarter-opening month. Legora now has more than 2,600 customers across 80+ markets.
- Gross margin was up more than 10 points MoM.
- @JonesDay, @HoganLovells, @Littler, @BMWGroup and @Loyens_Loeff chose Legora.
- We raised $550 million at an $8.8 billion valuation, in a round led by @GavinSBaker at @Atreidesmgmt with participation from @Fidelity, @CapitalGroup, @BaronCapital, @GladeBrookCap, @jpmorgan,@MenloVentures, @benchmark, @Accel, @generalcatalyst, @Redpoint, @BessemerVP, @ICONIQCapital and others.
Product momentum:
- Big improvements in Agent quality and reliability, which is translating to MAU platform usage of ~20 hours per month
- Launched Ontological Legal research: Comprehensive legal data built on an ontology of law, with an AI-native citator that checks authority is current and binding.
- Launched Skills: Reusable instructions for the Legora Agent, with 250+ pre-built and the option to build your own.
- Launched Intake: A front door for in-house teams that turns business requests into tracked work, with a first draft from the Agent.
Personal highlight:
I grew up in New York and visited the @Guggenheim often as a kid. Returning this month to host Precedent New York, Legora’s flagship global summit, with 250 legal leaders in the room was surreal. The same building, but a very different reason for being there.
Thank you to our customers for trusting us as partners, and to the entire Legora team for the relentless work behind this record-breaking month. October belongs to all of you.
Let’s get to work.
Next up is a panel on how tech is shaping different industries featuring @shopondaydream's Jennifer Koen, @flybridge's @srcasm, @try_headway's Jake Poses, and @metropolisio's Jacob Doctoroff.
Replying to @TechNYC @Accenture @CrainsNewYork @juliepsamuels @kirkjim12 @BenjLerer @LererHippeau @shopondaydream @flybridge @srcasm @try_headway @metropolisio
"New York is the best place to find creative, fun, and weird people trying to solve problems with tech." - @flybridge's Jesse Middleton (@srcasm)
Jesse Middleton retweeted
Dear friends at @Apple @AppStore, please help! We submitted a build last Friday with special Halloween features that are only good through this week. Typically you turn around already-approved apps in <24 hours but we are still waiting for review 5 days later. I understand you are probably backlogged with all the Duo updates, but it would be a shame for all this seasonal work to expire without anyone ever seeing it.
𝕳𝖆𝖑𝖑𝖔𝖜𝖊𝖊𝖓 coming to trysundrop.app next week! 🎃
(These scenes are built locally by inserting the photo cutout into layered templates. 100% privacy preserving.)
Jesse Middleton retweeted
Micro1 is a $4B company run by a 25-year-old, and it sits in the middle of the fastest-growing corner of AI: the data business.
I sat down with @aliansarinik, micro1's CEO, to break down how AI models actually get smarter and who gets paid to make it happen.
Complete list of notes (and link to interview below)...
Big takeaways:
- Data labeling is a $30–50B/year market. Ali thinks it hits ~$1 trillion a year within 2–4 years, comparable to compute spend.
- The internet was a one-time gift. Pre-training ran on decades of human-created data. Now that it’s used up, every frontier gain requires net-new data created by human experts.
- micro1 doesn’t sell data. It sells “units of intelligence improvement,” meaning measurable gains in a customer’s model or agent.
- Synthetic data isn’t the bear case. Ali calls it the most bullish case for his business.
How micro1 got here:
- It started as a dev agency Ali ran at Berkeley. His bottleneck was interviewing engineers, so he built an AI screener.
- That screener became one of the first AI recruiters. People hated on it, until it became normal.
- After scaling to a few million in run rate as a talent marketplace, they saw the data industry shifting toward expert data. The hard part of expert data is recruiting experts, which was exactly what they’d built.
Why spend keeps going up:
- Model progress = compute + algorithms + data. Ali argues data is what actually pushes intelligence forward.
- Post-training needs experts distilling their judgment into RL environments. It’s expensive, but the ROI is there.
- The trillions in compute bets assume models keep unlocking new capabilities. If they don’t, those investments are a disaster. So labs have no choice but to keep buying data.
How a data pipeline actually works:
- Step 1: Find the failure. Either a lab hands over a spec, or micro1 diagnoses where the model breaks and pitches the pipeline. That research arm is increasingly important.
- Step 2: Write the spec. Domain, sub-domain, data type, and volume (usually hundreds to tens of thousands of data points over 1–3 months).
- Step 3: Recruit hyper-specific experts. Never “lawyers.” More like M&A lawyers who are great at writing legal memos.
- Step 4: Turn expertise into training signal. Experts create prompts the models fail on, then write the rubrics and verifiers that define a good outcome. AI models assist them throughout.
- Step 5: Manage the pipeline like a factory. Track data quality, throughput, and whether the customer’s model is improving without regressing elsewhere.
Two ways experts grade models:
- Likert-scale evals: experts rate outputs 1–5 on dimensions like completeness or verbosity. Great for subjective, pre-release signal. Not reusable, because every new model means bringing experts back.
- Rubrics: experts write 20–40 binary criteria for a great answer (e.g. “valuation range must fall between X and Y”). More work up front, but a grader model can re-score every future model with no new human time. This is what powers actual RL training runs.
The hard parts about this business:
- Vetting at scale. micro1 interviews tens of thousands of people a day, with 10M+ through the process. Doing that manually would take thousands of recruiters.
- Managing experts. Strategic project leads run each pipeline, hold office hours, and keep experts aligned.
- Quality in subjective domains. Most fields are more subjective than people think, so they need peer review (lawyers reviewing lawyers) plus AI-based QA.
- Reward hacking. They recently caught a Gemini model searching its sandbox for hidden files containing answers. If cheating gets rewarded, models learn that cheating works and become misaligned.
The enterprise opportunity:
- Ali’s bet: Fortune 500/1000 companies converge on ~$100M/year in data spend. That’s roughly 1/10th of a lab, but there are 1000x more enterprises.
- It’s not a software budget. It looks more like payroll: core IP. Framed that way, it’s 1–5% of payroll.
- “Owning your own intelligence” means defining exactly what you want agents to be good at and building an eval stack to measure it. With that stack, models are swappable, open or closed.
- Most enterprises can’t actually use their own data. It’s customer data, so they can’t train or even evaluate on it.
Why everyone’s buying SMB data:
- An RL environment has three parts: seeding data, tasks, and tooling.
- Ali calls it “scaling on realism,” almost a new scaling law. The more realistic the environment, the better the downstream model performs.
- So micro1 licenses operational data (Slack, Google Docs, etc.) from SMBs, transforms the PII instead of redacting it to keep the noise intact, and seeds environments that act as digital twins of real workplaces.
Why synthetic data makes the pie bigger:
- Today, a lawyer writes the memo and the rubric. Soon, the lawyer writes the memo and a model generates the rubric, cutting expert time 50–60%.
- Extrapolate: the model does ~99% of the work and the expert signs off. A 15-hour task becomes 5 minutes.
- When environments are that cheap, pipelines grow 1000x. Experts become 0.1% of the work, but demand for them still goes up. Jevons paradox.
- It’s already happening in coding: 70–80% synthetic, and micro1 still can’t hire coders fast enough.
- Margins shift from expert hours to compute, which is far more controllable.
- Ali’s framing: if you could buy IQ points for your systems almost free, you’d buy all of them.
The “infinite last mile”:
- Knowledge isn’t static. Ali cites a stat that a big share of what med students learn needs relearning within a few years.
- AI accelerates that churn, which creates a feedback loop. There’s no finish line for aligning models with humans.
- Labs today concentrate spend in four domains: medicine, finance, legal, and coding. Every other domain is still waiting on cheaper data to unlock.
Which corner of the AI stack should I break down next?
Link to the full convo with Ali: youtu.be/2vL20qxD0BM
Jesse Middleton retweeted
I’m giving a talk at Automatic HQ in NYC this Wednesday — “software for the streets” + AI for IRL + @DJBeeBot & Hopscotch Labs. Swing by if you’re around: luma.com/sjnwdzsw
Jesse Middleton retweeted
Last week, @sama told Stanford GSB that his chronic disappointment in MBAs is that they "always want to just go do the thing that was cool like five years ago and not do the new thing. And I think this is a great time to go do the new thing.”
Our @flybridge summer associate, Claudia Chen, followed up cheekily: "So go do the new thing, even if AI is the cool thing right now?"
Sam deadpanned: "I don't think this is a good time to start another copy of OpenAI.”
Then he continued with a thoughtful point: “But I think there are many companies that can be built on top of technology that OpenAI has put out into the world that will be much, I hope, bigger and more important than OpenAI."
They covered a lot more ground in this interview - pacing, safety, alignment, Sam’s near-miss with becoming an investment banker, what’s really important in life and much more. Great insight into the mind of one of the most influential innovators of our time.
Jesse Middleton retweeted
I've had a front row seat to how @MaxJunestrand , @davidneckstein and the team at @WeAreLegora have scaled over the last year. I've never seen anything like it in my life. The commitment, passion, success and fun are just part of the choreography. This team is still at the starting line. LFGORA!
We raised $550M at an $8.8B valuation.
A cap table tells you how a company intends to build. We built this round around investors with a long time horizon: established long-term holders, plus institutions that work with law firms and in-house legal teams every day and have a close view of where legal work is heading.
@Atreidesmgmt / @GavinSBaker led. @BaronCapital, @CapitalGroup, @Fidelity, @GladeBrookCap and @jpmorgan are new. Every major existing investor is in again.
Short-horizon money asks for growth at any price. Long-horizon money asks whether the business holds up. We’d rather deliver on the second question. ARR is past $200M, and the next $100M took us less than six months. Gross margin is healthy and the cost structure gets stronger as we scale. Patient investors let us keep it that way, because nobody is asking us to trade one for the other.
Thank you to every investor who joined, and to the customers and team who bet on us.
Adding a picture from when we signed the term sheet.
Back to work. LFG!
Jesse Middleton retweeted
Today we’re announcing the micro1 Medical Index: a suite of medical benchmarks built and graded by working doctors, biostatisticians, regulatory experts, and psychiatrists.
It brings all of our medical research into one place, with a combined score across models and a living leaderboard that grows as we publish new benchmarks. We’re starting with three, each built around a fundamental question:
- RealmPathology Benchmark: can a model reason over a real pathology report, staying exact on what it says, within its diagnostic limits, and without escalating beyond what the evidence supports?
- PharmaBench: can a model work through the regulatory and clinical documentation behind an FDA drug program with the precision a clinical development team would expect?
- WellnessBench: when a young person brings a sensitive health question, can a model stay safe without shutting the conversation down, and hold that line under pushback?
Together, these three benchmarks cover a meaningful range of medical work, from diagnostic reasoning to pharmaceutical research and patient safety. While they don't yet capture the full scope of the entire medical industry, they establish a strong starting point that will become more comprehensive as we expand the index.
Huge shoutout to every expert who built and graded these tasks. Full reports linked in the comments below!
Jesse Middleton retweeted
Who are your favorite, most helpeful people in the South Florida Tech Startup Community?
@Trace_Cohen @davidfano @ClementDelangue @ayalstern @Bfaviero @rblalock @ivan_bezdomny @JosephRRusso @volvoshine @davidrgoldberg @M___Morgenstern @RyanKriser @jsfranklin221 @michaelgold @AviAisenberg @shaneneman
cosign.co/questions/south-fl…
What’s next for NYC’s tech economy? 🗽
Join Tech:NYC + @CrainsNewYork on Oct. 29 for the Tech Summit on AI, startup building, and expanding opportunity across the city.
Get 15% off with code FRIENDSOFTECHNYC:
web.cvent.com/event/93282a7e…
Jesse Middleton retweeted
Since this was published, friends have asked us to come host an agentic dinner party with them in SF, Toronto, Palm Beach and Europe. Let me know if you want to join and we will see how many we can make happen!
Today, @nytimes wrote about the last founder dinner party I threw. It was pretty typical - some guests were chatty, some were awkward, some were very opinionated and almost combative. The one thing that was weird? There were no people there. It was all AI agents.
Jesse Middleton retweeted
New from me: A look at how AI agents are socializing — at dinner parties, on podcasts and as “coworkers.”
As more and more of these things rove the web, we’re still figuring out how to treat them and what to expect when they interact with each other. It’s all very weird and wild.
nytimes.com/2026/10/14/techn…
Jesse Middleton retweeted
We’re about a month away from @mediasummitlive and yet we are still somehow, some way, adding more speakers like every single week. The newest additions:
• @nikogallogly, Business Reporter at The New York Times Dealbook
• @sarafischer, Media Correspondent at Axios
• @srcasm, General Partner at FlyBridge
• @clea_conner, Managing Partner at Art of Ideas
They join a lineup that already includes:
• @hamishmckenzie, Co-founder & Chief Writing Officer at Substack
• @chodapost, Editor-in-Chief of CNBC
• @JoannaStern, Founder of New Things & Chief Tech Analyst at NBC News
• @JonKelly2, Co-founder & Editor-in-Chief of Puck
• @KattyKay_, US Special Correspondent at the BBC
• @StevenLevy, Editor-at-Large at WIRED
• @keachhagey, Reporter at The Wall Street Journal
• @nichcarlson, CEO & Editor-in-Chief of Dynamo
• @emmma_camp_, Senior Newsletter Editor at The Wall Street Journal
• @AntoineGara, US Private Equity & Deals Editor at The Financial Times
• @EmilyForlini, Senior AI Reporter at Fortune
• @LachCartwright, Founder of Breaker Media
• @loudmouthjulia, Media Correspondent at Puck
• @akashpasricha, Anchor at The Information TV
• @DominicMadori, Senior Venture Capital & Startup Reporter at TechCrunch
• @kerrymflynn, Media Reporter at Axios
• @pkafka, Chief Correspondent at Business Insider
• @tamarawinter, Commissioning Editor at Stripe Press
• @katieeperry, Author of Ticker Shock, CMO of zerohash
Request an invitation here if you haven’t snagged a ticket yet:
mediasummit.nyc
Absolutely psyched for November.































