Rob Bailey

New York can/should not create a new tax on company building by eliminating QSBS NY needs to be a MORE welcoming place to build. Don't drag us back. I signed the letter (linked below)

How many VC funds have $1 of DPI? And how many have 1x DPI? Damn this thing takes forever.

Have Mamdani and Hochul considered focusing on cutting costs instead of all this tax/cash grab?

@rohindhar This is real.

Someone told me recently they were at an event for people who had recently moved to San Francisco and worked in tech And mind you, the stat they said was so outrageous it strains credibility and I can’t really vouch for it 😅 But of the 40 people at the event 38 of them had moved from New York

@MilesDigitek Hey to be clear I didn’t cook this. :) If I did there would he tahdig. :)

Beautiful spread. Norouz Mubarak. Cultural traditions outlast political regimes.

If you read just ONE article this weekend, make it this. A 5-minute read could genuinely change your career trajectory. While everyone else panics about AI job losses, this is the blueprint to ensure you thrive. I wrote this for friends & family first - now it's public:

GTM engineer is now an essential role in any marketing org

Happy Norouz to my Persian friends and followers! Inshallah this time next year Iran will be truly free. :)

Alexander Stubb is a great negotiator. I'm really impressed and hope his proposal is accepted. Go Finland!

@bri_guy_ny Great point.

When New York City‘s housing market suffered for a little bit because of a once in a lifetime global pandemic, we were subjected to stories about how New York was dead every single day. But now that NY is thriving and FL and TX are collapsing? Not a peep.

Lovable’s Head of Growth has shared 5 key lessons about Lovable’s $1M to $400M ARR journey Elena joined Lovable at ~$1M ARR > $100M by July 2025 > $200M by November 2025 > $400M by February 2026. I’ve been following Elena for a long time (since she was doing sessions at Reforge in 2023) and take detailed notes each time she is interviewed on any show!

The revenue timeline on this company reads like a typo. $10 million ARR in mid-2024. $55 million by April 2025. $95 million by July. $200 million by October. $400 million by February 2026. 40x in 18 months. Fal is the invisible backend of AI-generated media. They don't build models. They host other people's image, video, and audio models and run them faster than anyone else. Adobe, Canva, Shopify, and Perplexity all route generations through fal's inference engine. Burkay Gur and Gorkem Yurtseven started the company in 2021 building Python runtime tools for ML. When Stable Diffusion launched and every developer needed fast image generation, they pivoted to inference for diffusion models. That timing was perfect, but execution is what got them here. They crossed $50 million ARR with 25 employees. $100 million with 45. Their entire go-to-market team at nine figures of revenue was six people. At ~120 employees today, they're running roughly $3.3 million in revenue per head. Their engineering edge: the team comes from compiler design, programming languages, and database internals. They optimize inference the way database engineers optimize query execution, squeezing milliseconds out of every API call across billions of generations. That advantage compounds with volume. The fundraising pace matches. $125 million Series C in July 2025 at $1.5 billion. $140 million Series D in December at $4.5 billion. Now a $300 million round at a ~$7.3 billion blended valuation. Four rounds in a year. Sequoia, Kleiner Perkins, a]16z, GIC, NVIDIA all on the cap table. The bet behind all of this: models are commoditizing. A breakthrough lasts three to four months before replication or distillation. The infrastructure that runs models at scale gets stickier with every generation processed. $400M revenue, ~120 people, EBITDA positive. These are the economics of owning the inference layer when every app on earth wants to generate media.

I ACCIDENTALLY OPENED MY CTO'S PERSONAL NOTION WORKSPACE AND NOW I UNDERSTAND WHY HE SHIPS 5X FASTER THAN THE REST OF US. He is 48. I am 26. He manages 3 products and never works past 5 PM. I work 10 hours a day and barely clear my Jira board. In his workspace, one specific document explained everything: Most people panic when the workload scales. They work longer hours, burn out, and eventually drop the ball. High performers do not manage time. They manage boundaries. The document was a list of strict operating rules. Here are 18 systems you can steal.

@sonicaghi Great memory! Any NYC plans? We still have 2 Locandas here….

@RMB LOCANDA!

When in 2011 we raised our seed for Mashape. Was $1.5M (originally planned for $500k) on a $6.5M post. At that time it was one of the largest seed round ever.

@jeremypiven And just like that…..

Exclusive | Adrian Grenier addresses potential 'Entourage' revival — and reveals which co-star is 'ready to work' pagesix.com/2026/03/19/ent…

@oanaolt I’m actually working on something new. Not consulting shop. :)

@RMB Start a consulting shop you could do this outside of a vc

If a vc really built the muscle to help founders hire early, really great sales people, they’d win every deal Founders, especially technical founders really struggle with this And the generic recruiting firms fall short. It’s so disfuncțional, it’s unbelievable.

@rabois There also should have been better choices.

True.

That isn't what AI thinks. :-) According to ChatGPT: Taste doesn’t disappear — it just gets re-encoded. A. Taste = structured signal Agents will interpret things like: Clean API design Consistent naming conventions Well-structured schemas Good documentation Predictable behavior 👉 This is basically “engineering taste” So instead of: “This app feels premium” It becomes: “This system is clean, consistent, and low-entropy” B. Taste = trust / reputation layer Agents (or the systems configuring them) will use: Benchmarks Ratings Usage data Network effects “Top tools for X” registries 👉 This becomes algorithmic taste Very similar to: PageRank App store rankings GitHub stars (but more dynamic)

@RMB of course they do (subset of taste) the rub is that AI will only are about function.

Software chosen by humans is about taste. Software chosen by AI is about function.