sam lessin 🏴‍☠️

The state of VC in 2025. Where we are, where we are going, and where to invest ... docsend.com/v/w5bsh/wtfvc25

interesting AI discussion in the MOL app today...

"Centralize the Knowledge & Centralize the Team"... Small, mighty, and consolidated .. single knowedge base. is the way of 2026. In person, top 1%, drop everyone else and everything else. Very few people, very many tokens.

@yrechtman yah just like unlimited PTO ... expense "as much as you want" ;)

Best perk of working at slow is unlimited token budget

In the AI era, the power around 'Terms of Service' updates shifts... because you can just have your bot read them super fast and know what the company is actually doing... and then switching is usually almost free.

boom.

Coming this week on MOL

Master of Bots for Slow? I am interested in hiring a 'master of bots' to work on software and bot stack at slow / build internal (and maybe external apps, etc) with me. This isn't a for sure role, but I am for sure interested -- application below...

If you want to join (as a creator who wants a leg up on using bots OR as a bulider/TA bot expert looking to connect with creators) -- here is link ... this is gonna be GREAT - luma.com/bolmxe68

The number one place AI is really giving leverage? to creators in our creator fund who can now build software on their own and are optimizing the sh*t out of their busiesses with Cluade. So much so we are hosting a 2 day bootcamp for our creators & friends on bots for creators

One thing with AI is that all these slow companies are getting way more experimental in the products they are putting out ... case and point - check out from Tofu 'check if your identity has been stolen by North Korea to apply for jobs" :)

Lol - on techmeme today... just as I was discussing on the pod 2 weeks ago :) asymco.com/2026/03/10/the…

Why spend $100B on infra... when we can sell $100B of 'infra' to our users and make profit on it? lol apple.

I kinda wish they instrumented sublet and thunder... (also is Teewinot really 1,000 vertical feet?)

@DavidMKeyes hah ... for years i have been joking about the "e-cane" ... big battery and antenna in a from factor useful for self defense in the modern city ... once again the Israelis are a few steps ahead on implementation!

Khamenei's walking stick is how we got him. A Mossad agent named Mak El Halicha (censor cleared publication) spent years next to the Supreme Leader. We packed his stick with listening devices, gps, cameras and lidar (sorry Elon). We knew his every step. Literally.

Another trillion dollar use case ... of course the button is "brew/off"

@kiaran_ritchie nothing... you are missing nothing

I don't see how Anthropic, OpenAI or any of the model providers have any hope of defending their moats. And consequently, I think they're going to get wiped out. Right now, in early 2026 they have a meaningful advantage in terms of model capability. But far cheaper and open source models are not far behind. How long can they maintain a meaningful advantage? For the vast majority of use cases, we don't actually need much higher intelligence. It doesn't take 140 IQ to automate Turbotax or powerpoint. Eventually we will be saturated in cheap, local models that are "good enough". Of course some scientific labs and frontier research will always want the latest and greatest. But that market is orders of magnitude smaller than these company valuations can justify. What am I missing?

@wolfejosh amen.

everyone is saying "there won't be enough compute." I think the opposite the amount of capex being deployed for multi-gigawatt data centers does not make sense why the narrative is going to shift from centralized mega-clusters to localized edge inference...

@YakubuAgbese but Druckenmiller's whole point with that concept was to get ahead of heard being the most important thing / his point is all that matters is being early not late! (also btw this only makes any sense when you are in the public market and can also sell easily not just buy!)

@lessin Legendary investor Stan Druckenmiller says invest, then investigate And while it definitely sounds cool to say “I got a Ph.D and studied deep tech BEFORE it was cool” , truth is sometimes you gotta build the plane while you fly: invest with the herd first, develop an edge later

The VC Flight to ‘Deep Tech’ is almost certainly the wrong move for you Mr. Lazy-Narrative VC

good ragebait well done! you are totally wrong re cycle timing ... i have been super early on everything i have made real money on thematically and specifically -- several 100x and a few 1000xers ... BUT you are right about one thing ... i am only "rich" ... not super super rich-- because i have always enjoyed being early and the real money is made in AUM plowing dollars late -- being late and big is actually a better business model!

Sam is wrong. Even dogshit VCs know that you don’t make money being early, you make money going where the best founders are building Sam’s biggest win, an early check in Solana, came in 2018 at the END of the 2017 crypto wave after crypto markets crashed 2026 Sam would have shit all over 2018 Sam for being last in to a wave that had passed. In 2026, the best founders are starting cos at the intersection of AI x Real World or infra to power it. We are at the start of a multi-decade macro tail wind. Everyone is going to get hilariously rich and make so much money as long as they have enough bets anywhere in the space that one hits, like Sol did for Sam.