Jay Kapoor's banner

Jay Kapoor

@JayKapoorNYC • 12,536 subscribers

GP @VSCVentures, backing physical AI | Founder, Dirty Jobs: Invite-only summit for 300 builders of the dirty, dusty & dangerous | Past: @CarnegieMellon @NFL

Shorts

I’ve already seen approximately 2,657 angles of The Tip of Manhattan and I still want more.

I’ve already seen approximately 2,657 angles of The Tip of Manhattan and I still want more.

185,542 просмотров

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Deedy Das has been a VC for 2.5 years. In that time he's backed OpenRouter, Wispr Flow, Goodfire, Pangram, and more. He also passed on a company Nvidia later bought for $20B. On this episode of CLIMB, Deedy, Partner at Menlo Ventures, brought *all* his receipts. We also had a fun debate about the value of going open source. When he said open weights are not going to be the cheap way out, I pushed back so he priced it all out for me. KIMI K3, about $15 per million output tokens. Well, Opus is about $25. That's all the discount you are getting, and to get it you are now hosting and maintaining a 2.8 trillion parameter model yourself. His read on the labs shipping open weights is that they are not in the business of giving away free things. They'll give it away until they can charge for it. What if frontier model costs don't meaningfully come down? Maybe free models were never free? Hear his argument for yourself and tell me if you agree. EP 101 of CLIMB is live 👇

Deedy Das has been a VC for 2.5 years. In that time he's backed OpenRouter, Wispr Flow, Goodfire, Pangram, and more. He also passed on a company Nvidia later bought for $20B. On this episode of CLIMB, Deedy, Partner at Menlo Ventures, brought all his receipts. We also had a fun debate about the value of going open source. When he said open weights are not going to be the cheap way out, I pushed back so he priced it all out for me. KIMI K3, about $15 per million output tokens. Well, Opus is about $25. That's all the discount you are getting, and to get it you are now hosting and maintaining a 2.8 trillion parameter model yourself. His read on the labs shipping open weights is that they are not in the business of giving away free things. They'll give it away until they can charge for it. What if frontier model costs don't meaningfully come down? Maybe free models were never free? Hear his argument for yourself and tell me if you agree. EP 101 of CLIMB is live 👇

93,080 просмотров • 10 дней назад

The jobs that actually matter don't happen behind a screen. Neither does the future of AI. Dirty Jobs is back in 2026. VSC Ventures annual summit for 300 founders, operators, and partner-level investors bringing physical AI to the world's dirty, dusty, and dangerous industries. One day only. September 23rd in San Francisco. No panels, press, or vendor halls. Real discussions and debates with all the people pushing the frontier forward. For two years, you had to be invited to join. Today, we open applications to everyone. Free to attend. Every application reviewed by human eyes, approved by human hand. Leave your chatbots at home and come 'talk dirty' with the people building what matters. Apply to attend:

The jobs that actually matter don't happen behind a screen. Neither does the future of AI. Dirty Jobs is back in 2026. VSC Ventures annual summit for 300 founders, operators, and partner-level investors bringing physical AI to the world's dirty, dusty, and dangerous industries. One day only. September 23rd in San Francisco. No panels, press, or vendor halls. Real discussions and debates with all the people pushing the frontier forward. For two years, you had to be invited to join. Today, we open applications to everyone. Free to attend. Every application reviewed by human eyes, approved by human hand. Leave your chatbots at home and come 'talk dirty' with the people building what matters. Apply to attend:

81,067 просмотров • 23 дней назад

Podcasts that celebrate companies scaling from $0 to $50M ARR rarely interrogate *WHAT* internally enabled that hypergrowth. So I did. Genspark hit $50M ARR in <5 months and recently raised at $1.25B valuation led by Joseph Floyd at Emergence Capital In this episode with cofounder/COO Wen Sang: • Why they killed AI search at 5M users • How their agents ship finished work, not drafts • What breaks internally when you grow this fast This episode is a playbook on speed. Link in comments 👇

Podcasts that celebrate companies scaling from $0 to $50M ARR rarely interrogate WHAT internally enabled that hypergrowth. So I did. Genspark hit $50M ARR in <5 months and recently raised at $1.25B valuation led by Joseph Floyd at Emergence Capital In this episode with cofounder/COO Wen Sang: • Why they killed AI search at 5M users • How their agents ship finished work, not drafts • What breaks internally when you grow this fast This episode is a playbook on speed. Link in comments 👇

108,413 просмотров • 8 месяцев назад

Turns out every viral robot demo you've seen was... staged. The backflips, the coffee, the laundry-folding — the robot got handed the answers before the test. Jagdeep Singh says the whole industry's obsession with VLA models is lighting money on fire. You may know my guest as a CEO took QuantumScape public and built Infinera before that, betting against the "impossible" physics twice and proven right each time. Now he's taking another huge swing in Physical AI. The bet everyone else is making: pay humans to tele-operate robots through a task thousands of times, then train on the footage. Jagdeep calls this a dead end. "A drop in the bucket" next to what a robot actually needs to understand the physics of our world. Rhoda AI's approach is revolutionary, training on internet video instead. Turns out 80% of the internet is video, more than enough to teach a model how the physical world really moves. The payoff is that while the the teleoperation crowd burns 70,000 hours of data, some claim 500,000, Rhoda gets a robot doing real factory work on 10 to 20 hours. Proven fact. We're staring down labor shortages in every dirty, dangerous corner of the economy. Turns out the problem was never the robot. It was the number of hours it took to make one useful. Thrilled to be an investor in Jagdeep and Rhoda alongside great investors like Vinod Khosla and Navin Chaddha. New CLIMB is live below and linked in comments 👇

Turns out every viral robot demo you've seen was... staged. The backflips, the coffee, the laundry-folding — the robot got handed the answers before the test. Jagdeep Singh says the whole industry's obsession with VLA models is lighting money on fire. You may know my guest as a CEO took QuantumScape public and built Infinera before that, betting against the "impossible" physics twice and proven right each time. Now he's taking another huge swing in Physical AI. The bet everyone else is making: pay humans to tele-operate robots through a task thousands of times, then train on the footage. Jagdeep calls this a dead end. "A drop in the bucket" next to what a robot actually needs to understand the physics of our world. Rhoda AI's approach is revolutionary, training on internet video instead. Turns out 80% of the internet is video, more than enough to teach a model how the physical world really moves. The payoff is that while the the teleoperation crowd burns 70,000 hours of data, some claim 500,000, Rhoda gets a robot doing real factory work on 10 to 20 hours. Proven fact. We're staring down labor shortages in every dirty, dangerous corner of the economy. Turns out the problem was never the robot. It was the number of hours it took to make one useful. Thrilled to be an investor in Jagdeep and Rhoda alongside great investors like Vinod Khosla and Navin Chaddha. New CLIMB is live below and linked in comments 👇

40,462 просмотров • 3 месяцев назад

Entertain. Educate. Energize. That's the storytelling pattern. More than ever, leaders across every industry recognize how much storytelling matters to success. And it’s the same pattern behind every great story that sells, whether it’s F1 Movie, Barbie, celebrities sweating on Hot Ones or your company. Remember: People don’t hate being sold to, they hate being bored. So whether your pitching customers, recruiting "cracked" employees or convincing an investor. Storytelling that sells follows this same pattern. Watch my latest Fast Break to help you fix your Storytelling 👇

Entertain. Educate. Energize. That's the storytelling pattern. More than ever, leaders across every industry recognize how much storytelling matters to success. And it’s the same pattern behind every great story that sells, whether it’s F1 Movie, Barbie, celebrities sweating on Hot Ones or your company. Remember: People don’t hate being sold to, they hate being bored. So whether your pitching customers, recruiting "cracked" employees or convincing an investor. Storytelling that sells follows this same pattern. Watch my latest Fast Break to help you fix your Storytelling 👇

66,278 просмотров • 11 месяцев назад

Just 6 vintage years produced 85% of ALL venture returns since 2000. If you think you can time venture, Noah Lichtenstein came with receipts that'll prove this thinking B.S. Noah Lichtenstein runs Crossover VC a unique fund-of-funds with access to ~800 companies and he has 120+ data points tracked on every one. He's backed 23 managers like immad Yuri Sagalov, Andy Konwinski, Jack Altman who he says, can win with a median fund size of $40M. Emerging manager allocation is now down 81% from 2021. Money flooding to mega funds. Noah thinks that's *exactly* the right time to back operator-led Fund 1s. His math: with 10% entry ownership, a $40M fund has a lot more ways to return 3x than a billion-dollar fund that needs $30B in exits. His hottest take: early-stage funds might produce liquidity faster than growth-stage. $500K at seed, you sell down by Series C. Lead the A and sit on the board? Locked in til IPO. Noah admitted he doesn't do a lot of talk shows or podcasts. But when he shows up, man, he shows up with ALL the data. Full conversation on the new CLIMB. This one you won't want to miss

Just 6 vintage years produced 85% of ALL venture returns since 2000. If you think you can time venture, Noah Lichtenstein came with receipts that'll prove this thinking B.S. Noah Lichtenstein runs Crossover VC a unique fund-of-funds with access to ~800 companies and he has 120+ data points tracked on every one. He's backed 23 managers like immad Yuri Sagalov, Andy Konwinski, Jack Altman who he says, can win with a median fund size of $40M. Emerging manager allocation is now down 81% from 2021. Money flooding to mega funds. Noah thinks that's exactly the right time to back operator-led Fund 1s. His math: with 10% entry ownership, a $40M fund has a lot more ways to return 3x than a billion-dollar fund that needs $30B in exits. His hottest take: early-stage funds might produce liquidity faster than growth-stage. $500K at seed, you sell down by Series C. Lead the A and sit on the board? Locked in til IPO. Noah admitted he doesn't do a lot of talk shows or podcasts. But when he shows up, man, he shows up with ALL the data. Full conversation on the new CLIMB. This one you won't want to miss

27,739 просмотров • 4 месяцев назад

Major family offices made 41 direct startup bets last month, almost all AI. We've seen this movie before. Now it looks like we're getting a sequel. Noah Lichtenstein built Crossover VC to have a real data-driven approach and focus on getting top-tier private venture exposure through both up and down cycles His data confirms that just a handful of vintage years produce most of venture's returns, all time. Josh Kopelman made a very similar point recently -- FRC made 90% of its profits in a single 36-month window. Meanwhile the FO allocators who went all-in on 2021 SPVs, went dark on venture in 2022, and are now rushing back into AI because of the Fomo That's not investing. That's gambling with extra steps. You can't time venture. Noah explains why 👇

Major family offices made 41 direct startup bets last month, almost all AI. We've seen this movie before. Now it looks like we're getting a sequel. Noah Lichtenstein built Crossover VC to have a real data-driven approach and focus on getting top-tier private venture exposure through both up and down cycles His data confirms that just a handful of vintage years produce most of venture's returns, all time. Josh Kopelman made a very similar point recently -- FRC made 90% of its profits in a single 36-month window. Meanwhile the FO allocators who went all-in on 2021 SPVs, went dark on venture in 2022, and are now rushing back into AI because of the Fomo That's not investing. That's gambling with extra steps. You can't time venture. Noah explains why 👇

10,697 просмотров • 4 месяцев назад

Больше нет контента для загрузки