Gabriel's Recent LinkedIn Posts

Everyone is just talking their own book, keep your head down and keep going. Dario Amodei: "Engineers at Anthropic have stopped writing code. AI generates it. Most software engineering automated in 6-12 months." He runs an AI company. His revenue depends on you believing this. Sam Altman says AGI is imminent. He's raising capital at a $300B valuation. Lying or not? But they're not neutral. Narrative drives valuation → Valuation drives fundraising → Fundraising drives headlines → Headlines drive narrative → The loop compounds. AGI fear-mongering benefits people... That's why they do it. Ignore the hype cycles, ship something, talk to users. Repeat.

SoftBank sold 5% of Nvidia for $3B in 2019 and wired it straight into WeWork. The result? $245B left on the table…A $47B IPO that imploded… And an entire ecosystem that now thinks capital is the product. In this latest Lobster Capital piece, I break down: - The brutal math of a $100B fund - Why blitzscaling minted Uber & DoorDash but murdered everything else - And why today’s $10B+ AI war chests are playing the exact same high-stakes game If you raise, invest, or just care about the VC ecosystem, this one will make you rethink every term sheet you’ve ever seen. 🔗 Link in the comment section. What’s your take? Has “Capital As A Weapon” permanently broken early-stage? Drop your comments below.

Excited to back Perfectly (YC W26)! The data is clear: 70-80% of tech jobs are filled through referrals before they ever hit a public board. Yet, most candidates are still shouting into the void of "Apply Now" buttons. Victor, Gary, and Huimin are building Parker, an AI career super-connector that doesn't just "find" jobs, it pulls the right strings to get you in front of the right people. We’re moving toward a world where your career isn't limited by your current network, but by your potential. Parker makes that potential visible to the world’s best companies. Proud to support this team on their mission to make sure no top-tier talent goes unnoticed!

This is the most counterintuitive thing about early-stage valuations: If you raise $450K at pre-seed, investors take 15-20% equity. If you raise $900K at pre-seed, investors also take 15-20% equity. Your valuation doubled because you raised more money. Wait, what?... At early stage, investors don't calculate valuation based on your metrics. They (usually) anchor to a standard - 15-20% equity per round. The valuation is just reverse-engineered from that percentage. So if they're investing $900K and want 20%, your post-money valuation is $4.5M ($900K ÷ 20%). If they're investing $450K and want 20%, your post-money is $2.25M. Why does this happen? Because the ability to raise more money is itself the signal. A founder in San Francisco raising $2M pre-seed vs. a founder in Costa Rica raising $900K: the SF founder has better connections/investor access/market positioning. That's worth more. The amount you raise is (often) a credibility multiplier. 𝟭. 𝗗𝗼𝗻'𝘁 𝗿𝗮𝗶𝘀𝗲 𝗹𝗲𝘀𝘀 𝘁𝗼 𝗮𝘃𝗼𝗶𝗱 𝗱𝗶𝗹𝘂𝘁𝗶𝗼𝗻. Raising $450K instead of $900K doesn't save you equity. You still give up 15-20%. You just cut your valuation in half. 𝟮. 𝗜𝗳 𝘆𝗼𝘂 𝗰𝗮𝗻 𝗿𝗮𝗶𝘀𝗲 𝗺𝗼𝗿𝗲, 𝗱𝗼 𝗶𝘁. Raising $2M instead of $900K signals you're worth more. Your valuation goes up proportionally. 𝟯. 𝗧𝗵𝗲 𝘀𝘁𝗮𝗻𝗱𝗮𝗿𝗱 𝗯𝗿𝗲𝗮𝗸𝘀 𝗮𝘁 𝗲𝘅𝘁𝗿𝗲𝗺𝗲𝘀. If you raise $5M pre-seed, investors won't take 15-20%. They'll want more or price you differently. But within the $500K-$2M range, the standard holds. This is why pre-seed fundraising is a weird game. You're not being valued on fundamentals. You're being valued on fundraising ability. And fundraising ability is a proxy for everything else: network, market timing, founder quality, narrative strength. The higher valuation gives you leverage in the next round.

A third to half of every YC batch was rejected before getting in. People commonly apply 2-4 times before acceptance. That number is going up, not down. Why? Either YC is getting better at second looks, or the signal is getting harder to spot on first pass, or persistence itself is the signal. I think it's all three. But the founder takeaway is if you can't survive being rejected once, you're not built for this… Garry Tan "If you can't be rejected even once and you mortally hate YC after, maybe try a different profession. Being a founder requires insane resilience. One rejection? There’s going to be a thousand more." This connects to a Yoni Rechtman pivot framework I posted before. When founders ask for advice on pivoting, those pivots fail. When founders announce a decision, those pivots work. Asking = uncertainty. Announcing = conviction. Same with YC applications. If you apply once, get rejected, and spend 6 months complaining on Reddit, you don't have conviction. If you apply, get rejected, build for another 6 months, and reapply with 10x the traction, you have conviction. YC's rejection-to-acceptance pipeline is a feature - filters for resilience before Demo Day. If you're a founder who got rejected from YC, treat it like a pivot checkpoint. What changed between application 1 and 2? If the answer is "nothing," you haven't learned anything. If the answer is "we went from idea to $50K MRR," you probably get in next time. The mortal ego wound is the test.

YC companies don't die because they ran out of money. They have money. They die because they get tired, demoralized. Usually it's co-founder issues. Great co-founders fight all the time, about everything. But they do it in a way that preserves the relationship. There's never a conflict so important that it's worth more than the relationship itself. That applies to co-founders, marriage, close friendships. The pattern repeats.

Second-time YC founders don't fail by repeating old mistakes. They fail by overcorrecting for them so hard they create new ones. Quang Hoang did YC twice… Plato in 2016, then came back with Vybe in 2025 Same accelerator… Completely different eras… In this episode of The Lobster Talks, we break down: - How YC evolved - Why batch size changes founder dynamics - What second-time founders still underestimate - And how AI shifts the playbook If you’re considering applying to YC, or investing in YC companies, this is the real picture. 🔗 Link in the comment section

Last week, the US military used AI to execute an airstrike on Iran. The AI identified targets, ranked threats and designed the attack sequence. That technology came from Silicon Valley. Which is weird. Because ten years ago, Silicon Valley wanted nothing to do with weapons. Infact, In 2015, Google employees walked out when the company worked on Project Maven. Military AI for targeting drones. They said, we do not build weapons. We connect people. Google ended the contract. That was the culture. Defense was untouchable, You could not raise money for a defense company in Palo Alto. You just could not, VCs called it a widowmaker. But fast forward to today and a company called Anduril just raised four billion dollars at a sixty billion dollar valuation. If you don’t know them, they build autonomous weapons, drones that hunt other drones, submarines with no crew. Basically, the most modern tech in warfare. The investors? Thrive Capital and Andreessen Horowitz. The same firms that backed Instagram, Stripe, OpenAI. Defense went from radioactive to the hottest bet in venture capital. So what happened? Why is every major VC firm suddenly pouring billions into weapons? And maybe more importantly…. Should we be worried about this? 🔗 Link in the commetn section

AI creates two paths that didn't exist before: 𝗣𝗮𝘁𝗵 𝟭: 𝗧𝗵𝗲 𝗟𝗶𝗳𝗲𝘀𝘁𝘆𝗹𝗲 𝗨𝗻𝗶𝗰𝗼𝗿𝗻. You build to $1-2M ARR solo, stay profitable, never raise again. You have the option to sell at 5-10x ARR to a strategic whenever you want. That's a $5-20M exit you control entirely. You own 100%. That's generational wealth without a single pitch deck. 𝗣𝗮𝘁𝗵 𝟮: 𝗧𝗵𝗲 𝗕𝗹𝗶𝘁𝘇. You build to $1-2M ARR solo, then raise $10-20M and hire 50 people in 6 months. You have more leverage than any seed-stage founder in history because you already proved you can build and sell. You're not raising to find PMF. You're raising to buy speed. The middle path (raise $3M, hire 10 people, grind for 3 years) might be dead. You either own the outcome or you raise enough to dominate. For investors… if a founder is at $1M ARR solo and raising $3M, ask why. If they're raising $15M, ask if they can actually deploy it in 18 months. The delta between those two questions is the entire investment decision.

🚀 Excited to announce our upcoming webinar: YC W26: Investment Insights by Lobster Capital YC's Winter 2026 batch is one of the most consequential in Y Combinator's history, and we've been in the weeds on every company. Join me on Monday, March 23 at 11:00 AM PDT for a live deep dive into W26. We'll share which companies we've invested in, the themes we're most excited about, and the signals we used to cut through 200+ startups to find the ones that matter. What we'll cover: - The macro themes shaping W26 (agentic AI moving from demo to deployment, AI as default infrastructure, and the next wave of vertical SaaS) - Our favorite companies from the batch, and why - The traction signals that stood out before Demo Day - How we think about valuation and entry points at this stage This isn't a rehash of Demo Day slides, it's our unfiltered take on where the best opportunities are, and where we think the market is wrong. 🎟 Free to attend. Spots are limited. 🔗 Register here: