MORE OR LESS – With Still California Resident Parker Conrad
Parker is a great hang… Jess and I do More or Less with our special … still California resident Parker Conrad (phew). We talk about where the real leverage in AI is / being "top of wallet" agent context, prediction markets and Hayek, random reward compensation schemes, and how un-fun it is to work on nuclear missile systems — and of course we keep talking about Cuck financing, because it is both hilarious and real.
HOT TAKES
- Election Reactions … including the laws of irony says only Democrats can build data centers, why plastic surgeons and stylists are going to make a fortune on post-Mar-a-Lago face makeovers, pour one out for the DC operatives, owning data centers already built & space is good (irony again!), tendency towards disequilibrium … also most importantly, what becomes of "no kings!" — that was the most fun / engaging message the Dems had! Hard to keep that going when we have nicely proven that we do not… in fact… have kings… kinda loses some of its zing. Finally, is Wyoming the last "free" state… discuss.
- Some personal peptide evidence — this is interesting (maybe just to me) but peptides do really seem to have been pretty useful through marathon tune up this time, comparing with lean mass crash last time, through my sub-3 hour push.
- How Do We Solve for Retail? … the top 1% in intelligence, drive, etc. are all gonna be totally fine. Do not worry about them… worry about the fact that AI is swiping at "retail" / the easily accessible "mass" stories that appeal — mass meaning, mass jobs, etc… what do you do? You gotta figure out new retail solutions for meaning, entrepreneurship, etc. for the top 30%, not the top 1% … after all, in history is it second sons who tend to cause all the revolutions and chop off heads — all the access, none of the assets…
- Public ANT isn't actually an interesting trade — invest at a 1-2B valuation, hope it gets to 10B... that's a bounded 5x. Go bigger than that and it just means the rest of the economy is f-ed, ex the inflation story.
- Slow will do any seed deal at $1B post — VCs used to say "you pick the price, I pick the terms" … now you might as well pick the prices as well … happy to announce that we'll do any seed deal at $1B post, on request, with up to .001% of any core check. Sure, why not.
- Watching RDDT slide — all I can think about is how awful a deal they did with OpenAI for all their data. This is like Github all over again.
- Lion, 9, incredulous — "so you are telling me there used to have to be a piece of wire to everywhere you would want to make a phone call?" Every generation's "obviously true" thing was somebody else's miracle.
- Jessica's "queen of the valley" era — I always thought of myself more as an "intern"… but maybe I need to change my Twitter bio to consort? Lol.
- New York Bath House Documentary … bath houses in NYC are having a moment … but I keep thinking about the old one owned by two brothers — they had a falling out, now they don't talk, and they both operate separate brands and operations in the SAME SPACE where they alternate days… how is there not a documentary on this?
- VC isn't going to successfully replace Universities and Governments in Fundamental Research — maybe big companies will (maybe) … but there's this whole trend of VCs starting to really invest in "science" … personally, I don't buy it. Science / fundamental research is critical for society, but broadly not money-making — the big dollars, rigor, willingness to lose money, etc. REALLY doesn't fit with venture capital. I am short it working (though it is really fun).
USE OF AI TOOLS
- SLOW AGM in review — some of my partners have been pushing out their "cut" of our AGM deck … I feel it's appropriate that for mine, you get a live presentation from Max Lessin (redacting the hard-core numbers, of course — you have to pay us management fees for those). I'll just note that this presentation was a one-shot to my AI system including all the videos, insights, etc… be optimistic, disciplined, and terrified all at once … also I don't know why we do these anymore when Max Lessin can do it better :)
- Nod — Human Review for AI Agents — I started building / have a V1 of a "human in the loop" review queue for AI actions. Let me explain: everyone is using different tools, Muse, etc… and I HATE them going wrong, and I also HATE them making mistakes. The ironic solution is a service which inserts a "human" assistant to check key work before it gets done. Feed a "Nod" key to your agent of choice and broad instructions, and pay more to have an actual human "supervisor" make sure your bots aren't screwing up / tell the bots to check in there. TBH, this is half an "RFP" — so if you (or someone you're building with) is making this for real, it should be a good business (based on my Fin experience of a decade ago and the reality of agent mistakes). Hit me up.
- Religious AI Routing — this is also something I've been half working on, similarly interested. Some people want halal AI, others want Rabbinic, whatever… point your systems at a "religious" router, and it will make sure the answers you get back pass the right rules / panel of religious experts, etc. Could work with a Nod-style system and kick the problem to the Rabbis of your choice if you aren't sure. BTW — with all the real estate, tax status, and religious authority, how on earth is the Catholic Church not in this game?? Same thing, half RFP.
- Torah Bot & The Lessin Boy's Podcast … these are old cuts, but honestly they're working so well. I'm learning so much from Torah Bot personally, and love that all these Rabbis are now reading it and sending feedback. The Lessin Boy's Podcast now has a serious catalogue of really interesting episodes … the kids love it, I learn a ton from it, etc… and the pipeline to create full YouTube videos one-prompt — things like "the history of lego" and "the history of tennis" — just excellent.
Best,
Sam
P.S. TGR's new ski film this year has Jackson Hole (of course / home base) and Jay Peak (first ski place!) — this is very exciting :)
P.P.S. Also this guy is hilarious … one place I feel like AI is really gonna show up / is showing up is that feed is getting way more compelling. And I'm finding funnier and funnier and weirder content in so many places. We might be in a golden age of comedy even if no one feels it / it's too fragmented to experience all together (also — long comedy, which I think will be one of the last remaining domains where humans rule over AI, because what makes things funny is exactly the swerves on patterns that AI can't do).