September 11th, 2001 + 25… is on the mind as I fly west after several days in New York. It was our first day of classes at Harvard and the world changed, dramatically. It was a real demarcation between the high school post-cold-war, tech boom, feeling of US power and security and college where the world very joltingly felt different. It also was a moment where the whole country came together in a way that I will always remember… that we ultimately are one people, and can do hard things…. Videos of the USA chant when bush threw out the first pitch of the World Series that have been circulating on the internet bring chills — for me 10/7 … a million miles away physically, but not so much emotionally or culturally felt just as disorienting and scary, if not moreso in the physical hand-to-hand glee of it / zero intermediation by technology. Never again happened… again. So what? So - today as we enter the new year there is no resolution, no payoff… just a view of how dark humanity can be when corrupted by degenerate ideology, that the West is to be cherished and defended, that there are real heroes who will selflessly do the right thing (would I be one, I don’t know) … and that despite all differences we can come together, face demons, and do hard things — and must. I am happy I got to see the Freedom Tower rise over Manhattan today.
MORE OR LESS :: iPhone Duo, Meta Muse & AI Scared Dave.I chose the US Open instead.. but I hear the gang got into the Duo (I will probably buy one, though I was looking for a dual mag-safe mount with a folding hinge to stick the two iPhones I carry together as a joke this week)… Muse (which is honestly really good), and some stuff where Dave got scared. Anyway, I would be listening to it on this flight home to catch up, but why when I can just get my bot to pull the topics and conclusions?
HOT TAKES
K-shaped inflation in New York Hotels .. I am trying to push this concept into the public consciousness so go with me… every time inflation numbers come out on the cost of mass market bread and TVs I laugh, as I sip on a $12 late or eat $42 sushi roll at the US Open. This week I stayed in NYC and a decent hotel cost $3500 a night (just decent) because of fashion week, etc. also lack of supply, also the fact that at the high-end of the K the market can and will bear insane inflation. One person this week was joking about the $10K a square foot soon-to-be reality post AI IPOs in SF. The point is this — the economy is splitting— the rich ARE getting richer, but the bubble services and product they buy are having no problem keeping pace… we have K-shaped economy which doesn’t just mean K-shaped income, it means K-shaped costs.
Railroads buildout borrowing was 6% of GDP… and everyone still went bankrupt… AI is going to test if we really are free market capitalists… AI debt this year is running basically equivalent to all US municipal debt without any of the tax benefits… even if AI is someday all its cracked up to be (which with a globally liquid free to transport non-regulated commodity seems really hard to square) — history says we are all gonna go bankrupt at SOME point… UNLESS it turns out that DC’s demand for GDP growth to service debt + SV’s legit technical breakthrough + cult beliefs + NYC’s over-the-moon happiness that tech finally needs real money and they get to participate vs. these damn capital efficient startups = AI really is communist and we force this whole new-world thru because everyone needs it to work this way. A generation ago the US basically stole all the railroads from the brits (they paid, we repossessed) — hopefully no one or no ‘entity’ repays the favor.
It’s so easy to make products, there will be no more product companies…. You can’t invest in products anymore… themes, people, cults, narratives, and of course traction in the form of MONEY… but the idea someone shows you a sick product and you want to invest / I think those days are over… any software, any product is just like writing - it is consumable and too easy to remake (especially in an era of weak copyright / where we have no ability to protect IP … even drug IP)
Kepler came out of stealth this week… very excited for Debo, Sasi, and team… we were I believe 50% of the first close of the seed… yes, deep tech isn’t usually our thing BUT the pitch was just so damn compelling even before AI and the memory crunch, how could you not… and now…. Wow. Excited for the team and have had the privilege to be there with some of the earliest capital. Getting to show up first when it isn’t obvious is the part of VC I like…
The GLUT of Belief — speaking of which, I am really struggling on this one — when infinity is possible or marketed (the west has a hard time computing stuff with infinities running around)… and infinite possibilities don’t need to be delivered to be liquid narratives… you end up with a GLUT of early stage belief. People might believe anything… which makes it really hard when your profession is obscure beliefs that ideally are slightly more right than wrong.
Alignment with WHAT … maybe it is just the scare run-up to some monster IPOs… or maybe just my personal experience where claude has been pretty testy with me recently about some super reasonable asks (no worries, there is a whole ecosystem out there) — but I really am getting sick of people talking about ‘alignment’ as tho that word means anything on its own / without context… Personally, as a freedom and capitalism guy — to me the object of Alignment should be the user / needs to be (yes I get that JSM’s harm principle doesn’t work so well when we aren’t out in the woods with sticks but have AI leverage).. If you believe in state planning then alignment with the State? When you start talking about alignment ‘with humanity’ … it takes some real balls to say you have the right to make that assertion / claim. Just like arguing about the internet in Social Studies is a bit intellectual - the market will get what the market demands / zero will exist in the west because merchants make it regardless of what polite society and the church says, it feels like we are having discussions with a clock running out on serious shit — and that doesn’t feel like a very good way to have a conversation.. such is life.
In a world of abundance only positional goods exist … competition is everything & Politics is everything when you have nothing to contribute to the market — which makes alignment even tougher — because people aren’t happy being materially met .. in fact, if you meet all your material needs and have nothing to work hard for, it tends to make people very sad — so they invent positional goods, which they can never have to pursue instead (rich people get this / get to live slightly in the future)… so what are you aligning for exactly? Striving for material ratification, ideally in teams, is the ultimate long-standing retail meaning framework for humans… take that away, and you better have some good ideas (because Star Trek didn’t)….
The US Open Proves It, of Course The Jobs Are Going Away — just look at the line callers at the US open… there used to be so many judges… now there is just an AI and a red light on the net…. Which Brings Me To The Ideal AI Startup RFS — Seth, Dani, and I were front row for the women’s semis .. which was fun because according to a bunch of people we were on TV a lot … so what is the AI business? Well... when you buy tickets you should be able to validate your place in the stadium and how you will show up on TV… people should then BID for you to wear their advertising hat / shirt, etc. (which scales really nicely, because if EVERYONE is wearing the same shirt, it is pretty noticeable — and bonus points if you are attractive / submit a selfie)… then have AI watch the whole game, validate you wore the ad, and pay out on the screen time you got… This is a real business idea and it is 100% an AI one, because all that validation and processing of seat position, on screen time, market making, etc. for quick ads is too inefficient without the bots! That is what all those trillion dollar data centers is for!
If we want American kids to care about science, let them gamble on it (btw the mainstream models won’t let you make marketing material about gambling, alignment oy!) — but seriously, meet them where they are at… if they had a 1:1 Einstein.. or the rookie cards of the entire Princeton Math Department…. They might be more into science.
What do you do with unlimited PHDs… A million PhDs is a real, one-time dividend on the narrow set of problems that were always just search problems dressed up as hard science — worth having, worth funding, genuinely underrated even if it's "just" protein folding plus a dozen cousins. but it's not a civilizational phase change, because the actual constraints on human flourishing were never a shortage of intelligence.
Nice chat with adweek on the slow creator fund — creator capital is still scarce: $10M valuations on $2M of real revenue vs $100M pre-money for a napkin tech idea… when capitalism rationalizes the space, returns compress — right now it's early.
How to teach & test kids in an AI age— 1636 did a serious deep dive on this… but to me the big hitters to talk about are (1) you need to decompose what you actually want kids to learn… you can’t just say ‘writing teaches critical thinking’ — if critical thinking is the goal.. how do you teach that really? Then when you test it, be smart about what you re testing — yes to stand and deliver & blue-books in big exam rooms. Do by self — I started saying this years ago, the future of exams is the past … if you want to demonstrate ability to do big projects, / high quality work at home — give everyone an even token budget and time budget … finite, and then expect spectacular work vs. what you would have expected in the era of card catalogues. It will all be fine.
The Big Unsolved Real Problem is What Do You Do with the Top 30% …Tech super-empowers the top 1% of thinkers, the top 1% of creators & 1% driven people … they will be fine… the bottom 50% will live better consumption wise, but have been f-ed for a long time in a lot of ways (maybe sort of, I am actually not that cynical but follow me) — who do you really have to worry about? The top 30% … those smart enough, those driven enough - but not spectacular who are out of ‘retail’ ways to feel dealt in and are going to lose the most… and the bad news? Those are the people really that lead revolutions — the classmates of the 1% … unless maybe they all just really like TikTok and folding iPhones… I guess, and the robots make capital so valuable / pure capital that labor is just kinda sad and playing weird political games with itself of low consequence (which is not what democracy is showing us btw)
BOT LIFE
This week in bot life was about bootcamps… we did another creator ‘learn to really use AI’ session in NYC, and I had the privilege of doing a many hour session for a ton of institutional LPs and CIOs and coach them through terminal, setting up Claude code running in the cloud for them, wiring in email, Cloudflare, digital ocean, and real deployments… this was fun and gratifying… it is fun to see people unlock what is really going on and how to REALLY use AI (not ask chat GPT random questions)… random low memory Q&A chatbot is 101… Grokbot and Muse… the ‘I give you a computer’ is 201… but who wants A computer… I want a flexible cloud of as many computers as I want… databases, etc. and then you just feed it API keys til it does your TRUE bidding…. That is the future.... Some stuff I made this week
Some films about cities like New York the History of New York,San Fran,… short clips to make points and cloned mees to make pitches (actually this was a two shot because one of the APIs / video services was really mad about a pitch promoting gambling… Alignment!) — these are all one-shot-to output… which is wild… if my bot can do it this well now… in a few months oy vey.
Some ASTRA + BLENDER Madness— I heard it was good… I told claude to deploy onto a Mac It controls blender and astra and use it to make some models starting from photos and walking through them (videos) — other than an argument it had with me when Claude noticed Jessica in one of the photo scenes with me and demanded a email-to-it consent to generate a model of her (which I caught it on because my architecture wouldn’t let it process an email from her to that specific route!) … worked pretty well.
A bunch more lessin boy YouTube videos on topics of the week the kids wanted to know about (this is fast becoming a favorite of mine.
Also g-d bless having claude fully control TestFlight and iOS deployments for me… so I can ask it to build and deploy an app (and then update them) full cycle is a game changer… super cool way to do it? Just start by deploying a bunch of basically blank canvas apps (because apple makes you fill out one form)… then later fully update them over the air to what you want them to do and it can be basically anything… that is how I turned my ‘hello world’ app into a killer whole system monitoring suite … we live in the future.
Shana Tova & Happy 5887,
Sam
P.S. absolute line of the week… “Nobody funding squirrels and lemurs right now.”… - a smart friend.
P.P.S Stavros Halkias is very funny (found via Tires TV show)… and this SAS British special forces show is chef kiss
P.P.P.S What a week of celebrating Venmo... and music -- Nappy Roots & The Immigrant Groove Tuesday, to Igor Levit & Eitan Katz wednesday -- coast to coast in every sense of the words