tech and capitalism's last boss
New York, NY
Joined December 2011
- Tweets1.4K
- Following1.1K
- Followers256
- Likes8.6K
we’re just gonna live in a world where SV perpetually rediscovers financing structures, ads, economic cycles, capex implosions etc aren’t we?
but hey if it doesn’t work out meta could just put desks in the data center and transform into an office right?
Susan Li, Meta’s cfo, pulled Zuckerberg aside at an exec offsite in hawaii late last year. she wanted to make sure he understood that if this model didn't work, it would be very, very bad in financial terms. Fun fact is nobody had seen the model yet
and the bet was already poured in concrete. Prometheus in ohio is past a gigawatt and hyperion in louisiana scales to 5, and both were planned in 2023 or 2024, before the new lab even existed. @alexandr_wang called it the opposite of lean startup, you double down before you know anything
zuckerberg's answer to the "very bad" scenario is the most interesting risk take i've heard from a hyperscaler this year.
For it not to work out, "it's not just that we have to fail. AI overall has to fail." if meta's own model flopped, plenty of other people would want that compute and pay more than it costs to run, and meta still had other internal uses for it anyway
so the hedge was never muse. it was the GPUs.
Meta committed to sites with power before the product existed, knowing the floor was the open market for compute. you don't call idle capacity a backstop unless you expect a line of buyers, which is a pretty big tell about where they think rental prices are going
I find it fucking hilarious that Silicon Valley just re-discovered ads through this hector guy from Wispr.
The easiest way to build a happy marriage is to marry someone else’s wife. Get the run down on the problems, automate them away from the beginning. Live happily thereafter.
The easiest way to become a millionaire in 2027 is buying someone else's failed startup.
~$2M ARR. Founder burned out. Investors want out. You can buy it for less than the lifetime value of its future cash flow.
Cut inefficiency. Automate ops.
Rinse -> Repeat -> Compound
The Employed Philosopher retweeted
The easiest way to become a millionaire in 2027 is buying someone else's failed startup.
~$2M ARR. Founder burned out. Investors want out. You can buy it for less than the lifetime value of its future cash flow.
Cut inefficiency. Automate ops.
Rinse -> Repeat -> Compound
“hey listen if I can’t get my Boeing 737 to fly I can just live in it okay?!”
Yes, agree with Susan Li's view. There is no soft landing when it comes to this level of capital investment. If it works, great. If not, the idea that they'll be able to sell excess compute as a backup plan is not realistic because that means everyone else has the same excess compute. That would be a painful, industry wide glut.
many are saying…
This interview was the most bearish thing I’ve seen on AI lately and it had everything to do with Zuck laughing nervously on $200b saying it better work …
google? the open source AI company?
What if Spotlight actually understood the meaning inside your files?
DigUp is an open-source macOS menu bar app powered by EmbeddingGemma 2 that brings 100% local, multimodal semantic search to your Mac.
Instead of searching for file names, you can use natural language to:
📸 Find specific scenes in your photos
⏳ Jump to the exact timestamp in audio or video
💻 Search through your codebase by intent
Everything runs locally in a single vector space.
watched the zuck + alexander wang interview. makes me bearish on meta again..very reactionary and devoid of vision. frankly oculus was a much more coherent story than whatever the fuck they think they're doing now
1/ poking fun at other AI experiences being built by "nerds that have no idea what middle america looks like" then talking about muse spark and muse enchant...just shows they have no clue or taste.
2/ meta has more media data than anyone yet they're getting their assess kicked by bytedance. i dont recall bytedance spending billions on researchers either. culture problem above any technical or organizational challenges.
3/ getting the "data guy" to run your AI org then being surprised why the models trail competitors is just banal. we've long established that at the frontier data is a seat at the table not the strategy.
4/ their use cases of "finding obscure teas" are even more pathetic than "booking reservations". you're fucking meta! discovery is your business. you dont need 3 levels of indirection and a chat thread to get products surfaced to you. has nothing to do with LLMs. thats just misunderstanding the problem and the tool you need to solve it.
5/ the whole idea of justifying capex spend because "if this AI thing doesnt work out we have demand in other places is dogshit". thats like saying "if my boeing 737 doesnt fly, ill use it as a trailer to live in".
overall seems like they're clueless and the only saving grace so far has been distribution not real product vision.
ok but like…do you know the structure ? it’s not a vanilla 2/20 fund we’re talking about here. it’s more similar to venture than you think (or assume).
there are a lot of bad things to be said about pedigree becoming so powerful in SV.
but how else would we have found out that there are no moats?
the subtlety of elevenlabs’ moat here would have made any prior analysis practically useless. our whole concept of moats and what the conditions that they apply in is overly simplistic and outright dangerous for startups to consider prematurely.
Ben Horowitz explains how ElevenLabs survived the big AI labs.
When a16z first invested, ElevenLabs looked like the company most likely to be crushed. Of course the big labs would make their models talk. They did, and ElevenLabs kept growing. It was valued at $22 billion last week.
This is how Ben says it happened:
1. It got into the developer stack by building the best product for developers, which gave it a built-in set of customers.
2. That put it in the data feedback loop: every real voice use case trains its models on proprietary, real-world data.
3. It built the subtleties: speaking Spanish with the exact intonation of English, laughing in the right place.
4. Those details win the big media deals, which bring more feedback and better data.
You can throw more GPUs at your voice model, Ben says, and still not match a model tuned to the use cases that have value.
Full breakdown here: theaiopportunities.com/p/ben…
it’s hilariously pathetic how hyperbole has become some kind of status signal in SV. you can just say things.
Anthropic's Sholto Douglas says AI capex could rise from $1 trillion this year to $4 trillion in 2028 and begin doubling humanity’s GDP by the early 2030s
“Over the course of the last four or five years, we’ve been 2x-ing or 3x-ing the amount of compute capacity devoted to AI every year."
"And so I think that’s roughly, together, the hyperscalers are spending about a trillion dollars this year on capex related to AI. A very interesting question will be, will that trend line hold, and so will it go to $2 trillion next year and then $4 trillion in 2028?"
"And if that trend line broadly holds, and you can expand that to encompass the broader robotics industry, then that means that in the early 2030s, you start to get to the point where you’re actually effectively doubling the GDP of humanity, which, again, is a little bit of a ridiculous concept, but requires a lot of things to go right.”
man you have to admire Dario for sticking to his guns. the whole industry is chastising him and the guy doesn’t budge. whatever you think of his posture you gotta give it to him.
The Employed Philosopher retweeted
I have my Grokbot SUCKING OFF my Muse agent while my OpenClaw assistant WATCHES and JACKS OFF in a CORNER and my PRODUCTIVITY has SKYROCKETED
it should be no surprise to anyone that a founder in a consensus category, whose entire existence was forged by the VC industrial complex, doesn’t have a coherent, compelling vision for the product they are building.
this is akin to asking an actor in Hollywood what made their movie a blockbuster.
at the same time this doesn’t diminish his huge tactical prowess, without which he wouldn’t be where he is today.
there are loads of founders out there that have a better articulation for why instinct needs to exist, but are missing the ability to deliver.
SV in general fetishizes strategy, but unsurprisingly, it is mostly a myth constructed with the benefit of hindsight. the value creation happens in the trenches, one step at a time, mostly through actions that many in SV would deem “below them”, such as running scrappy growth hacks.
Silicon Valley is no different than Hollywood. founders are mostly packaged up and cast by VCs to create categories that are by and large consensus.
if everyone says an idea is non-contrarian then it must be…contrarian?
Almost every founder has to survive some version of what Gabe describes here
Although Rogo has raised over $300M, 40 investors passed on their Series A (except @rabois)
Gabe on why every pass at that stage is a pass on the founder:
"I met with Sequoia, Kleiner, Benchmark, everybody, and 40 people passed.
And it wasn't just, you got the email with the deck and it's not exciting.
It was, 'Oh, this is interesting. Let me meet Gabe. Oh, I kinda like Gabe. Let me spend an hour with him. Oh, Gabe, come to IC. Oh, Gabe, let's go to dinner. Oh, Gabe, come in for the weekend.'
You know what? We're gonna pass.
And it's so personal, because at that stage it has nothing to do with anything but you.
Keith Rabois came a month after everyone else had rejected us, and Keith was like, 'Gabe, this isn't a contrarian bet. It's basically just Harvey for finance. Why would I do it?'
And I said, 'Keith, if it's not contrarian, why did every single one of your friends just say it was a bad idea and not believe in me?'"