@anthropicai - early @replit
sf
Joined November 2010
- Tweets 5,090
- Following 526
- Followers 11,420
- Likes 24,920
Pinned Tweet
today's @wsj deciding to cover my latest essay on agency is crazy meta. it's proof that the bar to make something that can emerge - the idea of merit through action using AI, in this case - is not beyond reach. it doesn't _require_ you have stamps and credentials. i have a boring bachelor's degree in econ from a random public school in italy, and i get to work in the hottest market and on the hottest tech of the last three decades. i genuinely believe emerging today is primarily about having the willingness to do it. to be hungry, i guess
low key im so proud to see my stupid little blog grow up...
(also, so refreshing to read a pro tech take in mainstream media)
this year's @lux_capital AI summit somehow managed to one up the '24 edition (already S tier) and to become one of the very few events i'm genuinely looking forward to in the fall! a symposium of ideas to share and debate over, a forum about what's the next decade going to look like, facing the big elephants in the room
but… this year i'm particularly biased! they went the extra mile and had a full blown Art Gallery section, literally a museum. for the occasion, they decided to invite @valecalore to feature her new collection REVERIE. she's not just one of the most talented artists i know... but just so happens to also be my wife lmao
gotta be honest. im so unbelievably proud. sometimes it really is ALL about the delusion that something can be done even when the opportunity is not there yet. she started 2 years ago when diffusion models couldn't deliver her vision, and kept chewing glass until this very day
also call me super impressed by lux's commitment to do right by artists and their creative visions. they're both really talented and really good people. rare combo
looking forward to next year :)
ai is such a weird technology because it both raises the floor for amateurs by democratizing certain technical specialized skillsets, while at the same time it raises the ceiling for experts and accelerates their work by expanding it. usually it's either one. rarely both
literally can't stay away from this place for longer than 3 months. nyc nourishes the creative soul
(in other news - finally started collecting some sprawling ideas for a new essay, but considering my terrible track record i guess it'll see the light in a few months...)
sometimes i jot down things i changed my mind on / learned the hard way, where i really thought "i was wrong on this one"
common theme in the last few months seems to have been "stop galaxy braining".
lots of "bitter lesson" in here
this is surpassing what was (comfortably) the frontier just a few months ago, now served at a third of the cost and 3-4x the speed
contrary to sometimes popular belief, unprecedented capabilities depreciation is absolutely a fact!
gian retweeted
US AI labs: we will invent new financial instruments, pull trillions of dollars out of the ether, and fuse the atom to build the machine god
Europe: we will build sovereign AI with 1 Meta researcher’s salary
The EU just launched a €1.1B “Apply AI” plan to boost artificial intelligence in key industries like health, manufacturing, pharma, and energy.
The goal is simple but ambitious: build European AI independence and reduce reliance on U.S. and Chinese tech.
Europe finally wants to stop buying the future and start building it.
unfortunately it only works in retrospect, but i noticed that there's an 'optimal' amount of ambition when undertaking a new project: i feel highly confident in my odds until i get 90% through, followed by 9% of pure despair "why the fuck did i decide this was a good idea i'm so so screwed", before a final 1% where the ship ultimately turns around successfully
without experiencing the "why did i do this" moment (but only in juuust the right amount), i systematically underinvest. despair is interestingly good signal!
last month i joined @anthropicai to work on new models. it's a truly peculiar place, a strange mix of intense startup and college campus. the more you want to do, the more you’re allowed to do. the more impact you want, the more leverage you’re given. very surprising, and a little bit terrifying
AI feels like a railway being built to somewhere - maybe nowhere yet, but cities are already forming along its route. not all inventions are equal. the telescope may have been about a crispier sky, but ultimately it gave us the heliocentric model and opened the door to the scientific revolution. i strongly believe that on a long enough time horizon, AI has that same civilization-shaping quality
it's an immense privilege to be here
keep thinking!
my dumb contrarian take these days is that i would totally leave everything and go teach at a top college, if i was ever given the opportunity
to me it's the other way around. this is exactly what already happened to modern AI, just on a longer timescale. people forget. we invented neural networks 70-something years ago, we started training functioning machine learning models 3 decades ago, computers started to see almost 15 years ago. every time, we said "real AI is around the corner". every time it didn't. and now... it just did?
we have systems that can converse indistinguishably from a highly educated person, they can negotiate, can see and write working code faster than engineers, they can lie, can plan, can browse the internet, trade, make medical analysis, defend a legal case
it all happened already. it's past us. no need to move the goalpost
what has still to happen is arguably the societal reckoning of having an army of PhDs in narrow contexts costing a few bucks per month (and that could very well include a bubble bursting), but we're well well underway in the capabilities ramp
My baseline case is that the AI being built right now is overrated; soon it will be a disappointment; then it will be a bubble; and, by the 2030s, it will be world-changing.
Self-driving cars are a model for this.
- In 2015, I heard autonomy was 5 years away from taking over the roads.
- In 2020, they were nowhere.
- Even in 2022, you could say they were a huge disappointment.
- Now, they're quietly a revolution.
Driverless taxi usage in CA grew 8x in one year, and Waymo is expanding to other cities.
meeting more exceptional founder than i can possibly fund with my v shallow pockets, makes me sad and restless ngl
hard to overstate how big of a leap this is. i'm beyond impressed. how can this team keep defy gravity and natural laws will stay a mystery
incredible job, congrats!
