talk is cheap, build @class_lambda C + rust + erlang + julia + λ. amateur in everything: distributed systems, ML, compilers, cryptography and investment.
Buenos Aires, Argentina
- Tweets 17,572
- Following 3,445
- Followers 9,286
- Likes 46,648
Federico Carrone retweeted
board anónimo solo de texto. cuatro categorías para empezar. el archivo dura un mes ¿si o no?
83
18
546
61,331
I worry that LLMs are beginning to dismantle the routines, careers, and goals that give many people’s lives structure and meaning.
People talk constantly about AI’s economic consequences and far too little about its existential and spiritual ones.
In my own case, I’m optimistic about my future. I have multiple interests, a strong network, and the ability to systematically create new companies and projects. But this isn't the case with most people.
I’m finishing an essay about this for @421Net with concrete recommendations about what you can do.
Friend of mine, math professor from one of the top five depts in the US [and Tao's coauthor]:
"I want to go away, to my little country house, and stay there. My profession has basically been destroyed by AI."
Those who were protected all your lives: feel the bitter taste of being suddenly unprotected.
8
1,067
I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.
2,032
4,639
48,342
8,300,477
Federico Carrone retweeted
It's like watching the death of fiat in a slow-motion train crash.
74
154
1,404
63,307
Federico Carrone retweeted
When I start teaching complex analysis on Monday, I've been wondering what to say to my class about why math is still worth studying in the age of AI.
Here is Francis Su's (@mathyawp) answer: studying math develops human virtues like tenacity and creativity.
With the start of the semester just a week away, here’s what I decided to say to my upper division math course about how AI will or won’t change what they’re learning…
The Enduring Value of Math, in an Age of AI francissu.substack.com/p/the…
74
119
662
54,440
Federico Carrone retweeted
AI won’t erase the gap between people.
YouTube, Coursera, and Internet made world-class knowledge free. Most people still chose TikTok.
AI is no different. Everyone gets Claude Code, Codex, and ChatGPT. The best will use them better and compound faster.
AI isn’t an equalizer. It’s a multiplier.
170
578
4,993
245,627
i try to walk 1 to 2 hours per day. it's the best biohack i've found in my life. highly recommended.
1
4
756
another way of seeing things is that this completely changes the economics of mathematics.
we have never seen industrialized mathematical production. most of our society grew thanks to the invention of new mathematical language created by less than a thousand people in a few generations. it will be hard to predict what will happen if new inventions keep appearing. we don't need AGI to change the world. we just need probabilistic LLMs and formal verification. AI will industrialize math before AGI even exists.
most of the mathematicians i know barely work with others and work mostly on proving things. now we will be able to focus on inventing new languages, in compressing reality into a new expressive language. most of mathematics is just re expressing a problem into another problem that is easier to solve. it's a hack.
the last 400 years, property rights, free exchange of goods and services but mostly mathematics enabled the automation of society thanks to calculus and algebra. if we invent new tools to re express our world into simpler problems what will humanity be able to achieve?
historically, pure mathematics improved the foundations or theory of mathematics, which then let applied mathematics do new, cool things.
right now, we're in a strange situation where gradient descent and optimization techniques are doing symbolic reasoning. llms + formal verification are letting us turn theoretical mathematics into an optimization problem.
you have an enormous space of possibilities, the llm generates some of them, formal verification verifies them, rejects or accepts them, and then you improve and search again. a full feedback loop.
for now, llms are good at finding an answer in an already existing representation. the question is whether they will be able to create or invent new representations to turn a difficult problem into an easy one. that would mean creating a new language.
the history of xix and xx century mathematics was about creating new formal structures that enabled computation and helped solve concrete problems.
in this new century, ai and computation might enable us to create new ways to express problems and solve them thanks to these new languages.
rather than pure -> applied we might be able to close the loop with:
pure -> computation -> ai -> new math discovery -> pure
1
3
589
historically, pure mathematics improved the foundations or theory of mathematics, which then let applied mathematics do new, cool things.
right now, we're in a strange situation where gradient descent and optimization techniques are doing symbolic reasoning. llms + formal verification are letting us turn theoretical mathematics into an optimization problem.
you have an enormous space of possibilities, the llm generates some of them, formal verification verifies them, rejects or accepts them, and then you improve and search again. a full feedback loop.
for now, llms are good at finding an answer in an already existing representation. the question is whether they will be able to create or invent new representations to turn a difficult problem into an easy one. that would mean creating a new language.
the history of xix and xx century mathematics was about creating new formal structures that enabled computation and helped solve concrete problems.
in this new century, ai and computation might enable us to create new ways to express problems and solve them thanks to these new languages.
rather than pure -> applied we might be able to close the loop with:
pure -> computation -> ai -> new math discovery -> pure
3
1
1,017
Federico Carrone retweeted
That this gradient descent also became pretty good in symbolic reasoning of Aristotle, Leibniz an Frege is pretty amazing. Seems after all the synthetic-analytic distinction did not hold up well -- mathematical thought done by gradient descent, synthetic or analytic.
1
1
708
Federico Carrone retweeted
Turns out Bariloche 🇦🇷 is having it’s biggest festival of the year this weekend
The Snow Festival
I booked this week completely at random 😂 I guess I was just lucky!
9
10
109
4,978
or maybe you need a longer time horizon.
it takes at least a while decade to see the results of any new technology being applied to the whole world. i’ve been implementing different automated solutions using LLMs to replace humans and adoption takes time. getting thing right takes time but we’re seeing incredible results and it does impact the work force directly. the thing is that if for us, a deep tech company, it takes a year to deploy a solution, imagine how much time it will take for a normal non tech company to do so.
If you look at the last 3 years of AI progress, technologists have largely been right about the (massive) increase in capabilities, and economists have largely been right about the economic impact. I don’t expect that to change.
Ricardian comparative advantage ftw!
1
552