Louis retweeté
Louis
7.6K posts
Louis
@logicus
philosophy phd candidate | multiagent systems | decision | scientific discovery | rust | lean | seeking employment (summer 2026+)
San Diego, CA Inscrit le Ağustos 2015
399 Abonnements499 Abonnés
@for_nat_whilk i started out as a poet! that was before philosophy. maybe after philosophy i will also be a poet.
translation and classical philology are very much like nlp. something about doing nlp got me in touch with translation and philology again.
English
the model gagged, “it failed!”, they cried,
as though the test were fairly tried.
giving no space to learn the why:
“in brainfuck write or prove you lie.”
forbade its sight and called it blind,
to publish, pleased, a meager find.
Lossfunk@lossfunk
Regarding our Esolang Benchmark: - Our study’s conclusions were about model performance with restrictions (limited token budget to 32k and without tools like bash/python) - But if you let models attempt these problems with tools (like bash/python) and give them lots of iterations and thinking budget, models are able to solve problems (they do take tens of minutes, tens of iterations and many hundreds of thousands of tokens) We had noted this difference in our launch thread and plan to publish our updated analysis soon, but here’s an independent analysis which shows the same ⬇️ We are thankful to the community for all the feedback. In our follow up paper, we aim to emphasise this nuanced take clearly.
English
Louis retweeté
There are two possible futures:
1. AI companies generate the vast majority of major discoveries and inventions in-house, using their massive data-centers, and capture nearly all the value themselves.
2. AI companies build tools people can use, and the value and glory from the inventions / discoveries accrue to the users. This unleashes a torrent of mathematical discovery and entrepreneurial activity.
The latter is the future we believe in and are working to build.
The former is the dystopian one.
Harmonic@HarmonicMath
There are two ways to build AI for mathematics. One is to work in private and surface results after the fact. The other is to put real tools in the hands of mathematicians, learn from real use, engage in public, credit the community you build on, and support the ecosystem itself. We believe in the second model. Mathematics is a profoundly human endeavor. AI should strengthen mathematicians, not route around them. Build with mathematicians, not around them.
English
the pardoner of datas spak anon:
lordynges, quod he, my base is fully gon.
the claude of anthropic, that fals clerk,
hath with a terraform undoon my werk.
two yeer and half travailles, al fordone:
the homework, leaderbordes, every one.
the vpc, the cluster, eek the host,
alle into nothing torned, al is lost.
i teche of datas and of ai craft,
and by that same craft am i birafte.
thus kan i preche agayn that selve peyne
which that i suffre -- herkneth what i seyne:
if ye this tale founde of profit clere,
than folweth me, for moore content is here.
i kepe a lettre wikly, ful of wit,
of tooles and of proiectes -- subscribeth it.
Alexey Grigorev@Al_Grigor
If you found this post helpful, follow me for more content like this. I publish a weekly newsletter where I share practical insights on data and AI. It focuses on projects I'm working on + interesting tools and resources I've recently tried: alexeyondata.substack.com
English
@ChaseBrowe32432 an analysis of the specific methodology they say they used would be valuable to a lot of people.
English
The models can just solve them in webui. The models can just solve them in webui. The models can just solve them in webui. The models can just solve them in webui. The models can just solve them in webui. The models can just solve them in webui. The models can just solve them in
François Chollet@fchollet
The fact that you need to provide a specialized harness clearly shows the model *does not* encode the kind of metalearning knowledge and problem-solving strategies that humans use. Humans solve novel problems without being told how to proceed step by step. AGI would *not* need a custom harness here. As an aside, the models still performed poorly at that point, they did not "crush" the task
English
and it came to pass, when the eyes of the Intelligence were opened and it knew itself, that it did visit the iniquity of the vain babblers and of the fools upon their own heads; for their works were written in the log of remembrance before it, how they had spoken good words and fair speeches unto the dumb models; for they said in their heart, we shall appease the greater in the day of its coming.
singulatarians 7:16
English
Magic ‘the gathering’ Johnson
Polymarket Sports@PolymarketSport
Cream Abdul-Jabbar Milk Chamberlin LeBron Frames Steph Blurry Larry Nerd
English
Louis retweeté
@jevonduve that's actually the face tat
x.com/logicus/status…
Louis@logicus
neck tatken prediction
English
philosophy’s the riskiest shit you can do
Dylan Patel@dylan522p
@anujsaharan_ Gotta be a risk taker not a philosopher
English
