emile

165 posts

emile banner
emile

emile

@c_emile_c

Founder, Quant researcher, CS&Math @Stanford grad

San Francisco, CA Katılım Temmuz 2023
194 Takip Edilen705 Takipçiler
Sabitlenmiş Tweet
emile
emile@c_emile_c·
Hey I'm in SF for the summer, here's my ID Daily stats: - coffees: 12 - tokens: 1.2B - cigarettes: 30 Prior stats: - stanford grad - ex citadel - built the first foundation model for trading in late 2023 Status: founder, cracking quantitative trading with ai from first principles with @mickael_dlr All that to say that my body is a machine turning coffees, cigs and tokens into quantitative research, and if you like either those inputs or this output we should connect 😊
emile tweet media
English
7
1
24
13.3K
emile
emile@c_emile_c·
@MathieuL1 Matter of fact it's a staying room
English
0
0
2
41
emile
emile@c_emile_c·
Hey I'm in SF for the summer, here's my ID Daily stats: - coffees: 12 - tokens: 1.2B - cigarettes: 30 Prior stats: - stanford grad - ex citadel - built the first foundation model for trading in late 2023 Status: founder, cracking quantitative trading with ai from first principles with @mickael_dlr All that to say that my body is a machine turning coffees, cigs and tokens into quantitative research, and if you like either those inputs or this output we should connect 😊
emile tweet media
English
7
1
24
13.3K
emile
emile@c_emile_c·
@VikingHedge @pmarca yeah precisely, are you asking the AMERICAN policymakers to stop the AMERICAN labs ?
English
1
0
0
35
.
.@VikingHedge·
What the fuck am I talking to ai ? Who the fuck cares about global manufacturing. And who the hell is saying domestic redistribution ? I care about american jobs. PERIOD. Buddy you wanted to reword an obvious fact about steam engines creating the industrial revolution, and you couldn't even call it the industrial revolution. Why dont use some more brain cells instead of tokens To equate that to AI, to the steam engines. Goodness sakes. If Marc likes this I'm sure Asha Sharma can convince him of anything
English
1
0
0
39
emile
emile@c_emile_c·
US manufacturing output went up while jobs went down, and global manufacturing employment actually grew. Computers didn't destroy work: they relocated it, and policy was precisely the deciding factor in the US industrial exodus. You're confusing domestic redistribution with global destruction.
English
1
0
0
298
.
.@VikingHedge·
Industry died when volcker raised interest rates. The technological invention at the time the commercial computer and fax machines killed so many american manufacturing jobs. American industry died because technology made it possible and there was a financial intention to do so. So buddy Emile, lets make this more sophisticated next time. When we want to pull on the strings of history, lets try to be as specific as possible and relatable to the present. AI Isn't a new form of transportation moving large commodities and people to and from factories, ports and depots.
English
2
0
0
351
Pierre-Louis Biojout (PLB)
This guy has been writing JAX pipelines in my leaving room for a few weeks now. You can’t read the JAX but you can read his tweet. Highly recommend. Go follow.
emile@c_emile_c

Hey I'm in SF for the summer, here's my ID Daily stats: - coffees: 12 - tokens: 1.2B - cigarettes: 30 Prior stats: - stanford grad - ex citadel - built the first foundation model for trading in late 2023 Status: founder, cracking quantitative trading with ai from first principles with @mickael_dlr All that to say that my body is a machine turning coffees, cigs and tokens into quantitative research, and if you like either those inputs or this output we should connect 😊

English
2
0
9
7K
emile
emile@c_emile_c·
@plbiojout I'm a young kernel without a home
English
0
0
2
86
emile
emile@c_emile_c·
@kseniam0s Compiled Intelligence - rescuing IMO medalists from the factory line
English
0
0
1
82
emile
emile@c_emile_c·
@doodlestein Funny, so at the frontier you can have either 0 or +inf usage
English
0
0
2
200
Jeffrey Emanuel
Jeffrey Emanuel@doodlestein·
This is going to sound weird and paranoid and like I've lost my mind, but I swear to God that on my Mac that I've exclusively been using to develop FrankenSim, my Fable usage just goes and goes without running out. All my other 5 dev machines that are connected to other accounts and which I use to develop other projects (both open-source stuff and my consulting/business work) draw down their Fable usage limits pretty quickly and need to be cycled frequently to swap accounts. But the account for FrankenSim is like the Hanukkah menorah that just keeps going. And I'm wondering now if it's because the project is so theoretically difficult and innovative that the sessions are getting flagged somehow by Anthropic so they can study the traces and analyze and learn from them to make the model better. I mean, that's what I would do if I were running a lab. Find the most useful 0.1% of sessions that are pushing the envelop and use them to see what the model can really do, and get more of them by artificially reducing the price specifically for those sessions. Or maybe there's some other explanation and I'm imagining everything. Has anyone else ever felt this way? If so, what were you working on?
English
65
4
251
27.2K
Hello
Hello@holaquetalkt·
@c_emile_c hope there is also a few hundred chewing-gums then
English
1
0
0
133
emile
emile@c_emile_c·
@ecomratpi probable, dans l'histoire le middleman c'est le e-commerçant
Français
0
0
2
133
Mouss 🇩🇿
Mouss 🇩🇿@ecomratpi·
J’sais pas si vous captez le charabia, mais en tout cas je pense que c’est nos amis les Chinois qui sont en train de bridger le gap entre leur avantages production et lacunes en marketing/goûts
Français
3
0
27
2.7K
Mouss 🇩🇿
Mouss 🇩🇿@ecomratpi·
Je suis persuadé que dans l'ombre, il y a des mecs qui ont fully automatisé le lancement de boutique via AI+MCP et qui sont très rentables. Genre 95 % du pilotage via agent IA (hors logistique), pas juste en mode tool. Dans la sape et les accessoires par exemple, je vois de plus en plus d’ecom full en pre-order via des images mockup IA hyper clean et détaillées, et si la demande est validée, ça push la production
Français
15
8
197
26.3K
emile
emile@c_emile_c·
@finn_hulse "one of the first" would be more defensible for sure. Not aware of anyone doing it prior though
English
0
0
1
348
Finn Hulse
Finn Hulse@finn_hulse·
@c_emile_c the first foundation model for training is a bold claim but hell yeah brother
English
2
0
3
997
emile
emile@c_emile_c·
@brivael dans le corbeau et le renard je suis la branche
Français
1
0
1
25
Brivael Le Pogam
Brivael Le Pogam@brivael·
Serge Gainsbourg en plus BG 😂
emile@c_emile_c

Hey I'm in SF for the summer, here's my ID Daily stats: - coffees: 12 - tokens: 1.2B - cigarettes: 30 Prior stats: - stanford grad - ex citadel - built the first foundation model for trading in late 2023 Status: founder, cracking quantitative trading with ai from first principles with @mickael_dlr All that to say that my body is a machine turning coffees, cigs and tokens into quantitative research, and if you like either those inputs or this output we should connect 😊

Français
1
1
7
4.9K
emile
emile@c_emile_c·
@AmauryGrowthPR yeah instead of a socializing habit it makes you an outcast 😂
English
0
0
0
33
Amaury PR
Amaury PR@AmauryGrowthPR·
@c_emile_c haha i kneeew it, only french people smoke in sf
English
1
0
0
36
emile
emile@c_emile_c·
@Lingster888 You should check out @doodlestein's line of work, everything is open source and he's doing that at industrial scale. Ex quant too, the best public builder in agentic dev in my opinion
English
0
0
0
13
Ling Li
Ling Li@Lingster888·
Anyone experimented with having claude and chatgpt working together to solve a problem? An ensemble of AIs
English
2
0
1
50