Didier Girard
@didiergirard.bsky.social
Reposted by Didier Girard
Fascinating happenings in the world of automated coding software and chess board moves from some of the biggest AI companies, as analyzed by @runtime.news.
Also relevant, as Tom points out here, the shift to use of "terminal" interfaces for software development. (techcrunch.com/2025/07/15/a...)
Also relevant, as Tom points out here, the shift to use of "terminal" interfaces for software development. (techcrunch.com/2025/07/15/a...)
The huge stakes behind AI-driven coding
Today: several wild days prove why AI-driven coding is at the center of enterprise tech, Nvidia will once again be allowed to sell chips designed around export controls to Chinese customers, and the l...
www.runtime.news
July 16, 2025 at 3:23 PM
Fascinating happenings in the world of automated coding software and chess board moves from some of the biggest AI companies, as analyzed by @runtime.news.
Also relevant, as Tom points out here, the shift to use of "terminal" interfaces for software development. (techcrunch.com/2025/07/15/a...)
Also relevant, as Tom points out here, the shift to use of "terminal" interfaces for software development. (techcrunch.com/2025/07/15/a...)
Le futur sera conversationnel : notre site web repensé en 40h grâce à l'IA
Face à la refonte de notre site web, je me suis posé la question : comment produire un site institutionnel à l'ère de l'IA ?
Face à la refonte de notre site web, je me suis posé la question : comment produire un site institutionnel à l'ère de l'IA ?
July 15, 2025 at 8:03 AM
Le futur sera conversationnel : notre site web repensé en 40h grâce à l'IA
Face à la refonte de notre site web, je me suis posé la question : comment produire un site institutionnel à l'ère de l'IA ?
Face à la refonte de notre site web, je me suis posé la question : comment produire un site institutionnel à l'ère de l'IA ?
Reposted by Didier Girard
J'ai eu beaucoup de plaisir à faire ce duo avec @didiergirard.bsky.social à @devoxx.fr pour vous faire découvrir ce qui se passe sous le capot des #LLMs, mieux comprendre comment ils fonctionnent, là où ils sont bluffants mais également là où ils pèchent !
May 26, 2025 at 11:09 AM
J'ai eu beaucoup de plaisir à faire ce duo avec @didiergirard.bsky.social à @devoxx.fr pour vous faire découvrir ce qui se passe sous le capot des #LLMs, mieux comprendre comment ils fonctionnent, là où ils sont bluffants mais également là où ils pèchent !
Reposted by Didier Girard
À @devoxx.fr , @didiergirard.bsky.social interviewe José Paumard, Java DevRel : 20 min d’échanges passionnés sur Java, leur parcours commun et l’avenir de la tech. Une rencontre entre deux légendes 👉 youtu.be/ka9HO0Y765c #Java #DevoxxFrance #SFEIR #Sfeirdev
Interview José Paumard, Java DevRel - Devoxx France
YouTube video by SFEIR TV
youtu.be
May 27, 2025 at 8:08 AM
À @devoxx.fr , @didiergirard.bsky.social interviewe José Paumard, Java DevRel : 20 min d’échanges passionnés sur Java, leur parcours commun et l’avenir de la tech. Une rencontre entre deux légendes 👉 youtu.be/ka9HO0Y765c #Java #DevoxxFrance #SFEIR #Sfeirdev
Reposted by Didier Girard
A few pics from my installation at L'expo des 150 last month. Curated by the legendary Cabaret du Chat Noir. The expo brought together over 150 urban artist. #parisstreetart #streetart #cabaretduchatnoir #chatnoir #paris
May 11, 2025 at 4:47 PM
A few pics from my installation at L'expo des 150 last month. Curated by the legendary Cabaret du Chat Noir. The expo brought together over 150 urban artist. #parisstreetart #streetart #cabaretduchatnoir #chatnoir #paris
Reposted by Didier Girard
"Cages" 50x50cm currently on view at Le Lavo/matik, Paris.
April 30, 2025 at 3:30 PM
"Cages" 50x50cm currently on view at Le Lavo/matik, Paris.
Ses illustrations son top
P.S you may enjoy my illustrated calendars for planning your 2025 ⤵︎
quotedvisually.co.uk/collections/...
quotedvisually.co.uk/collections/...
April 25, 2025 at 5:18 AM
Ses illustrations son top
Reposted by Didier Girard
Barack Obama joins Bluesky
Barack Obama joins Bluesky
Social network Bluesky, an alternative to X built on open source technologies, has scored a big win in terms of attracting notable users to its platform. The company on Sunday confirmed that former president Barack Obama has joined its service. Bluesky OOO Rose Wang replied to a post where someone wondered if the account posting […]
tcrn.ch
March 23, 2025 at 3:22 PM
Barack Obama joins Bluesky
Reposted by Didier Girard
For 25% of the most recent batch of startups at YC, 95% of the code was AI generated 😬
March 6, 2025 at 4:27 PM
For 25% of the most recent batch of startups at YC, 95% of the code was AI generated 😬
Reposted by Didier Girard
【 Build with Gemini / Developers Summit 】
💥 REX, Archi, SRE et IA ! 💥
@didiergirard.bsky.social et Aurélien Pelletier de @sfeir.bsky.social partageront leur retour d'expérience sur la construction d'une solution #GenAI
cloudonair.withgoogle.com/events/build...
💥 REX, Archi, SRE et IA ! 💥
@didiergirard.bsky.social et Aurélien Pelletier de @sfeir.bsky.social partageront leur retour d'expérience sur la construction d'une solution #GenAI
cloudonair.withgoogle.com/events/build...
Build with Gemini
Build with Gemini
cloudonair.withgoogle.com
March 5, 2025 at 11:04 AM
【 Build with Gemini / Developers Summit 】
💥 REX, Archi, SRE et IA ! 💥
@didiergirard.bsky.social et Aurélien Pelletier de @sfeir.bsky.social partageront leur retour d'expérience sur la construction d'une solution #GenAI
cloudonair.withgoogle.com/events/build...
💥 REX, Archi, SRE et IA ! 💥
@didiergirard.bsky.social et Aurélien Pelletier de @sfeir.bsky.social partageront leur retour d'expérience sur la construction d'une solution #GenAI
cloudonair.withgoogle.com/events/build...
Reposted by Didier Girard
Weekend long read on how Hard Fork’s Kevin Roose has oversold new AI tools coding without knowing how to code, as a case study in a larger phenomenon of leading journalists hyping AI:
open.substack.com/pub/garymarc...
open.substack.com/pub/garymarc...
Decoding (and debunking) Hard Fork’s Kevin Roose
His latest New York Times piece tells us a lot about what he doesn’t really understand
open.substack.com
March 2, 2025 at 8:10 PM
Weekend long read on how Hard Fork’s Kevin Roose has oversold new AI tools coding without knowing how to code, as a case study in a larger phenomenon of leading journalists hyping AI:
open.substack.com/pub/garymarc...
open.substack.com/pub/garymarc...
Reposted by Didier Girard
Notre première étude Agoratlas !
Où on a trouvé des milliers de bots/trolls insultants Zelensky, et félicitant Trump.
La première partie de l'étude étant dans le lien cité, je vais profiter de mon thread perso pour vous parler de ma journée et du process derrière 😅
Où on a trouvé des milliers de bots/trolls insultants Zelensky, et félicitant Trump.
La première partie de l'étude étant dans le lien cité, je vais profiter de mon thread perso pour vous parler de ma journée et du process derrière 😅
THREAD - Dans les dernières 24 heures, près de 35 000 comptes francophones ont parlé de la rencontre Trump/Zelensky sur Twitter.
Nos analyses ont permis de mettre en évidence de nombreux comportements coordonnés visant à manipuler l'opinion publique en faveur de la Russie ⤵️
Nos analyses ont permis de mettre en évidence de nombreux comportements coordonnés visant à manipuler l'opinion publique en faveur de la Russie ⤵️
March 1, 2025 at 7:17 PM
Notre première étude Agoratlas !
Où on a trouvé des milliers de bots/trolls insultants Zelensky, et félicitant Trump.
La première partie de l'étude étant dans le lien cité, je vais profiter de mon thread perso pour vous parler de ma journée et du process derrière 😅
Où on a trouvé des milliers de bots/trolls insultants Zelensky, et félicitant Trump.
La première partie de l'étude étant dans le lien cité, je vais profiter de mon thread perso pour vous parler de ma journée et du process derrière 😅
Reposted by Didier Girard
Le César 2025 de la révélation féminine revient à Maïwene Barthèlemy pour “Vingt Dieux”
February 28, 2025 at 9:00 PM
Le César 2025 de la révélation féminine revient à Maïwene Barthèlemy pour “Vingt Dieux”
Reposted by Didier Girard
"Vingt Dieux", de Louise Courvoisier obtient le César 2025 du meilleur premier film
www.telerama.fr/cinema/vingt...
www.telerama.fr/cinema/vingt...
February 28, 2025 at 8:39 PM
"Vingt Dieux", de Louise Courvoisier obtient le César 2025 du meilleur premier film
www.telerama.fr/cinema/vingt...
www.telerama.fr/cinema/vingt...
Reposted by Didier Girard
agent ≈ model + tools, within a for-loop + environment
February 26, 2025 at 2:05 AM
agent ≈ model + tools, within a for-loop + environment
Reposted by Didier Girard
Three awesome pieces of Musk artwork (all non-US) I have seen floating around in the last few days. 0 to 1939 poster (EveryoneHatesElon from London UK), Musk puppet poster is (Yescka from Oaxacan Mexico), Tesla Tiananmen poster is (ZezVaz from Porto Portugal.
February 24, 2025 at 5:05 PM
Three awesome pieces of Musk artwork (all non-US) I have seen floating around in the last few days. 0 to 1939 poster (EveryoneHatesElon from London UK), Musk puppet poster is (Yescka from Oaxacan Mexico), Tesla Tiananmen poster is (ZezVaz from Porto Portugal.
Reposted by Didier Girard
Left my 170k X followers behind indefinitely, because I just can’t take the management’s derangement syndrome any longer.
@bsky.app can we please add an edit button so I will stop missing the other place?
@bsky.app can we please add an edit button so I will stop missing the other place?
February 20, 2025 at 5:29 AM
Left my 170k X followers behind indefinitely, because I just can’t take the management’s derangement syndrome any longer.
@bsky.app can we please add an edit button so I will stop missing the other place?
@bsky.app can we please add an edit button so I will stop missing the other place?
Reposted by Didier Girard
I've been hearing a lot about different programming workflows to make full use of LLMs, but I want in-depth accounts of how it works. This blog by Harper Reed is exactly what I've been looking for.
harper.blog/2025/02/16/m...
harper.blog/2025/02/16/m...
My LLM codegen workflow atm
A detailed walkthrough of my current workflow for using LLms to build software, from brainstorming through planning and execution.
harper.blog
February 20, 2025 at 4:02 AM
I've been hearing a lot about different programming workflows to make full use of LLMs, but I want in-depth accounts of how it works. This blog by Harper Reed is exactly what I've been looking for.
harper.blog/2025/02/16/m...
harper.blog/2025/02/16/m...
Reposted by Didier Girard
Remote interviews are not going to be useful anymore, joining a long list of talent identification techniques that AI is invalidating (like take-home essays, etc.).
I don’t think institutions are adjusting fast enough.
I don’t think institutions are adjusting fast enough.
February 16, 2025 at 1:18 AM
Remote interviews are not going to be useful anymore, joining a long list of talent identification techniques that AI is invalidating (like take-home essays, etc.).
I don’t think institutions are adjusting fast enough.
I don’t think institutions are adjusting fast enough.
Reposted by Didier Girard
Reposted by Didier Girard
AI can now compete with the most elite human programmers in competitive coding
o3 won gold in the International Olympiad in Informatics beating a specialized version of o1 designed for the contest
General-purpose reasoning done by big LLMs now beats hand-crafted strategies arxiv.org/abs/2502.06807
o3 won gold in the International Olympiad in Informatics beating a specialized version of o1 designed for the contest
General-purpose reasoning done by big LLMs now beats hand-crafted strategies arxiv.org/abs/2502.06807
February 12, 2025 at 8:40 PM
AI can now compete with the most elite human programmers in competitive coding
o3 won gold in the International Olympiad in Informatics beating a specialized version of o1 designed for the contest
General-purpose reasoning done by big LLMs now beats hand-crafted strategies arxiv.org/abs/2502.06807
o3 won gold in the International Olympiad in Informatics beating a specialized version of o1 designed for the contest
General-purpose reasoning done by big LLMs now beats hand-crafted strategies arxiv.org/abs/2502.06807
Reposted by Didier Girard
Au sein de collectifs militants, dans des bibliothèques, des musées ou même des bars : depuis quelques années, les ateliers proposant cette démarche issue de l’éducation populaire connaissent un succès à la mesure de leur simplicité d’accès.
L’arpentage, une technique de lecture collective en plein essor : “Le grand intérêt, c’est que ça décomplexe”
www.telerama.fr
February 12, 2025 at 9:31 PM
Au sein de collectifs militants, dans des bibliothèques, des musées ou même des bars : depuis quelques années, les ateliers proposant cette démarche issue de l’éducation populaire connaissent un succès à la mesure de leur simplicité d’accès.
Reposted by Didier Girard
Cursor AI has 3 Modes: Chat, Composer and Agent.
Each designed for different tasks.
Here’s how to pick the right one. 🧵
Each designed for different tasks.
Here’s how to pick the right one. 🧵
February 11, 2025 at 5:00 PM
Cursor AI has 3 Modes: Chat, Composer and Agent.
Each designed for different tasks.
Here’s how to pick the right one. 🧵
Each designed for different tasks.
Here’s how to pick the right one. 🧵
Reposted by Didier Girard
There is a lot of important stuff in this new paper by Anthropic that shows how people are actually using Claude.
1) The tasks that people are asking AI to do are some of the highest-value (& often intellectually challenging)
2) Adoption is uneven, but many highly compensated fields already are high
1) The tasks that people are asking AI to do are some of the highest-value (& often intellectually challenging)
2) Adoption is uneven, but many highly compensated fields already are high
February 11, 2025 at 4:30 PM
There is a lot of important stuff in this new paper by Anthropic that shows how people are actually using Claude.
1) The tasks that people are asking AI to do are some of the highest-value (& often intellectually challenging)
2) Adoption is uneven, but many highly compensated fields already are high
1) The tasks that people are asking AI to do are some of the highest-value (& often intellectually challenging)
2) Adoption is uneven, but many highly compensated fields already are high
Reposted by Didier Girard
Not an easy chart to read, but worth spending a minute on. It tells you a few key things:
1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper
2) The cost of GPT-4 capability dropped 1,000 fold in 18 months
3) Pace of improvement is swift
1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper
2) The cost of GPT-4 capability dropped 1,000 fold in 18 months
3) Pace of improvement is swift
February 6, 2025 at 7:27 AM
Not an easy chart to read, but worth spending a minute on. It tells you a few key things:
1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper
2) The cost of GPT-4 capability dropped 1,000 fold in 18 months
3) Pace of improvement is swift
1) Gemini 2.0 Flash Thinking sets a new high in price-performance, better than DeepSeek r1 (on ELO) and cheaper
2) The cost of GPT-4 capability dropped 1,000 fold in 18 months
3) Pace of improvement is swift