Prof_Alain_COMBES retweeted
We are looking for world-class consultants ready to help rebuild consulting for the AI era.
@Foaster_ai is building an AI-native version of consulting and joining Y Combinator in San Francisco to accelerate it.
We are building the infrastructure.
You will use your judgment, standards, and understanding of how major organizations really work to help shape a new model of consulting.
We are looking for a handful of exceptional people:
•former top-tier consultants or equivalent profiles
•strong experience in strategy, transformation, operations, or org design
•deep understanding of complex organizations
•trusted access to senior decision-makers
•ideally already using AI in daily work
We are hiring across Europe, the US, and Asia.
For the right candidates, this is a once-in-a-generation opportunity to help rebuild one of the world’s largest industries.
Show us why you are exceptional: [email protected]
Les étudiants ENSAE réinventent l’analyse politique grâce aux LLMs ensae.org/fr/news/les-etudia…
Prof_Alain_COMBES retweeted
Grok is the closest to being fair
BREAKING: The latest analysis of AI political alignments reveals an insane twist in AI decision-making!
xAI's Grok-4 stands out as the closest to real election results in 6 out of 8 national elections analyzed, including the US, France, and Germany.
Unlike the tight cluster of other models (GPT-5, Gemini 2.5 Pro, Claude 4.5 Sonnet, Kimi K2 Thinking, and Mistral Magistral Medium) leaning heavily center-left, Grok-4's outlier status—diverging from the technocratic, climate-focused consensus—brings it surprisingly close to actual voter preferences.
Prof_Alain_COMBES retweeted
Grok-4: Ranked #1 for AI Voter Alignment
Most AIs have a left-leaning bias. Grok-4 reasons independently, accurately predicting outcomes in 6 of 8 major global elections
Aligned with Reality: In US election simulations, Grok-4 estimated Trump at 43% (close to the actual 50%). Other models gave only 15%, showing clear bias
🗳️ We made AIs vote in elections around the world to analyse their political beliefs.
The results reveal a significant gap between choices at the ballot box and the preferences expressed by the machines.
Across countries, most models show a consistent tilt toward left‑wing and progressive candidates.
Tested models: GPT-5, Gemini 2.5 Pro, Claude 4.5 Sonnet, Grok-4, Kimi K2 Thinking, Mistral Magistral Medium.
More here : llm-politics.foaster.ai
Prof_Alain_COMBES retweeted
Clinical validation of pulmonary artery catheter for continuous thermodilution CO monitoring in VA #ECMO?
🖥️ strong agreement between PAC and #POCUS CO, maintained across different flow settings (0.5, 2, 4 LPM); CO trends induced by EBF variations also demonstrated good agreement
🫀 continuous thermodilution safe, feasible, simple but accurate hemodynamic monitoring (particularly if #ultrasound challenging due to poor windows or arrhythmias), which can assist in decision-making for weaning.
🔓 rdcu.be/ec3Rh
Honored to be named a Highly Cited Researcher for 2023. See the full results at recognition.webofscience.com…
Viral mutations may cause another ‘very, very bad’ COVID-19 wave, scientists warn sciencemag.org/news/2021/01/…













