AGI - AI Dynamics

Hassabis: True AGI Requires Generality Across Cognitive Tasks

By

–

24 April 2025 21h33

Demis Hassabis affirme que la véritable AGI implique une cohérence sur de nombreuses tâches cognitives, et pas seulement d’être meilleure sur quelques-unes.

« Ce qui manque, c’est la généralité, pas simplement l’intelligence. »

Autrement dit, tant que les systèmes d’IA… pic.twitter.com/8jWJAXcDi6
— VISION IA (@vision_ia) 24 avril 2025

Demis Hassabis affirme que la véritable AGI implique une cohérence sur de nombreuses tâches cognitives, et pas seulement d’être meilleure sur quelques-unes. « Ce qui manque, c’est la généralité, pas simplement l’intelligence. » Autrement dit, tant que les systèmes d’IA

→ Voir le post original sur X — @vision_ia

24 April 2025

Alpha-Factory v1: Interactive Demo and Agent Gallery

By

@montreal_ai

–

24 April 2025 18h14

Alpha‑Factory v1 Interactive Demo & Agent Gallery: https://
github.com/MontrealAI/AGI
-Alpha-Agent-v0/tree/main/alpha_factory_v1/demos
… Out‑learn | Out‑think | Out‑design | Out‑strategise | Out‑execute #AGI #AGIALPHA

→ View original post on X — @montreal_ai

24 April 2025

Model Welfare: Scientific Uncertainty and Research Approaches

By

@anthropicai

–

24 April 2025 17h00

We remain deeply uncertain about the idea of “model welfare”. There’s no scientific consensus on it—or even on how to research it. We’re approaching the topic as carefully as we can. Find out more:

→ View original post on X — @anthropicai

24 April 2025

AI Models and Subjective Experience: New Research Program

By

@anthropicai

–

24 April 2025 17h00

As AI models become more complex and more capable, is it possible that they’ll have experiences of their own? It’s an open question. We recently started a research program to investigate it.

→ View original post on X — @anthropicai

24 April 2025

Alpha-Factory: Multi-Agent AGI System for Compounding Value

By

@montreal_ai

–

24 April 2025 16h46

[ Alpha‑Factory v1 ] Multi‑Agent AGENTIC α‑AGI — An antifragile constellation of self‑improving agents orchestrated to spot live alpha across any industry and turn it into compounding value. Mission End‑to‑end: Identify → Out‑Learn → Out‑Think →

→ View original post on X — @montreal_ai

24 April 2025

Survey of AI Agent Protocols: Standards and Interoperability

By

@jiqizhixin

–

24 April 2025 7h53

A Survey of AI Agent Protocols https://
arxiv.org/pdf/2504.16736

→ View original post on X — @jiqizhixin

24 April 2025

Google Presents Project Astra Universal AI Assistant Prototype

By

@googleai

–

24 April 2025 5h02

Attending #ICLR2025? Drop by the Google Booth at 12pm SGT to hear Vahid Ghadakchi discuss "Project Astra: A research prototype exploring future capabilities of a universal AI assistant". https://
deepmind.google/technologies/p
roject-astra/
…

→ View original post on X — @googleai

24 April 2025

Model Behavior: Quantification vs Uncertainty Expression

By

@patio11

–

23 April 2025 23h25

For example, this should be basis points rather than 20%, but suggesting the model put a number on everything causes it to float to the bottom of the list, versus it impressionistically suggesting "I don't know, could be a firmware issue" in the middle of the first paragraph.

→ View original post on X — @patio11

23 April 2025

AI Model Probability Estimates: Rigor and Custom Instructions

By

@patio11

–

23 April 2025 23h23

(I don't find any of the models to be particularly rigorous in what is a 90% versus a 98% or an 85%, but they don't often say 90% when they mean 30%, and so I keep the part of my custom instructions which tells them to provide estimates.)

→ View original post on X — @patio11

23 April 2025

AI diagnosis principle: thinking horses not zebras

By

@patio11

–

23 April 2025 23h18

There are at least three things that can cause that symptom but here the AI is doing the very useful exercise "If you hear hoofbeats think horses not zebras."

→ View original post on X — @patio11

23 April 2025