Measuring progress toward AGI: A cognitive framework
#HackerNews #MeasuringAGI #cognitiveframework #AIprogress #DeepMind #innovation
Google DeepMind startet einen Kaggle-Wettbewerb für 200.000 US-Dollar zur Entwicklung neuer AGI-Prüfverfahren.
Bis April 2026 entstehen Evaluierungsmethoden für fünf kognitive Schwachstellen von Sprachmodellen. Das Framework trennt das Abrufen von Trainingsdaten von echter Problemlösung und zwingt die Modelle zu planvollem Handeln.
#Google #DeepMind #Kaggle #AGI #News
https://www.all-ai.de/news/news26/google-deepmind-200-000-dollar-f%C3%BCr-agi-tests
Pushmeet Kohli (@pushmeet)
Google DeepMind는 AI를 과학의 궁극적 촉매라고 소개하며, 대표 사례로 단백질 구조 예측 데이터베이스인 AlphaFold Database(AFDB)를 제시했습니다. AFDB는 전 세계에서 무료로 제공되며 330만 명이 넘는 연구자들이 활용해 왔다고 언급되어, AI 기반 생명과학 도구의 광범위한 실사용과 영향력을 강조하고 있습니다.
At @GoogleDeepMind, we believe AI is the ultimate catalyst for science. 🧬 The best example of this has been the AlphaFold database (AFDB) of protein structure predictions which has been used free of cost by more than 3.3 millions researchers across the world! Today, in
Corriere.it - Homepage: Intelligenza artificiale e biologia: AlphaFold di DeepMind verso la mappa completa dell’«interattoma»
Il database di AlphaFold si arricchisce con milioni di strutture di complessi proteici, aprendo una nuova fase nello studio delle interazioni molecolari
Artificial intelligence and biology: DeepMind’s AlphaFold on the path to the complete “interome” map.
The AlphaFold database is enriched with millions of protein complex structures, opening a new phase in the study of molecular interactions.
Il database di AlphaFold si arricchisce con milioni di strutture di complessi proteici, aprendo una nuova fase nello studio delle interazioni molecolari. La collaborazione tra EMBL-EBI, Google DeepMind, NVIDIA e Seoul National University punta a comprendere meglio malattie e sviluppare nuovi farmaci
Agenci AI i prompt injection – nowe zagrożenie w sieci ujawnione
Czy pozwoliłbyś, by obcy blog albo przypadkowy e-mail wydawał polecenia w twoim imieniu? W świecie agentów AI to się dzieje naprawdę – i na dodatek bywa niewidoczne dla człowieka.
Czytaj dalej:
https://pressmind.org/agenci-ai-i-prompt-injection-nowe-zagrozenie-w-sieci-ujawnione/
#PressMindLabs #agenciai #bezpieczenstwoai #deepmind #googlechrome #openai
DeepMind’s Genie 3 can turn a single text prompt into a playable 3D world at 720p/24fps, and SIMA 2 fills it with smart agents – this isn’t sci-fi, it’s shipping via Google Cloud.
From child prodigy to a race to be the one that puts humans in a zoo. Great look into our current trajectory. Who will wield the stick?
https://www.tribecafilm.com/films/thinking-game-2024
#DemisHassabis #AGI #AI #LLM #DeepMind #Google #Geopolitics #ProteinFolding #Science #ThinkingGame #MachineLearning #Energy
The Thinking Game chronicles the extraordinary life of visionary scientist Demis Hassabis and his relentless quest to solve the enigma of artificial general intelligence. After the Premiere Screening: A conversation with CEO and co-founder of Google DeepMind Demis Hassabis, and Director Greg Kohs, moderated by Darren Aronofsky.
Chubby (@kimmonismus)
Demis Hassabis가 세계 지도자들에게 AGI의 영향력이 산업혁명보다 10배 크고 10배 빠를 것이라고 말한 발언을 소개한 트윗입니다. 작성자는 Perplexity 컴퓨트를 사용해 이를 그래프로 시각화하고 글로벌 GDP에 미치는 영향을 보여주며 '아무도 준비되지 않았다'고 경고합니다. AGI 영향력 경고 및 시각화 사례입니다.
Yesterday I was proudly told by a senior academic now a senior researcher at #Google #Deepmind that they are hiring more philosophers and ethicists.
This is a BAD THING. Settingaside teh fact that these are insecure jobs since corporations may change their mind about this direction on a whim, and only have the thinnest veneer of #AcademicFreedom, this is a further example of the privatisation of the #PublicRealm
(1/)
Katherine Borges
Hacker News
Andreas Becker
sayzard
Italian News by RSS
PressMind Labs
TechGlimmer
CM Thiede
Tom Stoneham