The AGI Monitor
@theagimonitor.bsky.social
Monitoring global developments, news and scientific discourse on AGI. Your daily read on the news about artificial general intelligence.
#AGI #AI #Tech #Science #News
For Turkish: @bidunyabilim.bsky.social
Editor: @mucahidkose.bsky.social
#AGI #AI #Tech #Science #News
For Turkish: @bidunyabilim.bsky.social
Editor: @mucahidkose.bsky.social
Silicon Valley optimism on AGI timelines has shifted from 2060 to around 2040 (or sooner, per Altman and Clark), fueled by breakthroughs like DeepSeek and surging research investment.
April 30, 2025 at 2:06 PM
Silicon Valley optimism on AGI timelines has shifted from 2060 to around 2040 (or sooner, per Altman and Clark), fueled by breakthroughs like DeepSeek and surging research investment.
DeepMind CEO Hassabis warns that AGI and superintelligence will redefine humanity, urging a new generation of philosophers (“the next Kant or Wittgenstein”) to tackle questions of consciousness, morality and the human condition.
April 29, 2025 at 9:07 PM
DeepMind CEO Hassabis warns that AGI and superintelligence will redefine humanity, urging a new generation of philosophers (“the next Kant or Wittgenstein”) to tackle questions of consciousness, morality and the human condition.
DeepMind CEO Hassabis envisions AGI within 5–10 years, driving breakthroughs from disease cures to climate solutions, but warns of dual-use risks and the need for international safety standards.
April 29, 2025 at 8:52 PM
DeepMind CEO Hassabis envisions AGI within 5–10 years, driving breakthroughs from disease cures to climate solutions, but warns of dual-use risks and the need for international safety standards.
ARC Prize analysis: OpenAI’s o3 model scores 41-53% on ARC-AGI-1 (vs. preview’s 88%). Struggles on ARC-AGI-2 (<3% vs. human 60%). High compute costs don’t boost accuracy. Microsoft CEO Nadella slams “benchmark hacking” as AGI gap persists. #AGI #AI
OpenAI's o3 is less AGI than originally measured
A recent analysis by the ARC Prize Foundation finds that OpenAI’s o3 model delivers significantly weaker results on standardized reasoning benchmarks than its previously tested o3 preview version.
the-decoder.com
April 29, 2025 at 8:25 PM
Former OpenAI staff and leading AI experts, including Geoffrey Hinton and Stuart Russell, warn in an open letter that shifting control to a Public Benefit Corporation endangers legal safeguards (profit caps, independent oversight and the “stop-and-assist” clause).
AI pioneers warn OpenAI's corporate overhaul could betray its original mission for humanity
A group of former OpenAI employees, researchers, and nonprofit organizations is urging regulators to block OpenAI’s proposed corporate restructuring, arguing it threatens the company’s founding missio...
the-decoder.com
April 28, 2025 at 7:31 PM
Former OpenAI staff and leading AI experts, including Geoffrey Hinton and Stuart Russell, warn in an open letter that shifting control to a Public Benefit Corporation endangers legal safeguards (profit caps, independent oversight and the “stop-and-assist” clause).
DeepMind CEO Hassabis predicts AGI within 5–10 years but says today’s AI lacks consciousness and genuine imagination. He advises building intelligent tools to advance neuroscience before pursuing self-awareness.
Google DeepMind CEO predicts AGI in 5-10 years but says AI still lacks imagination
Demis Hassabis, the Nobel Prize-winning head of Google DeepMind, also predicted a breakthrough moment for humanoid robots in the “next couple of years”.
indianexpress.com
April 28, 2025 at 7:18 PM
DeepMind CEO Hassabis predicts AGI within 5–10 years but says today’s AI lacks consciousness and genuine imagination. He advises building intelligent tools to advance neuroscience before pursuing self-awareness.
Futurist Bernard Marr says five key hurdles stand in the way for true AGI: common sense and intuition; transfer learning across domains; seamless physical–digital interaction; the enormous compute and data demands; and societal trust requiring explainability and accountability.
The Difference Between ChatGPT And Generative AI
Discover the key differences between ChatGPT and generative AI. Learn how ChatGPT, fits into the broader category of generative AI and what sets it apart as a specialized tool for generating human-lik...
bernardmarr.com
April 28, 2025 at 7:11 PM
Futurist Bernard Marr says five key hurdles stand in the way for true AGI: common sense and intuition; transfer learning across domains; seamless physical–digital interaction; the enormous compute and data demands; and societal trust requiring explainability and accountability.
Blockchain-powered decentralized AGI, backed by $25 billion, aims to break big-tech monopolies by enabling transparent, secure AI development on distributed networks.
Blockchain Fuels $25B Decentralized AGI Revolution, Protecting Innovation From Big Tech - Business Sandesh
The world of artificial general intelligence (AGI) is on the cusp of a revolution, driven by the inn
businesssandesh.in
April 28, 2025 at 6:50 PM
Blockchain-powered decentralized AGI, backed by $25 billion, aims to break big-tech monopolies by enabling transparent, secure AI development on distributed networks.
Benchmarks like MMLU / ARC-AGI approximate AI “intelligence” via multiple-choice tests but miss real-world robustness. The new GAIA benchmark (466 multi-step, tool-using questions across three difficulty levels) evaluates web browsing, multimodal understanding, code execution and complex reasoning.
Beyond ARC-AGI: GAIA and the search for a real intelligence benchmark
Guest Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Intelligence is pervasive, yet its measurement seems subjective. At...
venturebeat.com
April 28, 2025 at 5:11 PM
Benchmarks like MMLU / ARC-AGI approximate AI “intelligence” via multiple-choice tests but miss real-world robustness. The new GAIA benchmark (466 multi-step, tool-using questions across three difficulty levels) evaluates web browsing, multimodal understanding, code execution and complex reasoning.
Reid Hoffman: AI will be vastly more powerful and integrated by 2028—even if true AGI remains elusive. Multimodal models today combine text, vision and audio; memory has grown from a few pages to millions of tokens, enabling book-length drafts or codebases in single prompts.
Opinion | Reid Hoffman’s surprising take on the future of artificial intelligence
Reid Hoffman argues that artificial intelligence will do something smarter than world domination — it will become an essential part of everyday life.
sfstandard.com
April 28, 2025 at 4:58 PM
Reid Hoffman: AI will be vastly more powerful and integrated by 2028—even if true AGI remains elusive. Multimodal models today combine text, vision and audio; memory has grown from a few pages to millions of tokens, enabling book-length drafts or codebases in single prompts.
HTF Market Intelligence projects the Global AGI market to grow at a 45% CAGR from 2025–2030, driven by tech-giant and government investments, rising compute power and data, and expanding applications in healthcare, finance and more.
Artificial General Intelligence (AGI) Market to See Revolutionary Growth | Microsoft, IBM
According to HTF Market Intelligence the Global Artificial General Intelligence AGI market to witness a CAGR of 45 during the forecast period 2025 2030 The Latest Released Artificial General Intellige...
www.openpr.com
April 28, 2025 at 4:45 PM
HTF Market Intelligence projects the Global AGI market to grow at a 45% CAGR from 2025–2030, driven by tech-giant and government investments, rising compute power and data, and expanding applications in healthcare, finance and more.
Google DeepMind’s recent paper outlines AGI risks: misuse, misalignment, mistakes, structural. Proposes safety strategies (strict access, system-level mitigations). CEO Demis Hassabis urges multidisciplinary approach (philosophers, economists) to address societal impact as AGI nears. #AGI #AI
Google Lists The Harm AI Can Cause, And How To Deal With It - Trak.in - Indian Business of Tech, Mobile & Startups
Concerns surrounding artificial intelligence, particularly artificial general intelligence (AGI), have been voiced by experts for years, with fears about its potential to cause serious harm to humanit...
trak.in
April 28, 2025 at 4:37 PM
EY analysis: AGI’s rise will shift enterprises from process-driven to outcome-focused strategies, requiring strategic AI-human collaboration. New roles (AI governance, ethics) and EU AI Act-style regulation critical.
Beyond automation: how AGI will reshape decision-making, innovation, and governance
Agentic AI and AGI - how they will change the way we make decisions
www.techradar.com
April 28, 2025 at 4:29 PM
EY analysis: AGI’s rise will shift enterprises from process-driven to outcome-focused strategies, requiring strategic AI-human collaboration. New roles (AI governance, ethics) and EU AI Act-style regulation critical.
New ARC-AGI-2 benchmark by François Chollet challenges AI’s true intelligence: Top models (OpenAI o1-pro, DeepSeek R1) score only 1-1.3%, vs. human 60%. Test prioritizes efficiency ($0.42/task limit) to curb brute-force computation. Arc Prize 2025 aims for 85% accuracy. #AGI #AI
“Nothing Could Solve It”: Brutal New AGI Test Exposes Alarming Limits as Top AI Models Fail Spectacularly - Rude Baguette
IN A NUTSHELL 🔍 The ARC-AGI-2 test challenges AI models to identify visual patterns and adapt to new problems. 💡 Unlike its predecessor, ARC-AGI-2 emphasizes efficiency, assessing both problem-solving...
www.rudebaguette.com
April 21, 2025 at 2:55 PM
U.S.-China tech decoupling may accelerate AGI development, driving AI race in healthcare, cybersecurity, climate science. Risks: fragmented markets, ethical gaps (privacy, bias). Experts urge global collaboration on standards despite rivalry to mitigate disruptions. #AGI #AI
The AI Race: How U.S.-China Rivalry Could Shape the Future of Artificial Intelligence
The AI Race: How U.S.-China Rivalry Could Shape the Future of Artificial Intelligence - Tumirador
salajobrazovanje.co.rs
April 21, 2025 at 2:47 PM
Google’s Sergey Brin urges full in-office work to accelerate AGI development, citing 60-hour & 5 day office workweeks as a “sweet spot” for productivity.
“We Won’t Surrender”: Sergey Brin Declares RTO Critical for Google to Dominate the AGI Race and Leave Rivals in the Dust - Rude Baguette
IN A NUTSHELL 🚀 Sergey Brin calls for Google employees to return to the office every weekday to boost productivity and win the AGI race. 💡 Google’s push for in-office work aims to enhance collaboratio...
www.rudebaguette.com
April 21, 2025 at 2:32 PM
Google’s Sergey Brin urges full in-office work to accelerate AGI development, citing 60-hour & 5 day office workweeks as a “sweet spot” for productivity.
Recently, OpenAI quietly released o3 and o4‑mini, which may showcasing AGI‑like capabilities: o3 demonstrates active visual reasoning (zooming, rotating and interpreting images to solve handwritten notes), while o4‑mini scored 99.5% on the AIME 2025 test.
OpenAI's o3 revolution: OpenAI may have inadvertently released AGI. Is this the moment AI became smarter than us?
The o3 and o4-mini models can think with images, code better than you, and locate your bistro with a photo of a sandwich. Is this artificial general intelligence - AGI?
citymagazine.si
April 21, 2025 at 2:15 PM
Recently, OpenAI quietly released o3 and o4‑mini, which may showcasing AGI‑like capabilities: o3 demonstrates active visual reasoning (zooming, rotating and interpreting images to solve handwritten notes), while o4‑mini scored 99.5% on the AIME 2025 test.
Hollywood faces new disruption as Silicon Valley races toward AGI, raising unsettling questions about AI’s creative capacity. Experts disagree on definitions and timelines, while inherent model trade‑offs between reliability and randomness limit genuine artistic innovation.
Can AI Be Creative? The Next Wave of Tech Is a Tipping Point for Hollywood
As AGI rises in the tech industry, Hollywood grapples with existential questions of creativity and whether AI can ever be creative.
www.thewrap.com
April 21, 2025 at 2:02 PM
Hollywood faces new disruption as Silicon Valley races toward AGI, raising unsettling questions about AI’s creative capacity. Experts disagree on definitions and timelines, while inherent model trade‑offs between reliability and randomness limit genuine artistic innovation.
OpenAI bolsters its generative AI portfolio with o3, which is capable of multimodal reasoning over images, code and web and efficient o4‑mini. CEO Sam Altman confirms GPT‑5, unifying Omni‑series reasoning and GPT capabilities toward AGI, will launch in months.
OpenAI prepares to launch GPT-5 amid advancements with o3 and o4-mini models
OpenAI prepares to launch GPT-5 amid advancements with o3 and o4-mini models OpenAIs latest models showcase significant advancements in artificial intelligence capabilities.
biz.chosun.com
April 21, 2025 at 1:43 PM
OpenAI bolsters its generative AI portfolio with o3, which is capable of multimodal reasoning over images, code and web and efficient o4‑mini. CEO Sam Altman confirms GPT‑5, unifying Omni‑series reasoning and GPT capabilities toward AGI, will launch in months.
Google DeepMind’s 145‑page proposal presents a comprehensive AGI safety framework, targeting misuse, misalignment, errors, and structural risks through capability restrictions, continuous monitoring, model‑ and system‑level security, interpretability, and uncertainty estimation.
DeepMind Invites AI Community to Collaborate for Safe and Responsible AI Development
AGI (Artificial General Intelligence) is considered the technological "holy grail" by companies like OpenAI or DeepSeek. Presented as an opportunity for humanit...
www.actuia.com
April 21, 2025 at 1:26 PM
Google DeepMind’s 145‑page proposal presents a comprehensive AGI safety framework, targeting misuse, misalignment, errors, and structural risks through capability restrictions, continuous monitoring, model‑ and system‑level security, interpretability, and uncertainty estimation.
AI 2027, a near‑term forecast by experts from OpenAI and The Center for AI Policy, predicts AGI by 2027, with ASI following months later. This scenario, grounded in technical milestones and exponential growth, underscores both profound opportunities and existential risks.
2027 AGI forecast maps a 24-month sprint to human-level AI
The newly published AI 2027 scenario offers a detailed 2 to 3-year forecast for the future that includes specific technical milestones.
venturebeat.com
April 21, 2025 at 11:38 AM
AI 2027, a near‑term forecast by experts from OpenAI and The Center for AI Policy, predicts AGI by 2027, with ASI following months later. This scenario, grounded in technical milestones and exponential growth, underscores both profound opportunities and existential risks.
AGI development accelerates unchecked amid global competition, warns article. DeepSeek breakthrough highlights regulatory failure; past ethics pledges (e.g., 2023 pause letter) dismissed as hypocrisy.
Point of no return | The Express Tribune
.
tribune.com.pk
April 18, 2025 at 9:05 PM
AGI development accelerates unchecked amid global competition, warns article. DeepSeek breakthrough highlights regulatory failure; past ethics pledges (e.g., 2023 pause letter) dismissed as hypocrisy.
Historian Niall Ferguson warns AGI could drive human decline akin to horses post-automobiles: “If AGI surpasses humans, we’ll shrink or go extinct.” Cites *Three-Body Problem*’s Trisolarans as cautionary AI “aliens.” Urges confronting AGI’s existential risks vs. promises. #AGI #AI
AGI Will Displace Humans Like Cars Displaced Horses: Historian Niall Ferguson
It is apparent that humans are well on the path to creating intelligences superior to their own, but it remains to be seen how relevant humans remain once these intelligent beings are around. Briti...
officechai.com
April 18, 2025 at 8:56 PM
AAAI survey: 76% of AI experts deem AGI via current methods “unlikely.” Researchers split: scaling compute vs. needing breakthroughs. Neurosymbolic architectures & embodiment (per Torras/Sierra) seen as crucial for common sense/agency. Meta’s test-trained Maverick model failed benchmarks. #AGI #AI
AI researchers doubt current models will lead to achieving human intelligence
A survey of hundreds of AI specialists from around the world reveals that a large majority believe that this technology needs a different approach to surpass its current capabilities
english.elpais.com
April 18, 2025 at 8:50 PM
Honor announces May 2025 devices with Google Cloud AI: Veo video-gen, Gemini office tools, and AI Eraser/Outpainting. Flagships to get DeepFake Detection. AGI ambitions under "Alpha Plan" aim to "open human potential boundaries" via on-device/cloud AI.
Honor, Google Cloud collaborations deepen to foster advanced AI
The company teases what consumers can expect this year from its "close collaboration."
www.androidcentral.com
April 18, 2025 at 8:39 PM
Honor announces May 2025 devices with Google Cloud AI: Veo video-gen, Gemini office tools, and AI Eraser/Outpainting. Flagships to get DeepFake Detection. AGI ambitions under "Alpha Plan" aim to "open human potential boundaries" via on-device/cloud AI.