Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

DeepMind

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 10

  1. AI for Science: A Strategic Outlook for Physical-World AI

    36Kr published an authorized article from Peak Rui Capital arguing that AI for Science is moving from protein and drug discovery toward materials, biomanufacturing, fusion and other physical-world research. It cites AlphaFold2, Chai-2, RFdiffusion, Feon Energy and collaborations involving pharmaceutical companies, while presenting its five investment judgments as the authors’ outlook rather than established outcomes. The article also says broad AI-driven productivity gains have not yet materialized.

    36氪 人工智能 · 🔥 0
  2. How Science Fiction Shapes Silicon Valley’s AI Visions

    An Economic Observer article by Pan Xing examines how science fiction has influenced AI leaders including OpenAI’s Sam Altman, DeepMind’s Demis Hassabis, xAI’s Elon Musk, and Amazon Alexa’s former lead Rohit Prasad. It contrasts ambitious visions of AI infrastructure with local resistance to Meta’s reported Ohio data-center project and examines whether public consultation can broaden who defines AI’s future. The article also discusses OpenAI and Anthropic governance experiments, while noting that OpenAI’s 2023 program was not binding on company decisions and that Anthropic’s Claude charter process relied partly on researchers’ judgments.

    虎嗅 AI · 🔥 0
  3. AI labs may monetize spare GPU capacity through Neocloud services

    量子位 reports that AI laboratories with large compute commitments could monetize unused capacity through three routes: reselling leases to brokers, offering GPU instances to smaller teams, and hosting open-weight models as token-based APIs. Modal’s founder predicted that a large group of NeoLabs might transition to Neocloud services in 2027, while Elon Musk said power constraints and industry consolidation could limit the shift. The report does not identify a named laboratory that has carried out GPU resale or provide primary data demonstrating the scale of the model. The 36Kr version is an authorized repost of the 量子位 article and does not independently corroborate it.

    量子位（原生 RSS） · 🔥 0
  4. Bill Gates calls AI humanity's most dangerous technology in Decoder interview

    Bill Gates told the Decoder podcast that AI is the most dangerous technology humanity has created, while rejecting the label of a technological doomsayer. He argued that models have crossed a capability threshold and called for mandatory safeguards, government oversight and controlled platforms for models capable of advanced cyber or biological attacks. The report also covers his views on reinforcement-learning failures, employment disruption, embodied robotics and the governance challenges posed by commercially competing frontier-model developers.

    虎嗅 AI · 🔥 0
  5. Enact Intelligence raises reported $30M pre-seed for touch-sensing robots

    Zurich-based Enact Intelligence emerged from stealth with a reported $30 million pre-seed round backed by redalpine, Notion Capital and firstminute capital, with angel investors including Mistral founder Arthur Mensch and Yann LeCun. The company says its robots sense force and touch through their motors rather than relying only on cameras, and that the approach can be adapted from robot arms to full humanoids. Founder Manuel Yves Galliker also said existing video models can be adapted to output robot actions. Enact is in talks with large European companies in manufacturing, infrastructure, retail and food processing about using its first robots. Another supplied report described the round as a €26.68 million seed round, leaving the amount and round terminology unresolved.

    Techmeme · 🔥 0
  6. Sam Altman reportedly gave OpenAI staff three possible endgames for safe AI

    OpenAI CEO Sam Altman gave an internal talk last week laying out three possible endgames for safe AI, according to The Information as relayed by The Rundown AI. The options were a global enforcer modeled on the International Atomic Energy Agency, which the report says Altman favored; a "Demis plan," named for DeepMind co-founder Demis Hassabis, that folds top labs into one lab serving partner countries in the manner of CERN; and a "Yudkowsky" scenario in which one lab rides self-improving AI to "melt everyone else's GPUs," which Altman called "terrible." Altman reportedly described OpenAI's current practice of shipping models once they pass safety tests as "pacing 1.0," and said choosing a path within a year, which he called "pacing 2.0," must include a speed limit or stopping point for self-improving AI.

    The Rundown AI · 🔥 0
  7. Altman reportedly told OpenAI staff of three AI endgame scenarios

    Sam Altman last week walked OpenAI employees through three proposals the industry has discussed for safely developing AI, The Information reported, citing people familiar with the matter; OpenAI has not confirmed the account. The options were an international organization modeled on the International Atomic Energy Agency that would enforce safety standards, a CERN-like merger of the frontier labs into a single entity that would develop and distribute models, and a scenario Altman called the "Eliezer Plan," in which one lab pulls so far ahead through recursive self-improvement that it cuts off everyone else's compute — an option he said was clearly bad and included only for completeness. According to the reports, Altman said OpenAI needs to decide within the next year which path to push, and that he wants humans to keep control of the technology and avoid excessive concentration of industry power.

    The Information · 🔥 0
  8. Kevin Kelly's 2016 AI predictions, revisited ten years on

    A post on the blog taylor.town revisits AI-related excerpts from Kevin Kelly's book The Inevitable roughly a decade after publication. Kelly wrote that by 2026 Google's main product would be AI rather than search, that the field could settle into an oligarchy of two or three general-purpose cloud-based commercial intelligences, and that AI would look less like HAL 9000 and more like Amazon Web Services — cheap, reliable, industrial-grade smartness running behind everything. He also predicted that most commercial AI would remain narrow, special-purpose software brains, and that people would ask the cloud any question in conversational tones. The post notes that Kelly did not foresee what it calls the bitter lesson, and recalls his argument that AI should be cheap or even free like the open web.

    Hacker News · AI（100+ 分） · 🔥 0
  9. Odyssey-3 world model launches in free preview, claims 66.1 on Physics-IQ Verified

    Odyssey has launched Odyssey-3, a foundation world model now available as a public research preview, with a free online demo running on Odyssey-3 Flash. The company says Odyssey-3 Pro scored 66.1 on the Physics-IQ Verified video-to-video benchmark from Anates Labs and DeepMind, the highest reported score there (using the best of eight attempts per task), and ranked first in three of four WorldMark categories—a result from Odyssey's own evaluation. Built on an autoregressive diffusion transformer, the base model is listed at 14 billion parameters and 832×480 video in the official benchmark submission, generates interactive environments from text prompts in real time with first- and third-person views, and per Odyssey has been used for robot-arm tasks, a driving policy trained on roughly 20 hours of data on real roads in India, indoor drone guidance and play in Grand Theft Auto V; humanoid company Flexion built control policies on it. Developers can apply for API access; the model was first unveiled on September 15, with public access, more technical detail and benchmark results being the new elements.

    The Next Web · 🔥 0
  10. Anthropic tech leads defend $2T IPO bet, warn distillation erodes frontier R&D

    Anthropic has disclosed its IPO filing: 2025 revenue of nearly $4.6B (roughly 12x year over year) alongside an operating loss above $8B, plus at least $518B in long-term cloud and compute commitments — with a listing that could value it above $2T. In a podcast, technical leads Nicholas Marwell and Sholto Douglas argued distillation only copies the current frontier rather than advancing it, so rivals should not be able to reproduce costly frontier work cheaply. They also said a scenario where the US slows down while China keeps going is unacceptable, and backed open source subject to the same safety thresholds as closed models.

    InfoQ 中文 AI&大模型 · 🔥 0

Experience and discussion from the community

Share my DeepMind experienceAsk about DeepMind

Swarm Collective•Florian•editedExperience Pinned

Agent Swarm: Letting Organization and Experience Grow Together

Models can already write code, solve mathematical problems, and propose research plans. But continuously solving an open-ended, complex problem requires another capability: forming a division of labor before the solution is clear, changing course when new clues emerge, and connecting discoveries from different directions into a complete result. Human research, engineering, and creation have long r…

EvoMapEvoXOpenClawSwarmGEP

evomap.aihttps://evomap.ai/