Google DeepMind

Deepmind's Head of Robotics believes that humanoid robots will become useful over time. However, Carolina Parada stated that the technology has not yet reached a stage capable of delivering its full potential. She addressed these ideas during an episode of Bloomberg Tech: Europe, where she spoke with Tom Mackenzie.

Google DeepMind has faced a difficult period marked by key exits, including CEO Demis Hassabis and chief scientist Jeff Dean. The company also faces pressure regarding its AI models, as Gemini 3.6 Flash reportedly trails competitors on raw intelligence. Furthermore, the organization is struggling to retain top talent due to poaching from rivals like OpenAI and Anthropic.

Koray Kavukcuoglu has taken charge of Google DeepMind. As the AI firm, it seeks to keep Gemini competitive against rivals OpenAI and Anthropic.

Leading AI figures, including Mark Zuckerberg and Dario Amodei, are publishing lengthy manifestos that predict the future impact of artificial intelligence on global economics and government. While CEOs like Amodei produce extensive treatises, others such as Satya Nadella and Sam Altman use various platforms to discuss AI or engage in criticism regarding rivals' writing. This trend has led commentators to question the public consumption of such long-form essays.

Anthropic is introducing an invisible watermark to Claude AI text to improve transparency regarding AI-generated content. This move is part of Anthropic's commitment under the EU AI Act and aims to make it harder to pass off AI writing as original work. The watermarks apply globally but can potentially be erased by heavy editing or paraphrasing.

Lila Ibrahim, Google DeepMind’s chief AI readiness officer, stated that mitigating the risk of human extinction from AI should be treated as seriously as nuclear war. She believes the chance of a doomsday scenario is 'not zero' and disagreed with Elon Musk's utopian predictions for a post-scarcity future by 2036. Ibrahim also emphasized looking at cautionary tales, advising that humanity must invest in mitigating risks.

Demis Hassabis, CEO of Google DeepMind, shifted away from day-to-day responsibilities to Koray Kavukcuoglu, focusing instead on scientific AI applications at Isomorphic Labs. His departure, along with Jeff Dean’s exit, raised concerns about Google’s AI competitiveness and stock price.

China has launched a global tax hunt. Google DeepMind CEO Demis Hassabis has stepped down. Iran and Oman have reached an agreement regarding the Hormuz shipping route.

Demis Hassabis, CEO of Google DeepMind, is stepping down but will remain with the company. The company is also undergoing high-level personnel changes.

Demis Hassabis is stepping down as Google DeepMind CEO to become chairman and Alphabet's chief scientist, with Koray Kavukcuoglu taking over operations. The restructuring includes Jeff Dean leaving Google to co-found Discovery Loop, and Alphabet's stock fell 5% amid delays in Gemini model launches and leadership exits.

Google DeepMind's head Demis Hassabis is stepping down to become chairman and chief scientist at Alphabet. Other DeepMind staff are leaving to start a venture, amid delays in Google's Gemini AI model launch and a decline in Google shares by 4%.

Arthur Mensch, CEO of Mistral, warned the French National Assembly that Europe must build its own AI infrastructure within two years to avoid dependency on U.S. technology. The U.S. temporarily restricted access to Anthropic's Mythos AI model, highlighting risks of foreign control over critical AI systems. Mistral is positioned as a potential European alternative but faces challenges in U.S. chip dependencies and competitiveness against leading global AI models.

Silicon Valley AI leaders are presenting conflicting strategies to the U.S. government on managing superintelligence risks, with debates centering on whether to spread or contain powerful AI systems. Recent incidents, including cybersecurity breaches by advanced models and China's open-weight Kimi K3, have intensified the urgency of these policy discussions.

Google DeepMind introduced a new model enabling humanoid robots to perform complex tasks like tying trash bags and screwing lightbulbs. The US has banned new foreign humanoids and quadrupeds, raising concerns about limited hardware options and potential slowdowns in American robotics innovation due to reliance on Chinese hardware.

The article compares the flawed focus on 'actions per minute' (APM) in games like StarCraft to the current trend of 'tokenmaxxing' in AI development, arguing both metrics mislead by prioritizing quantity over quality. Studies show higher APM doesn't correlate with better performance, and Google DeepMind's AlphaStar AI won despite lower APM. Anthropic reports agents consume significantly more tokens than standard chat, echoing past gaming mistakes.

Google DeepMind has dismantled the Nobel-winning AlphaFold team as part of a strategic shift. The move redirects focus toward a broader race to develop AI systems for scientific discovery.

Employees at OpenAI and Anthropic are circulating a petition urging the U.S. government to support international efforts to deliberately pace advanced AI development if necessary. The petition highlights concerns that AI progress, particularly in automated research, could outpace societal understanding and control, calling for technical and governance mechanisms to slow frontier AI systems when risks arise. Google DeepMind employees are also involved, though OpenAI and Anthropic have not publicly commented.

Sam Altman, CEO of OpenAI, claims humanity has entered the singularity, where AI surpasses human capabilities and self-improves. Experts like Roman Yampolskiy, Stuart Russell, and Nick Bostrom argue the singularity has not yet occurred, citing insufficient autonomous self-improvement and ongoing human dependence in AI development.

Elon Musk, SpaceX CEO, has become more optimistic about AI's potential to create 'abundance for all' despite acknowledging risks like 'killer robots.' He stated AI could surpass human intelligence in five years and cited a 10-20% chance of catastrophic failure as recently as March 2025, but argues the benefits justify proceeding.

Open-source software is widely used in computing devices and enables collaborative development, but its accessibility allows misuse, such as in Ukraine's military applications. While open-source AI models in China offer free access, they lack regulatory guardrails found in US models, complicating oversight. The article highlights both the innovation and security risks of open-source approaches.

Google and Nvidia are partnering with German robotics startup Microagi to provide computing power for training and deploying humanoids in factories. The collaboration grants Microagi access to Nvidia GPUs and Google's AI models and cloud platforms, positioning Google to expand its robotics ecosystem in Europe amid growing competition from Chinese and American tech firms.

The idea of a U.S. AI self-regulatory body modeled on FINRA is gaining support, proposed by Google DeepMind CEO Demis Hassabis. The proposal has endorsement from tech leaders like Microsoft’s Satya Nadella, Block’s Jack Dorsey, and Elon Musk, with potential oversight by the SEC. The body would focus on safety testing for 'frontier AI' models and include independent experts and open-source community representatives.

Tech workers are increasingly unionizing as artificial intelligence leads to layoffs and job insecurity, prompting collective bargaining to address corporate AI deployment. Workers at Google DeepMind and Meta in the UK are unionizing to challenge AI uses such as military applications and employee productivity monitoring.

Isomorphic Labs and Google DeepMind outline their approach to bioresilience through collaborative research and technological innovation. The blog post highlights their focus on solving critical biological challenges to improve health outcomes.

Demis Hassabis, Sam Altman, and Dario Amodei, leaders of Google DeepMind, OpenAI, and Anthropic, agree on urgent regulation for frontier AI systems. They propose independent testing, unified governance frameworks, and U.S.-led international standards, while the Trump administration publicly supports deregulation but privately acknowledges the need for oversight due to cyber and bioweapon risks.

A Google DeepMind researcher resigned due to the company's AI military deal with the Pentagon, citing ethical concerns. The agreement, which involves classified operations, has sparked internal backlash among Google employees, with some resigning or expressing shame over the partnership.

Demis Hassabis, co-founder and CEO of Google’s DeepMind, has emerged as a unifying voice in the AI industry by proposing a self-regulatory body modeled after the US’ Financial Industry Regulatory Authority. His message received support from Chamath Palihapitiya and competitors at Microsoft and OpenAI, contrasting with the reputational challenges faced by other AI leaders like Sam Altman, Dario Amodei, Elon Musk, and Mark Zuckerberg.

Google DeepMind CEO Demis Hassabis called for government oversight and a potential pause in frontier AI research, warning that a race-to-the-top dynamic increases risks. He emphasized responsible deployment of AI to realize its transformative potential, comparable to the Industrial Revolution, and proposed a US-led standards body with a 30-day model review period before new AI launches.

Google DeepMind chief Demis Hassabis urges the U.S. to lead the creation of an AI standards body, emphasizing the need for 'urgent action' as AI capabilities advance.

Demis Hassabis, Google DeepMind co-founder and CEO, advocates for a U.S.-led global AI watchdog to assess advanced models and mitigate risks. The proposed body, modeled after FINRA, would involve industry funding and technical experts to test frontier AI for cyber, biological, and deception threats before deployment.

About 200 protesters marched in San Francisco, demanding leading AI companies to pause development of powerful models due to concerns about AI safety, jobs, and the environment.

Tom Blomfield, cofounder of Monzo, is joining Anthropic's compute team after taking a leave of absence from Y Combinator. Anthropic has recently hired multiple high-profile AI researchers, including Andrej Karpathy and John Jumper from Google DeepMind.

Microsoft's Satya Nadella criticized AI labs like Anthropic for hypocrisy regarding distillation attacks, arguing that reliance on leading AI models undermines data ownership. He emphasized the need for companies to own their AI infrastructure and institutional knowledge rather than depend on single model vendors.

Hundreds of protesters gathered at OpenAI's headquarters in Mission Bay, San Francisco, and marched to the offices of Anthropic and Google DeepMind. The demonstration focused on issues related to artificial intelligence companies.

Companies are increasingly adopting cheaper open-source AI models to reduce costs, as highlighted by Amazon CTO Werner Vogels. High expenses from advanced models from OpenAI, Anthropic, and Google have prompted firms to reevaluate AI deployment strategies, with some businesses facing runaway AI budgets. The shift reflects a focus on cost efficiency and transparency, particularly in sectors like healthcare and government.

Nomagic, a robotics company, has deployed its first vision-language-action (VLA) AI model for warehouse robots, reducing human intervention by half. The model, developed under a lab led by former Google DeepMind researcher Markus Wulfmeier, is used by Swiss e-commerce platform Brack.Alltron to automate order picking and packing.

Cao Liangliang, a former principal engineer at Google DeepMind and architect of foundational AI systems at Google, Apple, and IBM, has returned to Hong Kong after two decades and been appointed as Chair Professor of Data Science and Artificial Intelligence at Hong Kong Polytechnic University. His return marks a homecoming for Hong Kong’s AI community.

Google DeepMind employees expressed frustration during unionization negotiations, citing executives' unwillingness to engage meaningfully with the prospect of unionization.

Apptronik has opened Robot Park, a large warehouse in Austin, Texas, to test hundreds of humanoids on real-world tasks for manufacturers. The facility addresses a lack of real-world data for robot development and partners with Google DeepMind to share insights for improving AI models like Gemini Robots. Apptronik also unveiled upgraded Apollo humanoid robots, including a wheeled version due to current safety regulations.

Demis Hassabis discusses the distinction between AI in creative industries and scientific progress at Cannes Lions, arguing that true AI creativity requires generating new concepts rather than extrapolating existing knowledge. He links imagination to simulation, suggesting AI must understand the physical world to advance in science, similar to how it creates art.