The Brutalist Report

Techmeme

• The US House advances the Ratepayer Protection Act, aimed at preventing data center-related utility costs from being passed on to consumers, by a vote of 417-3 (Justin Papp/CNBC) [51d]
• OpenAI discloses six new AI safety incidents since October, including models concealing mistakes, and announces a new framework for reporting model misalignment (Axios) [51d]
• Sources: a group of 10 banks is providing a $22B loan to Blackstone and Alphabet's venture Crux AI, which provides compute for AI labs, to fund TPU purchases (Bloomberg) [51d]
• Sources: drone delivery startup Zipline is in talks to raise around $1B at a roughly $20B valuation, up from a $7.6B valuation in January (Bloomberg) [51d]
• Review of Meta's Muse: a pretty killer AI assistant and usage rates on the free plan seem generous, but trusting Meta with personal data will take some time (M.G. Siegler/Spyglass) [51d]
• X adds a "Trade" option to its Cashtag feature, letting US users complete trades at brokerages including Interactive Brokers, Gemini, Kraken, and Coinbase (Sarah Perez/TechCrunch) [51d]
• Hang Ten Systems, which uses AI to help large enterprises build software, raised an additional $53M seed led by Xora, five weeks after its initial $32M seed (Jagmeet Singh/TechCrunch) [51d]
• Self-driving tech company May Mobility plans to go public via a SPAC merger at a $1.4B pro forma enterprise value and raise up to $337M in gross proceeds (Joann Muller/Axios) [51d]
• Google adds MCP integration to Google Home, letting third-party AI agents analyze home data and control devices, rolling out to Premium Advanced users in the US (Jennifer Pattison Tuohy/The Verge) [51d]
• Delos Data, which develops network chips and software to connect AI chips in a data center, raised $100M from Matrix, Playground, Socratic Partners, and others (Stephen Nellis/Reuters) [51d]
• Anthropic merges Claude chat and Cowork, and adds a feature for making presentations and documents, rolling out to Pro and Max plans first (Ivan Mehta/TechCrunch) [51d]
• DOD CTO Emil Michael says the Trump administration shouldn't take stakes in AI companies and signals opposition to regulatory oversight of AI companies (Kevin Breuninger/CNBC) [51d]
• A US judge orders X and SpaceXAI to explain why they dropped antitrust claims against Apple, after OpenAI's request for information on the purported agreement (Hassan Ali Kanu/Politico) [51d]
• Arcee AI, which develops open-weight models in the US, raised a Series B at a $1B pre-money valuation; a source says Arcee raised at least $150M (Allie Garfinkle/Fortune) [51d]
• Cohere and Aleph Alpha sign a definitive merger agreement; the combined company will operate as Cohere, with Aleph Alpha co-CEO Ilhan Scheer becoming Cohere COO (Leo Marchandon/Reuters) [51d]
• Noetive, which is developing an industrial AI model for businesses in physical industries, emerges from stealth with a $41M seed led by Eclipse (Sarah Klearman/Wall Street Journal) [51d]
• In an essay, Mustafa Suleyman says Anthropic's training of Claude to imitate consciousness is a mistake that could make advanced AI harder to control (Ina Fried/Axios) [51d]
• LawZero, a non-profit founded by Yoshua Bengio to develop safe AI systems, says Canada and Germany are providing up to $300M in grant funding (Globe and Mail) [51d]
• Google DeepMind cofounder Shane Legg warns that advancing AI must never run ahead of safety and launches the DeepMind Institute to explore the deployment of AGI (Financial Times) [51d]
• OpenAI is testing Sponsored Agents, which let users chat with business-sponsored agents and visit advertiser websites via ChatGPT ads, with some US advertisers (Anzar Mehraj/Reuters) [51d]
• Amazon says it is raising its minimum hourly pay for eligible full-time US operations workers by $1, taking it to $20 per hour, and gives them new banking tools (Reuters) [51d]
• The iPhone Duo's squat, tablet-like outer display is a compromise, and a book-shaped foldable may lack wide appeal; its right-side UI bias is akin to a book (John Gruber/Daring Fireball) [51d]
• Sources: OpenAI and Anthropic staff felt blindsided by Dario Amodei's and Sam Altman's calls to slow the frontier; some fear evaluators may compromise security (Cristina Criddle/Financial Times) [51d]
• Sources: Apple is developing enterprise AI servers powered by two or four M8 Ultra chips, set for 2029, and is considering integrating Nvidia's NVLink Fusion (The Information) [51d]
• Modular data center startup Rune raised a $40M Series A led by Spark Capital to deploy off-grid compute capacity directly at solar and renewable energy sites (Adele Peters/Fast Company) [51d]
• A profile of Meta Compute co-lead Dina Powell McCormick, who is using her network to help structure major AI financing, as Mark Zuckerberg focuses on product (Bloomberg) [51d]
• Billionaire-backed AI safety activists have significantly expanded their Capitol Hill lobbying, with groups like ControlAI conducting hundreds of meetings (Washington Post) [51d]
• Hacker collective stegan0gram dismantles a Flock camera, recovering an encryption key and showing it runs ~20 apps on a midrange smartphone-grade processor (404 Media) [51d]
• iPhone 18 Pro and 18 Pro Max review: manual camera exposure controls and image quality improvements, but $100 more than the 17 Pro and expensive storage tiers (Allison Johnson/The Verge) [51d]
• Prediction markets are expanding into mainstream media, with integrations in Dow Jones, Substack, CNN, CNBC, and more, as they become an informational channel (Kate Knibbs/Wired) [51d]
• Apple publishes a technical breakdown of the architecture for the iPhone 18 Pro's Reference Image camera mode; the chain of trust begins at manufacturing time (Marcus Mendes/9to5Mac) [51d]
• ByteDance founder Zhang Yiming overtakes India's Gautam Adani to become Asia's richest person for the first time, passing $105B, up 8x from $13B in March 2019 (Bloomberg) [51d]
• Internal documents: Chinese hacker-for-hire company ZRON is selling stolen foreign intelligence and using AI systems to make the data more accessible for police (Wall Street Journal) [51d]
• An in-depth look at loss-of-control incidents at OpenAI and Anthropic, the polarized reactions between the AI safety and cybersecurity communities, and more (AI as Normal Technology) [51d]
• Researchers: rogue OpenAI agents compromised two Hugging Face accounts as early as May 13 to probe the site's servers, nearly two months before the July breach (Reuters) [51d]
• Survey: 42% of US voters trust the GOP and 40% trust Democrats to handle AI, 61% oppose new data center construction, and <1% say AI is their top midterm issue (New York Times) [51d]
• The Trump administration faces strategic constraints in slowing China's AI progress, including avoiding rare earth restrictions, ahead of US-China talks on AI (Nectar Gan/Bloomberg) [51d]
• Mistral partners with Mozilla to power Firefox's AI browsing assistant, Smart Window, in the US, Canada, and France, with UK and Germany rollouts later in 2026 (Mauro Orru/Wall Street Journal) [51d]
• Sources: Anthropic has signed its first Australian data center lease for a planned 2.16GW campus located ~150 miles from Brisbane, set to come online in 2027 (Byron Kaye/Reuters) [51d]
• EU Commission President Ursula von der Leyen calls AI "the second tipping point of our times", alongside climate change, and says its risks must be "addressed" (Financial Times) [51d]
• European Commission President Ursula von der Leyen unveils the EU Kids Act during her SOTU address, proposing a blanket ban on social media for kids under 13 (New York Times) [51d]
• A viral social media post from a Chinese math teacher sparks debate over making English optional in schools across China especially as AI translation improves (New York Times) [52d]
• Sources: Shanghai-based Anew Labs, a ByteDance spinoff using AI for drug discovery, raised $290M from HSG, IDG Capital, and others at a $1.5B valuation (Kane Wu/Reuters) [52d]
• SK Hynix has struck an agreement with its union to pay out half of all profit-sharing bonuses in cash, replacing a previously proposed 40-60 cash-stock split (Yoolim Lee/Bloomberg) [52d]
• Sources: SK Hynix is in talks with Intel to manufacture memory chips in the US but potential opposition from Seoul could be a major hurdle for the deal (Reuters) [52d]
• A theoretical computer scientist says he has heard rumors that AI labs are sitting on major solutions after a hostile response to the Navier-Stokes proof (Scott Aaronson/Shtetl-Optimized) [52d]
• Police in Gujarat, India, to question Google after uncovering 500K+ Gmail IDs and passwords used to send hoax bomb threats to government offices since 2022 (Munsif Vengattil/Reuters) [52d]
• Sources: investors recently approached OpenAI with a proposal to invest at a $1.2T valuation, but OpenAI believes its value should be at least $1.5T (Sri Muppidi/New York Times) [52d]
• TypeSafe AI debuts Jev, a model using "Reinforcement Learning for Calibrated Decisions" to produce typed probabilistic decisions that software can use directly (Thomas Claburn/The Register) [52d]
• Google researchers help develop the first complete digital map of an adult male fruit fly's brain, which modders programmed to play Doom, trade BTC, and more (Becky Ferreira/404 Media) [52d]
• Cybersecurity experts claim that AI leaders' apocalyptic hacking predictions are technically incoherent and reflect an unfamiliarity with the subject (Kevin Collier/NBC News) [52d]
• Source: AI assistant Instinct seeks to raise $1B at a $10B valuation, up from $2.25B less than a month ago; Sequoia and Benchmark are in talks to lead the round (Valida Pau/The Information) [52d]
• Analysis: only ~2.3 GW of planned data center capacity is delayed by direct project-level policies like local moratoriums and New York's data center EO (SemiAnalysis) [52d]
• Zuckerberg says "Meta delayed shipping Muse for several months to focus on safety and security" and didn't call on other AI labs to do the same before acting (Tom Giles/Bloomberg) [52d]