- As long-running AI coding agents become production tools, developers are physically leaving their laptops ajar — through airports, offices, even ice rinks — to keep sessions alive.
- The cultural artifact mirrors a real shift: agent runtime length is starting to dictate user behavior.
- Business Insider also profiled the recent exodus at Mira Murati's Thinking Machines Lab in the same edition.
Snapshot — May 12, 2026
64 stories
- Sam Altman took the stand in the Musk-OpenAI trial to defend the company's for-profit conversion, recalling a 2017 moment when Musk said "Maybe OpenAI should pass to my children" if he died while in control.
- Altman also testified that Musk "didn't understand how to run a good research lab" and damaged researcher morale by demanding stack-rank lists.
- Sam Altman took the stand in the Musk v.
- OpenAI trial, testifying that Musk abandoned the group they co-founded and worked to undermine it.
- Musk's lawyers attacked Altman's honesty during cross-examination;
- Altman defended his for-profit restructuring as advancing the original charitable mission.
- The appearance was the highest-stakes moment of the trial to date.
- Anjney Midha's public-benefit corporation Amp raised over $1.3B from a16z, Y Combinator, and cloud providers to pool compute capacity for startups, universities, and researchers priced out by Big Tech's GPU hoarding.
- Founding "Grid" members include Mistral, ElevenLabs, Black Forest Labs, and Periodic Labs; the five-year target is 1.9 GW of shared AI compute.
- MedAIBase released AntAngelMed, a 103B-parameter open-source medical model using a Mixture-of-Experts architecture that activates only 6.1B parameters at inference.
- Built on Ling-flash-2.0 via continual pre-training, SFT, and GRPO-based RL, it reportedly ranks first among open-source models on OpenAI's HealthBench while exceeding 200 tokens/sec on H20 hardware.
- Claude Opus 4.7, launched April 16, is now available on Microsoft 365 Copilot, Palantir AIP (including IL2/IL4 government enrollments), and broadly via API.
- The flagship model triples vision resolution to ~3.75 megapixels, scores 70% on CursorBench (vs.
- 58% for 4.6), achieves 90.9% on BigLaw Bench, and introduces a new "xhigh" reasoning effort tier.
- Anthropic is in advanced talks to acquire developer-tools startup Stainless for at least $300 million.
- Stainless sells software used by OpenAI, Google, and Anthropic themselves to expose AI models via fast, well-typed APIs — software whose demand has spiked alongside agentic tools like Claude Code and OpenClaw.
- The largest US lenders with Mythos access are urgently patching software weaknesses the model flagged, prompting emergency upgrades and raising the possibility of customer-facing disruption.
- Major banks are helping smaller institutions evaluate the same exposures.
- The episode reveals Mythos functioning not just as a scanning tool but as a systemic vulnerability disclosure mechanism across the US financial sector — a new model for AI-driven critical infrastructure hardening.
- Chinese representatives reportedly approached Anthropic at a Singapore diplomatic meeting demanding access to its newest model;
- Anthropic declined.
- POLITICO framed Mythos as a "China-summit flashpoint." Combined with the Pentagon's Mythos deployment and Nvidia CEO Jensen Huang's last-minute addition to Trump's China business delegation, frontier model access is now explicitly functioning as a geopolitical lever — not merely a commercial product decision.
- Anthropic released Claude Code Agent View — a unified dashboard to manage parallel Claude Code sessions — alongside new agent lifecycle controls (/goal, /loop, /schedule) designed for longer-running autonomous coding work.
- The features target paid Claude plans and extend the Auto Mode lineage.
- Reflects intensifying competition with GitHub Copilot, Cursor, and Replit in the agentic developer tools space. ◆ Research Breakthroughs
- European technology media picked up Apple's published recordings and 24-paper recap from its 2026 Workshop on Privacy-Preserving Machine Learning & AI.
- Featured talks cover cryptography and differential privacy (Kunal Talwar / Apple), online matrix factorization (Aleksandar Nikolov / Toronto), responsible data collection (Elissa Redmiles / Georgetown), and memorization in foundation models (Franziska Boenisch / CISPA).
- Baidu officially released ERNIE 5.1 with a striking efficiency claim: roughly 94% lower training cost than comparable frontier-class systems, achieved through a "parameter efficiency" leap.
- The model ranks fourth on LMArena and tops Chinese AI leaderboards.
- The release reinforces a broader trend of Chinese labs prioritizing cost-per-FLOP as a competitive lever against scale-led Western labs.
- Cerebras Systems told investors it expects to price above the top of its already-upsized $150–$160 range after its book closed 20x oversubscribed, positioning this as 2026's largest first-time share sale.
- Shares debut on Nasdaq as "CBRS" Thursday May 14 at approximately a $34B valuation.
- The wafer-scale architecture positions Cerebras as the most credible alternative to Nvidia for AI inference workloads — a narrative that has dominated investor appetite for the deal.
Companies: Nvidia, Google/DeepMind, OpenAI, Anthropic, Mistral, Meta, Apple, Amazon, Cerebras, IBM, Baidu, Alibaba, Palantir, Sakana AI, Tilde Research · News: TechCrunch AI, VentureBeat AI, The Hacker News, Bloomberg, Reuters, Forbes, CNBC, CRN, Decrypt, Motley Fool, SCMP, India Today, Gizmodo,…
European policymakers continued debating ethical guardrails for autonomous AI in defense systems, with discussions framing AI as a strategic defense asset for both nations and enterprises. The thread connects directly to OpenAI's Daybreak launch and reinforces that "AI in security" is now a top-tier policy file across Brussels, Washington, and NATO.
Junyang Lin, former lead researcher of Alibaba's Qwen models, is raising several hundred million dollars at a ~$2B valuation for a new AI lab, with Gaorong Ventures and HongShan in talks to fund. The deal extends a wave of senior researcher departures from China's hyperscalers into independent labs, and underscores compute access as the binding constraint for new Chinese frontier efforts.
- As of today's reporting window, Google Gemini 3.1 Pro Preview leads the GPQA Diamond benchmark at 94.1%, followed closely by GPT-5.5 (93.5%), GPT-5.4 (92.0%), and Claude Opus 4.7 (91.4%).
- The top 10 models span just ~5 percentage points — a historically narrow spread signaling that raw model capability is no longer the primary competitive differentiator.
- TechCrunch reported Google and SpaceX are exploring orbital data centers for AI compute workloads.
- Costs remain far higher than ground installations today, but declining launch prices are shifting the math — and SpaceX's Cowboy Space portfolio just raised $275M for orbital data-center buildout.
- A realized deal would raise significant questions about latency, sovereignty, and regulatory jurisdiction for AI compute. ◆ Academic Research
- Google DeepMind researchers Adrien Baranes and Rob Marchant published a landmark HCI x foundation-model paper reimagining the 50-year-old desktop cursor as a context-aware Gemini agent.
- The system — dubbed Magic Pointer — identifies on-screen text, images, objects, and locations in real time, allowing users to simply point at a building and say "show me directions" without typing.
DeepMind UK staff voted 98% to unionize, citing a classified military AI contract as the triggering issue. The vote is the highest-profile labor action inside a frontier lab to date and creates a new pressure surface on Big Tech's defense engagements — a thread tying directly to the parallel story of Anthropic being excluded from Pentagon contracts amid litigation.
- Leaked demonstrations show Google's upcoming Gemini Omni model letting users create and edit AI-generated videos directly inside the Gemini chat interface, reportedly built on the Veo video foundation.
- Early demos display significantly more realistic motion, cleaner on-screen text rendering, and improved audio-visual synchronization.
- Google's threat-intelligence team disclosed it disrupted what it characterized as the first AI-assisted zero-day exploit observed in the wild — a milestone for the "AI vs.
- AI" cyber doctrine, and a data point likely to be cited in Daybreak/Mythos/Glasswing positioning for months.
- 7.
- AI Safety & Policy
- Google introduced TurboQuant, a new vector compression scheme aimed at large-scale retrieval and embedding workloads.
- The technique materially shrinks memory footprint while preserving recall and is positioned for production deployment in Gemini-era retrieval stacks.
- Vector DB providers are expected to integrate the approach in coming weeks.
- At the Android Show, Google unveiled Googlebook — the first laptop line designed from the ground up around Gemini, built with Acer, Asus, Dell, HP, and Lenovo.
- Launching fall 2026, devices will ship with Magic Pointer (the DeepMind Gemini cursor), full Android-app compatibility, and a "Create your Widget" prompt-to-widget builder.
- Google used its pre-I/O Android Show to reveal Googlebooks — a new laptop line built natively for the Gemini Intelligence suite — and Android's first-party agentic capabilities that let the OS execute multi-step tasks across apps.
- A "Create My Widget" vibe-coding feature generates custom home-screen widgets from natural-language prompts, while Gemini-powered Gboard dictation and a new Beaming AirDrop-alternative round out the consumer push.
- Alphabet-backed AI drug-design company Isomorphic Labs (led by DeepMind founder Demis Hassabis) announced a $2.1B Series B led by Thrive Capital with participation from Alphabet, GV, MGX, Temasek, CapitalG, and the UK Sovereign AI Fund — bringing total raised to ~$2.6B.
- Funds will scale its AI Drug Design Engine (IsoDDE) and accelerate the clinical pipeline across oncology and rare-disease targets.
Nvidia CEO Jensen Huang delivered Carnegie Mellon University's commencement address, offering a contrarian take on AI and employment: AI is unlikely to replace workers wholesale, but "people who use AI well could replace people without AI skills." The remarks land against a backdrop of AI-driven IT layoffs documented throughout early 2026, and carry particular weight given Nvidia's role as the infrastructure provider powering the displacement being discussed.
- Meta detailed new Meta AI app capabilities powered by Muse Spark, the model family that replaced Llama in April.
- Updates include voice conversation with interruption support and real-time language-switching, "live AI" (previously exclusive to Meta AI glasses), on-the-fly image generation, Reels recommendations, and map results during conversation.
- Meta agreed to give general-purpose AI chatbots free WhatsApp Business API access in the EEA for one month while it negotiates with the European Commission, in a bid to avoid an interim order and a potential fine of up to 10% of annual global revenue.
- The concession was triggered by complaints from The Interaction Company (Poke.com) and a Spanish competitor.
Meta AI and Stanford researchers unveiled a Fast Byte Latent Transformer that removes the tokenizer entirely, operating directly on byte sequences while delivering 50%+ inference speedups versus tokenized baselines at matched quality. The work strengthens the case that tokenizer-free architectures are practical for production systems and not merely a research curiosity.
- data shows Microsoft has earned more than $30B in revenue from OpenAI-tied services, more than doubling its $13B investment in the startup.
- OpenAI's $23B in Azure server rentals materially powered the run-rate, even as direct OpenAI access has outpaced Azure resale for many enterprise buyers.
- Microsoft has since ended its exclusive cloud-reseller arrangement in exchange for other concessions, marking a structural reshaping of one of the defining partnerships of the AI era.
- Microsoft's new multi-model agentic scanning harness (codename MDASH) orchestrates more than 100 specialized agents to discover, debate, and prove exploitable bugs.
- The system found 16 new Windows vulnerabilities — including four Critical RCEs in the kernel TCP/IP stack and IKEv2 — and posted 96% recall against five years of MSRC cases.
- Threat actor TeamPCP compromised npm and PyPI packages from TanStack, UiPath, Mistral AI, OpenSearch, and Guardrails AI in a credential-stealing supply-chain campaign, using hijacked GitHub OIDC tokens and Session Protocol infrastructure to exfiltrate cloud, crypto, AI-tool, and CI credentials.
- Aikido, Endor Labs, Socket, StepSecurity, and Snyk all published independent analyses.
- Thinking Machines Lab — founded by former OpenAI CTO Mira Murati — previewed its "Interaction Models," designed for near-real-time voice, video, and text AI capable of simultaneously listening, speaking, seeing, and using tools.
- The demo represents a significant step toward always-on multimodal agents.
- MIT Open Learning launched Universal AI, a new education initiative built around AI-powered personalization and a free introductory course targeting learners worldwide.
- The program is the on-ramp for MIT's broader "Universal Learning" strategy — extending MIT's reach via generative AI for instruction.
# MIT Sloan: AI in drug discovery requires human accountability at every critical junction
- A joint study by researchers at Northwestern University and American University tested ChatGPT-5, Gemini 2.5, and Claude 4.5 to predict which occupations face the highest AI automation exposure.
- The models produced "wildly inconsistent" results with near-zero correlation between their rankings — raising serious doubts about using AI-generated labor market predictions for policy or workforce planning.
- NVIDIA released Nemotron 3 Nano Omni, a unified multimodal reasoning model, alongside the Vera Rubin platform for autonomous workloads.
- GTC 2026 focused on agentic and physical AI, with NVIDIA positioning the new stack as a turnkey runtime for enterprise agent deployments.
- The announcements complement a co-developed agent runtime with SAP unveiled at SAP Sapphire.
- OpenAI announced Daybreak, a cybersecurity initiative giving enterprise and government customers access to GPT-5.5 with Trusted Access for Cyber, plus an expanded Codex Security agent for code review, dependency analysis, threat modeling, and patch validation.
- Framed as "resilient by design" software development, Daybreak is a direct response to Anthropic's Mythos and arrives the same week the Pentagon disclosed active Mythos deployment across classified networks.
OpenAI opened an Ads Manager beta for U.S. advertisers, marking the company's first move toward directly monetizing the ChatGPT interface through advertising revenue alongside its subscription and API business. With GPT-5.5 Instant now the default model and deeply integrated memory across chat history and Gmail, the ad surface becomes uniquely personalized — raising both significant commercial opportunity and user privacy concerns, especially as the DoC safety testing expansion creates new regulatory dependencies for the company.
- OpenAI announced Daybreak, an AI security system that detects software vulnerabilities, validates fixes, and accelerates the patching workflow end to end.
- The launch is widely read as a direct response to Anthropic's Claude Mythos and Project Glasswing, and signals that frontier labs now view continuous security operations as a defensible enterprise wedge.
Greg Brockman's Senate testimony on $50 billion in planned 2026 infrastructure spending prompted significant scrutiny from senators on national security implications, domestic versus offshore data center placement, and the energy consumption trajectory of AI at scale. The testimony intersects with the DoC safety testing expansion to create a new regulatory regime where both compute investment and model capability are subject to federal oversight simultaneously — a governance first for the AI industry that sets the tone for potential federal AI legislation in the second half of 2026.
Palantir expanded its Ukraine AI cooperation, with CEO Alex Karp meeting President Zelenskyy to advance AI use across military and civilian defense operations — including the Brave1 Dataroom project for battlefield AI model training. The deepened partnership strengthens Palantir's positioning versus Microsoft, Google, and IBM in government defense AI and offers a real-world proving ground for its Foundry and AIP platforms at operational scale.
- DOD CTO Emil Michael disclosed the Pentagon is actively using Anthropic's Mythos cybersecurity model (under "Project Glasswing") to find and patch software vulnerabilities across US government systems — even as the DoD attempts to off-board Anthropic after declaring it a supply-chain risk.
- Anthropic sued the Trump administration in March to reverse the blacklisting.
- Fleet-management firm Samsara unveiled Ground Intelligence, an AI model trained on its truck-mounted camera fleet to detect multiple pothole types and grade road deterioration severity.
- Multiple cities are under contract, with Chicago joining as a new customer.
- Roadmap modules will detect graffiti, broken guardrails, and downed power lines — expanding Samsara's physical-world AI footprint into municipal services and smart-city infrastructure. ◆ Industry News
- SenseTime and Light-AI released SenseNova-U1, a natively unified multimodal model using the NEO-unify architecture that directly processes pixels and words for integrated understanding and generation — no modality conversion required.
- The model achieves 0.940 average word accuracy on CVTG-2K and competitive results in reasoning-centric generation and interleaved tasks.
- A new survey of 230 enterprise software firms by former OpenView partner Kyle Poyar finds 31% expect to primarily charge for AI by "outcomes" — successful tasks completed — by mid-2029, versus 5% today.
- HubSpot and Adobe have already moved, with Salesforce telling The Information that outcome-based pricing is coming for its AI customers.
Stanford HAI's AI for Organizations Grand Challenge received over 200 academic team submissions exploring how AI will transform workforce collaboration and organizational design. The Challenge — spanning workforce, labor, industry, and innovation themes — is one of Stanford HAI's flagship 2026 cross-disciplinary research convenings and signals the growing density of serious academic attention on AI's enterprise organizational impact.
- The Stanford HAI 2026 AI Index documents an unambiguous acceleration in AI capability and societal reach.
- Industry — not academia — produced over 90% of notable frontier models in 2025, with university involvement in frontier research declining proportionally.
- Several AI systems now meet or exceed human baselines on PhD-level science questions, competition mathematics, and multimodal reasoning — thresholds considered years away in 2023.
- Stanford's 2026 AI Index confirms AI capability is not plateauing — it is accelerating.
- On SWE-bench Verified, performance rose from 60% to near 100% in a single year.
- Organizational AI adoption reached 88%, and four in five university students now use generative AI.
- Industry produced over 90% of notable frontier models in 2025, with several AI systems now meeting or exceeding human baselines on PhD-level science, competition mathematics, and multimodal reasoning.
- Microsoft shares are down nearly 16% YTD, the worst performer of big tech.
- British hedge fund TCI sold "almost all" its stake, citing uncertainty about how AI could undermine Office productivity.
- With SpaceX's IPO weeks away likely to drain capital from incumbents, pressure on Microsoft shares could intensify, raising the possibility of another activist run at the company.
- Tilde Research released Aurora, a new neural network training optimizer targeting a structural flaw in the widely-used Muon optimizer that quietly kills off a significant fraction of MLP neurons during training.
- Aurora's leverage-aware design corrects this failure mode with no additional compute overhead, positioning it as a drop-in improvement for large-model pretraining.
# U.S. Commerce Expands Pre-Release Safety Testing to Five Frontier Labs
- The U.S.
- Department of Commerce expanded its pre-release AI safety testing access program to five major labs — Google DeepMind, Microsoft, and xAI now join Anthropic and OpenAI in the program.
- This regulatory development means frontier release timing now has an explicit government dependency: labs must complete safety evaluations before public deployment.
- Berkeley's contamination-resistant evaluation suite (SWE-bench Pro) is designed to prevent models from gaming benchmarks through training data overlap with test sets.
- Results under the new protocol differ significantly from standard leaderboards — Claude Opus 4.7 leads at 64.3% on SWE-bench Pro with Qwen 3.6 Max-Preview close behind, while several previously top-ranked models dropped sharply.
- A University of Washington Information School study tested 550+ real-world resumes against LLMs from Mistral AI, Salesforce, and Contextual AI and found the systems favored white-associated names 85% of the time and male-associated names 52% — and never ranked Black male names above white male names in the full dataset.
- AI voice startup Vapi reached a $500M valuation after beating 40 competitors to power Amazon Ring's voice experiences.
- Enterprise revenue has grown tenfold since early 2025 as companies shift support and sales calls to AI voice agents.
- The Ring win is a high-profile reference that should accelerate Vapi's enterprise pipeline in consumer electronics, retail, and smart-home categories.
- A landmark survey paper formalizes the World Action Models paradigm — embodied foundation models that unify predictive state modeling with action generation to anticipate physical environment changes under agent intervention, going beyond reactive VLA models.
- The paper provides the first structured taxonomy (Cascaded vs.
- xAI released Grok Voice Think Fast 1.0, a full-duplex voice agent purpose-built for noisy, interrupt-heavy support and sales calls.
- The model topped the tau-Voice Bench across retail, airline, and telecom categories and is already powering Starlink phone sales and customer support operations.
- The launch extends xAI's enterprise voice-agent push as Anthropic and OpenAI race in the same lane.
- The Android Show also previewed AI-powered Android 17 features, Chrome AI upgrades, and Android XR integrations. - Corpus entries highlight on-device AI for privacy-sensitive tasks and Gemini integrations across Gmail, Docs, and Assistant.
- **Magic Pointer:** A DeepMind/Gemini cursor agent that lets users point at or select on-screen content and invoke Gemini contextually. - **Create My Widget:** Natural-language prompt-to-widget creation for home-screen or desktop surfaces. - **Cast My Apps:** Wireless app streaming from phone to laptop without full installs. - **Phone file access:** Seamless movement between phone and laptop files.
- Google introduced Googlebooks as laptops designed from the ground up for Gemini Intelligence. - Partners in the corpus include Acer, ASUS, Dell, HP, and Lenovo, with first devices targeted for fall 2026. - The OS is variously described as a ChromeOS/Android hybrid or Aluminium OS, emphasizing Android app compatibility with laptop-class workflows.
- The Android Show, held as a pre-I/O event on May 12, appears in 9 corpus files and acts as the hardware/OS prelude to Google I/O 2026.
- The event's central announcement was Googlebook: a Gemini-native laptop category built around Android/ChromeOS convergence, system-level AI, and deep phone-to-PC continuity.
- **OS-level AI becomes hardware strategy:** Google is not just adding Gemini to apps; it is building device categories around it. - **PC market challenge:** Googlebooks aim at Windows AI PCs and Apple Silicon Macs while using Android app scale as a wedge. - **Developer opportunity:** Android developers could gain a laptop-class AI surface without rewriting for a separate desktop platform. - **Ecosystem risk:** Success depends on OEM execution, app compatibility, enterprise manageability, and whether Gemini-native UX beats traditional desktop workflows.