Today's AI Pulse
Anthropic shuts down Claude Fable 5 and Mythos 5 after US export‑control warning
Anthropic disabled access to its Claude Fable 5 and Claude Mythos 5 models after the Trump administration, via the Commerce Department, warned that foreign nationals could exploit the systems. The move followed alerts from Amazon and other firms about a potential jailbreak that could aid cyber‑security attacks. Anthropic declined the administration’s request for a voluntary shutdown but complied with the order.
Also developing:
By the numbers: AMD acquires AI memory‑optimization pioneer Mext
Blog Prize for the Big Questions About AI
A $20,000 blog essay prize has been launched to surface deep thinking on AI’s biggest strategic questions. Participants must answer one of four prompts—ranging from AI scaling dynamics to how massive philanthropic funds should be deployed—within 1,000 words by May 10, 2026, 11:59 PM PST. First‑place receives $10,000, second $6,000, third $4,000, and winning entries will be published on the organizer’s blog. The contest also serves as a novel hiring funnel to find a research collaborator for ongoing AI policy and technical analysis.
Hire AI Leaders Who Actually Use GenAI, Not Just Talk
I reviewed a lot of job descriptions for senior AI roles recently. What shocked me? Almost none asked candidates to show how they actually use GenAI themselves. Not opinions. Not buzzwords. Not “vision.” Actual usage. That is 𝗶𝗻𝘀𝗮𝗻𝗲. 🤯 We are entering an era where...
AI’s Hidden Costs: What K-12 Needs to Know About Energy and Infrastructure
School districts are shifting from AI pilots to district‑wide deployments, prompting a new focus on the physical infrastructure needed to support these tools. Leaders must grapple with rising electricity demand, the rapid expansion of on‑site and cloud data centers, and...
Cirrascale to Offer On-Prem Google Gemini Models
Cirrascale Cloud Services will deliver Google Gemini large‑language models on‑premise through Google Distributed Cloud, using Dell‑built appliances equipped with Intel CPUs and Nvidia GPUs. The offering supports fully air‑gapped or connected deployments, letting government, defense, finance, healthcare and education customers...
TikTok Is Testing an AI 'Remix' Feature, but You Can Opt Out
TikTok is piloting an AI "Remix" tool that lets viewers generate custom images from a creator's video using generative AI. The feature is currently experimental and limited to a small group of users, with no guaranteed full rollout. Creators can...
Weber Shandwick's AI Guru Moves to FGS Global
Artificial intelligence is rapidly reshaping how brands manage media, reputation, and content creation, from automated monitoring to predictive crisis modeling. In response, communications firms are building dedicated AI advisory practices to help clients translate technology into strategic advantage. FGS Global’s...
GE Appliances Deploys 800 AI Agents to Turbocharge Manufacturing and Supply Chain
GE Appliances has installed more than 800 AI agents across its factories, logistics hubs and supplier network using Google Cloud’s Gemini Enterprise. The rollout, which includes a 25% drop in backorders and minutes‑long shift summaries, marks one of the most...
Applied Digital Lands $7.5 B Hyperscaler Lease for Delta Forge 1 AI Data Center
Applied Digital Corp. signed a 15‑year, $7.5 billion lease with an undisclosed U.S. hyperscaler for 300 MW of AI‑focused capacity at its Delta Forge 1 campus. The deal lifts total contracted lease revenue above $23 billion and sent the stock up more than 12%...
Tesla FSD Anticipates and Pulls over for Emergency Vehicles
Right after pulling over for one emergency vehicle going my direction, Tesla's FSD pulls over again for emergency vehicles coming from the other direction. And turning right in front of me. This feature seems to work relatively reliably and before...
Iridius Raises $8.6M Seed Round
Iridius, a Seattle‑based compliance‑by‑design AI platform, announced an $8.6 million seed round. The financing was led by Chalfen Ventures with participation from Osage Venture Partners, Accenture Ventures, and Rock Yard Ventures. The capital will fund product development, engineering hires, and market...
Publishers and Authors Clash Over AI‑Generated Text as $1.5 B Anthropic Settlement Fuels Debate
Hachette’s cancellation of the horror novel “Shy Girl” after AI‑authorship allegations and Anthropic’s $1.5 billion settlement with authors have intensified a dispute between writers and publishers over AI‑generated content. The conflict pits detection tools that flag false positives against authors demanding...
X Square Robot Launches Wall‑B AI Model, Pledges Home Robot Rollout in 35 Days
X Square Robot unveiled Wall‑B, a new embodied AI foundation model for home robots, and said it will begin placing robots in everyday households within 35 days. Backed by major Chinese tech investors, the company highlighted its World Unified Model...
AI Data Launches AI CityReform on IDX to Slash Japanese Urban Redevelopment Cycles
AI Data Co. unveiled AI CityReform on IDX, an AI‑driven platform that unifies planning, consent and execution for urban redevelopment projects in Japan. The solution claims to cut the traditional 1‑2 year cycle to a few months, offering cost savings...
Harvey Finds GPT‑5.5 Boosts Legal AI Accuracy to 91.7% in New Benchmark
Harvey’s new benchmark analysis reports that OpenAI’s GPT‑5.5 model achieved a 91.7% overall score, edging out GPT‑5.4 by 0.7 points. The gains, highlighted by Head of Applied Research Niko Grupen, are seen as practical advances for transactional and litigation work....
Descartes Launches René, a New AI Agent for Fleet Data Intelligence
Descartes Systems Group introduced René, an AI‑driven conversational agent built into its new Fleet Data Intelligence platform that taps the Global Logistics Network’s execution data. The tool lets dispatchers ask natural‑language questions and instantly receive real‑time and long‑term insights without...
Researchers Launch APOLLO, a 25‑Billion‑Event AI Model to Forecast Disease
Scientists at Mass General Brigham introduced APOLLO, a transformer‑based foundation model trained on 25.2 billion medical events from 7.2 million patients. The system outperformed existing tools on 322 clinical tasks, including a 0.92 AUROC for schizophrenia onset, signaling a new era for...
AI Smart Glasses Will Help Visually Impaired Runners Take on the London Marathon
Visually impaired runners are using AI‑powered Oakley Meta Vanguard smart glasses to navigate the London Marathon, receiving live audio cues about landmarks, distance and pace while still hearing their human guides. The glasses combine a front‑facing camera, AI analysis and...
Attention, Please: TVision Deal Enhances Connected TV Measurement
xpln.ai has entered a strategic partnership with TVision to integrate the latter's person‑level attention data into its Connected TV (CTV) modeling platform. The collaboration transforms granular, second‑by‑second viewer engagement signals into scalable, predictive models that can be applied across linear...
UK Cyber Chief Warns of State‑Backed Attack Surge as £90M Funding Boosts Defences
National Cyber Security Centre head Richard Horne warned that the UK will face a surge in state‑backed cyber attacks and announced a £90 million ($115 m) funding package to accelerate AI‑powered defences for public bodies and SMEs. The warning comes as the...
Jeonbuk National University Deploys $12 Million AI Platform to Overhaul Campus Services
Jeonbuk National University introduced the JUMP AI management platform, a $12 million investment that unifies academic, administrative and career services across its Jeonju campus. The rollout, showcased at an AI System Sharing Conference, marks one of the largest AI infrastructure projects...
Google Launches Renewable‑powered Data Centre in Upper Austria for European AI
Google unveiled a new data centre in Upper Austria that will run on 100% renewable electricity, bolstering the company's AI cloud capacity across Europe. The project, part of Google’s broader clean‑energy push, aims to meet growing demand for sustainable AI...
AIs Hunt for Signs of Intelligent Life
Physicists are harnessing Nvidia GPUs and transformer‑based AI to sift through the massive data stream from telescopes like James Webb, rapidly uncovering faint, early‑universe galaxies that human analysts cannot process in time. UC Santa Cruz researchers adapted large‑language‑model architectures to enhance image reconstruction,...
AI Adoption in Financial Services Accelerates Globally
A new EY Global AI Sentiment Survey of 18,000 respondents across 23 countries shows that 49% have used artificial intelligence to guide savings or investment decisions in the past six months. Gen Z (68%) and millennials (65%) lead adoption, while...
What The First Wave of AI Failures Should Teach Every Organization
U.S. companies have poured $471 bn into AI over the past decade and five tech giants plan to add another $700 bn this year, driving annual spend toward $1.1 trillion by 2029. Yet a National Bureau of Economic Research study finds 95% of...
Aitech Launches Rugged SBCs for Military, Aerospace AI
Aitech unveiled two rugged single‑board computers, the U‑C8600 and U‑C8601, built on Intel’s 14th‑generation Core Ultra platform. The boards combine a multi‑core x86 CPU, integrated GPU and a neural processing unit, delivering roughly 2.5× CPU and 2× GPU performance over...
AI Briefing 4/24/26: A Robot Athlete, Chatbots on the Record, and the First AI Governance Forum
Sony AI unveiled "Ace," a robotic arm that makes table‑tennis decisions in 20 ms and has already beaten elite players, demonstrating AI’s move into real‑world physical tasks. More than a dozen leading U.S. law firms warned clients that any text entered...
Adverse Media Screening Gains Traction as Financial Firms Tighten Compliance
Dow Jones' Alexa Colquhoun explained how AI‑enhanced adverse media screening is becoming a core component of KYC programs for banks and asset managers. Regulators are urging firms to integrate news‑based risk signals, and the technology promises to cut false positives...
New Briefing Note: AI & Privilege: Part One, Internal Investigations – Key Considerations for Professional Services and Financial Services Firms
Norton Rose Fulbright released a briefing note on AI and privilege, focusing on internal investigations for professional and financial services firms. The note reviews core privilege principles and how they intersect with generative AI tools. It highlights key considerations for...
‘Look, No Hands’: China Chases the Driverless Dream at Beijing Car Show
At the Beijing Auto Fair, Chinese automakers showcased a wave of autonomous‑driving technologies as domestic EV sales slump 17% YoY. Huawei announced a $11 bn (80 bn yuan) five‑year fund for self‑driving software, while Xpeng and Xiaomi demonstrated AI‑powered command and in‑car...
Mark Cuban Touts Pittsburgh as AI Hub, Urging Entrepreneurs to Seize Opportunity
Mark Cuban, the billionaire Shark Tank investor, praised Pittsburgh’s emerging AI ecosystem as a new engine for entrepreneurship, citing the city’s low costs, research talent and supportive civic leaders. He urged young founders to specialize in building AI agents for...
AI Data Centers Could Lower Power Prices — Not Up Them
AI‑driven data centers are expanding rapidly, sparking fears that they will push up residential electricity bills. However, experts argue that the real cost driver is how utilities scale capacity and balance supply, not the presence of the facilities themselves. By...
Classiq Certifies Expert‑Level Quantum AI Agents for Production Applications
Classiq announced the certification of its first‑generation expert‑level quantum AI agents, enabling users to convert natural‑language goals into structured, executable quantum programs. The new agentic layer sits atop Classiq’s model‑based platform, promising repeatable, enterprise‑grade quantum development across sectors such as...
Taboola Unveils Realize+, Agentic AI Platform to Automate Campaign Management
Taboola announced the launch of Realize+, an agentic AI system that automates key performance‑marketing tasks such as budget allocation and ad generation. The platform aims to lift conversion rates beyond search and social while cutting manual workload, a move that...
Mistral’s Leanstral Wants to Kill Off Human-in-the-Loop Code Checks, but Is It Blowing in the Wind?
Mistral AI unveiled Leanstral, an open‑source code‑generation agent that couples large‑language‑model output with Lean 4 formal verification to produce mathematically proven code. The system employs a 119‑billion‑parameter mixture‑of‑experts model, activating only 6.5 billion parameters for efficiency, and is offered via a free...
Elon Musk Labels China's $8.4 B Orbital Data Center Push "Interesting"
Elon Musk posted a single word—"Interesting"—in response to news that Beijing‑backed startup Orbital Chenguang secured $8.4 billion in credit lines to build a gigawatt‑scale orbital data center by 2035. The reaction sparked a social‑media surge, highlighting the emerging U.S.–China rivalry over...
Salesforce's Agentforce AI Agent Hits $800M ARR, 169% YoY Growth
Salesforce reported that its AI‑driven Agentforce product generated $800 million in annual recurring revenue, a 169% year‑over‑year increase. The surge, driven by 29,000 transactions in 15 months, is reshaping investor confidence in AI‑enabled SaaS solutions.
Sony’s ‘Ace’ Robot Arm Beats Pro Table‑Tennis Players, Showcasing Real‑Time AI Speed
Sony unveiled its AI‑driven robot arm, Ace, which has begun regularly beating expert table‑tennis players. The system combines a mirror‑based camera, ultra‑low‑latency actuation and reinforcement‑learning algorithms, marking a first for human‑level performance in a mainstream sport.
SS&C Technologies Posts 9% Revenue Rise, Boosts AI Automation and Capital Returns
SS&C Technologies reported adjusted revenue of $1.648 billion for Q1 2026, a 9% increase driven by strong performance in its Global Investor & Distribution Solutions and GlobeOp units. The company also raised its 2026 guidance, underscoring the impact of AI‑enabled automation...
A Student’s Perspective on AI in Schools (Opinion)
High‑school student Elliot Rogers writes to Education Week urging schools to treat artificial‑intelligence as a core subject rather than a cheat‑sheet. He argues that AI tools are already reshaping homework and test‑taking, and that outright bans punish innovation. Rogers calls for...
Japan's LDP Calls for Penalties on Deepfakes and AI‑Generated Piracy
Japan's ruling Liberal Democratic Party has urged the government to add punitive measures to the 2025 AI law, targeting deepfakes and AI‑generated piracy. The proposal seeks to compel uncooperative operators and protect domestic AI sovereignty.
AI in Oncology Takes Off, Tackling HIV and Liver Disease, Pharma’s Recent Gains
At AACR 2026 in San Diego, AI-powered platforms took center stage, signaling a rapid shift toward data‑driven oncology. Parallel research revealed CRISPR screens that mapped CD4+ T‑cell genes that either promote or block HIV infection, while synthetic‑biology engineers demonstrated implantable...
ServiceNow and Google Cloud Unite AI Agents for Autonomous Enterprise Operations
ServiceNow and Google Cloud announced an expanded partnership at Google Cloud Next, introducing a suite of AI agents that can autonomously detect, diagnose and resolve enterprise issues across 5G networks, retail operations and IT systems. The solutions leverage Google’s Gemini...
Find Out What’s New in the Gemini App in April's Gemini Drop.
Google unveiled the 10th edition of Gemini Drops, adding a native macOS client, personalized image generation via Personal Intelligence and Nano Banana, and a Notebook LM feature for organizing chats and research. The update also introduces a free Lyria 3 Pro music tool...
Claude Labels Menu‑bar Launcher “Dumb” But Solves UI Split
Ha! This is new — I love it? Claude Opus 4.6 just called one of my ideas dumb. (I asked it to argue both sides of an idea.) "The menu bar app is dumb — it's a process launcher...
Data Debt Undermines AI Value Amid Organizational Silos
Data debt: AI’s value killer hidden in plain sight https://t.co/4x1ZfQU3h9 Which is news to no one who has been paying attention. But data has owners inside of most orgs. And they often don’t want it to serve everyone. Org culture is...
AI Is Turning Compute Into the New Oil
The article argues that artificial intelligence is reshaping compute into a commodity as valuable as oil. While AI appears as software—models, interfaces, apps—its real engine is massive, specialized processing power. Growing demand for GPUs and custom chips is straining supply...
New AI‑Powered Devices Will Quietly Redefine Workspaces
From AI whiteboards and smart monitors to mini edge workstations and wearables, the next wave of devices will quietly transform how and where work gets done. #FutureOfWork #AI https://t.co/rNDUA2uhNx
GPT-5 Autonomously Outperforms SOTA Protein Synthesis by 40%
More on our work with @OpenAI where GPT-5 ran our autonomous lab at @Ginkgo and beat state of the art by 40% in cell free protein synthesis 👇
Why Can’t Newsroom Leaders Just Be Normal About AI?
Recent AI experiments at major U.S. newsrooms have backfired, from the Cleveland Plain Dealer’s poorly received AI‑generated videos to McClatchy’s “content scaling agent” that threatens reporter bylines. A startup, Nota News, collapsed after publishing AI‑plagiarized local stories, highlighting the risks...
ChatGPT 5.5 vs Claude Opus 4.7: I Tested Both
Anthropic unveiled Claude Opus 4.7 on April 16, while OpenAI released GPT‑5.5 a week later on April 23. Both models are now commercially available, with Claude emphasizing higher‑resolution image handling and long‑form autonomous work, and GPT‑5.5 offering a fully native multimodal architecture for...