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
Fraud and the False Optimism of AI for Science
The article debates whether AI‑generated research constitutes fraud or a legitimate productivity tool. It contrasts optimistic views that see AI as a natural extension of scientific methods with pessimistic concerns that outsourcing idea generation erodes judgment and misattributes credit. The author argues that the ethical line hinges on the degree of human oversight and the intent behind using AI. Ultimately, the piece calls for clear standards to distinguish acceptable assistance from deceptive authorship.
Today Is the Day Anthropic Promised That Fully Autonomous Employees Would Be Tearing Through the Business World
A year after Anthropic’s CISO Jason Clinton warned that AI‑powered virtual employees would be roaming corporate networks, the promise remains unfulfilled. Security breaches, unreliable performance and a string of failed hype‑driven demos have kept autonomous agents out of production. Earlier...
Paper Mache Businesses
The New York Times highlighted Medvi, a $1.8 billion AI‑driven healthcare firm, as a solo‑founder unicorn, but investigative reports reveal it relies on deep‑faked patient images, bogus doctor endorsements, and outsourced medical work. The author labels such ventures “paper‑mache businesses,” where AI strips...
I Put Perplexity Vs. Claude to the Test: Here’s My Verdict
The article pits Perplexity AI against Anthropic’s Claude, testing their free versions across writing, research, coding, and multi‑turn conversation. Perplexity shines in real‑time, source‑cited research, while Claude excels at deep contextual reasoning, coding assistance, and creative prose. Both services share...
Pulse Releases Multilingual Table Benchmark and Unified Metric
Pulse just open-sourced PulseBench-Tab, a frontier benchmark for table extraction across 9 languages, alongside T-LAG, a new scoring metric that evaluates structure and content in one unified score. After processing billions of pages for Fortune 50 enterprises, the largest investment firms,...
AI Board Features CEOs, Musicians, and Artists
One of the more compelling AI use cases: An AI board of directors, which includes CEOs from Canva, Figma, Universal Music Group, plus Billie Eilish and Kendrick Lamar. Here’s SoundCloud CEO Eliah Seton:
Gupshup Launches Superagent
Gupshup unveiled Superagent, an autonomous AI platform that orchestrates customer conversations across all major messaging and voice channels. The tool lets users design campaigns, build journeys, provision infrastructure, process transactions and self‑optimize with simple prompts. It operates on WhatsApp, RCS,...
SAP and Google Cloud Expand Partnership to Deploy Multi-Agent AI
SAP and Google Cloud have deepened their alliance by linking SAP Engagement Cloud, SAP Customer Experience and Joule solutions with Google’s Gemini Enterprise AI platform. The integration enables marketers to deploy multi‑agent AI that accesses unified data across both ecosystems...
Precisely Launches EngageOne RapidCX
Precisely announced EngageOne RapidCX, a governed‑AI platform that adds real‑time visibility and decision support to customer communications. The new release introduces a Next Best Action Agent for email, a metrics‑driven homepage with five performance widgets, and a Communication Tracker that...
AI Tool Rips Off Open Source Software Without Violating Copyright
An AI‑driven tool called Malus claims it can recreate any open‑source project from scratch, producing functionally identical code while avoiding attribution and copyleft licenses. The service leverages the clean‑room design principle established in the 1982 IBM BIOS reverse‑engineering case, arguing...
NAN120: How Network Engineers Can Thrive in an AI-Driven World
In this episode, senior solution engineer Ashwin Chosi discusses how the surge in AI workloads is reshaping network engineering, emphasizing the need for scalable, resilient infrastructure and intelligent automation. Drawing on his experience at AWS, Microsoft Azure, and Keysight, he...
AI Mirrors Nuclear, Electricity, Press: Power, Infrastructure, Democratization
Just riffing on some AI metaphors/analogies floating around right now as a way of thinking through things. Here are some dominant ones: Nuclear weapons / Cold War arms race. AI as a dangerous new force that concentrates enormous power in the hands...
Google's MCP Unifies All Cloud Services for Agents
May not seem sexy but I love love love Google going all in on MCP. Even dummies like me can create amazing apps that leverage MCP. All GCP and Workspace services are now exposed via MCP. Developers can address GKE,...
I Put Grok Vs. ChatGPT Head to Head and One Stood Out
A side‑by‑side test of Elon Musk’s Grok and OpenAI’s ChatGPT reveals distinct trade‑offs. ChatGPT delivers consistently polished text, robust coding, and deep research, while Grok shines with faster responses, a witty personality, and real‑time X data pulls. Both platforms offer...
AI Threatens Jobs and Economy, CEOs Warn Uncertainty
We’re all sitting uncomfortably with AI right now. It’s coming from the top down. The CEOs of the big AI companies caution us that this technology may very well take all of our jobs. Or that if it doesn’t live...
AI Surveillance Threat Grows Despite Lack of Concrete Evidence
There’s plenty of precedent for AI being used for mass surveillance. While there’s not yet any smoking-gun evidence that the US government (or anyone else) is using LLMs to conduct surveillance in the way that could constitute a crime against...
Inside the UMG-Backed Patent Portfolio Targeting AI Music Derivatives: A Technical Blueprint for the Walled Garden Model?
Universal Music Group, through its affiliate Music IP Holdings and a partnership with Liquidax Capital, has built a portfolio of more than 60 AI‑music patents aimed at controlling AI‑generated derivatives. The core inventions require artist approval, embed digital watermarks, and...
Next Leap to Harness Engineering: JiuwenClaw Pioneers ‘Coordination Engineering’
OpenJiuwen released JiuwenClaw 2.0, introducing AgentTeam—a multi‑agent collaboration layer that the community dubs “Coordination Engineering.” In tests the system autonomously assembled a ten‑agent team to produce a 200‑page technical PowerPoint in under 20 minutes, with no human intervention. AgentTeam features a Leader...
Apple Chooses Google Gemini to Power New Siri
Apple getting airtime at @googlecloud Next as a Gemini customer is notable enough. @ThomasOrTK naming Apple as the company that picked Gemini to power the new Siri made it land harder. Google won a marquee customer and reminded everyone in...
AI Boosts Self‑Representation, Yet Courts Face Overload
AI is augmenting the ability of people to represent themselves in court. That's great from the perspective of access to justice. It also poses challenges for the courts. Speaking to judges in one court system recently, it was clear that...
Nvidia Backs AI Company Vast Data at $30 Billion Valuation
Vast Data announced a $1 billion Series F round that lifts its valuation to $30 billion, with Nvidia among the investors. The funding, led by Drive Capital and Access Industries, also includes Fidelity, NEA and other backers. The startup, which provides AI‑focused data‑management...
Transforming AI’s Energy Demand Into a Grid Asset
AI needs energy. A lot of it. But the more interesting question is what happens next. In this Earth Day episode of TEQ, Carolina Milanesi sits down with Christopher Wellise (@cwellise), VP of Sustainability at @Equinix, to unpack whether the infrastructure...
Agent-to-Agent Orchestration Guarantees Predictable, Compliant Workflows
Love that we are spending more time on agent to agent orchestration. This is not an easy task. Thomas Kurian said " We enable agent to agent orchestration, allowing agents to seamlessly delegate tasks from one to another, including support for complex, generative and deterministic orchestration patterns. This helps ensure that for your critical workflows, such as those that need compliance, your agents can follow well specified paths every single time, guaranteeing predictable outcomes. Orchestration flows and agents can respond to events. These events can be real time, they can be schedule based, they can be trigger based, or they can be batch inference job. " #GoogleCloudNext
Groundcover Adds Agentic AI Observability in Google Cloud
Groundcover has launched a generally‑available AI observability layer for Google Cloud that natively supports Vertex AI. The update requires no extra instrumentation and keeps all telemetry inside the customer’s own cloud environment. It adds full‑trace visibility across multi‑step agentic workflows...
Discord Group Hacks Anthropic Model by Guessing URL
Bloomberg reports a certain group got access to Mythos by guessing the URL for the new model. Guessing? Really? From Google: The unauthorized group is a private Discord channel of AI enthusiasts who specialize in tracking and testing unreleased large language models...
AI Decisions Demand Trusted Context, Not Blind Guesses
+1 to this Karthik Narain "But here's the reality, reasoning without context is just a guess, and when you expect your AI to make decisions and your agents to take actions, you cannot afford to guess trusted context, turns an intelligent gas to a decisive action." #GoogleCloudNext
AI Turns Weather Satellites Into High‑Resolution Ocean Current Maps
Researchers at UC San Diego’s Scripps Institution of Oceanography and UCLA introduced GOFLOW, an AI technique that transforms weather‑satellite thermal images into high‑resolution ocean‑current maps. Published in Nature Geoscience on April 22, 2026, the method captures currents as small as...
AI Art: Genuine Feeling or Clever Manipulation?
If you are emotionally moved by an AI generated song or video, is this a generative experience or are you simply being manipulated?
Ask Your RLM Questions for Upcoming MIT Explainer
Ask us your questions about Recursive Language Models (RLMs) & where LMs are underutilized or inefficient. We’ll pick some to feature in an upcoming explainer w/MIT PhD student Alex Zhang (@a1zhang), who recently developed RLMs. For more on his work: https://t.co/pFtCO03Hcq...
NVIDIA and Google Cloud Collaborate to Advance Agentic and Physical AI
NVIDIA and Google Cloud announced new AI infrastructure at Google Cloud Next, featuring Vera Rubin‑powered A5X bare‑metal instances and a suite of Blackwell GPUs that promise up to ten‑fold reductions in inference cost per token and massive scaling to nearly...
Real Hacker Reveals AI's Threats and Countermeasures
Talking with a real hacker will freak you out. Thanks @theonejvo for freaking me out about how AI could be used to attack everything in our modern society. And what we can do about it. https://t.co/0znpRiZ0tz
Cerebral Valley
in two weeks, we're going to figure out how people will talk to machines in this new world. absolutely stacked lineup of speakers and discussion group leaders for the Cerebral Valley Voice Summit https://t.co/H44Iinj6XW
Dreame Nebula NEXT Auto Teams with UC Berkeley to Accelerate AI‑Defined Vehicles
Dreame Nebula NEXT Auto has expanded its collaboration with the University of California, Berkeley, bringing engineers and researchers together to speed up AI‑defined vehicle development. The partnership focuses on safety, control and full‑stack AI integration as the company moves toward...
Use Claude Code to Vet Startup Ideas Like VCs
Before you tie yourself to your startup idea, have a conversation with Claude Code about it. It is so good in evaluating business opportunities. If you want to look at your business from the lens of a VC, you can also...
True Intelligence Goes Beyond Routing and Reactive Telemetry
"Routing messages is not the same as reasoning about them, connecting assets is not the same as understanding them, and reactive telemetry is not the same as choreographed intelligence." #AI #RiskManagement https://t.co/PYomJYHDkY
RE/MAX Agent Deploys AI‑Powered Real‑Estate GPTs Across Bucks and Montgomery Counties
Tim Dimmick, a RE/MAX broker in Souderton, Pennsylvania, has launched a suite of localized ChatGPT tools for Bucks and Montgomery County towns, delivering real‑time, community‑specific guidance to homebuyers and sellers. The AI assistants aim to boost client knowledge and streamline...
Governance Is the Overlooked Barrier in Enterprise AI
I've been closely following agentic technology. What is the one thing that is holding back enterprises...but most vendors don't discuss? Governance. @GoogleCloud is addressing governance head on with Agent Identity, Agent Gateway and Agent Observability. #CIO #Cloud #AI #GoogleCloudNext
Gemini Enterprise Projects Eliminate Context Pollution for Complex Tasks
Ok a new term for me - context pollution. @ThomasOrTK said " We're introducing Gemini enterprise projects. Projects give your agents a high fidelity workspace and the ability to use deep think to solve your most complex business challenges...
Gemini Embedding 2 Is Now Generally Available.
Google announced that Gemini Embedding 2 is now generally available through the Gemini API and Vertex AI. The multimodal embedding model, which handles text, images, video, and audio, moves from preview to production‑grade stability and performance. Early adopters used it for e‑commerce...
Engineers Supervising AI Become Key Differentiators, Says Pichai
The framing from @sundarpichai on engineers who supervise AI work goes beyond job-loss optics. It signals which skills will actually matter and positions Google’s engineers as genuine differentiators in an AI-first stack. https://t.co/9My9nlVayH
MIT CSAIL Pushes AI Reliability over Raw Power at ICLR
This week, MIT CSAIL will join other top ML researchers at ICLR to tackle a shift in focus from more powerful AI to more reliable systems 🧵 Our papers at the conference show how to potentially make AI models stronger critical...
Vodafone Business and Google Cloud Deploy AI Concierge and MDR Services for SMBs
Vodafone Business and Google Cloud announced two new SMB solutions—a managed detection and response security service and an AI Concierge powered by Gemini—marking the latest rollout in their $1 billion, ten‑year partnership. The offerings launch first in Germany, with AI Concierge...
Nvidia CEO Says AI Will Be a Permanent Micromanaging Boss Who Never Stops Nagging You
Nvidia CEO Jensen Huang told a Stanford panel that AI will act like a relentless, micromanaging boss, constantly nudging employees but dramatically raising productivity. He argued that this surge in efficiency will ultimately create more jobs than it eliminates, countering...
OpenAI Teams With Accenture and TCS to Speed Global Codex AI Rollout
OpenAI announced a partnership with consulting powerhouses Accenture and Tata Consultancy Services to fast‑track the global rollout of its Codex AI coding platform. The collaboration embeds OpenAI experts in client firms and leverages the consultants' integration capabilities, aiming to move...
Grant Thornton Survey Finds 79% of CRE Boards Back AI, Governance Lags
Grant Thornton released its 2026 AI Impact Survey, revealing that 79% of construction and real‑estate (CRE) boards have approved major AI investments, far above the cross‑industry average. The report warns that governance frameworks and incident‑response plans are far behind, creating...
Graphium Health Unveils Capture AI to Automate Patient Data Capture
Graphium Health introduced Capture AI, a generative‑AI platform that automates patient data capture and processing, promising to eliminate costly system integrations for hospitals and health networks. The launch targets a long‑standing bottleneck in digital health workflows.
Employment Hero Launches AI‑powered HeroForce in New Zealand to Cut $12 Bn SME Compliance Costs
Employment Hero rolled out its AI‑driven HeroForce platform in New Zealand, promising to automate payroll, HR and compliance for small‑to‑medium enterprises and address an estimated $12 billion in duplicate administration costs. The launch marks the first large‑scale co‑employment model in the country,...
Self‑Hosted AI Agents Keep Data Private and Always On
AI agents that run inside your own setup rather than someone else’s cloud is the unlock. Your data stays with you, a full team of agents works 24/7, and it pulls from your internal docs plus places like Reddit and...
Dreame Unveils Nexis AI‑Powered Smart Swim Goggles in Silicon Valley
Dreame AI Glasses Division launched the Nexis smart swim goggles at a Silicon Valley showcase, debuting a MicroLED‑based HyperView HD display, AI‑driven coaching and a 12 mm ultra‑thin comfort fit. The product aims to create a data‑rich underwater experience and expand...
Blackbaud Q2 2025 Beats Forecast, Citing Strong Nonprofit and Education Software Growth
Blackbaud reported $281 million in organic revenue for Q2 2025, a 6.8% year‑over‑year increase, propelled by higher transaction‑solution sales and expanding AI‑driven features for nonprofits and schools. The company lifted full‑year guidance and highlighted a 40% larger contract‑renewal cohort, underscoring momentum in...