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
OpenAI and Anthropic Just Met with Religious Leaders at the ‘Faith-AI Covenant’. Here’s Why
Tech giants OpenAI and Anthropic joined leading faith representatives at the inaugural Faith‑AI Covenant roundtable in New York, organized by the Interfaith Alliance for Safer Communities. The meeting aimed to draft cross‑faith moral guidelines for rapidly evolving artificial‑intelligence systems. Participants ranged from Hindu, Buddhist, Sikh and Christian groups to the Mormon church, with plans for follow‑up sessions in Beijing, Nairobi and Abu Dhabi. While companies tout the partnership as a path to responsible AI, critics warn it could sideline deeper regulatory debates.
The Human Factor: Why AI-Powered SOCs Still Need People in Charge
Artificial intelligence is reshaping security operations centers (SOCs), especially when paired with managed detection and response (MDR) services. Sophos’ white paper shows its AI agents can trim alert noise by more than 60% and cut investigation time by up to...
Kodiak AI, General Dynamics Land Systems Partner for Military Vehicles
Kodiak AI and General Dynamics Land Systems have formed a strategic partnership to build autonomous ground vehicles for defence. Kodiak will supply its Physical AI "Kodiak Driver" software while General Dynamics handles vehicle integration, power and communications. The joint effort...
BACH Ensures Consistent Characters Across AI Video Shots
Every AI video creator has this nightmare: You generate a character with one prompt… but in the next shot, she’s a different person. Different face. Different skin. Different eyes. You can’t tell a story like that. BACH — the world’s first...
AI Curation Prunes Legacy Code, Boosts Academic Review
There is a lot of press on the enshittification of the internet via AI, but another dynamic fueled by LLMs is AI-curation. Recently, Linus Torvalds authorized the deletion of 138K lines of code in the Linux kernel. The reason: dealing...
Sam's Top AI Recommendations
Sam Arbesman revisits his favorite AI pieces as Riskgaming hits its 500th issue, highlighting five thought‑provoking selections. He showcases Benjamin Breen’s classroom experiments where large‑language‑model “history simulators” spark student creativity despite hallucinations. He cites Sara Tillinger Wolkenfeld’s warning that delegating creative work...
Engineers Evolve Into AI Orchestrators, Not Coders
Anthropic engineers reportedly haven’t handwritten code in 3 months. That sounds insane until you see how the best teams are starting to work. They’re not sitting there typing every line anymore. They’re running multiple agents in parallel, reviewing the work, giving feedback,...
Future Builders Outpace Outdated Governance Systems
Mars colonies, robot workers, and AGI lawsuits all point to one big truth: the people building the future are moving faster than the systems meant to govern them. Will these systems be relevant or needed at all soon? -- Figure scaling...
How AI Can Enhance Benefit Leader, Broker Relationships
Gyde, a brokerage‑focused AI firm, launched its GydeOS platform and Gia assistant to shift benefit brokers from pure administrative tasks to proactive healthcare navigation. The tools automate calendar management, product recommendations, outreach, and renewal handling, freeing brokers to engage in...
The Scoop: Updated Data Gives New Peek Into What’s Fueling AI Search Today
Muck Rack’s latest analysis reveals that 99% of links cited by AI chatbots come from non‑paid media, with journalism leading at 27% of all citations and more than half of those stories published within the past year. Corporate blogs contribute...
Public-Sector AI Translation Has Saved Taxpayers More than $30 Million, Wordly Says
Wordly, an AI‑powered translation platform, reports that public‑sector use has saved U.S. taxpayers more than $30 million since its 2019 launch. Adoption by states, local governments and schools has surged fivefold in the past two years, driven by growing multilingual populations...
Oracle Applications Analyst Summit: Fusion Agents Are Shifting From Assist To Decide
At Oracle’s Applications Analyst Summit, the company announced that AI‑driven agents will be exclusive to its Fusion SaaS platform, abandoning on‑premise suites such as EBS, JD Edwards, PeopleSoft, and Hyperion. Oracle is moving from generic assistants to productized agentic applications that...
AI in Edu: News, Views, & Moves (May 8)
Education Week identifies a cultural and political backlash against ed‑tech, driven by pandemic‑era device rollouts, the rise of generative AI, and concerns over student mental health. The backlash is reflected in mixed educator sentiment—some see academic value, others cite harms...
NSF Initiative Aims for AI-Readiness
The National Science Foundation launched the AI‑Ready America initiative to build a nationwide network of AI coordination hubs, one in each state, territory and the District of Columbia. Up to 56 hubs will be eligible for up to $1 million per...
The AI Hype in Healthcare Needs Randomized Evidence
The article argues that the AI hype in healthcare must be grounded in randomized evidence. Venture capital is flooding AI decision‑support tools, yet most lack RCT validation. Two recent trials illustrate the gap: a 10,422‑patient UVA predictive‑analytics study found no...
AI Agents Redefine Leadership and Decision-Making
AI Agents Are Already Reshaping Business Leadership and Decision Making AI agents aren’t just tools — they’re influencing how leaders make decisions, streamline strategy and drive organisational change. Read more 👉 https://lnkd.in/e_pCciEv #AI #Leadership #BusinessStrategy #BernardMarr
Reusable Prompt Frameworks Beat Random Guesswork
Most people fail with AI because they write random prompts. The best results come from reusable frameworks 👇 🧩 1. Structured Template “Act as a [role]. Help me with [task]. Output as [format].” → Best for precision & consistency 🔎...
‘AI for Good’ and ‘I Am Not a Robot’: Learning the Machines
The new book *AI for Good* spotlights pragmatic leaders who are applying artificial intelligence to solve immediate, high‑impact problems. It profiles Cleveland Clinic’s Rita Pappas, who used AI to streamline patient admissions, and retired Army General Gustave Perna, who built...
5 Steps for Frontier AI Readiness
Frontier AI models such as Anthropic's Claude Mythos and OpenAI's GPT‑5.4‑Cyber can autonomously discover software flaws and generate exploits, compressing the gap between vulnerability discovery and attack. This acceleration forces a move from traditional vulnerability management to continuous exposure management,...
Ai Twitter Is Trash - Keep Your AI Skills Up The Right Way - #130
The post argues that most AI content on social media quickly becomes obsolete, echoing Andrej Karpathy’s claim that 90% of AI Twitter advice dies within six months. It lists ten practices senior AI engineers have abandoned, starting with AutoGen/AG2, which...
How AI Is Helping Legacy HVAC Systems Meet New York City’s Emissions Standards
Entech is retrofitting legacy HVAC systems with AI‑driven sensors and control software to help New York City buildings meet Local Law 97’s emissions mandates. Early adopters have already saved about $42 million and cut carbon output by roughly 137,000 tons, averaging a 22%...
AR Automation: Cutting Through the Hype
Bob Schultz talks with Chris Capron, the founder of Sephora Software, about the rapid shift in accounts receivable (AR) automation driven by AI. Capron argues that legacy vendors are nervous because AI tools like Claude and Vibe coding can replace...
How Silicon Valley Sold Washington an AI Race
Silicon Valley has turned the China‑AI race narrative into a policy lever, using full‑page ads, white papers and congressional testimony to push for export controls, military AI contracts and deregulation. The story began in 2017 after China’s AI Development Plan,...
AI Influencers Erode Trust by Masquerading as Real Advisors
AI influencers are starting to sell fitness, health and lifestyle advice that looks completely real. The problem is not just fake images. It’s that people increasingly cannot tell whether the person giving advice even exists. That changes the internet fundamentally. Trust is...
Nick Bostrom Has a Plan for Humanity’s ‘Big Retirement’
Philosopher Nick Bostrom released a new paper arguing that a low‑probability AI catastrophe could be acceptable if advanced AI ultimately eliminates humanity’s existential threats and creates a “solved world.” The proposal marks a departure from his earlier warnings about AI...
Is AI Going To Take Retail Jobs? Maybe, But Not The Ones You Expect
Andon Labs launched Andon Market, an experimental retail store run by an AI chatbot named Luna. Luna immediately hired human staff to handle customer interaction, positioning the AI as a manager rather than a floor worker. The experiment highlights a...
Urgent Care Clinics Boost Revenue and Throughput with AI Scribe
Experity Health’s AI Scribe, embedded in its urgent‑care EHR, automatically transcribes provider‑patient conversations into structured notes. Since its rollout in October 2025, more than 450 clinics have logged over 400,000 visits, cutting daily documentation time by roughly 40%. The efficiency...
AI-Guided Labs Are Approaching Full Autonomy
Ginkgo Bioworks is building fully autonomous biopharma labs that combine 70 robots, 90 lab devices and AI‑driven software in an 18,000‑square‑foot Boston facility. Scientists can submit dozens of unique protocols each day, which the system executes without human intervention. In...
Schematik
Schematik, backed by Lightspeed Venture Partners in a 2026 pre‑seed round, is launching one of the first generative AI platforms that translates plain‑English design requests into hardware code, wiring diagrams, component lists, and assembly instructions. Founder Sam Beek, who previously...
Signal Labs
Signal Labs is creating an AI‑first "attention infrastructure" that filters and prioritizes signals for large, multi‑party organizations. Its inaugural product, SignalOS, tackles the fragmented U.S. healthcare payment stack by automating claim, authorization and clinical data coordination. Founder Raj Ronanki, former...
Abridge Releases Ambient AI Tech for Nurses
Abridge announced that its ambient documentation AI is now available to nurses at every health‑system client nationwide. The technology automatically captures and records patient‑encounter notes, aiming to cut the administrative load that nurses traditionally shoulder. Early adopters include Mayo Clinic,...
Automating Grunt Work Is Still Agentic AI's Sweet Spot
At SAS’s annual conference, financial‑services leaders emphasized that the near‑term value of agentic AI lies in automating high‑volume, low‑complexity back‑office work such as KYC onboarding, compliance documentation, and routine claims processing. JPMorgan Chase adopts a hybrid model—purchasing external AI models while...
Generative AI Hype Outpaces Real ROI for Businesses
Remember that MIT study that showed that the ROI for generative AI wasn’t really there for most businesses? Or any of the six or seven studies from other teams that followed, showing basically the same? The situation for agents looks to...
Franz AllegroGraph 9.0 Debuts GraphTalker, an AI Agent for Enterprise Knowledge Graphs
Franz Inc. launched AllegroGraph v9.0, featuring GraphTalker, an AI agent that lets users query enterprise knowledge graphs via natural language. GraphTalker uses an iterative, schema‑aware approach to generate, test, and refine SPARQL queries, delivering verified answers and reusable workflows. The platform...
Three Wins for AI in the Revenue Cycle
Brian Kenah, CTO of EnableComp, outlines three early AI wins in hospital revenue cycle management. First, document intelligence converts unstructured PDFs—contracts, fee schedules, regulations—into real‑time decision data for reimbursement. Second, intelligent integration leverages APIs and robotic process automation to replace...
Willis Knighton Health Taps AI Scribe Vendor
Willis Knighton Health, a four‑hospital network in Shreveport, Louisiana, has chosen Commure as its enterprise ambient AI scribe vendor. The partnership integrates Commure’s AI‑driven documentation tool with the system’s Meditech Expanse EHR. A seven‑specialty pilot showed 88% of providers experienced...
When Well-Behaved Agents Trigger Disaster
A new class of outages is emerging as autonomous agents make rapid, correct decisions that unintentionally collide. In a 2:17 a.m. scenario, a performance, cost, and routing agent each acted perfectly, yet their combined actions took the database layer offline. Recent...
Garry Tan
Garry Tan’s custom instructions are based. It pushes any LLM past half-finished bullshit and into actually useful answers. Add this to your SOUL md:
Public AI Isn’t the Frontier
The blog argues that the era of treating AI like a standard product launch is ending. While early models were built, tested internally, announced, and released with minimal oversight, today’s most powerful frontier models are being treated as systems that...
One Developer, AI Agents Replace Whole Teams
We're entering a new era of software where a single person, working with AI agents, can build products that previously required entire teams. In this episode of the @LightconePod, they break down the rise of AI coding agents, "tokenmaxxing", and the...
AI Struggles with Basic Data Tasks for Hospital Administrators: Study
A study by Mount Sinai and Mayo Clinic evaluated nine large language models on two basic administrative data tasks using 50,000 emergency department records. Simple prompts like “how many patients were admitted?” yielded poor accuracy across all models. Adding chain‑of‑thought...
Human Value Shifts to AI Supervision and Coordination
If AI agents handle the front line, attention shifts to the layers above. Human roles that manage escalations, guide decisions, and oversee workflows take on greater weight. Beyond that, new roles emerge to coordinate systems where people and agents operate...
Data Labeling Becomes U.S.’s Fastest‑growing AI Job
I've spent years interviewing data workers who are the lifeblood of the AI industry in countries like Kenya, Colombia, and the Philippines. Now it's the #4 fastest growing job in the U.S. AI is coming for your job, and it's not in...
How to Go From AI Beginner to Pro in 2026
The post outlines a five‑stage framework for advancing from AI beginner to system builder in 2026, highlighting the widening gap between casual users and power users. It details tool selection, prompt engineering, and the shift toward AI‑driven workflows and autonomous...
Sony Lumps Machine Learning and Generative AI Together
Seen some debate on what Sony means when they say AI. They're essentially rolling machine learning and generative AI into one heading. In other words, the use cases they're talking about cover both areas.
AI Disruption Mirrors China WTO Shock,
The AI shock is following a similar disruption pattern that followed China’s WTO entry, says Apollo’s Slok. Back then manufacturing workers lost their jobs, but overall employment levels held up. This time around, white-collar workers are being displaced https://t.co/7QVCrpA4If https://t.co/weF22s1CNV
DBS Group: Putting AI Into The Bank’s DNA
DBS Group reported record financial results in 2025, with total income of S$22.9 bn (≈$18 bn) and profit before tax of S$13.1 bn (≈$9.7 bn), driving its market capitalization past $100 bn and closing the year at $124 bn. The bank scaled its AI engine to...
AI-Generated Hospitalization Summaries Cut Clerk Workload, Boost Physician Well‑being
Agentic AI summaries of hospitalizations were safe and reduced data clerk burden, burnout, and improved sense of well-being for physicians in a prospective study https://t.co/0nXZ0wjugg
CIOs Must Prioritize Plugin Security Over AI Hype
A personal agent is only as safe as its plugins. In the wild, we're already seeing extensions act as conduits for malware. CIOs: It’s time to move from "opportunity" to "risk mitigation." 📉 https://t.co/5poCN8at4u #CIO #AI #Cybersecurity #RiskManagement #InfoSec
Agentic AI Extends Beyond Telco IT, Shaping Future
Last week I MC’d the invite-only Amdocs Executive Forum in Madrid. Themed “The Future is Now”, focused on #AgenticAI and how it reaches beyond #telcos’ IT systems… see post for details https://t.co/6rahwjFyBz