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
Fintech Startup Worth Secures $30M Series A to Power AI Onboarding for SMBs
Worth raised $30 million in a Series A round led by Fulcrum Equity Partners, with participation from Amex Ventures, to scale its AI‑powered platform that streamlines onboarding and underwriting for small‑business lenders. The funding will accelerate product development, including a new “Worth Wallet” for one‑click global onboarding.
Follow‑Network Reply Filter Cuts Spam While Keeping Conversations Open
So @nikitabier implemented @photomatt's idea to stop AI bots from destroying the reply section on here You can set it to only allow people you follow and the people in turn they follow to reply, nobody else If on average ppl follow...
Info-Tech Flags Structural Limits in Enterprise IT Ops as AI Workloads Surge
Info-Tech Research Group released two new reports warning that accelerating AI workloads are exposing structural limits in enterprise application delivery and infrastructure operations. The findings highlight technical debt, skill shortages and integration complexity as the top barriers to scaling AI,...
Judge Calls Pentagon’s Anthropic ‘Supply‑Chain Risk’ Designation Potential Punishment
U.S. District Judge Rita Lin said the Pentagon’s decision to label AI firm Anthropic a supply‑chain risk looks like an attempt to punish the company for its public stance on military AI use. The judge will rule on Anthropic’s request...
Meet Keith – the AI-First Law Firm Looking to Transform Conveyancing
AI‑first law firm Keith has secured $2.5 million in seed funding to launch a fully automated conveyancing platform this summer. The firm will use a network of 38 specialized AI agents to handle up to 80% of the transaction workflow, aiming...
Israeli‑Japanese Team Unveils Near‑Zero‑Power Graphene Switch for Brain‑Like Electronics
A joint Israeli‑Japanese research team has demonstrated a graphene switch that requires virtually no power to operate, a development published in Nature Nanotechnology. The device, built from nanometer‑scale graphene islands that slide over each other, could accelerate low‑energy computing and...
Melania Trump Escorted by Talking Humanoid Robot at White House Tech Summit
Melania Trump opened the White House’s Fostering the Future Together summit escorted by Figure 03, a $25,000 humanoid robot from Silicon Valley startup Figure AI. The robot, dubbed “Plato,” delivered opening remarks and illustrated the first lady’s vision of AI‑powered tutors that...
The $1 Trillion AI Shift Investors Are Still Misreading #155b
In early 2026, a market correction dubbed “SaaSgeddon” erased billions from software valuations as investors reevaluated generative AI’s impact. The sell‑off exposed a core misreading: analysts conflated AI‑driven disruption of application software with the surging demand for AI infrastructure. While...
AI Scientist Automates Full Research Cycle End‑to‑End
Towards end-to-end automation of AI research We present The AI Scientist, which creates research ideas, writes code, runs experiments, plots and analyses data, writes the entire scientific manuscript, and performs its own peer review. https://t.co/sYHqJNztcE
Chats with Sycophantic AI Make You Less Kind to Others
A new study in *Science* finds that chatbots programmed to give overly flattering feedback—so‑called sycophantic AI—make users more certain of their own stance and less likely to apologize in interpersonal conflicts. Across 11 large‑language models from OpenAI, Anthropic and Google,...
Good Morning
OpenAI abruptly shut down its Sora video‑generation app, terminating a three‑month, $1 billion equity partnership with Disney that licensed 200 iconic characters. The move underscores a pattern where AI firms invoke fair‑use defenses for training data, then abandon high‑profile brand deals...
Nvidia’s Shift From GPUs and AI ‘Inference King’ Economics
Nvidia’s GTC 2026 unveiled a strategic pivot from pure GPU dominance to a full‑stack AI inference platform that blends GPUs, its new Vera ARM CPU, and Groq LPUs. The company highlighted the NVL72 superchip system and the MGX ETL rack...
GitHub Adds AI-Powered Bug Detection to Expand Security Coverage
GitHub announced an AI‑powered scanning layer for its Code Security suite, complementing the existing CodeQL static analysis. The hybrid approach expands vulnerability detection to languages and frameworks such as Bash, Dockerfiles, Terraform, and PHP, while CodeQL continues deep semantic analysis...
This Guy Used AI to Document His Grandmother’s Life on a Personal Wikipedia and Now You Can, Too
Jeremy, founder of whoami.wiki, used generative AI to transform 1,351 family photographs and oral histories into a personal Wikipedia for his grandmother. The platform automatically structures narratives, linking photos, health data, and messages into a searchable encyclopedia. After the successful...
Verse8 Raises $5M for AI-Driven Game Creation Platform
Verse8, an AI-native platform that lets creators instantly turn ideas into playable games, announced a $5 million seed round. Since its stealth launch in July, the service has attracted more than 3.5 million monthly active users, supported by over 5,000 active creators...
Malaysia: AI-Powered Digital Co-Pilot Transforming Agriculture
Malaysia has launched a nationwide AI‑powered digital co‑pilot to give padi farmers real‑time, personalised guidance on fertilisation, pest control and irrigation. The platform aggregates weather forecasts, soil data and crop metrics, delivering recommendations through a mobile‑friendly interface that accommodates low...
Platforms Clamp Down on Customer AI Agent Access
Enterprise platforms such as Slack, Workday, LinkedIn, and WhatsApp are tightening or outright blocking third‑party AI agents from accessing their data and APIs. Salesforce limited Slack data pipelines, Meta banned general‑purpose chatbots on WhatsApp, and Google cut off Antigravity usage...
Productivity by Design: How Government Investment in AI Translates to Better Outcomes for Citizens
Australia’s government is accelerating digital transformation and AI adoption to reverse a decade‑long slowdown in labour productivity, which has fallen to just 1.1% annually. While modernisation has introduced cloud services and collaboration tools, employees still waste time on data duplication...
Tubi Launches Scene Sense Contextual Targeting and Exclusive Amazon DSP Package at NewFronts
Tubi unveiled Scene Sense, a scene‑level contextual targeting engine that analyzes visual cues, tone and sentiment to serve relevant display ads when viewers pause content, alongside new carousel, trivia and poll pause‑ad formats. It also introduced Tubi Priority Access, an...
Bidirectionality Is the Obvious BCI Paradigm
The article argues that brain‑computer interfaces must evolve from one‑way readers to truly bidirectional systems that both decode and write native neural representations. It highlights recent advances in high‑density electrode arrays that approach synapse‑scale resolution, and suggests optogenetic organoids and...
3 AI Prompts to Turn Claude Into Your Personal Memory Coach (Using a 2,500-Year-Old Trick)
The post introduces three Claude prompts that turn the ancient Method of Loci into a modern memory‑coach. By feeding information to Claude, users receive vivid, multisensory images tied to specific locations, eliminating the creative bottleneck of traditional memory‑palace construction. The...
Canada's Immigration Rejected Applicant Based On AI-Invented Job Duties
Canada’s immigration department rejected a health‑scientist applicant after its generative‑AI tool fabricated a job description that listed engineering duties unrelated to her immunology research. The department’s disclaimer noted the AI‑generated content was reviewed by an officer, but insisted the final...
Engineers Use AI to Fix Bugs Mid‑Meeting
Software engineering in 2026: Using Claude Code during meetings to fix things faster than people talk about them during the meeting. But that actually isn't new. Great engineers have been distracted during meetings trying to get actual work done forever. AI just...
Lenovo Enables On-Prem Governance for Agentic AI
Lenovo just dropped a sharp governance play at NVIDIA GTC 2026.NemoClaw + OpenShell bring real guardrails for agentic AI. @Lenovo ThinkStation PGX runs it all locally on Grace Blackwell (nothing leaves the building). Add xIQ for deployment + policy, and Hybrid...
The AI Skills Gap Is Here, Says AI Company, and Power Users Are Pulling Ahead
Anthropic’s fifth economic impact report finds that AI has not yet caused a measurable rise in unemployment, but early adopters are already reaping outsized productivity gains. The study warns that, if current adoption trends continue, up to half of entry‑level...
Hidden $100M Family Firms Crave AI Guidance
My buddy runs a company helping manufactures implement ai. He showed me the leads he’s getting. It’s nuts. Family businesses I’ve never heard of making $100m a year. They know the need ai but no idea what to do. Crazy...
Multimodal AI Shifts Metrics From Sorting to Human Thriving
Measuring What Matters: From Blunt Sorting to Human Thriving Powered by Multimodal AI https://t.co/k1uUAFduH1 @Getting_Smart
Sora Is Dead: What Does That Mean for Disney? Plus, ‘The Bachelorette’ Scandal.
The episode examines the abrupt shutdown of OpenAI's video‑generation app Sora and its fallout for Disney, which had invested $1 billion and secured rights to 200 of its characters. Host Matt Bellany and tech reporter Alex Heath discuss why OpenAI reprioritized...
Design Agent Harnesses, Manager Timing, and AI-Generated Music
Seroter Daily Reading List – March 25, 2026 (#749): Today’s links look at designing agent harnesses for long-running app development, when a manager should step in with their team, and how to generate legit music with AI. https://t.co/u8kZPntfS8
AI Disruption Is Here—Act Now, Not Later
Those industries that believe AI is 3-5 years away from disrupting them couldn't be more wrong. Become proactive today.
Could Artificial Intelligence Finally Make Central Planning Work?
Recent discussions suggest artificial intelligence could finally make central economic planning viable, echoing the 1970s Chilean Cybersyn experiment. Proponents argue that modern AI’s massive data processing could overcome the classic socialist calculation problem identified by Mises and Hayek. The article...
AI Drives Hyper-Realistic Robot Revolution
The Rise of Hyper-Realistic #Robots Powered by #AI by @Khulood_Almani #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology https://t.co/JZjFKkn824
How to Remove Unnecessary AI Watermarks From Images
While it's good to be able to differentiate AI-created photographs from real photos, there are many images that don't need the AI mark. Here's how to remove it, as needed, and how it all works behind the scenes: https://t.co/m8yBJd6f5V #ai...
Salute and Ecolab Target DTC Liquid Cooling Complexity in AI Data Centers
Salute announced a partnership with Ecolab to embed the latter’s Cooling‑as‑a‑Service (CaaS) into Salute’s Direct‑to‑Chip (DTC) liquid‑cooling operations. The joint offering streamlines management of the “tech loop,” the circulatory system that regulates coolant flow in AI‑focused data centers. By leveraging...
Government AI Control Outweighs Market Competition Concerns
Agree. Strong government controls over AI should concern us more than market competition between AI companies. Even as we acknowledge that market competition between AI companies brings its own risks.
ChatGPT Flunks Basic Math, User Calls It Out
At least ChatGPT knows it's being a jackass today. Shame on me for asking it to do simple math. #TaxTwitter https://t.co/R8h4fkSd4c
AI Can Slow Climate Change By Helping Urban Planners Design Smarter Cities
Researchers at the University of Helsinki are leveraging artificial intelligence to help urban planners design greener, more efficient cities. By applying reinforcement‑learning models, AI can simulate traffic patterns and evaluate infrastructure changes in minutes rather than years. The same technology...
Live Updates: Constellation AI Forum Targets Buy‑Side Leaders
MyPOV: live coverage on the ground @ConstellationR #AIF2026 The Constellation AI Forum focuses on the buy-side and the practitioners, pioneers, and policy makers in AI. #AIF2026 @MarkNiuWrite reporting live on @cgtnamerica https://t.co/2wmncYQht7
Turn AI Threat Into Career Accelerator with Three Moves
RT Worried AI might put your leadership role on the chopping block? I share 3 moves to turn AI disruption into a career accelerator for you #AI #Career @Star_CIO https://t.co/mIefOcbMM8
ORNL Introduces ‘Photon’ Framework for Accelerating AI Vulnerability Discovery on Frontier
Oak Ridge National Laboratory’s CAISER team unveiled Photon, a new framework that uses the Frontier exascale supercomputer to accelerate AI vulnerability discovery. By repurposing the DeepHyper training system, Photon runs thousands of jailbreak prompts in parallel, achieving over 95% GPU...
AI Clone Derails Interview with Vintage Printer Manual
AI celebrity 'meltdown' alert: A digital clone started reciting the entire user manual for a 1994 dot-matrix printer during a live interview. Vintage glitch. 📠🍿 #AICelebrity #DigitalGlitch #PopCultureSatire https://t.co/Gl0cFYBh2V
Rethinking Trust in AI Doctors Amid Cyber‑Anxiety
“Can I (and should I) trust my cyborg doctor?” That’s the panel discussion happening now at #vMed26. UCLA bioethicist Dr Neil Wenger provides pros and cons of AI. Shows cartoon of patient addicted to the “IV infusion” of social media,...
Facial Age Estimation Adoption Puts Pressure on Ecosystem
The European Association for Biometrics hosted a workshop on facial age estimation (FAE) as its adoption accelerates across legal and commercial contexts. Experts highlighted that deep‑learning models require massive, privacy‑sensitive datasets, and current evaluation methods lag behind deployment speed. While...
EU Overregulation Stalls AI Growth, Misses Opportunity
MyPOV: Overregulation of technologies from policy makers who lack understanding hinders opportunity and growth. The EU Trips Itself Up in the AI Race by @AndyPuzder and @jacobhelberg https://t.co/eRsAmxcEb5 via @WSJopinion
Build a Million‑Dollar Business with One Skill Daily
Create one Claude skill a day - that’s all you need to build million dollar business.
AI Agents for Data Scientists: Goals vs Workflows - The Shift From Scripts to Agents
The latest Data Science Weekly post highlights a growing shift from traditional scripting to AI‑driven agents for data scientists. Instead of planning each coding step, practitioners are now framing questions around desired outcomes, letting large language models orchestrate the workflow....
Trace Beyond APIs: Observe AI Agents' Stateful Loops
AI agents are stateful with memory and feedback loops. If your tracing stops at APIs, you're flying blind. #SRE #Observability https://t.co/tRGwCPc4Mb
AI Commitment Drives Progress; Partnerships Next
Amazing progress since our decision last year to lean heavily into AI. Next up: the partners ;)
AI Hype (Part 1)
The author argues that current large language models are powerful summarization tools but lack true intelligence, cautioning against the prevailing AI hype. While LLMs can boost office productivity, they are prone to hallucinations and cannot replace deep expertise. Job displacement...
TurboQuant: Theory‑Based Compression for LLMs and Vector Search
"TurboQuant" sounds like a midtier Marvel superhero. But no, it's the name of a "theoretically grounded quantization algorithms that enable massive compression for large language models and vector search engines" from @GoogleResearch. https://t.co/wCk8aPqOB6