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
How AI Is Being Used in Dentistry
Artificial intelligence is rapidly reshaping UK dentistry, with roughly 40% of practices already deploying AI tools. Chat‑powered virtual receptionists now handle bookings and patient queries, cutting response times from hours to seconds and slashing missed appointments. Clinical AI analyzes X‑rays and scans, flagging decay, infection or gum disease with near‑human accuracy while also generating realistic before‑and‑after visualisations for cosmetic procedures. These technologies free dentists from routine admin, improve diagnostic confidence and are reported to lift patient volumes by as much as 55%.
Chrome's New Tab Shifts Toward AI‑First Experience
The 'New tab' page in Chrome is looking more and more like Gemini... Check those options for Deep Search and Create images right under the search box. And those take you directly to AI Mode after entering your prompt. Again,...
Unauthorized Users Reportedly Gain Access to Anthropic’s Mythos AI Model
Anthropic’s Claude Mythos, a frontier AI model for vulnerability detection, was accessed by an unauthorized Discord‑based group on the day it was announced. The group guessed the model’s online location and exploited a third‑party vendor’s integration to gain unrestricted access. Anthropic...
Bruce Interviewed About Agentic AI
Bruce sat down with Sidebar Advisors founder Niraj Chhabra to discuss the rise of agentic AI and its looming impact on law firms. The interview, now available on Spotify, Apple Podcasts, and Buzzsprout, delves into how autonomous AI systems could reshape...
ChatGPT Images 2.0 Is a Generational Leap
OpenAI unveiled ChatGPT Images 2.0, a major upgrade to its image generation model that can produce high‑resolution graphics with accurate typography, dense copy, and UI elements. The system handles layouts up to 2 K resolution, supports varied aspect ratios, and for Plus/Pro...
Gemini Can Now Look Through Your Google Photos to Create AI Images, with a Few Caveats
Google has added a new Gemini feature that can scan a user’s Google Photos library to supply reference images for AI‑generated pictures, leveraging the Nano Banana 2 image model. The capability is part of Google’s Personal Intelligence push, which links Gemini...
AI‑written Emails Erode Trust in Every Message
Knowing that any email I get might be researched, drafted, and sent by an AI seems to diminish the value I assign to _all_ the emails I am getting. Just realized that while browsing through my inbox just now. This likely isn’t...
Oracle Expands Google Cloud Partnership with Natural Language Database Agent
Oracle announced the Oracle Artificial Intelligence Database Agent for Gemini Enterprise, a natural‑language interface that lets users query Oracle Autonomous Database on Google Cloud without writing SQL. The agent is offered through Google Cloud Marketplace and will be free for...
10 GitHub Repositories To Master Claude Code
Claude Code has quickly become a leading agentic coding assistant, capable of reading codebases, editing files, and running terminal commands. To unlock its full potential, developers are turning to a curated set of GitHub repositories that provide structured agents, custom...
Google Announces Innovations in Mega-Scale Networking for the Agentic Era
Google unveiled Virgo Network, a mega‑scale datacenter fabric that interconnects up to 134,000 chips and delivers 47 petabits per second of bi‑directional bandwidth. The architecture flattens network layers, giving AI accelerators more than four times the prior‑generation bandwidth and sub‑millisecond telemetry...
Google Delivers Connective Tissue for Autonomous AI Agents to Access Data without Restrictions
Google Cloud unveiled the Agentic Data Cloud at Cloud Next 2026, a platform that transforms static data warehouses into a dynamic reasoning engine for autonomous AI agents. The offering combines a universal context engine, agent‑first developer tools, and a cross‑cloud...
With Gemini Enterprise Agent Platform, Google Brings Agentic Development and Control Under One Roof
Google Cloud unveiled the Gemini Enterprise Agent Platform at Cloud Next 2026, consolidating Vertex AI’s tools into a single hub for building, scaling, governing, and optimizing AI agents. The platform introduces Agent Studio’s low‑code interface, an upgraded Agent Development Kit...
Data Readiness Determines AI ROI Success
#Ad Your AI strategy is only as strong as the data behind it. Cloudera’s latest report, The Data Readiness Index 2026, shows why data access, governance, quality and infrastructure have become critical factors in determining whether AI initiatives deliver real...
Layer 1A Is Table Stakes. The Real AI Infrastructure Question Is Above It.
Enterprises migrating AI workloads discover that the real lock‑in lies not in storage (Layer 1A) but in the reasoning and context layers above it, which the author calls “borrowed judgment.” Google announced a Knowledge Catalog that productizes this non‑portable layer, using...
On Capitol Hill, a Debate over Who AI at Work Is Really Working For
The House Education and Workforce Subcommittee held its sixth hearing on AI in the workplace, pitting employer advocates who want a federal override of state AI‑employment laws against worker groups demanding state‑level protections. Witnesses from the CHRO Association, Ogletree Deakins,...
On Capitol Hill, a Debate over Who AI at Work Is Really Working For
The House Education and Workforce Subcommittee held its sixth hearing on artificial intelligence in the workplace, chaired by Rep. Ryan McKenzie. Titled “Building an AI‑Ready America,” the session examined AI’s economic impact on both workers and employers. Lawmakers heard from CEOs,...
Google Maps Is About to Get a Big Dose of AI
Google announced a suite of generative AI tools for its Maps and Earth platforms, targeting enterprise customers. The new Maps Imagery Grounding feature lets users generate realistic Street View scenes from text prompts, while Aerial and Satellite Insights accelerates satellite‑image...
Cloud Next ‘26: Momentum and Innovation at Google Scale
Google Cloud announced record‑breaking AI throughput, processing 16 billion tokens per minute—up 60% from the previous quarter—and earmarked over half of its 2026 machine‑learning compute budget for cloud services. The company unveiled Gemini Enterprise Agent Platform, reporting 40% quarter‑over‑quarter growth in...
Google Cloud Next ‘26
Google Cloud Next ‘26 showcased a sweeping AI‑centric roadmap, highlighting that 75% of its customers now run AI products and 330 firms processed over a trillion tokens in the last year. Token processing surged to 16 billion per minute, a 60% jump...
Google’s Gemini Can Now Run on a Single Air-Gapped Server — and Vanish when You Pull the Plug
Cirrascale Cloud Services has deepened its partnership with Google Cloud to ship the Gemini model on‑premises via Google Distributed Cloud. The solution bundles a Dell‑certified appliance equipped with eight Nvidia GPUs and confidential‑computing safeguards, allowing a fully private, air‑gapped deployment....
Google Cloud Launches US$750M Agentic AI Partner Fund at Cloud Next
Google Cloud announced a $750 million Agentic AI Partner Fund at Cloud Next 2026, aimed at accelerating AI value identification, prototyping, and deployment across its global partner network. The program will provide resources, Gemini Enterprise tools, and embedded forward‑deployed engineers to...
SONAR Sitrep: AI Data Center Build-Out Fuels US Freight Surge
The AI data‑center construction boom is becoming the largest privately‑funded infrastructure effort in U.S. history, spending the equivalent of the Interstate Highway System’s annual budget every five weeks. Between 2026 and 2031, the build‑out will generate roughly 18.9 million additional truckloads...
Google Cloud Announces Eighth-Generation TPUs, Boasting AI Training and Inference Leaps
Google Cloud unveiled its eighth‑generation tensor processing units, the TPU 8t for training and TPU 8i for inference. The 8t chip is 2.8× faster than the prior Ironwood generation, delivering 121 FP4 exaflops and supporting pods of up to 9,600 chips. The 8i...
CoreWeave Expands Multi-Cloud AI Stack at Google Cloud Next
CoreWeave announced a suite of multi‑cloud AI tools at Google Cloud Next, including a private interconnect that slashes deployment times from months to days. The company also launched SUNK Anywhere, a unified Slurm‑on‑Kubernetes control plane that spans CoreWeave, Google Cloud,...
We're Launching Two Specialized TPUs for the Agentic Era.
Google announced two new Tensor Processing Units—TPU 8i and TPU 8t—built for the emerging “agentic” AI era. TPU 8i focuses on ultra‑low‑latency inference, enabling autonomous AI agents to reason and act quickly. TPU 8t is optimized for training, offering a massive shared memory pool...
5 Big Google Cloud Security And Wiz Announcements At Next 2026
Google Cloud announced three AI‑powered security agents—Threat Hunting, Detection Engineering, and Third‑Party Context—now in preview, alongside a speed‑boosted Triage and Investigation agent that cuts manual analysis from 30 minutes to one minute. The company also expanded the Wiz platform, acquired...
The Most Interesting Startups Showcased at Google Cloud Next 2026
Google Cloud Next 2026 unveiled a $750 million fund to accelerate AI agents on its platform, offering cloud credits, engineering support, and deployment rebates to partners of all sizes. The program targets startups and consulting firms, enabling proof‑of‑concept projects with Gemini...
CoreWeave, Google Cloud Link up for AI Training, Inference
CoreWeave announced three cross‑cloud AI capabilities—CoreWeave Interconnect, SUNK Anywhere, and LOTA Cross‑Cloud—designed to simplify AI training and inference across multiple cloud platforms. Interconnect will directly link CoreWeave’s AI cloud to Google Cloud in a private preview, while SUNK Anywhere lets...
An ‘AI Scientist’ Can Tackle Drug R&D. What Does that Mean for Pharma?
AI agents are moving from analytical tools to autonomous coworkers in pharma, with Owkin’s K Pro platform acting as an “AI scientist” that can answer complex research questions in hours rather than weeks. The system pulls together literature, gene‑expression data, and...
What Amazon and Anthropic’s Deeper Partnership Means for Enterprise AI
Amazon and Anthropic have deepened their partnership, committing Anthropic to spend more than $100 billion on AWS technologies over the next decade while Amazon expands access to the Claude AI platform. The deal signals a shift in enterprise AI from pure...
Israel’s Most Promising Startups in 2026
The 2026 Calcalist Top 50 rankings show Israel’s startup ecosystem pivoting from lightweight software tools to autonomous AI workers and core infrastructure. Agentic AI accounts for 24% of the list, while AI‑infrastructure firms make up 18%, and quantum computing rises to...
EBay Publish AI Activate Prompt Library
eBay has launched an AI Activate Prompt Library as part of its AI Activate programme, which gives UK small businesses free access to ChatGPT and tailored training. The library offers plug‑and‑play prompts covering listing optimisation, customer communication, market research, performance...
Trapped in an Elevator? AI System at Namo Bharat Stations Can Raise Alarm in 60 Seconds
The National Capital Region Transport Corporation (NCRTC) has deployed an AI‑powered system that can detect commuters trapped in elevators at Namo Bharat stations within 60 seconds and automatically alert the Operations Control Centre. The solution uses existing CCTV cameras, processes...
What Cisco Is Doing to Avoid ‘Compliance Theater’
Cisco’s Chief People Officer Kelly Jones says the company’s AI rollout is driven by continuous employee sentiment measurement and leadership modeling. Over 18 months of surveys and focus groups revealed that AI‑using staff are more engaged, confident, and likely to...
What Cisco Is Doing to Avoid ‘Compliance Theater’
Cisco is shifting from superficial compliance checks to a robust AI governance framework that embeds risk assessment, transparency, and continuous monitoring across its HR and product teams. The company has launched an AI Ethics Council, mandatory model‑impact reviews, and a...
Zuora Adds AI Agents for Quote-to-Cash Workflows and Auditability
Zuora has launched AI agents integrated into its quote‑to‑cash platform, creating a finance‑grade intelligence layer that operates within existing controls, permissions, and audit frameworks. The agents enable read‑only simulations, contract analysis, revenue allocation comparisons, and audit‑ready responses across billing, revenue...
Claude Mythos Finds 271 Firefox Vulnerabilities
Anthropic's Claude Mythos AI model uncovered 271 vulnerabilities in Mozilla Firefox, prompting the release of Firefox version 150 which patched over 40 CVEs, including three directly credited to the AI. While most findings were low‑severity issues not assigned CVEs, the...
Vodafone Rolls Out Google-Powered Security and AI Tools for SMEs
Vodafone Business has introduced two new services for small‑and‑medium enterprises – a managed detection and response (MDR) security suite and an AI‑driven virtual front‑desk called AI Concierge – both built on Google Cloud technologies under its 10‑year, $1 billion partnership. The...
Best AI Agents for Influencer Marketing in 2026
Influencer marketing is transitioning from manual, fragmented tools to autonomous AI agents that can manage the entire campaign lifecycle. A 2026 benchmark shows 36.7% of marketers already use AI for creator discovery, with smaller shares applying it to content generation...
Vodafone Business, Google Cloud Target Enterprise AI
Vodafone Business announced a $1 billion partnership with Google Cloud to bring advanced cybersecurity and generative AI services to small‑business customers. The first offering is a managed detection and response (MDR) service powered by Google Security Operations, debuting in Germany. Simultaneously,...
Agentic ITSM: Understanding the 4 Levels of AI Maturity in IT Service Management
The article outlines a four‑stage maturity model for AI in IT service management, ranging from simple document retrieval (Stage 1) to fully autonomous, cross‑system agents (Stage 4). It argues that most vendor claims of "agentic AI" actually describe Stage 2 or 3 capabilities,...
Comment: Should AI Be Used to Design Adverts?
The article debates AI’s role in advertising, weighing creative benefits against environmental and ethical concerns. It cites a UK government report forecasting AI‑driven water use to jump from 1.1 bn to 6.6 bn cubic metres by 2027, and highlights mixed consumer reactions...
Harbin Engineering University Launches Sociology and Artificial Intelligence Dual Degree Programme
Harbin Engineering University has received Ministry of Industry and Information Technology approval to launch a dual bachelor’s degree that combines sociology with artificial intelligence. The program is designed to produce graduates who can apply AI tools to social governance, public...
AI Designs RISC‑V CPU Core in 12 Hours
How Did An AI Agent Build a RISC V CPU Core In Just 12 Hours https://spectrum.ieee.org/ai-chip-design?share_id=9374945
Executives Praise AI, Workers Lose 51 Days Annually
"though 81 percent of executives think their AI deployments have 'significantly improved productivity,' their workers are actually wasting eight hours per week cleaning up after AI’s messes, which is the equivalent of 51 work days a year." https://t.co/lAaa8wpyg3
Merck to Partner with Google Cloud on AI Initiatives
Merck & Co announced a partnership with Google Cloud to accelerate its artificial‑intelligence capabilities, allocating up to $1 billion over several years. The deal will fund AI infrastructure, licensing of Google’s Gemini Enterprise platform, and dedicated cloud engineers working alongside Merck...
Prioritize Ethical, Efficient AI over Mere Feasibility
We’re entering a phase where “can we build it?” matters less than “should we build it this way?” Energy, cost, and latency are becoming very important societal decisions. LLMs are powerful, but they’re not always the right abstraction. Good (AI) engineering is knowing...
What The Record Venture Funding Quarter Actually Means For Your Startup’s Fundraise
Crunchbase reported a record $300 billion poured into startups in Q1 2026, with four mega‑unicorns accounting for 65% of the total. Despite the headline concentration, early‑stage funding jumped 41% year‑over‑year and AI/ML deal count rose to 6,678 in 2025. Redpoint’s data shows...
AI Investor Glasswing Adds 14 Experts to Its Tech Advisory Councils
Glasswing Capital, a venture firm focused on artificial‑intelligence startups, announced the addition of 14 senior executives and domain experts to its technology advisory councils. The new members bring deep experience from leading AI research labs, cloud providers, and enterprise software...
The Sequence AI of the Week #847: Everything You Need to Know About Claude Opus 4.7
Claude Opus 4.7 arrived with modest benchmark gains—SWE‑bench 87.6%, MCP‑Atlas +14.6 pp—but its real shift lies in the API. The release removes all sampling‑level knobs (temperature, top_p, top_k, thinking.budget_tokens) and forces the sole "adaptive" thinking mode. In their place come semantic controls:...