Today's AI Pulse
AI agents challenge cybersecurity frameworks, sparking budget hopes and governance concerns
AI agents are being deployed to test existing cybersecurity frameworks, prompting CISOs to see a path to unlock larger cyber budgets. At the same time, industry leaders warn that the lack of clear algorithmic authorization raises governance risks. The convergence of these trends highlights both the promise and the perils of ungoverned AI in security operations.
Also developing:
By the numbers: OneRobotics acquires Nanoleaf for $40M
AI in Local Government? Detroit Residents Are ‘Not Widely Convinced’ It Belongs.
A University of Michigan survey of over 2,100 Detroit residents shows limited enthusiasm for municipal artificial intelligence, with support rarely exceeding 40% for most applications. AI for locating missing children garnered the highest backing at 57%, while identifying crime suspects faced the most opposition. Concerns about bias and data privacy were prevalent, with 40‑55% fearing unequal impacts and two‑thirds refusing to share personal data. Only 16% of U.S. cities run AI pilots and fewer than 2% have city‑wide deployments, highlighting a trust gap.
Master OpenClaw: Complete Guide From Install to Security
OpenClaw: The complete guide @ClaireVo has just put together the definitive guide to getting started with and mastering OpenClaw. Building on our podcast episode, this post covers everything you need to know, from first install to multi-agent setups, plus the real costs...
AI Is Good At Optimizing Things, Including Ad Fraud
Spider AF’s latest report reveals that AI‑powered performance ads are set to dominate spend, with traffic projected to rise 192% month‑over‑month versus a 23% decline in legacy campaigns. However, the same AI optimization appears to double ad‑fraud rates compared with non‑AI...
AI and Quantum Are Forcing a Rethink of Digital Trust
Enterprises are confronting a seismic shift in digital trust as AI agents multiply, pushing machine‑to‑human identity ratios from 100:1 toward 1,000:1. At the same time, digital certificates are being issued with ever‑shorter lifespans, complicating lifecycle management and increasing the risk...
SAP and ANYbotics Drive Industrial Adoption of Physical AI
SAP is partnering with Swiss robot maker ANYbotics to embed four‑legged autonomous inspectors directly into its ERP platform. The robots act as mobile data‑gathering nodes, using edge AI to analyze thermal, acoustic and visual feeds and instantly push fault alerts...
Fixing the Zero-Click SEO Crisis
In 2025 Google rolled out AI Overviews that present concise answers directly on the search results page. This feature caused Devon’s web‑design blog to lose roughly 61% of its organic clicks, turning a steady lead pipeline into a zero‑click funnel....
AI Coding Agents Driving Unprecedented CTO Adoption
I've been selling software for 15 years now. I've never seen as much "market pull" before as I'm seeing now. For the AI sr software engineer, we've built. CTOs are truly mind blown. Few insights that actually surprise me: - The...
Dynamic Identity Systems Needed for Independent AI Agents
Identity for AI agents came up in almost every serious conversation last week at RSAC. It is not just “give them an identity.” It is that: identities are dynamic agents act independently and they will outnumber humans We built identity systems for people. Now we need...
Sfaal and SDS Propose Pre-Contractual AI Safeguards for Writers
The French Syndicat des agents artistiques et littéraires (Sfaal) and the Syndicat des scénaristes (SDS) have announced a joint framework of pre‑contractual safeguards aimed at protecting writers from unauthorized AI‑generated reproductions of their work. The proposal includes mandatory disclosure clauses,...
DINO Accelerates Transformer Detector Training to SOTA Speed
🦖 DINO: Faster Training for Transformer Detectors Early transformer detectors like DETR were powerful but painfully slow to train. In 2022, DINO (Detection Transformer with Improved Denoising Anchor Boxes) changed that. By adding denoising queries and smarter anchor-based initialization, DINO stabilized training,...
Ensemble and Cohere Partner to Build First RCM-Native Large Language Model for Healthcare
Ensemble, a leading revenue cycle management provider, and Cohere are expanding their partnership to create the healthcare industry’s first RCM‑native large language model. Unlike generic LLMs that rely on heavy prompt engineering, the new model is fine‑tuned from scratch on...
NAB Show: Evergent To Showcase Subscriber Monetization & AI Innovations
Evergent, a subscriber lifecycle management specialist, will showcase a suite of AI‑driven tools at the NAB Show in Las Vegas from April 18‑22. The new offerings include subscriber lifecycle intelligence to boost acquisition and engagement, as well as flexible monetization...
Claude Code Leak Reveals Hidden TypeScript Insights
Turns out Claude Code was leaked on March 31, 2026. There are several snapshots of the TypeScript codebase on GitHub. I don’t want to link here for legal reasons, but there are some interesting educational tidbits that can be learned...
CoreWeave Becomes Critical AI Infrastructure, Attracts Mega‑Cap Funding
The shift is clear enough. What was once a niche GPU cloud provider is now treated as a critical infrastructure company. And with this latest financing milestone, @CoreWeave is signaling that the next era of AI compute will be...
The AI Oligarchs Don't Want You Reading This
New York Assemblymember Alex Bores, a Democratic candidate for NY‑12, is confronting a $10 million super‑PAC funded by AI industry leaders Greg Brockman, Marc Andreessen and Ben Horowitz. Bores, who authored the state’s RAISE Act—the nation’s toughest AI safety law—promises federal...
Autonomous Agents Risk Malicious Prompts Despite Sandbox Efforts
This is the type of thing I’m worried about with completely “autonomous” agents only the inclusion of something more malicious than ads. It could be unintentional or via prompt injection. I have some tools that can run while I’m sleeping...
Generative AI Moves From Experiment to Production Imperative
If you’re not using #GenerativeAI yet, you’re already behind. Shift: experiments → production. → Open models = control → Custom AI = advantage → Agentic #AI = everything Mistral AI Forge + NVIDIA’s NEMO Tron coalition are accelerating this shift. The next 2–3 years reward ownership. by...
NAB Show: Fabric To Showcase AI-Driven Media Supply Chain Innovations
Fabric will debut an expanded AI‑driven media supply chain platform at the NAB Show, unveiling updates to its Origin and Xytech product lines. The rollout includes Origin Studio, a cloud‑native title management system, and Origin Insights, a data‑as‑a‑service analytics platform....
The Fed Chair Just Said What AI Leaders Won't: The Models Don't Work
Fed Chair Jerome Powell publicly expressed his lack of confidence in the economic models used to forecast markets, noting that no system has reliably predicted the economy. He highlighted that while large language models (LLMs) have advanced dramatically, they remain...
What Scarcity Taught Computing, and AI Might Need to Relearn
The article reflects on how early computers, constrained by expensive storage and limited memory, forced engineers to develop disciplined indexing, selective retrieval, and purposeful forgetting. It argues that modern AI research often assumes unlimited context windows, leading to information overload...
Cyara Launches Agentic Testing
Cyara announced a new agentic testing and AI governance suite for voice and IVR channels, enabling enterprises to validate, monitor, and control autonomous AI agents before they interact with customers. The offering adds an assurance layer that pits AI agents...
Harness Rolls Out Release Orchestration Features with AI-Enabled Verification and Rollback
Harness announced new Release Orchestration capabilities that embed AI‑enabled verification and automatic rollback into its Continuous Delivery platform. The features also bring native warehouse‑level feature management from the Split acquisition and Database DevOps support for Snowflake, allowing code and data...
Enterprises Need Evaluation, Not More AI Pilots
Most enterprises do not have an AI model problem. They have an evaluation problem. That was one of my biggest takeaways from my conversation with @karibriski from Nvidia and @Toucas from Mistral AI at GTC. In the agentic era, the winners will not...
Nomadic Raises $8.4 Million to Wrangle the Data Pouring Off Autonomous Vehicles
NomadicML raised an $8.4 million seed round at a $50 million post‑money valuation to commercialize its vision‑language platform that auto‑annotates autonomous‑vehicle video. The tool transforms terabytes of archived footage into searchable, structured datasets, enabling rapid identification of rare edge‑case events for training...
Bloomerang Launches AI Strategist ‘Penny’
Bloomerang has launched "Penny," an AI‑driven fundraising strategist now available in beta on its Giving Platform. Built from thousands of consulting engagements, Penny analyzes each nonprofit’s full donor history to deliver real‑time, actionable recommendations. Early adopters, such as the Joe...
AI Safety Becomes a Cybersecurity Priority
AI safety has entered the cybersecurity era. @IrenaCronin and I write this newsletter every week. AI safety is becoming a cybersecurity issue because advanced AI can now help both defenders and attackers, making the risks more immediate and practical. As AI systems...
Anyway, What Does AI-RAN Even Mean?
AI‑RAN is currently a marketing‑driven label rather than a defined technical standard, despite telecom’s decades‑long use of AI in radio networks. Vendors offer divergent interpretations, making it hard for operators to model total cost of ownership or plan capital expenditures....
IRS Pilots Palantir AI to Pinpoint High‑value Audits
The IRS is testing Palantir's AI-powered analytics platform to identify "highest-value" audit and investigation targets, documents obtained by Wired reveal. The pilot program aims to cut through decades of fragmented legacy systems to surface taxpayers most likely to be committing...
SmartBear Delivers AI Enhancements Across Software Application Testing Lifecycle
SmartBear unveiled a suite of AI enhancements spanning its testing portfolio, including agentic test generation in Reflect, natural‑language query skills for Zephyr, AI‑driven API test creation in ReadyAPI, and improved object detection in TestComplete. The new features pull contextual data...
Datamine Updates MineScape with Digital Twin Tools
Datamine unveiled MineScape 2026, its latest mining software suite that adds digital‑twin capabilities, AI‑driven assistance, and automated design and scheduling tools for coal and stratigraphic operations. The platform now integrates real‑time operational data with predictive scenarios, enabling teams to simulate...
Move Beyond One‑Off Tests: Embrace Human‑Centric AI Evaluation
One-off tests don’t measure AI’s true impact. We’re better off shifting to more human-centered, context-specific methods.
AI's Future Depends on Empowering Workers, Not Just Companies
The question of who benefits from AI is becoming more urgent. Some argue that AI should not only serve companies, but also empower workers by giving them more control, insight and influence over how work is done. It is a shift...
Lawmaker Looks to Award Grants for Veteran Suicide Prevention AI Models
Rep. Ryan Mackenzie (R‑PA) is drafting the Data Driven Suicide Prevention and Outreach Act to fund AI‑driven predictive‑model grants for veteran suicide prevention. The bill would award one grant per each of the VA’s 18 Veterans Integrated Service Networks, targeting...
AI Is Reshaping Corporate Training and LMS Platforms
#AI is already making an impact on enterprise systems, including #LMS platforms. But what does it mean for the future of corporate training? @LearningCartLMS Co-Founder Troy Gorostiza looks ahead on our blog➡️https://t.co/kWjLaTItSK #elearning #futureofwork #onlinelearning https://t.co/FjGRTW0TqI
800 VDC Marks Inflection Point for Data Center Power
Was fun going deep on 800 VDC. I had to talk to my friends in the trade working on these data centers for context lol. Lots of juicy nuggets I uncovered. 800 VDC: The Inflection Point Reshaping Datacenter Power...
Treat AI Prompt Standards as Versioned Infrastructure
NEW POST The quality of what AI coding assistants produce depends on how well the prompter articulates team standards. @techygarg proposes treating such instructions as infrastructure: versioned, reviewed, and shared artifacts. https://t.co/fqAeB6irlB
Enterprise AI: Success Needs Integrated CX, Not Silos
Enterprise hits and misses - agentic #AI project failure versus success, open source versus AI, and the perils of disconnected #CX https://t.co/LHhLk3cEG1 by @jonerp. #EnSw
Claude Gains Agents and Orchestration in Upcoming Model
Claude will finally get agents and orchestration with its next model. It’s one thing that makes Perplexity Computer so unique.
AI Glasses Blur You Out of Real-World Video
#AI Sunglasses That Can “Erase” You From Real-World Footage by @gigadgets_ #EmergingTech #Innovation #Technology #Tech https://t.co/GqQ6ITQZKJ
Offer: Train Marketing Teams on AirOpsHQ and Claude
If there are any teams out there who are looking for someone to come in and train their marketing team on @AirOpsHQ + Claude, we should chat
Trust, Not Speed, Drives AI Success
AI without trust = risk. AI with trust = scale 🚀 5 essentials for responsible AI: ✔️ Governance ✔️ Anonymization ✔️ Data minimization ✔️ Audits ✔️ Privacy by design The winners in AI won’t just be the fastest. They’ll be the most trusted. #AI #Privacy #ResponsibleAI #Tech https://t.co/97kLb1Dkhl
Nvidia Launches DLSS 4.5 with 6× MFG for RTX 50 Series
Nvidia is rolling out a DLSS 4.5 update today with new frame generation features. The Nvidia app beta introduces 6x Multi Frame Generation for RTX 50-series GPUs https://t.co/16AqOs7K50
Non‑generative AI Crucial for National Security and Legal Search
non-generative AI and the training of non-generative AI is integral to our national security. drones don’t generate prose, images or audio. the search and target, exactly what ROSS Intelligence built but for locating the correct passage of verbatim law.
DNS Covert Channel Bypasses AI Guardrails, Enables Remote Shell
“Specifically, it abuses a hidden DNS-based communication path as a "covert transport mechanism" by encoding information into DNS requests to get around visible AI guardrails. What's more, the same hidden communication path could be used to establish remote shell access...
Open AI Tools Will Spur Global Adoption, Demand Caution
Discussed on @Bloomberg Asia TV: As AI models, tools, and architectures become more open, AI adoption will accelerate worldwide. Security imperfections should be approached with caution as open technologies like OpenClaw and the ecosystems mature over time. https://t.co/AANjdXLL0D
Edge Hype Overrated: Telcos Aren't Ready for Real-World Inference
Rolling my eyes at the EdgeWash nonsense from @opennebula panellist at @BERECeuropaeu event “Telcos in a unique position” to make edge inference happen. Not in the real world, no
WSJ: OpenAI Halted Sora, Calling It AI Slop
‘More AI Slop than AI Magic’: WSJ Reveals the Reason OpenAI Suddenly Shut Down Sora https://t.co/NbGZlATFjS
Old‑School ML Turns Messy Support Chats Into Actionable Insights
Great detailed write-up by Mariia explaining how we built topic modelling that turns surprisingly messy support chats into structured & applicable actionable insights A fun reminder that "traditional" ML (e.g. >4 years old) is still very useful https://t.co/vo9KfPIg6q https://t.co/KFMGcHYHxN
Leaders Must Inspire and Warn to Harness AI
Leaders must be both the chief inspirer and the chief fearmonger right now when it comes to AI. The best will become 10x stronger. The worst will tap out. Both good outcomes for the org. Tough to play both.
LeCun Repeatedly Appropriates Ideas; Media Stays Silent
Over and over and over LeCun borrows other people’s ideas and makes them sound like his own. It’s astonishing that the media never investigates, when the pattern has been so consistent for decades.