Today's AI Pulse
Odyssey lands $310M Series B, valued at $1.45B
Odyssey announced a $310 million Series B round that values the AI lab at $1.45 billion. The round was led by Natural Capital and included backers such as Amazon, AMD Ventures, GV, EQT and IQT. In parallel, Odyssey struck a preferred‑cloud deal with AWS to use Trainium silicon for large‑scale model training.
Also developing:
Measuring Claude 4.7's Tokenizer Costs
Anthropic’s Claude Opus 4.7 introduces a new tokenizer that consumes roughly 1.3‑1.45 times more tokens than the 4.6 version, especially for English prose and code. The shift raises per‑session costs by about 20‑30 percent because cached prefixes and user inputs expand proportionally. Benchmarks show a modest gain in strict instruction‑following accuracy—about five percentage points—while the price per token remains unchanged. Users must weigh the higher token usage against the incremental reliability boost when deciding whether to upgrade.
Napier AI Names Top Markets for AI-Driven AML Compliance
Napier AI’s AI/AML Index 2025‑2026 spotlights the UK, US and France as the most advanced markets for AI‑driven anti‑money‑laundering compliance, while Hong Kong lags due to high upfront costs. Regulators in Europe, especially Germany and France, are issuing explicit guidance...
Search Evolves Into Agentic Manager with Gemini AI
This week we saw so many things get set up as we transition to a new era of the web - the agentic web. I did something different with my newsletter this week. I asked Gemini to look at the transcripts from...
The New Talent Imperative: How Leading Organizations Are Getting Serious About AI Skills
Leading firms are realizing that AI tool investments alone won’t deliver value without a skilled workforce. A new virtual event on May 21 will reveal original data on how organizations define, source, assess, and develop AI capabilities for non‑technical employees. The...
AI Videos Outshine Hollywood's Diminishing Production Value
I originally thought this was some kind of fan-made AI video, which is both a testament to how good the AI video generators have become and a grim sign of how bad the “big” expensive Hollywood movies look now
In the AI Propaganda War, Iran Is Winning
Iran is leveraging inexpensive AI tools to produce high‑volume, hyper‑realistic propaganda videos that target U.S. audiences and regional rivals. By automating deep‑fake creation, Tehran can flood social media with disinformation faster and cheaper than traditional state media. The campaign includes...
Iran’s AI Memes Are Reaching People Who Don’t Follow the News – and Winning the Propaganda War
Iran’s state‑linked Explosive Media has deployed AI‑generated Lego‑style videos with rap soundtracks, amassing billions of views since February. Dubbed “slopaganda,” the clips blend disinformation and antisemitic tropes with humor, targeting users who avoid traditional news. By mimicking the tone of...
Top 32 MCP Servers to Make Claude Super Powerful
The post explains that Claude’s power is limited by its lack of access to external resources, and that Model Context Protocol (MCP) servers act as secure gateways to tools like Google Drive, GitHub, live web search, and databases. By connecting...
AI Wristband Reads Muscles, Directs Robots Like Hands
MIT engineers built an AI wristband that controls robots by reading your hand muscles. It works by using an ultrasound to capture images of the muscles and tendons in your wrist. An AI algorithm then translates those images into the exact position...
Leading in the Age of AI: When Management Becomes the Differentiator
The April 28 virtual workshop, "Leading in the Age of AI," examines how artificial intelligence reshapes managerial responsibilities. Featuring leaders from DoorDash, McKinsey, Mento, and Charter, the session highlights that AI fluency alone won’t drive results; managers must translate technology into...
Marsh Aims to Be ‘AI Winner’ by Focusing on Gains in Growth, Productivity, Efficiency
Marsh is positioning itself as an AI leader by embedding artificial intelligence across growth, productivity and efficiency initiatives. The insurer‑broker reported Q1 2026 revenue of $7.6 billion, up 8%, while AI‑driven tools have already cut document‑processing time by 20% and accelerated...
Factory Raises $150M at $1.5B Valuation to Scale AI Coding Agents for Enterprises
Factory, an AI‑driven coding platform founded in 2023, raised $150 million in a Series B round at a $1.5 billion valuation. The round was led by Khosla Ventures with participation from Sequoia Capital, Insight Partners and Blackstone, and veteran investor Keith Rabois joined...
Air Force Advances Secure AI Capabilities with IL5 Approval for VIA’s eJARVIS
The U.S. Air Force has granted Via Science’s eJARVIS AI platform Impact Level 5 Authority to Operate, upgrading its security clearance from the IL4 level granted in May 2024. The IL5 rating permits the system to process more sensitive Controlled Unclassified...
Air Force Advances Secure AI Capabilities with IL5 Approval for VIA’s eJARVIS
The U.S. Air Force has granted Impact Level 5 (IL5) approval to VIA’s eJARVIS artificial‑intelligence platform, allowing it to operate on classified networks. The clearance marks the first AI system certified for high‑sensitivity data within the service, enabling real‑time analytics for...
One Doc to Rule Them All: The Golden Ring Framework
The author, a former VP of Sales Excellence at OwnBackup, warns that most revenue teams are floundering with AI, lacking clear metrics and ownership. He argues that IT and Salesforce admins shouldn’t be the sole architects of AI systems; sales...
The $5 Photo Shoot: How a Small Austin Jewelry Brand Stopped Waiting and Started Producing
A husband‑and‑wife jewelry brand in Austin used the AI image generator Nano Banana to create lifestyle product photos in seconds, paying only five cents per image. In one afternoon they produced over 40 new assets that previously required costly photo...
Human‑scale Data Aims to Close Drug Trial Gap
90% of drugs that enter clinical trials fail. A big part of why: the models they're based on weren't trained on human biology. Mouse data, non-representative cell lines, sparse perturbation coverage. The gap between in silico predictions and what actually happens...
AI Won’t Fix Your Leadership Communication, but It Might Expose It
Meta unveiled an AI‑driven avatar of Mark Zuckerberg to spearhead enterprise communication, positioning the tech as a consistency, reach, and signalling tool. The op‑ed argues that while AI can streamline message distribution, it cannot replace the core need for clear,...
AI Privacy Isn’t What You Think It Is
The post argues that AI chat interactions are not truly private because the data traverses extensive infrastructure such as logging, storage, and monitoring pipelines. It shows how AI enriches the type of data captured, turning ordinary queries into detailed behavioral...
CoChat Launches AI Collaboration Platform to Combat Shadow AI
CoChat debuted in early April 2026 as an AI collaboration platform aimed at curbing the rise of shadow AI within enterprises. By consolidating access to leading large language models (LLMs) and autonomous agents, it eliminates fragmented, unmanaged AI silos. The...
MCIOB Develops Own AI Software for SMEs Using AI
Steve McKenna, director of the SME contractor Gemstone, used Anthropic’s Claude Code to create Construction AI, a full‑featured, multi‑tenant SaaS construction management platform. The system boasts over 700,000 lines of code, 186 secured database tables and 596 API routes, delivering...
Employers Say They Struggle to Find Workers with the Right AI Skillset
Employers are increasingly reporting difficulty finding entry‑level talent with practical AI skills, with 53% naming the gap as their top hiring challenge. While 78% of higher‑education leaders claim they meet employer expectations, only 28% of employers agree that universities are...
AI Coding and Design Competition Intensifies as Anthropic and OpenAI Expand Agent Capabilities
Anthropic is set to launch Opus 4.7, a model that adds generative design tools that could directly compete with Figma, prompting co‑founder Mike Krieger to leave the board. OpenAI counters with a revamped Codex platform that embeds autonomous agents capable of...
What San Francisco’s AI Billboards Say About the State of the Industry
A recent San Francisco Chronicle survey found that roughly 50% of the city’s outdoor billboards now promote artificial‑intelligence applications, platforms, and infrastructure. The ads target senior IT leaders, emphasizing productivity, scalability and agentic features rather than safety or ethical concerns....
Every Old Vulnerability Is Now an AI Vulnerability
Microsoft patched CVE‑2026‑26144, an XSS flaw in Excel that now exploits the Copilot Agent to silently exfiltrate spreadsheet data. The vulnerability demonstrates how AI agents can amplify traditional bugs, granting them autonomous, privileged actions beyond the original exploit. Security experts...
AI Cuts B2B Site Build From 8 Weeks to Days
What typically used to take 8 weeks took me a few days in 2026. In a today’s live AI session at 10AM PT, I’m sharing how I built a full B2B marketing site in 3 days with AI for a new...
India Fintechs Scramble for Access to Anthropic Mythos
Indian fintechs including One97 Communications, Razorpay and Pine Labs are lobbying Anthropic for early access to its restricted AI model Mythos, which can identify software flaws far beyond human experts. Anthropic has kept Mythos under wraps, deeming it too dangerous...
At Roundtable on AI, Members of Congress Express Angst and Fears of ‘Destruction’
Congressional leaders gathered for a House Oversight subcommittee roundtable on artificial intelligence, where lawmakers voiced deep anxieties about the technology’s rapid evolution. Concerns ranged from federal workers using AI chatbots on sensitive data to AI‑generated deepfake pornography and the potential...
Roblox Expands AI Assistant with Agentic Tools for End-to-End Game Development
Roblox has upgraded its Roblox Assistant within Roblox Studio by adding agentic AI tools that guide creators from concept to launch. The new Planning Mode can analyze code, ask clarifying questions, and produce structured development roadmaps. Integrated mesh generation and procedural...
Physical AI Matters More Than Humanoid Robots
Forrester released a new report titled "Physical AI Perceives, Reasons, And Acts In The Real World," outlining four core capabilities—world modeling, perception, reasoning, and actuation—that together enable AI to interact with physical environments. The report argues that these capabilities matter...
Four Architecture Layers Power the Agentic Enterprise
What does a CIO need to know about architecting the agentic enterprise? #sponsored 4 architecture layers every agentic enterprise needs: → Context: harmonized data → Work: composable business logic → Agency: reasoning + execution → Engagement: humans + agents, side by side The model is not...
HIMSS Advocates for Consistent Nationwide AI Regulation
HIMSS is urging the U.S. government to adopt a single set of AI guardrails that ensure safety and trust across the healthcare sector. The organization’s public‑policy principles, outlined by Jonathan French, call for nationwide standards to eliminate a patchwork of...
Runway CEO Advocates AI-Driven Film Production Model to Scale Output and Reduce Costs
Runway CEO Cristóbal Valenzuela told the Semafor World Economy Summit that artificial intelligence can reshape film production by lowering costs in scripting, planning and visual effects. He argued studios should spread budgets across several projects rather than betting on a...
Everyone's Talking About AI Agents. Almost Nobody Has Actually Built One. Here's How
A recent poll shows only 3% of readers have actually built a self‑running AI agent, despite 65% claiming to experiment with them. The gap narrows as Model Context Protocol hits 97 million monthly installs and low‑code platforms like Make.com simplify integration,...
Secure-by-Design: 3 Principles to Safely Scale Agentic AI
Enterprise adoption of agentic AI is shifting from experimental copilots to autonomous agents that execute decisions across environments. This expands the attack surface, prompting a secure‑by‑design shift championed by a CrowdStrike‑NVIDIA blueprint. The article outlines three core principles: treating AI...
Match AI Risks to the Right Organizational Zone
Not all risks belong in the same part of the organization. Some call for exploration, others for disciplined execution, and a few require full commitment across the enterprise. Treating them the same way is a common source of confusion. The...
ChopGrad Cuts Gradient Cost, Boosts Training Efficiency
ChopGrad: Making Training More Efficient by Cutting Gradient Complexity In this episode of Artificial Intelligence: Papers and Concepts, we explore ChopGrad, a novel technique aimed at improving the efficiency of training deep learning models by selectively simplifying gradient computations. Instead of...
AI Boosts PhD Work, Yet Threatens Core Skills
AI and the PhD student: friend or foe? Graduate students increasingly use artificial-intelligence tools to draft, code and search... but many fear it could erode the very skills a doctorate is meant to build. https://t.co/b8cdoCI6Bh https://t.co/scTkfsXuCW
Lazy Creators, Not AI, Produce the Content Slop
There's a lot of AI slop out there. Is the problem AI or lazy execution? The same people churning out slop with AI were probably churning out slop without it too. https://t.co/Xum0OCHduP
VCs Reveal 2026 Agentic AI Investment Trends
ICYMI, we’re back with another Crunchbase Predicts webinar. 👏 We’re diving into agentic AI & what top investors are betting on for 2026. Join to hear how VCs evaluate opportunities, where AI is gaining traction, & what could define the next wave...
Anthropic CEO to Discuss Pentagon Dispute with White House
Anthropic CEO to meet White House chief of staff amid Pentagon dispute, Axios reports - https://t.co/JeVknmFC2Q
AI Transforming Contact Centers: Insights From Synovus
Love these short customer discussions. In this episode, I chat with @synovus' Senior Director of Customer Care, Data Lloyd about AI's impact in the contact center. #CIO #CMO #AI #CCaaS #CXNation @Five9
DeepSeek Seeks First External Funding to Boost AI Compute
DeepSeek in talks to raise outside capital for first time, seeking to beef up its financial war chest so it can better compete in the costly battle to develop leading AI models.... That will buy lot of compute... https://t.co/8YciEvxn9J
Thought Leadership Drives AI-Powered Search Success
I'm so sorry about my eyebrows at the beginning of this video...I get really excited talking about thought leadership as an AI search strategy. https://t.co/oh0ICuLXiB
Experts Weigh in on Lingering Mythos Concerns
The release of Mythos scared an awful lot of people - but now the dust has settled a little, how worried should we be? @sciam asked me to take the temperature of the experts https://t.co/C967TRaGLC
Data‑center Delays Jeopardize US AI Expansion
Data centre delays threaten to choke AI expansion Almost 40% of such builds in US risk hold-ups, including projects tied to Microsoft and OpenAI $NBIS $CRWV https://t.co/6uooabWhYh
Launching Chit Chat Stocks Compute: Podcast Neo Cloud for AI
Officially launching Chit Chat Stocks Compute, our journey to building a podcast Neo Cloud for AI workloads
New Open‑Source SRE Extension Boosts Gemini CLI
Want a little Site Reliability Engineering help from your agentic CLI? @palladius just open sourced a new SRE extension for the @geminicli. It's got a handful of useful skills for investigations, anomaly detection, and mitigations. https://t.co/euwVVEOK6r https://t.co/dnjWoIdlmD
AI Is a Business Revolution—Adopt Now or Fall Behind
I’ve seen businesses take weeks to execute, now it’s seconds. That’s not evolution, that’s a revolution. I use AI across everything: finance, marketing, even my own photography. If you’re not leveraging these tools, you’re already behind. This isn’t optional anymore....
Convolution Powers All Image Filters in 60 Seconds
Blurring. Sharpening. Edge detection. They ALL come down to one operation: convolution. Here's a 60-second visual breakdown of how a kernel slides across an image to produce filters — pixel by pixel. If you're learning CV, bookmark this. #ComputerVision #Convolution #ImageProcessing #OpenCV #DeepLearning #CNN