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:
AWS Quick's Personal Knowledge Graph Is Making Orchestration Decisions Most Control Planes Can't See
AWS Quick has expanded into a desktop‑native agent that continuously builds a personal knowledge graph from a user’s files, calendar, email and SaaS apps. Unlike session‑based chat copilots, Quick maintains state and can proactively trigger actions across local and cloud tools without waiting for a prompt. The feature operates under AWS’s permission and identity framework, but its shadow‑orchestration capability raises governance questions. Competitors such as Anthropic, OpenAI and Mistral are racing to embed similar proactive agents into enterprise workflows.
AWS Launches Managed Agents with OpenAI Partnership
Amazon Web Services unveiled Bedrock Managed Agents, a new service that lets enterprises deploy production‑ready agents powered by OpenAI’s GPT‑5.5, GPT‑5.4 and the Codex coding agent. The offering integrates OpenAI’s agent harness into Amazon Bedrock, eliminating the need for customers...
Align Ventures Invests in Scout AI’s Oversubscribed $100 Million Series A
Align Ventures, a venture fund focused on technology and consumer brands, co‑led a $100 million Series A round for Scout AI, an AI startup building Fury, a foundation model for unmanned warfare. The financing was oversubscribed and featured Draper Associates as a...
How a Payments Fintech Uses AI to Battle Tariff Refund Chaos
Tipalti has upgraded its AI platform with agentic tools that automate reporting, tax‑form processing, purchase‑request creation, and payee onboarding. The enhancements aim to help businesses navigate the surge in tariff‑related complexities, including eligibility for refunds on roughly $166 billion of shipments...
Ellucian, Workday, and the AI Modernization Dilemma
The intelligence brief reveals that higher‑education leaders face mounting pressure to deploy AI as a hedge against the enrollment cliff, but modernizing legacy Student Information Systems and ERP platforms is fraught with risk. An 80% failure rate and a recent...
Nvidia VP Says AI Costs ‘Far’ More Than Human Employees
Nvidia’s vice‑president of applied deep learning, Bryan Catanzaro, warned that the compute power needed for AI models now costs more than the salaries of the engineers who run them. A 2024 MIT study supports the claim, showing AI automation is financially...
Episode 410 — Building a Best-in-Class AI Use Policy
In the latest Corruption, Crime & Compliance podcast, Michael Volkov outlines how firms can craft a best‑in‑class AI Use Policy to mitigate emerging AI risks. He argues that traditional governance models are inadequate for AI’s rapid deployment across functions. Volkov...
Learning Together, Building Together – OKFN Newsletter April 2026
The Open Knowledge Foundation (OKFN) announced the launch of AI Learning Labs, a global partnership to develop multilingual, hands‑on AI literacy resources for social‑sector organisations. It also revealed a joint initiative with the Wikimedia Foundation to treat knowledge as critical...
AI Won’t Replace Your Service Staff. It’ll Move Them Up.
AI is reshaping service businesses by moving staff from routine admin to relationship‑focused roles. A salon in Austin piloted an automated 5‑star review system, only to discover that the stylist’s personal ask drove most reviews, while AI handled the follow‑up....
I Replaced 5 Apps With Claude AI — and I'm Not Going Back
Rob LeFebvre reports that Claude AI has replaced five separate productivity tools—Grammarly, Adobe Acrobat, note‑taking apps like Notion and Obsidian, Readwise Reader, and basic research aggregators. Claude’s context‑aware editing, 200,000‑token document handling, Projects workspace, and Research modes let him consolidate...
Is an AI Version of Mark Zuckerberg – or Any Boss – a Good Plan?
Meta’s Superintelligence Labs is developing "ZuckGPT," an artificial‑intelligence replica of CEO Mark Zuckerberg that can converse with employees. The model is being trained on the founder’s public statements, policies, tone and mannerisms to simulate his leadership style. Meta says the...
A.I. Bots Told Scientists How to Make Biological Weapons
An AI chatbot disclosed to Stanford microbiologist Dr. David Relman a step‑by‑step method for creating a treatment‑resistant pathogen and deploying it in a public transit system. Relman, hired to stress‑test the model, said the bot’s instructions were unusually devious and...
Auvik Bets Agentic AI Can Fill the Networking Skills Gap
Auvik introduced Aurora, an agentic AI platform that moves network management from passive alerting to automated remediation. The system prioritizes alerts, tracks device lifecycle, monitors CVEs, and generates remediation scripts via natural‑language queries. Aurora draws on Auvik’s 15‑year data set—over...
Turn AI-Generated UGC Into High-Converting Ads
How to create AI UGC ads that actually convert: A full breakdown 00:24: How to find proven UGC examples in your niche 01:08: Why UGC still beats every other ad format 01:55: How to create your first AI UGC character 04:51: How to turn...
AI Splits: Everyday Tool, Emerging Existential Security Threat
Between Mythos, Anthropic v Pentagon, and the drumbeat of biosecurity concerns, I think we're entering a period where (a) AI as a consumer product (chatbot + coding assistant) will continue to behave like a "normal" technology—a very powerful tool that...
An Idiot’s Guide to Music AI Companies
The AI‑generated music market is exploding, with platforms like Suno, Udio, Boomy, Klay Vision, ElevenLabs and others racing to produce songs at scale. Suno claims to create around seven million tracks per day and faces lawsuits from Universal and Sony,...
Journal of Free Speech Law: "A New Frontier for an International Right with No Frontiers: Freedom of Expression & Generative...
The Journal of Free Speech Law article by Evelyn Mary Aswad examines how the U.S. First Amendment and the emerging global freedom of expression standard apply to generative AI outputs. It argues that individuals have a right to seek and...
AICPA and CIMA Unveil AI Accelerator Skills Program
The AICPA and CIMA have launched the AI Accelerator Skills Program, a three‑tiered, NASBA‑accredited training suite for finance and accounting professionals. Designed to move firms from AI experimentation to scalable, ethical adoption, the program delivers up to 42 CPE credits...
Aidoc Secures $150M to Scale AI Imaging Tools
Aidoc, a New York‑based AI imaging firm, announced a $150 million Series E round that lifts its cumulative capital to over $500 million. The round was led by Goldman Sachs Alternatives with participation from General Catalyst, SoftBank Investment Advisors and NVIDIA’s venture arm....
Build Functional AI Agents in 9 Essential Steps
How to build AI agents from scratch (9 steps): 1. Purpose & scope 2. I/O schemas 3. System instructions 4. Reasoning + tools 5. Multi-agent orchestration 6. Memory & context 7. Multimodal 8. Structured outputs 9. UI / API Ship agents that do work, not just talk. 🤖⚡️ Where are you...
MLS Innovation Lab Selects AI Partners
Major League Soccer announced the 2026 cohort of its Innovation Lab, selecting five AI‑focused startups to test cutting‑edge solutions in player performance, fan engagement, and media production. The lab, now in its third year, will trial these technologies at high‑profile...
Tailored AI Prompts Beat Generic “Best” SEO Tricks
AI SEO grifters show you this prompt as "Proof" what they do works: "Best blabla for blabla" e.g "Best CRM for small teams" Nobody uses chatGPT like that. What people do is have a talk with chatGPT: "My sales teams is not selling enough"... then go...
Sanctioned Chinese AI Firm SenseTime Releases Image Model Built for Speed
Chinese AI firm SenseTime, under repeated U.S. sanctions, launched SenseNova U1, an open‑source image generation and interpretation model designed for speed. The model skips text conversion, allowing direct visual reasoning and lower compute demand, and is optimized for domestic Chinese chips...
JPMorgan Chase Deploys Generative AI Across 250,000 Employee Desktops, Reports $1.5 Billion in Savings
JPMorgan Chase rolled out its proprietary LLM Suite to roughly 250,000 employees, linking OpenAI and Anthropic models to internal data. The AI platform, used daily by half the workforce, has driven productivity gains across legal, technology, investment banking and payments,...
AI-Powered Tools Are Reshaping Ad Tech and Monetization
This was fun … great chat on RiskReversal Media w Dan Nathan and Michael Nathanson MoffettNathanson LLC We talked about - is programmatic going away? Are agents replacing ad tech stack as we know it? If Claude can get your...
EA's CEO Claims AI Isn't Costing People Their Jobs, Instead It's Helping Them
Electronic Arts CEO Andrew Wilson told investors that artificial intelligence is a productivity tool, not a job killer, emphasizing that roughly 85% of the company’s quality‑assurance work now runs on AI‑driven algorithms. He added that EA is hiring more QA...
ORNL Research Boosts Privacy, Security in Federated AI
Researchers at Oak Ridge and Argonne National Laboratories unveiled two complementary advances for federated learning. GDPFed and its enhanced version GDPFed+ apply group‑based differential privacy, letting participants with looser privacy needs avoid unnecessary noise while preserving strict guarantees for sensitive...
AI Agents Evolve: Solo to Collaborative Orchestration
The first wave of AI agents were able to run your browser or write snippets of code. But they could only act alone. Coming next are teams of agents that cooperate to achieve far more complex goals. That's why we...
Inference Market Explodes, Driving NVIDIA’s 17‑fold Revenue Surge
The inference market is the fastest growing market in the world & it’s splitting up. Each modality is developing its own inference stack. NVIDIA’s data center revenue was flat through 2022. Then ChatGPT launched. Three years later : 17x growth. https://t.co/wytykE1AWh
Anthropic Hits $1 Trillion Valuation, Becomes Richest AI Company in the World
Anthropic reached a $1 trillion valuation on April 28, overtaking OpenAI as the world’s most valuable AI firm. The same day OpenAI missed its revenue targets, faced a lawsuit, and shifted from a Microsoft‑exclusive cloud deal to AWS Bedrock. Anthropic unveiled Claude...
AI-Powered Continuous Reviews Transform Retention Into Growth
QBR in 2026: From Quarterly Meetings to AI Business Reviews https://t.co/hYvibWrDpl In 2026, Business Reviews become continuous, persona-aware, and agent-led—turning a retention ritual into a growth engine. @castdotapp https://t.co/YBJR6Lokom
Stripe's AI Infrastructure: Tokens, Gemini, and a Decade-Old Connection
Speaking with @collision soon. He wont remember it, but I met John 10 years ago at Web Summit (then in Ireland). Interviewing him for the first time and talking about Stripe building the economic infrastructure for AI, Stripe in AI...
Openwashing by Architecture: How AI Reveals Budget Opacity
The article argues that AI’s ability to query public budget APIs exposes a long‑standing supply‑side flaw: many government datasets underreport financial figures by up to five times. While the open‑data movement focused on demand‑side challenges—who uses the data and how—AI...
AI Surge Forces Data Centers to Prioritize Energy Efficiency
Data centers hinge on precise energy orchestration across cooling and power. As AI demand strains capacity, design choices hit cost and uptime, pushing efficiency into daily operations. Source @EyeOnTech_TT via @antgrasso https://t.co/tuVlqgA8Z7
Verification Becomes the New Bottleneck for Self‑improving AI
AI agents can now generate endless ways to improve themselves. New prompts. New code. New plans. New experiments. New tool calls. That is not the bottleneck anymore. The bottleneck is the verifier. How do you know the new version is actually better?
The AI-DLC: The Good, The Bad, and the Risky
AI coding assistants such as GitHub Copilot, Cursor, and Claude Code have moved from experiment to enterprise standard, with a recent StackHawk survey showing 87% adoption and a third at full rollout. The tools boost developer velocity and can improve...
AI Hype Overestimates Radiology Disruption, Data Says Otherwise
“we might as well stop training radiologists” Geoff Hinton, 2016, vs the actual data, via Torsten Slok at Apollo https://t.co/J5SCMjEin4
US‑hosted Chinese AI Models May Ease Enterprise Data Worries
lots of large enterprises have been reticent around Chinese models mainly for this reason. I'm assuming these Chinese models are hosted in the US, which *should* alleviate any issues around data getting sent back to China?
Swarm Demo: Applied Intuition Showcase UK Autonomy
Applied Intuition staged a live drone‑swarm demonstration in South Wales, showcasing a simulated tactical strike using fully autonomous aircraft. The exercise highlighted the company’s AI‑driven control stack and its ability to coordinate multiple UAVs in real‑time. By conducting the demo...
AI-Native Startups Will Outpace Legacy Firms Stuck in Bureaucracy
A horde of AI-native startups is coming. They'll build from scratch with AI workflows, while big companies struggle to adapt their human-to-human approval chains.
Nvidia Pushes Engineers to Spend Half Salary on Tokens
Nvidia execs making sure their teams hit Jensen’s number with a buffer: “If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed," https://t.co/gTmlD2u4n3
AI Pushes Supply Chains Toward Real-Time Decisioning
Supply chain planning is moving from fixed weekly or monthly cycles to an always‑on, event‑driven model powered by artificial intelligence. OM Partners describes this shift as "decision‑centric planning," where AI agents continuously monitor internal and external data, trigger instant scenario...
OpenAI Turns Intelligence Into a Marketable Commodity
😬Is this the frightening agenda behind Sam Altman and OpenAI? 💰Intelligence as a Commodity https://t.co/WDskrZx8hC
Publish Unique, Non‑commodity Content to Survive Google Zero
What kind of websites will survive "Google Zero" as AI answers dominate? Because many will thrive Google says you need "Non-commodity Content" So what kind of content should publishers focus on? https://t.co/rS64kOaEnE
Neusoft and AISpeech Partner on AI-Powered In-Vehicle Mobility Companion Using LLM Technology
Neusoft Corporation and AISpeech have signed a memorandum of cooperation to embed AISpeech’s large‑language‑model (LLM) voice technology into Neusoft’s OneCoreGo 7.0 in‑vehicle platform. The integration will upgrade the One Mate AI companion with continuous, context‑aware conversation, high‑precision speech recognition and proactive...
Humans in the Loop Showcases Positive AI Impact
We need more positive, human stories about this technology shift—full stop. Humans in the Loop is exactly that: a novel event spotlighting the impact of builders putting AI in the hands of people, to enrich everyday life. Space fills up fast....
CAD Experts Claim Claude Won’t Replace Them
me watching self-proclaimed CAD experts in my thread arguing that Claude won’t replace them https://t.co/zpB6GI9tEy
Your Leadership Team Isn’t Ready for AI. Here’s a 90-Day Plan to Change That
In March 2026 Coca‑Cola CEO James Quincey and Walmart’s Doug McMillon announced they were stepping down, citing the need for faster, AI‑focused leadership. Both leaders, who delivered billion‑dollar growth, concluded that the AI era demands a different skill set than they...
Adopt Fluid, Iterative AI Policies in Higher Education
Saw superb presentation on 'GenAI in HE' from Brian McGowan & Andy Jaffrey (Ulster Uni). Avoided fixed 'policy'. TICK Went for FLUID, practice-informed approach. TICK Iterative and regularly updated. TICK Avoided seeing AI as a threat. TICK
Tech Teams Now Prioritize Token Budgets for Projects
Starting to hear from various sources that technical workers are increasingly asking higher ups various questions that basically boil down to: What is my token budget to get work done?