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:
Opportify Gains Early Adoption for Email Insights to Stop Sign-Up Fraud
Opportify has launched its Email Insights solution, a risk‑based intelligence platform designed to stop fraudulent sign‑ups at the point of entry. Unlike traditional validators that rely on simple syntax or MX checks, Email Insights scores each address on domain stability, authentication signals, behavioral anomalies, and known abuse indicators. Early adopters report fewer invalid contacts, lower bounce rates, and a sharp drop in bot‑driven registrations. The cloud‑native service integrates via APIs and SDKs while adhering to Zero Trust security principles.
Kairos: AI Interns for Everyone
AI startup Kairos launches an AI‑powered personal intern that can log into applications, navigate websites, and execute complex multi‑step tasks without human oversight. The platform includes a dedicated browser, native integrations with over 20 tools such as Gmail, Notion, and...
Oversonic Robotics Launches Humanoid Robot RoBee in US Market
Italian robotics firm Oversonic Robotics introduced its latest cognitive humanoid, RoBee, to the United States this month, aiming to broaden its footprint in healthcare and advanced manufacturing. The launch coincided with a strategic North American partnership targeting medical and aerospace...
Prompt Injection Threat Turns AI Agents Against Employers
When AI Agents Turn Against You: The Prompt Injection Threat Every Business Leader Must Understand As organizations deploy #AIagents to handle everything from customer service to financial decisions, a critical #security #vulnerability threatens to turn these digital workers against their...
Tencent Hunyuan Releases HPC-Ops: A High Performance LLM Inference Operator Library
Tencent Hunyuan has open‑sourced HPC‑Ops, a CUDA‑based operator library that accelerates large language model inference on NVIDIA GPUs. The library provides high‑performance kernels for Attention, Grouped GEMM and fused MoE, supporting bf16 and fp8 precisions via a compact C++/Python API....
XGSynBot Unveils Dual-Arm Mobile Robot at CES Showcase in Las Vegas
XGSynBot showcased an AI‑powered dual‑arm mobile robot at a private CES event in Las Vegas. The platform merges force‑controlled manipulators, an omni‑directional chassis and a multi‑sensor vision system to serve industrial and commercial tasks. Marketed as a universal, cross‑scenario solution, it...
Grammarly and QuillBot Are Among Widely Used Chrome Extensions Facing Serious Privacy Questions
Incogni’s 2026 privacy risk report examined 442 AI‑powered Chrome extensions and found that over half collect user data, often with deep‑level permissions. The study highlighted that scripting and activeTab permissions let extensions read and modify any web page, exposing emails,...
Clawdbot's Power Matches Any Computer—Limited only by Imagination
Asking what you can *actually* do with Clawdbot / @moltbot is the same as asking what you can actually do with a computer. It’s capped by whatever you think of asking it. It’s controlling the whole computer.
Audits for AI Systems that Keep Changing
ETSI released TS 104 008, a continuous‑auditing based conformity assessment (CABCA) specification for AI systems. It shifts assurance from periodic reviews to ongoing cycles that automatically collect evidence from logs, model parameters, and data samples. The framework operationalizes regulatory requirements into machine‑readable...
How Miu Miu Beauty Used AI to Ensure a Successful Launch
Miu Miu Beauty, launched by L’Oréal Luxe in 2025, deployed AI early to streamline internal workflows and pre‑test creative assets for its debut fragrance Miutine. The brand leveraged AI tools such as Vizit and L’Oréal GPT to predict visual engagement and...
Lumana's Vision for Agentic AI: When Cameras Stop Recording and Start Acting
Lumana is redefining video surveillance with an agentic AI platform that moves cameras from passive recorders to active decision‑makers. Its VIA‑1 system learns locally from each camera’s environment, delivering over 90% reduction in false alerts for customers such as Salesforce...
Risky Business #822 -- France Will Ditch American Tech over Security Risks
In this episode Patrick Gray and Adam Boileau review a week of cybersecurity headlines, highlighting France’s decision to replace US collaboration tools like Microsoft Teams and Zoom with a sovereign platform, and China’s alleged “Salt Typhoon” operation that spied on...
Trustless AI Agent Standard Could Hit Ethereum Mainnet on Thursday
Ethereum is set to launch ERC‑8004, a new smart‑contract standard that enables trustless communication between AI agents, on mainnet Thursday around 9 AM ET. The protocol introduces a unified method for registering, discovering, and validating AI agents without altering the underlying network....
I Tested Nvidia's DLSS 4.5 — and It's so Great, I Can Barely Spot the Difference Between Upscaler Modes
Nvidia's DLSS 4.5 upscaler delivers image quality so close between Quality and Performance modes that differences are barely perceptible, according to tests on a Lenovo LOQ 15i Gen 10 laptop with an RTX 5060 GPU. The second‑generation transformer model improves sharpness, reduces ghosting, and...
‘This Train Isn’t Going to Stop’: Shocking Sundance Film Shows Promises and Perils of AI
The AI Doc: Or How I Became an Apocaloptimist premiered at Sundance, featuring AI luminaries such as Sam Altman, Yoshua Bengio and Elon Musk while probing whether the technology poses existential risk or unprecedented opportunity. Director Daniel Roher frames the...
Software and AI Profit From Margins, Not Price
Jeff Bezos said, "Your margin is my opportunity." Does this apply in software & AI? Yes—but not on price. 🧵
Show HN: Pinecone Explorer – Desktop GUI for the Pinecone Vector Database
Pinecone Explorer is a native macOS desktop application that lets users browse, query, and manage Pinecone vector databases. The tool supports dense, sparse, and hybrid index types, offering advanced search, filtering, and built‑in reranking capabilities. It connects to both serverless...
Founders Reveal Real AI Metrics, Pricing, GTM Strategies
At SaaStr AI 2026, no vanity metrics Founders are sharing real numbers: What AI features actually convert. What pricing actually works. What GTM motions actually scale. May 12-14. SF Bay Bring your Granola. Or even a notebook. https://t.co/BCPo2TiqOt
Nemotron-Personas-Brazil: Co-Designed Data for Sovereign AI
NVIDIA released Nemotron‑Personas‑Brazil, an open dataset of six million fully synthetic Brazilian personas grounded in official IBGE census and labor statistics. The collection spans 20 fields, 1.5 k occupation categories, and covers every Brazilian state, delivering culturally authentic Portuguese narratives. Built...
Everything You Need to Know About Viral Personal AI Assistant Clawdbot (Now Moltbot)
Moltbot, the lobster‑themed personal AI assistant formerly known as Clawdbot, went viral after a rapid GitHub star surge and a forced rebrand due to Anthropic’s trademark claim. Developed by Austrian coder Peter Steinberger, the open‑source tool automates tasks like calendar...
A Coxswain on Your Shoulder
Tom Tunguz created an AI meeting‑review agent that acts like a coxswain, silently analyzing recordings of pitch meetings, one‑on‑ones and investor calls. The system scores each conversation on a five‑point rubric—active listening, empathy, questioning, clarity and technical depth—providing concrete feedback...
E202: Recent Advances in LLMs and How They Will Impact Science and Pharma Research
In this episode, Javier Tordable, founder and CEO of Pauling.ai and former Google technologist, explains how modern large language models have evolved into autonomous AI agents capable of executing multi‑step scientific workflows, from literature synthesis to hypothesis generation and computational...
'Doomsday Clock' Ticks 4 Seconds Closer to Midnight as Unregulated AI and 'Mirror Life' Threaten Humanity
The Bulletin of the Atomic Scientists moved the Doomsday Clock to 85 seconds before midnight, the closest point ever, citing escalating nuclear tensions, stalled climate action, and unregulated AI and synthetic “mirror life.” The report warns that the United States,...
Anduril Has Invented a Wild New Drone Flying Contest Where Jobs Are the Prize
Anduril has launched the AI Grand Prix, an autonomous drone‑racing contest that tests participants' software‑writing skills rather than piloting ability. The event, held in Ohio, offers a $500,000 prize pool and direct hiring opportunities that bypass the company’s normal recruiting...
Memorable AI Names Drive Early Virality and Adoption
some thoughts on naming AI products naming is important, and has impact, especially on early virality. babyagi was originally called "task-driven autonomous agent" and don't think it would have performed as well autogpt had a more descriptive name, and had a lot...
Free AI-Powered Tech News Beta Launches on Pulse
Pulse is being covered on the front page of Pulse right now. How meta. 👀 Our beta launched today. Up to 50 registered users so far. Grab a free beta account and follow some topics, publishers, creators, companies, and investors...
SUPERAGENT AI Launches Insurance Quoting AI Agent
SUPERAGENT AI unveiled the world’s first Quoting AI Agent, a tool that automates the traditionally labor‑intensive insurance quoting process. The agent gathers customer data, navigates multiple carrier rating engines, and delivers optimized quotes without human re‑keying. Early‑access sign‑ups are open...
DSGym Offers a Reusable Container Based Substrate for Building and Benchmarking Data Science Agents
DSGym, a collaborative effort from Stanford, Together AI, Duke and Harvard, introduces a reusable container‑based framework that evaluates data‑science agents through real code execution. The suite standardizes tasks, agents and environments, offering 972 analysis and 114 prediction challenges spanning finance,...
Mastercard Launches ‘Agent Suite’ to Help Merchants Deploy Agentic AI
Mastercard announced the Agent Suite, a portfolio of customizable agentic AI services slated for release in Q2 2026. The offering combines AI agents with consulting, leveraging Mastercard’s payments infrastructure, data assets, and a global network of over 4,000 advisors. Initial use...
Interdisciplinary Communities Unlock Practical Wisdom in Responsible AI
Once in a while I write a quick post of appreciation for someone who has inspired me or my work. Today I wanted to highlight Jacob Metcalf who I went to back in 2016 as he (and his org, Ethical...
Amazon Alexa+ Is Now Available to Everyone. Here’s How to Turn It Off (2026)
Amazon has rolled out its AI‑enhanced Alexa+ to all Prime members, automatically replacing the original Alexa on supported devices. The upgrade, announced at CES, is still in Early Access and remains free for Prime users, though a $20 monthly fee...
How Porn Stars Can Survive in the Age of AI
The AVN Adult Entertainment Expo highlighted a turning point for the porn industry as artificial intelligence reshapes content creation. Performers face mounting threats from AI‑generated deepfakes that can siphon revenue and damage reputations. In response, stars are adopting subscription services,...
The State-Led Crackdown on Grok and xAI Has Begun
A coalition of at least 37 U.S. attorneys general has opened investigations and issued an open letter demanding xAI halt the creation of non‑consensual sexual imagery by its Grok chatbot. A Center for Countering Digital Hate report found Grok generated...
LLM-as-a-Courtroom
Falconer introduced an “LLM‑as‑a‑Courtroom” system that automatically updates documentation when code changes, framing the decision process as a legal trial with prosecutor, defense, jury, and judge agents. The architecture extracts exact code and document excerpts, constructs arguments, and lets multiple...
Home Depot Introduces AI Tool to Help Contractors Build Project Material Lists Faster
Home Depot has added a Material List Builder AI to its Project Planning platform, allowing professional contractors to generate detailed, job‑specific material lists in minutes using natural language, voice or existing documents. The tool groups items by project phase, shows...
AI-Powered Content Suggestions Keep Docs Fresh Automatically
Support teams have always found it incredibly hard to keep docs/help centres up to date with policy & product changes, but it's more important now than ever before. Docs are load-bearing infrastructure in the AI era. Hence one of my...
OpenAI CEO Altman Admits He Broke His Own AI Security Rule After Just Two Hours, Says We're All About to...
OpenAI CEO Sam Altman admitted he violated his own rule by granting the Codex model full access to his computer within two hours, citing the agent's reasonable behavior. He warned that the convenience of AI agents can lead users to...
Google’s More Affordable AI Plus Plan Rolls Out to All Markets, Including the US
Google has rolled out its lower‑cost AI Plus subscription to every market where its AI plans are offered, including the United States at $7.99 per month. The expansion adds 35 new countries, bringing the plan to emerging regions after an initial launch...
Bing Webmaster Tools Testing New AI Performance Report
Microsoft is testing a new AI Performance report within Bing Webmaster Tools, offering beta‑level visibility into how often Bing Chat and Copilot cite a site. The report displays daily citation counts, cited pages, grounding queries, and query intent, but it...
GrubMarket Introduces AI Monitoring Tool for Food Distributors and Produce Operators
GrubMarket has launched the Monitoring AI Agent, an artificial‑intelligence tool that continuously watches operational and financial data for food distributors, wholesalers and produce operators. The agent integrates with ERP, accounting, ecommerce and payments systems, sending real‑time alerts when KPIs, inventory...
AI Witnesses Promise Faster Justice, Threaten Bias
Will our courts have AI witnesses? AI witnesses are AI systems that turn logs and sensor data such as car crash records, phone traces, building access logs, and video into a timeline and reconstruction that can influence court cases and...
OpenAI Turns Into Venture Hub, Monetizing Research Tools
>be openai >announce plans to take a cut of openai-aided discoveries >release prism, free tool for research teams >rate limit prism >announce prism+, more free credits, latest models >just promise small equity if you build co >openai becomes next yc >sama used to run yc
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, Ex-Anthropic Team Lead
In this episode, David Duvenaud—former Anthropic alignment‑evals lead and computer‑science professor—explores the "gradual disempowerment" thesis that fully capable AI will economically and politically marginalize humans, threatening liberal democracy. He argues that democracy arose from nations needing productive, educated citizens, but...
Built‑in Agent Swarm Streamlines Multi‑Model AI Review
The agent swarm in K2.5 is a good idea. Anyone using AI seriously has pasted entire outputs into a second…third…fourth model for review and clean up. Wonder how well it works built in.
Disconnected Systems Cause Bad CX; AI Agents Can Fix
Behind almost every bad customer experience with a big company or government is tech systems that aren't connected to each other. This is one of the biggest opportunities for AI agents to improve CX.
Continual Learning Transforms Quickly—Imagine a Year's Impact
This is what continual learning brings us: a body and a voice in a night or two. What is it going to do in a year?
AI Auto‑Responses: A Sales Strategy Mistake
I can't think of a worse sales strategy than auto-responding to social media posts with AI-generated comments that pitch your company. Bad look.
Replit Agent Powers Offline AI Voice Assistant on Mobile
these guys got @replit agent to build a fully local and offline AI voice assistant for mobile 👀
Robot Masters 1,000 Tasks After <24‑hour Demos
Scientists have developed an #ImitationLearning-based method that enabled a robot to learn 1000 distinct manipulation tasks from less than 24 hours of human demonstrations. Learn more in Science #Robotics: https://t.co/T9EKphwp70 https://t.co/R3TFjfdUgd
Helix 02: Fully Autonomous, Whole-Body Dishwashing Powerhouse
Introducing Helix 02 It's our most powerful model to date - it's using the whole body to do dishes end-to-end and it's fully autonomous https://t.co/jtUVCRIsLh