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:
Groundcover Eyes Visibility Gap in Agentic AI Monitoring by Targeting Multi-Step Workflows
Groundcover announced an expansion of its AI Observability service, adding native support for Google Vertex AI and targeting multi‑step agentic workflows. The platform uses a patented eBPF sensor to automatically capture every LLM interaction, token usage, and tool call without requiring instrumentation, keeping data inside the customer’s own cloud. New features provide span‑level token tracking, full execution traces, and configurable focus levels, enabling engineers to correlate AI behavior with infrastructure performance and cost. The BYOC design reduces outbound network traffic by more than 95%, preserving security and multi‑cloud flexibility.
Kooiman Supplies Camera-Guided Hoeing Machine Farming GT
Dutch firm Kooiman Mechanisation has begun supplying the Farming GT, a camera‑guided hoeing machine from German maker Farming Revolution. The system leverages a database of over 100 crops to differentiate crops from weeds, using GNSS RTK and real‑time camera correction...
$4.5M Seed for Realm to Advance AI in Enterprise Sales
Realm announced a $4.5 million seed round led by Frontline Ventures, with participation from HubSpot Ventures, Cal Henderson and Alex Bouaziz. The funding will accelerate the startup’s AI platform that structures fragmented revenue data to automate deal‑critical documents such as RFP responses...
AI Drives Surge in ‘Bug Bounty’ Reports, but ‘Slop’ Is Rising Too
Bug bounty programs across the crypto and open‑source sectors are experiencing a dramatic surge driven by AI tools that can quickly generate vulnerability reports. Cosmos Labs reported a 900% increase in submissions, averaging 20‑50 daily, while HackerOne logged 85,000 valid...
People.ai Becomes Backstory, Redefining How Revenue Teams Get Answers About Their Deals
People.ai announced it is rebranding as Backstory, shifting its focus from raw data and analytics to an AI‑powered Revenue Answers Platform. The new platform stitches together signals from emails, calls, CRM records, intent data and news to generate real‑time answers...
Rethinking Sensors for Physical AI: Why Machines Need to See the World Differently
Traditional sensors have been built for human perception, translating physical signals into formats humans can interpret. As AI moves into the physical world, machines now consume sensory data directly, requiring a shift to machine-centric sensing. The article argues that sensors...
How AI Will Transform The Legal Business Model
Olivier Chaduteau, founder of the AI‑native Day Two consultancy, explains how embedding artificial intelligence has dramatically lifted his firm’s revenue compared with the pre‑AI Day One model. He outlines three adoption phases—initial skepticism, superficial licensing, and the current Phase 3 where...
GPT-ImageGen-2 Converts Entire Ulysses Into Comic
GPT-ImageGen-2 did this in one shot, with just the prompt "turn all of Tennyson's Ulysses into a comic, across as many pages as needed. make it great, include the full text" 10 pages, though it did use what seems to be...
Valeo Deepens Google Cloud AI Deal at Cloud Next
Valeo, a leading automotive Tier 1 supplier, is rolling out Google Cloud’s Gemini for Workspace to all 100,000 employees worldwide, extending a partnership that began in 2008. The deployment follows a broader launch of Gemini Code Assist, which the company says...
South Korea Expands AI Cybersecurity to Safeguard Cloud-Based Education Systems
South Korea’s Ministry of Education and KERIS are expanding an AI‑driven cybersecurity platform to protect private‑cloud environments used by schools and universities. The AI‑based Automated Cyber Intrusion Detection and Notification System recorded roughly 480 million threat indicators in 2025, confirming 86 000...
Watershed Launches New AI Agents to Clean “Messy” Sustainability Data
Watershed, a climate‑solutions platform, unveiled AI‑driven agents that clean and structure messy sustainability data. The agents automate unit conversions, duplicate removal, missing‑value handling and can produce disclosure‑ready ESG reports with suggested decarbonization actions. Alongside the tools, Watershed launched an eight‑week...
Meta's Employee Surveillance for AI Sparks Internal Outcry
Sources: Meta's move to track how employees use their computers to train AI caused such a strong reaction that a follow-up note was posted internally acknowledging “there has been a lot of concern about this." “This makes me super uncomfortable. How...
SUSE Ascendant
SUSE, the long‑standing open‑source Linux vendor, is emerging at the nexus of digital sovereignty, cloud‑native adoption, virtualization disruption and AI‑driven infrastructure. The company’s 2020 acquisition of Rancher and later NeuVector gave it a Kubernetes‑centric platform and built‑in container security, positioning...
Meta Employees Freak Out at Training Their Replacements
Meta has rolled out a mandatory software that captures employees' keystrokes, mouse movements, and screen content to feed its generative‑AI models. The initiative triggered a wave of internal dissent, with staff voicing privacy and surveillance concerns. An internal follow‑up memo...
AI Mastery Demands Skill, Not Laziness
The developers who say "AI makes you lazy" haven't tried using it for a full week. Using AI well is harder than writing the code yourself. You have to know what to ask, when to push back, and when to throw...
AI Lab Behind Coca-Cola’s AI Christmas Ad Launches in the UK
The AI lab that produced Coca‑Cola’s AI‑generated Christmas commercial is opening a dedicated UK operation. The new office will be overseen by co‑founder Johnny Rohrbach and UK managing director Daniel Lipman. The move aims to bring generative‑AI creative capabilities to...
Linexa Closes €2M Pre-Seed to Advance AI-Driven Manufacturing
Munich‑based Linexa announced a €2 million (≈$2.2 million) pre‑seed round led by Project A, with several angel investors joining. The startup is building an AI‑driven platform that decodes legacy industrial automation across multiple vendors, turning opaque machine‑control logic into a unified data...
Antony Pegg: Introducing the AI DBA Workbench: PostgreSQL Monitoring That Diagnoses, Not Just Reports
pgEdge has launched the AI DBA Workbench, a self‑hosted PostgreSQL monitoring platform now in public beta. The solution combines a zero‑agent collector with 34 built‑in probes, a three‑tier anomaly detection engine, and an LLM‑driven AI assistant named Ellie that can...
New Algorithms Help Surgeons Make High-Stakes Transplant Decisions in Minutes
Researchers at the ISHLT meeting unveiled AI tools designed to speed heart‑transplant decisions, aiming to cut the 15‑30 minute evaluation window. The flagship model, TOPHAT, analyzes 20 donor variables to predict a center’s likelihood of accepting a heart, while a...
Amazon Launches Alexa+ in Mexico in First Non-English Market
Amazon has rolled out Alexa+ in Mexico, marking its first launch in a non‑English market. The assistant is trained to recognize Mexican Spanish slang and cultural cues, such as “chido” and the flexible meaning of “ahorita.” Amazon achieved this by...
Insilico Tops AI Drug Discovery, Drives Oncology Breakthroughs
I had to skip this year's #AACR2026 but I am watching it from the distance. Our posters and team did a very good job - it looks like our clinical-stage oncology assets are finally being better understood by the pharma...
New AI-Powered Tool to Unlock Decades of Planning Data
The Ministry of Housing, Communities and Local Government, together with the Department for Science, Innovation and Technology’s AI Incubator, has built an AI‑driven tool called Extract to digitise decades of planning records trapped in scanned PDFs, historic maps and paper...
The AI Agent Race Is on – and Google Wants to Win It
Google CEO Sundar Pichai used the Las Vegas cloud conference to unveil Gemini Enterprise, a unified branding for the company’s AI suite aimed at enterprise customers. The move repositions Vertex AI as the core platform for building custom AI agents...
RAM Price Relief? SK Hynix Plans $13-Billion Korean Fab
SK Hynix announced a $12.85 billion investment to build a new advanced‑packaging fab in South Korea, with construction starting this month. The plant will focus on high‑bandwidth memory (HBM) chips that power AI accelerators, addressing a surge in demand driven by...
OpenGradient Receives $9.5M in Funding
OpenGradient, a decentralized compute layer for verifiable AI, announced a $9.5 million funding round to scale its network of GPU and Trusted Execution Environment (TEE) nodes. The round was led by a16z crypto with participation from Coinbase Ventures, SV Angel, and...
Can Europe Ever Compete with California?
At Salesforce’s TDX developer conference, the company showcased a new AI‑first strategy aimed at turning customers into "agentic companies" through tools like Slackbot agents, agentic scripting and integrated large‑language‑model services. The event featured deep collaborations with Anthropic and OpenAI, underscoring...
Trusted Media Brands Gets An AI Assist To Sell Cross-Platform Audiences
Trusted Media Brands (TMB) is deploying generative AI from Jasper.ai to stitch together disparate audience data across print, web, social, streaming and newsletters. The AI parses massive performance datasets, turning them into concise narratives that help sales teams answer RFPs...
From Governance to Enablement: How Healthcare CIOs Can Stop Killing AI Innovation
Healthcare CIOs risk stifling AI projects by treating governance as a barrier. Tony Pastorino argues that rebranding oversight as data enablement and forming cross‑functional teams can unlock responsible innovation. He advises separating idea generation from compliance checks and focusing on...
HR Perspectives by John Dawber: “Adoption Improves Dramatically when AI Becomes Tangible in Daily Work Rather than an Abstract Future...
Novo Nordisk’s Global Business Services leader John Dawber stresses that AI adoption in pharma accelerates when it becomes a tangible part of everyday work rather than an abstract concept. He argues that hiring should focus on learning agility and growth...
BLOG: Will AI Eat Your Job in Estate Agency? No – and Here’s Why
The blog argues that AI will not replace estate agents because the core of the job remains people‑centric—valuations, viewings, and relationship building require physical presence. While portals such as Rightmove and Zoopla have already disrupted marketing, AI is likely to...
CNN Hunts Real Patient Stories on ChatGPT Health Use
It must be a milestone if CNN is actively looking for patient stories of using ChatGPT for health issues. Using generative AI is not newsworthy enough to be covered frequently. Also, there are many patient stories out there proving that...
OpenAI Open-Sources Euphony: A Browser-Based Visualization Tool for Harmony Chat Data and Codex Session Logs
OpenAI has open‑sourced Euphony, a browser‑based tool that transforms raw Harmony chat JSON and Codex session logs into interactive conversation timelines. The web app auto‑detects four data structures, rendering them with metadata panels, JMESPath filtering, focus mode, grid view, and...
IFLYTEK Positions Singapore as Regional AI Hub at GITEX
iFLYTEK, a Chinese AI and speech‑technology leader, reaffirmed Singapore as its Southeast Asian command center at GITEX Asia 2026. The Singapore office, launched in 2024, has sold more than 10,000 hardware units and generated roughly RMB 200 million (about $28 million) in revenue,...
OpenAI's $1.5B DeployCo Bet Signals AI‑PE Shift
OpenAI plans up to $1.5B into DeployCo JV (initial $500M equity; $10B valuation). Macro: AI capital shifts into PE. Key factor: equity leverages deal flow. Trading insight: overweight AI infrastructure/PE-linked names. — Viktor Kopylov, PhD, CFA. More insights: t.me/si14Kopylov
AI Code Generator Hits Free‑tier Limit Instantly
Claude: Just say the word and I’ll generate the code me: Go Claude: You’ve reached your free limit https://t.co/tvcsf3CuE0
New AI Lab Core Automation 'Nerdsniped' Researchers From Anthropic, Google DeepMind
Core Automation, an AI startup founded by former OpenAI vice president Jerry Tworek, announced its launch on X, branding itself as "the world's most automated AI lab." The company has quickly attracted top talent, including former Anthropic researcher Rohan Anil...
Musk's $60B Cursor Deal Validates AI Coding; Kilo Emerges
Elonmusk paying $60B for cursor just validated the entire ai coding category. Kilo is the open source play for what comes next. built for autonomy. built to last.
AI Draws a Quirky Map of Alberta
I asked an AI engine to draw me a map of Alberta. It has a few issues, but it's definitely a map. https://t.co/jr4P1F9iic
Conversational AI Is Rewriting the Customer Service Playbook
Customer expectations for instant, personalized service are at an all‑time high, prompting brands to move from reactive scripts to proactive, AI‑driven interactions. Conversational AI leverages natural‑language understanding, intent and sentiment detection to anticipate problems before customers reach out. By integrating...
AI Coding Tools Empower Domain Experts in Non‑tech Firms
One thing I loved about this episode is @levie's hypothesis that non-tech companies have been limited by access to software engineering talent previously. Now those companies' extremely talented domain experts, combined with tools like Claude Code, unlock untold opportunity.
Scaling Up: More Mac Studios Join NVIDIA DGX Setup
Current local setup: 2x NVIDIA DGX Mac Mini Mac Studio (not pictured) Soon to come: More Mac Studios https://t.co/4GtfgqJXAX
Token Cost Conundrums
Tokenization is the billing backbone for AI APIs, but each model uses its own proprietary tokenizer, turning identical prompts into different token counts. This lack of a standard unit creates opaque pricing, making cost forecasting difficult for enterprises. Benchmarks show...
All-Day Smart Glasses Powered by Claude Code
The Holy Grail would be able to have smart glasses and Claude Code while walking around with those all day.
Report: In the S&P 500, Disclosure of AI Risks Surges From 12% to 83%
A new Conference Board report shows that AI risk disclosures among S&P 500 companies exploded from 12% in 2023 to 83% by the end of 2025. Despite this surge, board-level AI expertise remains scarce, rising only from 1.5% to 2.7% while...
‘No Accountability, No Checks and Balances, No Responsibility’: How Indigenous Peoples Think About AI
The Relational Futures project surveyed Aboriginal and Torres Strait Islander peoples to capture their first qualitative baselines on artificial intelligence. Participants voiced deep mistrust, warning that AI can exacerbate existing inequities in welfare, health and aged‑care systems such as the...
Marketing Strategists Search for a Solution to AI’s All-Too Predictable Outputs
Marketing strategists are confronting a "sameness trap" as generative AI tools like ChatGPT and Claude deliver increasingly predictable, average outputs. Agencies such as Zeal and Bodacious have built custom bots and "second brain" workflows to push AI beyond the mean,...
Viewpoint — Digital Freedom: How AI Could Undermine Chinese Communism
The article argues that advanced, uncensored large‑language models (LLMs) could erode the Chinese Communist Party’s information monopoly. By moving dialogue into private, AI‑driven conversations, the traditional Great Firewall’s choke points become ineffective. The piece highlights how AI systems, built on...
MSI Unveils the Cubi NUC AI 3MG: A 0.51-liter Mini PC Designed to Handle AI, Edge Computing, and Office Tasks...
MSI introduced the Cubi NUC AI 3MG, a 0.51‑liter mini PC aimed at business and edge AI workloads. It ships with Intel Core Ultra 9 386H, offering up to 100 TOPS of AI acceleration, dual 2.5 GbE, Wi‑Fi 7, and support for four displays. The...
Google Is in Talks with Marvell About Two New AI Chips—Putting Pressure on Broadcom
Google is reportedly in talks with Marvell to develop two custom AI chips—a Memory Processing Unit and a new TPU optimized for inference. The discussions come after Google signed a long‑term AI‑chip supply agreement with Broadcom that runs through 2031....
Securing Acting AI: Prioritize Safety Over Hallucinations
Best practices for building agentic systems “Given LLMs’ propensity to hallucinate and deviate from expectations, security is perhaps the most important element of building safe agentic systems.” “You’re no longer securing software that suggests, you’re securing software that acts.” https://t.co/LrCrRC8taq #AIGovernance