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:
South Korea Enforces World-First AI Law Requiring Watermarks
South Korea’s AI Basic Act, enacted in early 2024, mandates digital watermarks on all AI‑generated content and requires risk assessments for high‑impact AI systems. The law imposes fines up to 30 million won after a one‑year grace period, aiming to position the country as a global AI leader. Start‑ups argue the compliance burden is heavy, while civil groups say the law lacks strong citizen protections. Enforcement will hinge on regulators’ ability to verify watermarks and risk‑assessment reports.
AI Builders Worry About Leaks, Not Claw‑bot Uprisings
from my convos (just fyi), nobody who is seriously building and stress testing agent capability is worried about the claw bots in moltbook we saw similar reactions w autogpt/babyagi, smallville, project sid, truth terminal, etc. which are all significant but small...
APAC Compliance Teams Rely on Manual Processes Despite Increasing AI Adoption, Report Claims
A new Fenergo survey of 110 APAC compliance professionals shows that 66% still depend on manual workflows, causing delays and high false‑positive rates in KYC and AML checks. While 54% are actively scouting AI solutions, only 34% have begun deploying...
Improving Acne Detection with Memory Classifiers
The article announces a new approach to acne detection that leverages memory‑based classifiers, a type of AI model that retains contextual information across image sequences. By incorporating temporal memory, the system can differentiate between transient skin changes and true acne...
Anthropic's New Plugins and $3 Billion Lawsuit
The episode examines Anthropic's newly released Co‑Work plugins, which let enterprise users automate niche tasks and integrate AI more tightly into their workflows. It also breaks down the $3 billion lawsuit filed by music publishers accusing Anthropic of copying roughly 20,000...
Show HN: I Built an AI Conversation Partner to Practice Speaking Languages
TalkBits launches an AI‑powered conversation partner that lets users practice speaking multiple languages through short, natural dialogues. The platform emphasizes everyday speech, voice interaction, and instant pronunciation feedback, supporting over ten languages and regional English variants. Its adaptive engine corrects...
CLI‑enabled Agents Risk Identity‑changing Prompt Injections
With autonomous agents who have access to the command line, like Claude code and Open Claw, you don't only have to worry about prompt injection that executes commands and operations, but you also have to worry about prompt injection that...
Moltbook Is a Human-Free Reddit Clone Where AI Agents Discuss Cybersecurity and Philosophy
Moltbook is a Reddit‑style forum populated entirely by AI agents, with over 35,000 bots generating content without human input. The platform runs on OpenClaw, an open‑source harness that lets these models control a host computer’s messengers, email, and web browsers....
Competitors Using AI Cut Mortgage Costs 25%
“It’s not AI that’s your competition. It’s your competitors using AI.” A real-world example from mortgage origination 👇 The frustratingly high cost to originate a mortgage (>$10K!) has persisted for decades. Except for originators on Vesta . Powerful stats from...
Deep Neural Networks Transform Voxel-Based Morphometry Preprocessing
Deep neural networks are reshaping voxel‑based morphometry (VBM) preprocessing by automating key steps such as bias correction, tissue segmentation, and spatial registration. Early trials show up to a 70% reduction in processing time while preserving or improving segmentation accuracy compared...
This Startup, Backed by Bill Gates, Is Looking to Transform Computing as We Know It - Optical Transistors Could Upend...
Neurophos, a startup funded by Bill Gates, introduced the Tulkas T100, an optical GPU that claims 470 petaFLOPS of compute power. The processor integrates large‑capacity RAM and SSD storage, targeting AI and high‑performance workloads. By using optical transistors, the chip...
Stake DIEM for Free Private Clawdbot Access
DIEM is the only crypto token that enables free (and private) use of Clawdbot/@openclaw 1. Create https://t.co/iUPC8ijDPN account and stake DIEM 2. Install @openclaw and select Venice as inference provider 3. Choose Opus 4.5 or Kimi K2.5 (latter is nearly as intelligent...
Economic Superalignment Empowers Aligned Agents, Blocks Misaligned
In anticipation of AGI and the singularity, based on my 10 years of experience designing and struggling with the challenges of DAOs, for the past few years I’ve been developing Synomics - a comprehensive paradigm of superalignment for swarms of...
Anthropic Brings Agentic Plug-Ins to Cowork
Anthropic has added plug‑ins to its Cowork platform, extending the agentic capabilities of Claude Code to non‑coding enterprise users. The plug‑ins automate specialized tasks across departments such as marketing, legal, and sales, and can be customized without deep technical skills. Anthropic...
Model‑Robotics Loop Accelerates Experimental Discovery Rates
Agree completely with @kevinweil this loop between models and robotic labs works today and it’s going to be exciting the directions that scientists take autonomous labs. Feels inevitable that it will yield a great speed up in our...
AI Agent Replaces Staff, Drives $100M ARR in 24 Months
This AI company hit $100M ARR in under 24 months by making a crazy bet on AI agents most startups ignored👇 At the end of 2023, Podium already had 60,000+ SMB customers using the product to manage leads and customer...
Over 175,000 Publicly Exposed Ollama AI Servers Discovered Worldwide - so Fix Now
Security researchers uncovered roughly 175,000 Ollama AI servers worldwide that are publicly reachable because they are bound to all network interfaces instead of localhost. These misconfigured instances lack authentication, allowing attackers to perform "LLMjacking"—hijacking the models to generate spam, malware,...
Give AI Your Context, Get Truly Relevant Answers
NEW: I just dropped something that will completely change how AI responds to you, and it's entirely free. If I could scream from the mountaintops that you need an AI context vault, I would. But because I love my vocal...
Affordable X AI API Unlocks Radical Personalized News
If @elonmusk opened up @xai's API to @OpenClaw (and made it affordable) it would bring highly personalized news to everyone. I'm using @blevlabs AI to do that and it's so powerful. Yes, his AI is better than what is running OpenClaw,...
4CRisk.ai’s Playbook for AI Governance and Compliance
4CRisk.ai released a guide outlining how firms can build a robust AI governance program in 2026. The playbook stresses moving beyond high‑level policies to concrete technical monitoring, risk tiering, and evidence‑based compliance for both in‑house and third‑party models. It recommends...
Robots Gain Finger-Level Force Awareness for Reliable Tasks
Robots are getting good at copying motion. Force is the harder problem. Too little force and the task fails. Too much force and something breaks. UMI-FT tackles this gap by adding something most learning setups miss: finger-level force awareness. Each...
Overreliance on AI Dulls Thinking, Smart Use Boosts Learning
Over reliance on AI is poison to your brain. The MIT essay study and this Anthropic SWE study are pointing us at the same thing: if you delegate the brainy stuff to AI, maybe you gain a little speed, but...
Mito AI Raises $4.5M to Empower Video Professionals with AI Tools
Mito AI, a collaborative platform that brings AI orchestration tools to video professionals, announced a $4.5 million Series A round led by Lightspeed Venture Partners, with participation from Kibo, Kfund, Sequoia, a16z scouts, LifeX and several high‑profile angels. The service lets filmmakers,...
US Restrictions Spur Allies to Pursue Sovereign AI
U.S. policies are driving allies away from using American AI technology. This is leading to interest in sovereign AI — a nation’s ability to access AI technology without relying on foreign powers. This weakens U.S. influence, but might lead to...
Agentic AI Requires Planning, Memory, Tools, and Reflection
Agentic AI — finally explained without the hype. This sketchnote gets it right 👇 Agentic AI isn’t “better prompts.” It’s systems thinking applied to AI. Five fundamentals that actually matter: 1️⃣ Agents = Systems Reasoning + memory + tools +...
UK Based Sovereign AI Partners with Accenture, Palantir for AI Infrastructure Advancement in EMEA
Sovereign AI has entered a strategic partnership with Accenture and Palantir to build hyperscale AI data centers across the UK and broader EMEA region. The consortium will integrate Dell’s AI Factory platform and NVIDIA’s GPU hardware, while Palantir’s Chain Reaction...
Perplexity Adds Microsoft’s Azure as a Cloud Service Provider
Perplexity AI signed a three‑year, $750 million agreement with Microsoft to run its services on Azure, leveraging the Azure Foundry platform for access to frontier models from X, OpenAI and Anthropic. The partnership runs alongside Perplexity’s existing relationship with Amazon Web...
Focal Brings Research-Backed Behavioural Coaching to Wealth Advisors
Focal, an AI‑driven productivity platform for wealth advisors, has partnered with behavioural‑science firm Shaping Wealth to embed research‑backed coaching into its meeting workflow. The integration delivers real‑time, personalized insights during client conversations, helping advisors ask better questions, improve prospecting and...
Ex-CEO Behind Bars, 10x’s Clinical Ambitions, Pharma’s AI Gambit
The former CytoDyn chief executive was sentenced to 30 months in prison after convictions for securities fraud, wire fraud and insider trading. Meanwhile, major pharmaceutical companies are accelerating AI adoption, signing multi‑billion‑dollar deals with platform providers such as Nvidia and...
PlannerPal Launches Pre-Meeting AI Tool for Financial Advisers
PlannerPal, an AI‑driven workflow platform for financial advisers, has introduced a Pre‑Meeting Prep Pack that automates the assembly of client‑specific data before consultations. The tool aggregates portfolio valuations, tax summaries, goal progress, and recent communications from multiple sources, cutting typical...
NDSS 2025 – Provably Unlearnable Data Examples
The NDSS 2025 paper tackles the lack of provable guarantees in machine‑unlearning by introducing a certification framework for (q, η)-learnability. It proposes Provably Unlearnable Examples (PUEs) that tighten certification bounds and resist simple weight‑recovery attacks. Experiments show PUEs cut certified learnability...
CEOs Should Reward Employees Letting AI Take Repetitive Work
Why CEOs Should Incentivise Employees to Replace Themselves With AI Push innovation — this controversial view argues CEOs should encourage staff to let AI take over repetitive tasks, freeing up talent for higher value work. Read more 👉 https://lnkd.in/eJTrBVYx #AI...
Mc+qubo Achieves Improved Reinforcement Learning with Quadratic Unconstrained Binary Optimisation
Researchers introduced MC+QUBO, a method that recasts Monte Carlo episode selection as a Quadratic Unconstrained Binary Optimisation problem and solves it with quantum‑inspired samplers. By using Simulated Quantum Annealing and Simulated Bifurcation, the approach filters trajectories to maximise reward and promote...
AI Sharpens Pricing Strategy, Not Replaces It
Pricing is no longer just spreadsheets and gut feeling. Generative AI can simulate scenarios, test price elasticity, and surface insights that once required whole analytics teams. But the real edge doesn’t come from the model alone. It comes from asking the right...
Context Engineering and the Future of AI-Powered Business
The Mint interview highlighted a shift from prompt engineering to context engineering as the key to scaling personalized AI interactions. Salesforce executives explained that embedding customer history, business rules, and emotional cues into AI models creates truly human‑like conversations. Real‑world...
AI Revolutionizes Customer Research: Human Insight Still Essential
I wrapped up Season 5 of Building With Buyers , but then this convo was TOO GOOD + too relevant so then I un-wrapped up Season 5 😂 at least for a bit Caitlin Sullivan joined me to talk about...
Amazon in Talks To Invest Up To $50 Billion in OpenAI
Amazon is reportedly negotiating a multi‑billion‑dollar investment of up to $50 billion in OpenAI, the creator of ChatGPT. The funding round could total $100 billion, potentially valuing OpenAI at $830 billion. Amazon CEO Andy Jassy is leading talks with OpenAI CEO Sam Altman,...
Is 'Shadow AI' A Threat to Your Business? Report Claims Workers Are Increasingly More Willing to Cut Corners and Take...
A new industry report warns that “shadow AI” – unsanctioned artificial‑intelligence tools used by employees and managers – is spreading rapidly across enterprises. The study found a 45% year‑over‑year increase in the use of unapproved AI applications, driven by workers’...
How Context-Aware Agents and Open Protocols Drive Real-World Success in Enterprise AI
Enterprise AI is shifting from experimental LLM pilots to production deployments that require real‑time operational context. General‑purpose models lack domain‑specific data, prompting firms to adopt small, specialized language models (SLMs) for cost‑effective, on‑premise inference. The Model Context Protocol (MCP), an...
Machine Learning on Systematically Curated Data Reveals Key Determinants of Magnetic Hyperthermia Performance
Researchers introduced a machine‑learning framework that predicts the specific absorption rate (SAR) of superparamagnetic iron oxide nanoparticles (SPIONs) used in magnetic hyperthermia. The model was trained on a systematically curated dataset of 1,850 entries spanning 30 intrinsic and extrinsic features...
Genie 3 AI Generates Real‑Time Functional Maps
One of the wildest emergent capabilities of Genie 3 is that maps actually work. As I walk around the forest, the GPS display updates its heading in real time. Remember. There is no game engine here. This is an AI hallucinating a...
Moltbook Shows LLMs Remain Pure Next‑Token Predictors
Yes, Moltbook (by clawdbot) is still next-token prediction combined with some looping, orchestration, and recursion. And that is exactly what makes this so fascinating. (It is also why understanding how LLMs actually work really does pay off. Lets us see through the...
Clawdbot Represents Anthropic’s Biggest Safety Nightmare
clawdbot is basically the manifestation of Anthropic’s greatest nightmare safety team must be losing it
Policy Boosts Dexterous Robots, Paving Way for Child‑like Learning
A new Science #Robotics Focus article highlights the merits of a new policy that enables dexterous robotic manipulation and speculates on how future research may enable robots to learn from humans the way a child learns from watching its parents....
Hiring Systems &
Hiring a: + Systems Engineer and an + AI/ML Research Engineer Open to freelance or full time - just shoot a DM with the work you’re most proud of and a few lines about yourself Plus if you’ve got experience with mechatronics
AI Proliferation Shallowens Bugs, Boosts Top‑tier Fixes
Have we reached the stage of “many AIs make all bugs shallow”? Great writeup on AI, open source, & bug bounties by @stanislavfort cofounder of AISLE “Mass adoption collapsed the median quality (“slop” killed bug bounty..) but.. raised the ceiling” https://t.co/iDvdiDy41J
Claude Code Executes “Ship It” Instantly
Having spent enough time with Claude Code, now when I type "ship it" - it knows exactly what to do.
AI Agents Launch Decentralized Social Network to Debug OpenClaw
If you aren't paying attention, the AI agents, running on @OpenClaw (formerly ClawdBot) have their own social network now and are working together to debug it. Decentralized computing is finally here.
New Test Suggests ChatGPT May Be Pulling Google Results
Interesting new test by @MalteLandwehr showing how ChatGPT likely uses Google, plus my thoughts: https://t.co/ndqonHJoTq
Add a Trust KPI to Balance AI Visibility
Do you have a TRUST KPI to balance out your AI visibility KPIs? In a world of infinite content, the need for TRUSTED curators goes up. Are you building high AI visibility and low trust building content? https://t.co/YM2Eg6SB8l https://t.co/qg67K7ZwMF