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:
CEO Interview with Dr. Mohammad Rastegari of Elastix.AI
Elastix.AI, led by former Meta and Apple AI scientist Dr. Mohammad Rastegari, is building a reconfigurable FPGA‑based inference platform that promises dramatically lower cost and power consumption for large‑scale AI models. The company claims up to ten‑fold reductions in inference spend and five‑fold gains in energy efficiency compared with traditional GPU clusters. By coupling advanced model optimization with hardware that can be re‑programmed after deployment, Elastix.AI aims to eliminate the rigid, costly refresh cycles that currently dominate AI infrastructure. Its business model spans token‑as‑a‑service, software subscriptions, and hardware leasing or sales.
A Used RTX 3090 Is Still the Best GPU for Local AI in 2026, and It's Not Even Close on...
The used Nvidia RTX 3090 remains the top‑value GPU for local AI workloads in 2026, offering 24 GB of VRAM and solid performance for $600‑800. Its VRAM‑per‑dollar ratio eclipses newer cards like the RTX 4090 and RTX 5090, which command multi‑thousand‑dollar price tags. The...
Insurers Split: Covering vs Excluding AI Agent Errors
As businesses increasingly rely on autonomous AI agents, some insurers are offering specialized coverage for AI errors, while others introduce exclusions, signaling a shift toward explicit policies addressing AI-related risks. aibusiness
What Is Ollama, How to Use It, and Why It Matters (10 AI Prompts to Test Local Models)
Ollama is a lightweight platform that lets users run advanced large‑language models directly on their own machines, bypassing cloud APIs. It packages popular open‑source models into a simple command‑line interface, enabling instant local chatbots. By keeping inference on‑device, Ollama eliminates...
Accelerate SaaS R&D 5‑10× with AI Bootcamp
Our first-ever AI Coding Bootcamp starts on April 15. It is for SaaS companies who want to 5-10x their R&D process speed. It will be 90 minutes per week for 8 weeks. Learn more at the link in the first...
AI Is Now Essential for Every SaaS Product
Spending lots of time with Claude Code in the last months (+ OpenClaw more recently) has made it abundantly clear to me that every piece of SaaS hugely benefits from being infused with AI. All not new, this has become clear...
The Power to Shape AI
In this weekend "big‑think" episode of the AI Daily Brief, host Ethan Mollick (via his essays "The Shape of the Thing" and its update) explores how humanity’s agency can shape the rapid evolution of AI agents. He traces the transition...
Chat AI Becomes the Ultimate Natural Language Interface
I'm surprised myself HOW much I'm gravitating towards Claude in the terminal for more and more things. 2 weeks ago I vibe-coded a simple web app dashboard that aggregates data from various bank accounts. Today I replaced it with a few...
Using AI Fuels Optimism; Fear Stems From Avoidance
I've noticed lately that the people most afraid AI is going to take their job and disrupt their earning potential have spent the least amount of time actually using AI. And those using AI more and more tend to be...
Google App Starts Rolling Out Quick Access to AI Mode History
Google is rolling out a persistent AI Mode history button in the Android Google app, replacing the Labs beaker icon on the Home tab. The new UI slides recent AI chats into view, mirroring the Gemini app’s menu and aligning with...
ATO Issues Agency-Wide Guidance on Authorised Use and Liability of AI
The Australian Taxation Office has released agency‑wide guidance governing the authorised use and liability of artificial intelligence. The policy encourages staff to leverage AI tools, such as those that draft briefs for ministers, while imposing strict governance, risk‑management and accountability...
The Bitter Lesson
The "Bitter Lesson" argues that raw scale—more data, compute, and larger models—consistently outperforms clever, hand‑crafted algorithms. Historically, breakthroughs from Deep Blue to AlexNet illustrate this pattern, and modern large language models reinforce it. AI developers spend months fine‑tuning prompts only to...
AI Agents Make Personal Research Virtually Costless
Building experimental stuff nowadays is so worthwhile. The opportunity cost is near 0: you just have these agents run research and analysis and notify you when they're done. I know @karpathy's autoresearcher does this but feel like you can take...
FlyGuys Reports Seventh Consecutive Year of Growth, Cements Position as Infrastructure Layer for Enterprise Reality Data
FlyGuys announced a 40% year‑over‑year revenue increase, marking its seventh consecutive year of double‑digit growth. The company now operates a network of more than 20,000 FAA‑certified drone pilots that deliver standardized reality data for AI‑driven applications across construction, energy, agriculture...
Context Engineering Emerges as ‘Next Battleground’ for Enterprise AI
Elastic’s chief product officer Ken Exner warned that enterprise AI is moving from prompt‑centric experiments to “context engineering,” where the quality and relevance of data feeding large language models determines success. He outlined the evolution from early hype, rapid pilot...
AI Uncovers Drug for Dog's Cancer, Pharma blocksAI Uncovers Drug for Dog's Cancer,
Dog diagnosed with cancer → Owner asks ChatGPT for solution → gives tumor sample to University to be DNA sequenced → Owner puts sequence through AlphaFold program to identify targetable protein and finds a helpful drug → asks drug company...
Meet OpenViking: An Open-Source Context Database that Brings Filesystem-Based Memory and Retrieval to AI Agent Systems Like OpenClaw
OpenViking is an open‑source context database from Volcengine that reimagines AI‑agent memory as a virtual filesystem. It maps resources, user data, and agent skills into hierarchical directories accessed via a `viking://` protocol, replacing flat chunk‑based RAG with directory‑aware retrieval. The...
AI Onboarding Saves Founder Hours, Boosts Team Readiness
Every new hire at a wealth management firm means weeks of personal training by the founder. Victor Medina knew this bottleneck was unsustainable—but didn’t know how to solve it without risking client experience. Inside the AI Business Lab® Mastermind, he built...
Nvidia's Evolution: From GPU‑Centric to AI‑Optimized CPU Platform
Like I said after GTC 21, when Grace was first introduced -- @nvidia has been methodically rewriting what kind of company it is, one conference at a time. The CPU is taking center stage at this GTC, and that's not...
AI Generates Nude Images that Outrank Real Photographs in Sexual Appeal, Study Finds
A recent study in the Archives of Sexual Behavior found that AI‑generated nude images of women are perceived as more aesthetically pleasing and sexually attractive than real photographs, even though the photos are judged more realistic. Researchers surveyed 649 Czech...
AI Writes My Newsletters in Minutes, Keeps My Voice
Wrote my next 4 months of newsletters this morning. Took me 90 minutes in total with ChatGPT/Claude. In the past, this would have taken me 10+ hours. Here was the process: 1. Export all previous newsletters 2. Upload file into a Claude...
AI-Powered Morning Brief Boosts Daily Productivity
My morning brief is by far my favourite skill I have built. Every morning AI delivers to me: 1. Calendar scan — What's on today, what needs prep, what's going to eat my deep work block. 2. Task triage — Pull today's...
Anthropic Is Doubling Claude's Usage Limits During Off-Peak Hours for the Next Two Weeks
Anthropic announced a two‑week promotion (March 13‑27) that doubles Claude’s usage limits during off‑peak hours, specifically a five‑hour window from 8 AM to 2 PM ET. The boost applies automatically to Free, Pro, Max and Team plans across web, desktop, mobile, as well...
AI Code Refactoring Requires Relentless Prompting and Patience
Spent last night refactoring terribly written software 🤖 into manageable pieces and refactoring requirements. Even though the code had a number of correct examples after I got it to deploy a few resources with the correct pattern, the agent continued...
Skipping Negative Impact Costs Wins, AI Shows
I’ve written 1,394 LinkedIn posts about Discovery. But according to the data… I still have more to learn: Last week, we released AI Skill Intelligence on a limited basis. It measures your proficiency against industry-standard revenue skills. So I ran my own discovery calls...
How Materials Informatics Aids Photocatalyst Design for Hydrogen Production
Researchers used machine‑learning interatomic potential (MLIP) calculations to screen dopants for orthorhombic Sn₃O₄, identifying aluminum as a stable dopant. Experimental hydrothermal synthesis confirmed the predictions, with 5 % Al‑doped o‑Sn₃O₄ delivering 16‑times higher hydrogen production under visible light. The study demonstrates...
Tesla's Optimus Robot Shows Off Dancing Moves
Elon Musk and TESLA Optimus Dancing l? https://t.co/Hku90HZNGg #tesla #optimus #humanoidtech #humanoid #robot #Robotics #AI #TechRevolution #TechInnovation #ArtificialInteligence #PhysicalAI @AlbertoEMachado @Eli_Krumova @postoff25 @Khulood_Almani @anand_narang @NutritiousMind @baski_LA @TanyaSinha_ @devaang @AlAmadi1 @jeancayeux @enilev @efipm @mvollmer1 @Nicochan33 @RagusoSergio @FrRonconi @Shi4Tech @sallyeaves @LaurentAlaus @Fabriziobustama @smaksked @MargaretSiegien...
AI Can't Replace Core Software, Only Simple Prototypes
Nope, @elonmusk, not even close. Software has not been eaten by AI. Nobody has replaced an operating systems, or Excel or GPS or their favorite browser or favorite AAA video game entirely with AI. The results just aren’t high enough calibre. Sure,...
The Industrialization of Algorithm Design: AI-Driven Research for Systems
UC Berkeley researchers introduced AI‑Driven Research for Systems (ADRS), a closed‑loop framework where large language models iteratively generate and refine system algorithms using simulators as hard verifiers. The approach treats code generation as an evolutionary search, allowing the LLM to...
ChatGPT Aids Owner in DIY Dog Cancer Vaccine
#DidYouKnow that ChatGPT helped an owner save his dog's life with a DIY cancer vaccine? (Daily Horizon News) #MedTech #JVGpost https://t.co/XCTSYrV2BF
PE Firms Eye Anthropic AI to Boost Portfolio Innovation
The market is counting out private equity-owned (and private credit-financed) companies, even though they could be the firms that pursue AI most aggressively. A consortium of PE firms are in talks to integrate Anthropic's AI into their portfolio companies. https://t.co/ajG448PXen
Microsoft Quietly Scraps Plans to Bring Copilot to Notifications and Settings on Windows 11 as It Moves to Reduce AI...
Microsoft has abandoned its plan to embed Copilot‑branded AI features—such as Settings integration, notification suggestions, and an autonomous File Explorer menu—into Windows 11. Announced in 2024 and demoed publicly, the features never entered preview and have now been officially scrapped. The...
Future Breakthroughs Will Come From Beyond Parametric Models
The next major breakthrough will branch out at a much lower level than deep learning model architecture. It will be a new approach. A better model architecture can lead to incremental data efficiency & generalization gains, but it won't fix...
Google, OpenAI, Anthropic Poised for Recursive AI Breakthrough
The failures of both Meta and xAI to maintain parity with the frontier labs, along with the fact that the Chinese open weights models continue to lag by months, means that recursive AI self-improvement, if it happens, will likely be...
Horace Dediu on Apple Sitting Out the AI Spending Race
Apple is deliberately avoiding the $650 billion AI‑infrastructure spend that hyperscalers like Amazon, Google, Microsoft and Meta are pouring into data centers. Instead, it keeps its capital budget around $14 billion, licensing Google’s Gemini model for roughly $1 billion a year and embedding...
Google Photos AI Enhance Revives Old Pics Instantly
The "AI Enhance" feature in Google Photos on Pixel phones can dramatically improve even old photos with just a tap or two. Here's how to get started with this powerful new tech... https://t.co/WVe3OXidnW #photography #android https://t.co/IQ5H6LHJwV
30% of AI Model Power Still Fossil‑fuel Driven
Wait until he finds out that 30% of the energy that drives his models are fed by fossil fuels
Autonomous Maritime Defense AI: Next VC Goldmine
1/ Fully automonous + maritime defense AI Institutional Investor: "Could Be the Next Big Investment Opportunity in Venture Capital... "... https://t.co/NoMGiNKai6
AI Should Preserve, Not Replace, SMB Human Edge
RT SMBs don't need "more AI" to win - they need AI that protects their human edge: relationships, culture, and trust. I dig into how small businesses can use tech to make work more human, not less. #AI #SMB @Star_CIO https://t.co/Gb3ialWSOO
Applied Intuition Poised as AI‑physical World Leader
Lux family co Applied Intuition 🇺🇸 will be one of the largest, most important companeis at the intersection of AI + physical world.
Master the AI Stack, Not Just Prompts
AI is layered 🧊 AI → Machine Learning → Neural Networks → Deep Learning → Generative AI GenAI is the visible tip. The real power sits underneath. Master the stack, not just the prompt. #AI #ML #DeepLearning #LLM https://t.co/lpckDcmRFN
Leaders Boost AI Adoption by Engaging Teams in Use‑Case Design
Leaders can drive AI adoption by involving teams in defining use cases. Task groups with identifying challenges and brainstorming AI solutions for their specific organizational pain points. #AIStrategy #Leadership #Innovation https://t.co/CTf0pDNecw
Use AI to Redesign Your Org, Not Just Emails
Most leaders are asking: “How can AI speed up my email?” The better question: “How should AI reshape my entire org chart?” Task-level thinking produces incremental improvements. Structural thinking produces transformational ones. We do the structural thinking inside the AI Business Lab® Mastermind....
Nvidia GTC Preview: Top 10 Highlights & Cadence Question
Preview - Will @Nvidia keep its yearly cadence? Join me for the full preview at noon PST - My top 10 Countdown of Nvidia's GTC conference in #SanJose https://t.co/U4ScbXTjgy
AI's Potential Exceeds Companies' Adoption Limits
The #AI Adoption Gap: AI Can Do More Than Companies Allow by @niritcohen @Forbes Learn more: https://t.co/sGpSukEUm1 #ArtificialIntelligence #MachineLearning #ML https://t.co/f2Lkigc08I
AI Will Slash Service Costs to Near Zero
Everyone talks about AI taking jobs... ...Hardly anyone talks about the price collapse coming to services. Legal help Financial advice Tutoring Software development Medical diagnostics Design Many things that cost thousands of $ today could become nearly free. https://t.co/Kr5quDRfV2
Seeking Custom LLM‑Powered Voice Assistant with App Integration
Is anyone making an Alexa-like speaker/mic product that connects to your LLM of choice + relevant critical apps?
AI Infrastructure Commitments Under Scrutiny Amid 340% Lease Surge
Datacenter leases: up 340% to $700B globally. But here is the question nobody in the room wants to answer: "If we had to audit every AI infrastructure commitment we made this year, how many survive scrutiny?" Sometimes the hardest meeting prompt is the...
Dynamic Evaluation Framework for LLMs in Patient Care
Our new @NatureMedicine paper proposing a path for better, dynamic evaluation of large language models for patient care https://t.co/VHvyxDFKuc CES-clinical environment simulator @pranavrajpurkar https://t.co/Q9tyeupEk3
AI Will Augment Executives, Upskill Secretaries
Graphing out where AI complements human work (e.g. chief executives) & where it could pressure them to learn new skills (e.g. secretaries). More details: https://t.co/59AK5OVnoQ Credits: xY2j-Ib2p9--NmEX-43- on Reddit & EBRD https://t.co/0qXOgPYYnN