Today's AI Pulse
Pinterest inks $4 billion AWS cloud pact to power AI visual search
Pinterest announced a $4 billion multi‑year agreement with Amazon Web Services, the largest infrastructure deal in its history, to power AI‑driven visual search and discovery through 2031. The contract expands use of AWS Trainium chips for large‑language and vision‑language models and adds a migration to a Kubernetes‑based architecture.
Also developing:
AI Juggernaut Rumbles on Even as Markets Whipsaw
AI‑related credit markets are surging despite Middle‑East tensions and rising energy prices. Morgan Stanley still forecasts $400 billion of high‑grade AI debt issuance this year, with hyperscalers alone expected to sell over $100 billion in new bonds after already raising more than $80 billion in Q1. CoreWeave added $1.75 billion of high‑yield junk bonds to fund its AI compute partnership with Meta, while private‑credit lenders are mobilising billions for data‑center projects. Investor appetite for AI‑linked financing remains strong, even as broader credit markets show occasional outflows.
Custom Claude Widgets Put AI First on Your Home Screen
This changes how you use AI every day. @Anthropic just gave Claude iOS full widget customization. Now you decide what shows up first not the app. Quick access = faster thinking = better execution. What would you pin on your home screen?
How to Use Claude Like the Top 1% of Users
The guide shows that elite Claude users treat the model as an infrastructure layer, not a one‑off chatbot. By establishing permanent files—about‑me, voice‑profile, anti‑AI style—and a Cowork folder structure, Claude loads full context before each session. Advanced prompting techniques such...
AI Disclaimers Undermine Trust in Political Ads
From the American Association of Political Consultants Foundation: The "disclaimer effect" is real. When viewers saw an AI disclosure on a political ad, their trust and receptivity to the message dropped sharply — even when the ad contained no actual...
'A Self-Inflicted Hit': Washington State Just Rolled Back Sales Tax Exemptions for AI Data Centers Worth Hundreds of Millions
Washington Governor Bob Ferguson signed SB 6231, tightening the state’s sales‑tax exemption that previously covered equipment upgrades for existing data centers. The change leaves new facilities untouched but raises costs for refurbishing AI‑focused hardware. The move follows a national pattern where...
Attack on Titan Studio Slammed for AI Use and It Won't Be the Last Time
WIT Studio, the creator of Attack on Titan, faced fan outrage after using generative AI in the opening sequence of Ascendance of a Bookworm, prompting an apology and a promise to replace the clip. The studio, which posted a ¥170 million...
Venice Keeps AI Chats Local, Adds Verifiable Encryption
"Every major AI platform — ChatGPT, Claude, Gemini, Grok, Perplexity — stores your conversations on centralized servers. That data can be reviewed, classified, subpoenaed, hacked, or sold. You're trusting the company, its employees, its vendors, and every government that asks. Venice...
You Can Get a Free Gemini API Key Right Now with No Billing Required — Here's What to Do with...
Google’s Gemini API now offers a completely free tier through AI Studio, requiring no credit‑card or billing setup. Users can obtain an API key in a few clicks, receiving a string that unlocks Gemini 2.5 Pro, Flash, and Flash‑Lite models with modest...
People Question Need for AI-Generated Artemis Images
I don't understand why there are AI photos from Artemis. I mean, the real pictures are super awesome - there's just no reason for fake. Is there?
First US Newsroom Strike For AI Protections Staged by ProPublica's Journalists
ProPublica journalists, represented by the ProPublica Guild, staged a 24‑hour strike on Wednesday, marking the first U.S. newsroom walkout explicitly demanding safeguards against AI‑driven layoffs. About 150 union members picketed the New York headquarters, with coordinated pickets in Chicago and...
List of FREE AI Courses Offered Directly by the AI Companies Themselves.
Major AI developers—including Anthropic, Google, NVIDIA, OpenAI, Hugging Face, Microsoft, Meta, IBM, Amazon AWS, and DeepLearning.AI—have launched free, self‑paced courses on their own platforms. The list compiled by Sifu Yik aggregates direct links to each provider’s training portal, covering fundamentals...
Project Glasswing: What Power Companies and Grid Operators Need to Know
On April 7, Anthropic unveiled Project Glasswing, a coalition of 12 technology leaders deploying the Claude Mythos Preview AI model to automatically discover and patch software vulnerabilities. The model has already identified thousands of zero‑day flaws, including a 27‑year‑old bug in OpenBSD and chained...
Think ChatGPT Can Replace Your Lawyer? Think Again
Artificial intelligence tools such as ChatGPT, Gemini and Claude are being marketed as legal assistants, but they are not licensed to practice law. Recent high‑profile incidents—Krafton's $250 million contract dispute in South Korea and OpenAI’s $10 million punitive lawsuit in Illinois—show that...
AI Accelerates, Human Insight Delivers Real Value
The reason that we're using AI is because it accelerates the iterations that we're looking to do within our companies. AI alone won't provide the use case. AI + human skill and creativity = the use case that we are all looking...
AI Generates Custom Classroom Infographics for Any Subject
📚 From plant cells to plot diagrams, infographics can support learning in lots of different subjects. 🤖 In this new blog post, I’m sharing how AI can help you create classroom infographics tailored to your curriculum goals. 🔗 If you’re ready to...
Cadence (CDNS), Nvidia (NVDA) Expand Agentic AI Collaboration
Cadence Design Systems announced an expanded partnership with Nvidia to deliver a suite of agentic AI and physical AI‑accelerated tools for electronic design automation. The offerings, optimized for Nvidia Grace CPUs and Blackwell GPUs, will run on Cadence’s Millennium M2000...
Customize ChatGPT Voice with OpenClaw's Talkable Claw
I'm having a Her moment over here with my OpenClaw. It's literally the craziest thing in the world to be able to just talk to your claw and have fully customizable voice. Did you ever want to improve your ChatGPT...
Atlassian (TEAM) Introduces AI-Powered Visual Capabilities in Confluence
Atlassian announced AI‑powered visual capabilities for Confluence, highlighted by the Remix feature now in open beta. Remix instantly transforms text‑based pages into charts, infographics, scorecards and presentation‑ready summaries while keeping the visuals linked to the original document. The company also...
I've Had a Pixel for Years and Just Found the Photo Editing Tool that Made Me Ditch My Mirrorless
Google introduced a conversational editing feature called "Help me edit" in the Google Photos app, debuting on the Pixel 10 series and now available on iOS and Android. The tool lets users describe desired changes in plain English, automatically selecting the...
OpenClaw Offers Unmatched Freedom Over Locked AI Agents
The most underrated thing in AI agents right now is: OpenClaw/Hermes Agent is just more free than other locked down AI agents (the standard out-of-box Claude/ChatGPT route) "Free the Claw" is not a vibe I understood until I tried it. Now...
Temperature Tuning for Non-Nerds
Renee, a Portland‑based boutique PR firm owner, discovered that using AI on default settings produced hallucinated quotes in press releases and robotic pitch emails, doubling her workload. A March report estimates AI hallucinations cost firms over $67 billion worldwide, prompting regulators...
AI‑Designed Enzymes Break Evolution’s Limits for Biomanufacturing
Nature never designed enzymes for industrial manufacturing. It designed them to survive. That mismatch is one of the biggest bottlenecks in biomanufacturing — and most of the industry has been working around it with incremental fixes for decades. Matthew Thompson from Biomatter...
AI Labour Shift: 10‑Year Transition More Manageable Than Expected
The AI Labour Shift: Why a 10-Year Transition Might Be Better Than You Think by @Timothy_Hughes https://t.co/gnz4oSh31V @DLAIgnite #Sales #SalesTips #SalesLeader #Marketing #MarketingStrategy #Leadership #ArtificialIntelligence #Innovation #Technology #Tech #TechNews https://t.co/P1D2C0JDFz
Small Models Also Found the Vulnerabilities that Mythos Found
Researchers tested a suite of inexpensive, open‑weight language models on the same code snippets Anthropic highlighted for its Mythos system. All eight small models flagged Mythos's flagship FreeBSD exploit, including a 3.6 billion‑parameter model that costs roughly $0.11 per million tokens....
MIT Sloan Highlights Cutting-Edge AI Applications
Putting #AI to work: The latest from MIT Sloan Management Review by Brian Eastwood @MITSloan Learn more: https://t.co/atM4GstWdI #MachineLearning #ArtificialIntelligence #ML https://t.co/v1vi7yURk6
AI Hype Fails to Boost Productivity, Says Fortune
My take on AI productivity in @FortuneMagazine: “Forget the AI bubble. Productivity was weak, and if AI had delivered, productivity would be way up. The Silicon Valley guys say GDP is going to 5% or 6%, but it’s just not happening.” SO...
Google Introduces AI-Powered Restaurant Booking as Search Evolves Into Action
Google has launched an AI‑driven restaurant reservation feature in the UK, embedding end‑to‑end booking capabilities into its AI Mode search interface. Users can ask the AI to locate suitable venues, check availability and confirm a table without leaving the search...
Ten Crucial AI Questions Higher Ed Must Address
A little earlier than usual, but this week's Substack newsletter asks ten questions about AI and higher education that I think we should be grappling with, and worry that we re not: https://t.co/GF5ktxRFbS
AI Supply Chain Reset Signals Massive Growth Ahead
Excellent take. This is a reset before the next leg up. Every part of the AI supply chain has massive growth ahead of it. 👏🏻🚀
Small Models Also Found the Vulnerabilities that Mythos Found
Anthropic unveiled Claude Mythos Preview and Project Glasswing, pledging $100 M in usage credits and $4 M to open‑source security groups while claiming the model autonomously discovered and exploited thousands of zero‑day bugs. AIS AI researcher Stan Fort tested the showcased vulnerabilities on inexpensive,...
Learn to Direct AI, Don't Let It Replace You
The most important career decision you’ll make in the next 3 years: Will you manage AI or be replaced by it? The winners will be the people who learn to direct AI, not compete with it. That window is open right now, but...
Balancing Fast AI Automation with Human Accountability
It’s rare to find another founder as obsessed with AI workflows as I am, so diving into Lars Maat’s approach to team-wide adoption was a total blast. His take on balancing rapid automation with human accountability is exactly what agencies...
Catching Illicit Distributed Training Operations During an AI Pause
MIRI’s Technical Governance Team proposed an international treaty that would require registration of any AI chip cluster exceeding the compute power of 16 H100 GPUs. The original definition left a loophole: a distributed network of many small nodes could evade...
Explore 10 AI‑Science Frontiers on SciRouter Now
Something new, the 10 Frontiers of AI x Science which you can build in now on SciRouter https://t.co/KfLM56A2Kc
One Claude Command Beats 4,000 Hours of Pipelines
me at 2AM realizing I wasted 4,000 hours building Zapier-to-Notion pipelines when a single terminal command to Claude literally organizes my entire life https://t.co/LYdqS5hGMW
Marketing Teams, Stop Using Claude Cowork Like a Chat Window
The post argues that most marketing teams treat Claude Cowork as a smarter chat window (Level 1) and miss out on deeper productivity gains. Level 2 transforms Cowork into a virtual department using Skills, Subagents, scheduled tasks, and shared project context. By codifying...
Mythos: Strong Yet No Breakthrough, Just Linear Progress
Mythos, while a very powerful model, appears to be a relatively linear improvement over previous models. AI progress is real, impressive, and continuing. But Mythos does not appear to be a discontinuity.
Governed AI in Core Workflows Drives SaaS Disruption
RT The real SaaS disruption isn't chatbots on top of legacy screens. It's governed, explainable AI embedded directly into core workflows. #AI #CIO @Star_CIO https://t.co/QCJX9x5Zwh
ServiceNow Rolls Out Generative AI Across Entire Portfolio, Boosting Sales Automation
ServiceNow announced that its complete product lineup now includes generative AI, data connectivity, and autonomous workflow capabilities. The upgrade, highlighted by the new Context Engine and AI Control Tower, aims to cut deployment time for sales and CRM teams and...
Iran Conflict's Ripple Effects on AI Development
How is the Iran war impacting AI? I sketch out some first order effects and possible consequences here: https://t.co/HFHGbn2REK https://t.co/Y8xVdGXhVy
Upgrade to Personal AI: OpenClaw, Hermes, GBrain
If you are consistently hitting the limits of what the packaged AI software world can give you, then you owe it to yourself to try the personal AI software world Try OpenClaw/Hermes Agent, and you can skip ahead to my...
Claude Hits Search Budget Limit Mid‑chat, Is This New?
Claude just told me it "ran out of search budget" halfway through the first message in a fresh chat. Is this new?
Public API Keys Misused as Live Gemini AI Credentials
Developers’ public API keys now function as live Gemini AI credentials, enabling attackers to run costly and unauthorized operations. https://t.co/Oo1InL5G8f
AI's Strength Lies in Integrating Five Essential Pillars
This diagram nails it. Modern AI isn’t one thing — it’s 5 interacting pillars: GenAI • LLMs • RAG • Agents • Agentic systems The edge isn’t tools. It’s how you connect, control, and govern them. Miss one layer → fragile AI. https://t.co/QvhRkNaHxJ
Each New Accelerator Builds on the Previous
The accelerator accelerator accelerator CPUs are the OG GPUs accelerate CPUs TPUs accelerate GPUs (Tensor cores) LPUs accelerate TPUs Maybe something simpler could work
Claude May Be Ending Human Critical Thinking
🚨Sometimes I wonder if Claude's greatest impact on humanity is to make us STOP thinking. https://t.co/eljL3RT2Ju #AI #LLMs https://t.co/PqjgBrQBCr
One-Hour AI Program Simplifies Classroom AI Introduction
🎥 Just shared on YouTube 👉 Looking for an easy way to introduce AI to students? Hour of AI offers a one-hour program with resources and partners to teach various AI aspects.... https://t.co/8IISiq4WYo #HourOfAI #AIEducation #STEM #EdTech #ComputerScience https://t.co/9jevgJf7lf
Anthropic's Compute Shortage Limits Mythos Rollout
So how much of the limited (virtually non-existent) Mythos rollout a function of Anthropic’s compute scarcity?
AI Designs Bee Garden, One‑trip Home Depot
Today's project: Can AI help Ben create his new bee garden structure and only have to go to Home Depot one time. https://t.co/1BpRXN0qBJ
Claude Stalls at Step Nine, Hits Daily Limit
Claude is really great at getting to step 9 out of 10, then saying I hit my daily limit