EDISON TKP

“Can’t reach Claude.” Looks like a simple error message, but it highlights a bigger issue in how most people use AI. If your workflow depends on one model, one app, or one provider, you don’t really have an AI system. You have a single point of failure. That’s the real lesson.

Check out the latest article in my newsletter: Claude Agents, Sub-Agents, and AI Teams: How I’m Building Faster Than Ever linkedin.com/pulse/claude-a… via @LinkedIn

Check out the latest article in my newsletter: System Prompts Are the Real Product Spec (So I Collected 140+ of Them) linkedin.com/pulse/system-p… via @LinkedIn

@anthonyloke This “mrt train will stop longer than usual” message happens frequently now. Without explanation or timing, it creates real disruption for commuters. Can we expect clearer updates and better operational standards?

A new chapter for browsing just launched. OpenAI just introduced ChatGPT Atlas — a browser built natively around AI, not as an addon. This is what changes: No more copy-paste between tabs and chatbots AI becomes the primary interface, not a sidebar Research, summarize, extract, and act in one flow Atlas is now live on macOS. Windows, iOS, Android are coming soon. This might become the first browser people choose for thinking, not just viewing. Have you tried it yet? **Curious to hear first impressions.**👇 linkedin.com/pulse/introduc… via @LinkedIn

Check out the latest article in my newsletter: AI Agents are finally getting “real-world usable” — thanks to a silent missing piece: Agent Skills. linkedin.com/pulse/ai-agent… via @LinkedIn

💡 Switzerland is moving one step closer to a programmable financial future. The Swiss Bankers Association, together with UBS, PostFinance, and Sygnum, has successfully tested a Deposit Token Proof of Concept on Ethereum. ✅ Cross-bank payments with tokenized deposits ✅ Escrow-like smart contract transactions ✅ Full compliance with AML/KYC and SIC settlement ✅ Interoperable, programmable, and legally enforceable 👉 This marks a milestone in bridging traditional banking with blockchain-native finance—without compromising regulatory integrity. The next phase? On-chain master records, CBDC integration, and zero-knowledge proofs for privacy and scalability. 🔗 A glimpse into how regulated banks can power the next era of digital money. #Blockchain #CBDC #Tokenization #DigitalFinance #Switzerland linkedin.com/pulse/deposit-… via @LinkedIn

🎓 The Future of Hiring Isn’t About Degrees Anymore LinkedIn CEO Ryan Roslansky recently said it best: “The future of work doesn’t belong to those with the fanciest degrees — but to people who are adaptable, forward-thinking, ready to learn, and ready to embrace new tools.” In the AI era, your mindset matters more than your diploma. Here are the 4 traits every employer will value most 👇 1️⃣ Adaptability – Change fast. Don’t resist it. 2️⃣ Forward Thinking – See where things are heading, not just where they are. 3️⃣ Ready to Learn – Stay curious, keep evolving. 4️⃣ Embrace AI – Use it. Don’t fear it. 💡 The takeaway: Your degree might open a door — but your adaptability keeps you in the room. #AI #FutureOfWork #Leadership #CareerGrowth #LinkedIn linkedin.com/pulse/future-h… via @LinkedIn

New Grok release

New @GoogleResearch paper builds a personal health assistant that reads a user’s data, answers health questions, and coaches daily habits. It evaluates the system on 10 tasks with 7,000+ human annotations and 1,100 hours from experts and users. The assistant covers 4 needs, interpreting a user’s data, giving clear facts, offering wellness advice, and handling symptom questions. It uses 3 specialist agents, a data science agent, a domain expert agent, and a health coach agent. The data science agent plans analyses and writes code to turn wearable streams and medical records into clear numbers. The domain expert agent joins those numbers with medical knowledge and user context to explain findings and reason about likely causes. The health coach agent runs multi turn chats that set goals, surface barriers, and track progress using motivational interviewing. An orchestrator decides which agent to call, coordinates their work, and returns one joined answer. In studies the agents and the full system improved diagnosis quality and increased user rated coaching helpfulness versus a strong baseline. Overall, this is a practical template for turning a general chatbot into a personal health assistant grounded in personal data. ---- Paper – arxiv. org/abs/2508.20148 Paper Title: "The Anatomy of a Personal Health Agent"

Last week we released Claude Sonnet 4.5. As part of our alignment testing, we used a new tool to run automated audits for behaviors like sycophancy and deception. Now we’re open-sourcing the tool to run those audits.

Almost no one is talking about this! OpenAI quietly dropped 4 technical guides on: - How to Stay ahead in the age of AI - AI in the Enterprises - Identifying and Scaling AI - Practical guide to building AI agents It’s way better than you’d expect: 👇

open-source platform for building and managing AI agents with multi-model integration

R.I.P n8n? OpenAI just dropped AgentKit a full-stack framework that makes building, deploying, and optimizing AI agents as easy as drag-and-drop. It’s agent orchestration with memory, reasoning, and evals. Here’s what just changed ↓

OpenAI employee: "the agent builder we released today was built end-to-end in under six weeks, with Codex writing 80% of the PRs" this matches the AI-2027 report's 2026 forecast: - coding automation goes mainstream - agents work like teammates - AI R&D is 50% faster from algorithms, not just compute

Types of AI Agents

Announcing my new course: Agentic AI! Building AI agents is one of the most in-demand skills in the job market. This course, available now at deeplearning.ai, teaches you how. You'll learn to implement four key agentic design patterns: - Reflection, in which an agent examines its own output and figures out how to improve it - Tool use, in which an LLM-driven application decides which functions to call to carry out web search, access calendars, send email, write code, etc. - Planning, where you'll use an LLM to decide how to break down a task into sub-tasks for execution, and - Multi-agent collaboration, in which you build multiple specialized agents — much like how a company might hire multiple employees — to perform a complex task You'll also learn to take a complex application and systematically decompose it into a sequence of tasks to implement using these design patterns. But here's what I think is the most important part of this course: Having worked with many teams on AI agents, I've found that the single biggest predictor of whether someone executes well is their ability to drive a disciplined process for evals and error analysis. In this course, you'll learn how to do this, so you can efficiently home in on which components to improve in a complex agentic workflow. Instead of guessing what to work on, you'll let evals data guide you. This will put you significantly ahead of the game compared to the vast majority of teams building agents. Together, we'll build a deep research agent that searches, synthesizes, and reports, using all of these agentic design patterns and best practices. This self-paced course is taught in a vendor neutral way, using raw Python - without hiding details in a framework. You'll see how each step works, and learn the core concepts that you can then implement using any popular agentic AI framework, or using no framework. The only prerequisite is familiarity with Python, though knowing a bit about LLMs helps. Come join me, and let's build some agentic AI systems! Sign up to get started: deeplearning.ai/courses/agenti…

🤯 ChatGPT just got a “Buy Button.” OpenAI + Stripe launched the Agentic Commerce Protocol (ACP) — an open standard that lets AI agents not just recommend, but actually buy for you. ✅ Shop via ChatGPT with Shopify & Etsy (first wave). ✅ Instant checkout, no redirects, no friction. ✅ Open protocol → any business can plug in. This is the start of the multi-agent economy — where your AI doesn’t just answer but acts. Think travel booking, subscription management, bill payments… all handled by AI. My take? This is the Amazon One-Click moment for AI. 👉 The question isn’t if AI will handle commerce. It’s when your business will integrate into it. #AI #AgenticAI #Stripe #OpenAI #FutureofCommerce linkedin.com/pulse/rise-age… via @LinkedIn

🚨 Big News for Solana DeFi 🚨 USYC, a tokenized share class of a U.S. government money market fund, is now live on Solana 🪙 This brings institutional-grade, yield-bearing collateral onchain — a first for the ecosystem. Why it matters: 📈 Daily NAV updates via oracle ⚡ Instant USDC redemptions (T+0/T+1 ready) 🛠️ Wallet allow-listing & institutional infrastructure With USYC, eligible institutions can hold collateral that keeps accruing yield while being deployed in money markets, margin accounts, and automated vaults. This changes the game: onchain collateral doesn’t have to be static anymore. 👉 Link in comments to learn more. #Solana #DeFi #USYC #Tokenization #Yield linkedin.com/pulse/breaking… via @LinkedIn