Was ist neu?

AI Is Making Intelligence Cheap — 10 Skills That Will Matter Instead Dario Amodei, CEO of Anthropic, published an essay arguing we need to pace the frontier of AI development. Sam Altman, Elon Musk and Demis Hassabis broadly agreed. But underneath that debate sits a more interesting question: what happens to humans if intelligence itself becomes cheap? For thousands of years intelligence was scarce, and we built schools, degrees and entire careers around the equation "I know something valuable that you don't." AI is breaking that equation — the same way calculators made arithmetic cheap and phone cameras made taking a photograph worthless while taste became everything. This video covers what becomes scarce when intelligence becomes abundant, the 10 skills that grow more valuable, and a five-layer skill stack for what to actually learn. Dario Amodei — "We Must Pace the Frontier": https://darioamodei.com/post/we-must-pace-the-frontier CHAPTERS 0:00 One morning, this 0:10 "We need to slow down AI" 0:44 Altman, Musk and Hassabis agree 1:10 The end of intelligence scarcity 1:42 Calculators, cameras and commodities 2:20 So what becomes scarce? 2:38 The 10 skills that stay valuable 3:45 The five-layer skill stack 4:17 What will you be uniquely valuable for? THE 10 SKILLS 1. Problem framing — who decides which problem is worth solving? 2. Judgment — AI gives options, humans carry consequences 3. Domain expertise — AI plus deep knowledge is powerful 4. AI orchestration — redesign the workflow, not the chatbot 5. Systems thinking — understand what surrounds the task 6. Communication — attention is the scarce resource 7. Human relationships — people work with people 8. Taste — creation gets democratised, taste does not 9. Creativity — connect ideas that shouldn't connect 10. Agency — "let me try this and let you know" If intelligence becomes a commodity, what human skill will be worth the most? Tell me in the comments. #AI #FutureOfWork #CareerAdvice #ArtificialIntelligence #Skills2026 My Linkedin: https://www.linkedin.com/in/jacinthp/ --------------------------------------------- Github: https://github.com/jacinthpaul --------------------------------------------- Website: https://www.programstrategyhq.com/ ---------------------------------------------

Claude Certified Associate Exam Prep — Full Course All 7 Domains, One Video. A complete refresher for the Claude Certified Associate (Foundations) exam — all seven exam domains covered in one video, in the order they build on each other. Nine sessions: a core concepts primer, then one session per exam domain, weighted the way the exam is weighted. Watch straight through, or jump to the domain you're weakest in using the chapters below. CHAPTERS 00:00 Core concepts primer 00:00 Prompting and task execution (14% of exam) 00:00 Output evaluation and validation (21%) 00:00 Product and model selection (12%) 00:00 Workflow integration and solution design (16%) 00:00 Configuration and knowledge management (12%) 00:00 Governance, risk and responsible use (15%) 00:00 Troubleshooting and optimization (10%) 00:00 Closing and best wishes WHAT YOU'LL COVER · Writing prompts that hold up under real task conditions · Checking and validating model output before it ships · Choosing the right Claude product and model for a job · Designing Claude into an existing workflow · Projects, system prompts and knowledge setup · Responsible use, risk and governance basics · Diagnosing and fixing weak results WHO IT'S FOR Anyone sitting the Claude Certified Associate exam, plus anyone using Claude at work who wants the fundamentals in one pass. No prior certification required. #Claude #ClaudeAI #ClaudeCertified #AICertification #AnthropicClaude #PromptEngineering #AIforWork #ExamPrep My Linkedin: https://www.linkedin.com/in/jacinthp/ --------------------------------------------- Github: https://github.com/jacinthpaul --------------------------------------------- Website: https://www.programstrategyhq.com/ ---------------------------------------------

Courtesy: pyoderma.hubb

GPT-6 Astra launched September 3, 2026, and OpenAI's president called it the start of the AGI era. So what actually qualifies as AGI — and does Astra clear the bar? OpenAI's own definition: "highly autonomous systems that outperform humans at most economically valuable work." Note what it leaves out — consciousness, understanding, any single test. Five criteria usually invoked: generality, transfer, autonomy, on-the-job learning, agency. Astra scores on most. The one that keeps getting quoted — 99.9% on ARC-AGI-3 — came from a provider-adapter harness. ARC Prize's standard harness put the same model at 62.7%. Same model. Same test. Different scaffolding. Sources: OpenAI, ARC Prize, Axios #AGI #GPT6 #OpenAI #ArtificialIntelligence #AInews

Four products, one stack. LangGraph is the orchestration runtime — graph execution, state, persistence, interrupts, retries. LangChain's create_agent is the fast path on top of it. LangSmith gives you traces, datasets, evaluators and cost/latency telemetry. LangGraph Platform runs it long-term: task server, queues, cron, threads API, Studio. Start at the layer your problem lives on. To get started with AI Native Software Engineering, check out the full video — link in comments. #LangChain #LangGraph #LangSmith #AIAgents #AIEngineering

AI-Native Software Engineering is not about simply using ChatGPT, Claude, GitHub Copilot, or another AI coding tool. It's about putting AI and AI agents at the core of your software development process. Specs go in. Verified, tested work comes out. You own the judgment in between. You provide clear specifications, acceptance criteria, AI agents and workflows, tests and evals, security guardrails, and human judgment and accountability. The AI helps build. You design the system that verifies whether the work is good enough to ship. Full breakdown in the pinned comment: what AI-Native Engineering is NOT, AI-Native Engineer vs AI Engineer vs ML Engineer, the 8 skills to learn, 6 projects to build, MCP, context engineering, spec-driven development, evals and AI security. Subscribe to Program Strategy HQ for AI-Native Software Engineering, AI agents, spec-driven development, MCP, AI evals and enterprise AI adoption. #AINativeSoftwareEngineering #AINativeEngineering #SoftwareEngineering #AIAgents #SpecDrivenDevelopment #AIEngineering #MCP #AIEvals #AgenticAI #ClaudeCode #shorts