Generative AI

The second wave of AI agents just hit. This one's all about alignment. Capital has poured into generative AI at record speed, pushing global investment from $5B in 2021 to nearly $30B in 2024. Thousands of startups are building agents that code, research, and run operations. Most are built for one job at a time. When systems can’t coordinate, businesses lose time, data, and trust. Millions of micro-decisions stay trapped in silos instead of moving as one network . What’s shaping the market right now? • Capital: AI is now one of the largest destinations for global investment. • Agents: Every company is deploying them, few can align them. • Orchestration:The layer that makes agents collaborate at scale is still being built. GenAI.Works is building the infrastructure that helps agents cooperate, share context, and move work forward as one. First wave: Defined by breakthroughs in models and hardware. Second wave: About to be be shaped by systems that work together. This is where coordination becomes the real measure of progress. Are you ready to be a part of the next era of AI before it’s too late? Read the full newsletter to learn how to invest. #ai #agenticai #future #technology #generativeai

Harvard finds how we waste time to fix AI's work, built to save time. That’s workslop. It is the kind of output that looks finished but feels empty when you read it. You were promised efficiency but you end up fixing tone, structure and substance while the clock keeps running. The productivity gain that turns into invisible rework is not innovation. It is a digital busywork with better branding. Harvard research found: → 40% of employees say they’ve received workslop in the past month. → It comes from peers (40%), managers (18%) and even execs (16%). → Each one costs around 2 hours to fix, about $186 per employee per month. → For a 10,000-person company, that’s nearly $9M a year in lost productivity. AI isn’t your co-pilot if you’re flying solo cleaning up its mess. How many times have you rewritten an “AI draft” that was supposed to save you time? 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MKlfV0 Reg CF offering via DealMaker Securities. Investing is risky. Image Credits: Kate Niederhoffer et al Research paper link in the comments. #llm #ai #artificialintelligence #aiagent #machinelearning

Google indexing just killed 82% of Reddit’s AI visibility. In a single week, ChatGPT’s citations of Reddit dropped from 29.2% to 5.3%. No scandal. No outage. No internal change. Just a silent Google tweak that made Reddit harder for AI models to see. Wall Street noticed and Reddit’s stock fell 15% in days. Reddit’s $60M deal with Google was built on training access and when that visibility vanished, investors reacted as if a revenue stream had disappeared. What it shows: → AI visibility now dictates market perception. → Model access drives real financial value. → A single line of code can move billions in cap. → Visibility has shifted from marketing metric to market metric. If AI systems can’t see you, maybe the market won’t either. Are we looking at the first AI visibility crash? 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MJCQf0 Reg CF offering via DealMaker Securities. Investing is risky. #artificialintelligence#openai #google #economy #llm

AgentKit is literally everywhere but how do you actually build with it? OpenAI has just launched AgentKit, a framework that helps you create AI agents that can think, act, and use real tools, all in one place. It gives developers everything needed to design, test and deploy without juggling prompts or APIs. What’s inside: → A visual agent builder to design workflows → A code SDK for adding logic in Python or Node → A connector registry for APIs and databases → ChatKit for embedding agents into apps and websites How to start building: 1. Plan Your Agent: Define its goals, target users and which tools or APIs it needs. 2. Use Agent Builder: Create workflows visually, connect data sources and set rules. 3. Add Logic with the SDK: Use Python or Node to define behavior and handoffs between agents. 4. Integrate Tools: Plug in APIs, databases or web actions through the Connector Registry. 5. Test and Refine: Use built-in evaluation and tracing to monitor and improve performance. 6. Deploy and Embed: Add your agent to an app or website with ChatKit’s prebuilt components. AgentKit makes it easier to turn ideas into working agents that can power real products. Is this the point where AI shifts from experimentation to true infrastructure? Link to the official tutorial in the comments. 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MJ7By0 Reg CF offering via DealMaker Securities. Investing is risky. #openai #agentkit #ai #agentbuilder #cheatsheet

OpenAI. xAI. Anthropic. All running on NVIDIA. All funded by NVIDIA. xAI is raising $20 billion, with NVIDIA contributing up to $2 billion in equity. OpenAI, Anthropic, and xAI all build on NVIDIA’s hardware and now share its capital. The supplier of compute is becoming the architect of the entire AI economy. • Control hardware and you control innovation speed • Invest downstream to lock in long-term GPU demand • Use equity to secure influence across competing ecosystems • Price scarcity as strategy, not limitation • Blur the line between vendor and venture capitalist What looks like collaboration is starting to resemble a supply-chain empire. When the chipmaker funds the model-makers, where does competition actually live? 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MHL8j0 Reg CF offering via DealMaker Securities. Investing is risky. #artificialintelligence #startups #venturecapital #gpu #nvidia

Over the past year, the conversation around GenAI has evolved. It has moved from building individual agents to thinking about agentic architectures that can reason, adapt, and operate at scale inside real-world enterprise systems. But the big question is: What’s actually working, and what does it take to go from pilot to production? Alex Wang, Head of Education at GenAI.works, will be joined by Arun K. Subramaniyan, CEO of Articul8 (formerly at AWS, Intel, and GE), to dig into exactly that, with real-world examples. They will also discuss: ✅ The tradeoff between domain-specific GenAI and general-purpose models ✅ What hyper-personalization really looks like in the enterprise ✅ How to work with multimodal and messy enterprise data ✅ And more on agentic system design, trust, and deployment at scale Whether you're building agent workflows, leading an AI team, or exploring GenAI adoption inside a company — this one’s for you. 𝘕𝘰𝘵𝘦: 𝘞𝘦 𝘩𝘢𝘥 𝘵𝘰 𝘳𝘦𝘴𝘤𝘩𝘦𝘥𝘶𝘭𝘦 𝘵𝘩𝘪𝘴 𝘴𝘦𝘴𝘴𝘪𝘰𝘯 𝘥𝘶𝘦 𝘵𝘰 𝘤𝘪𝘳𝘤𝘶𝘮𝘴𝘵𝘢𝘯𝘤𝘦𝘴 𝘣𝘦𝘺𝘰𝘯𝘥 𝘰𝘶𝘳 𝘤𝘰𝘯𝘵𝘳𝘰𝘭. 𝘞𝘦 𝘴𝘪𝘯𝘤𝘦𝘳𝘦𝘭𝘺 𝘳𝘦𝘨𝘳𝘦𝘵 𝘢𝘯𝘺 𝘥𝘪𝘴𝘳𝘶𝘱𝘵𝘪𝘰𝘯 𝘵𝘩𝘪𝘴 𝘮𝘢𝘺 𝘩𝘢𝘷𝘦 𝘤𝘢𝘶𝘴𝘦𝘥 𝘢𝘯𝘥 𝘨𝘳𝘦𝘢𝘵𝘭𝘺 𝘢𝘱𝘱𝘳𝘦𝘤𝘪𝘢𝘵𝘦 𝘺𝘰𝘶𝘳 𝘶𝘯𝘥𝘦𝘳𝘴𝘵𝘢𝘯𝘥𝘪𝘯𝘨. 𝘛𝘩𝘦 𝘴𝘦𝘴𝘴𝘪𝘰𝘯 𝘩𝘢𝘴 𝘣𝘦𝘦𝘯 𝘮𝘰𝘷𝘦𝘥 𝘵𝘰 𝘵𝘩𝘦 𝘧𝘰𝘭𝘭𝘰𝘸𝘪𝘯𝘨 𝘵𝘪𝘮𝘦, 𝘢𝘯𝘥 𝘸𝘦 𝘭𝘰𝘰𝘬 𝘧𝘰𝘳𝘸𝘢𝘳𝘥 𝘵𝘰 𝘴𝘦𝘦𝘪𝘯𝘨 𝘺𝘰𝘶 𝘵𝘩𝘦𝘳𝘦: 🗓️ Oct 07 – 9 PM PT (Oct 08 – 12 PM MST)

OpenAI just killed 1,000+ AI startups — in one update... At DevDay, Sam Altman detonated a bomb: AgentKit. What’s inside: -> Agent Builder A full visual canvas for designing multi-agent workflows with nodes, logic, and versioning. Preview runs, set guardrails, and collaborate with your team—all in one place. -> Connector Registry A central hub for data governance. Manage Dropbox, Google Drive, SharePoint, Teams, and third-party connectors from a single dashboard. -> ChatKit Embed AI-powered chat directly into your app or website. Custom UI, brand colors, real-time streaming—your product gets a full agentic chat layer in minutes. -> Evals 2.0 Measure and optimize your agent’s performance with datasets, automated grading, and reinforcement fine-tuning. OpenAI has effectively turned agent development into a no-code playground—for enterprises and indie builders alike. Why pay for duct-taped workflows when OpenAI hands you production-ready agents, tested, deployed, and live — instantly? Question is, who adapts — and who dies? Share your thoughts in the comments. ___________________________________________________________ Join our 13M+ community of AI enthusiasts and 250+ partner companies, and get the opportunity to own a piece of it ➜ https://hubs.li/Q03CNgFT0 Reg CF offering via DealMaker Securities. Investing is risky. #OpenAI #AgentKit #Automation #AI #Startups

Why Python is the Heart of Data Science In data science, Python turns raw data into stories. Its rich libraries and simple design make finding insights faster and smarter. → Easy to Learn: Python’s readable syntax makes it accessible for beginners and powerful for experts. → Rich Libraries: Tools like Pandas, scikit-learn, TensorFlow and PyTorch make working with data and machine learning faster and simpler. → Strong Community: A global community ensures constant innovation, tutorials and open-source resources. → Seamless Integration: Python connects with databases, APIs and visualization tools for smooth data workflows. Are you leveraging Python to unlock the full potential of your data? Credits - Laurent Pointal 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MvR9s0 Reg CF offering via DealMaker Securities. Investing is risky. #cheatsheet #python #machinelearning #artificialintelligence #ai

An AI-written Deloitte report just flunked a $440K audit. The welfare audit produced for the Australian government contained fake citations and misquoted legal cases, errors that slipped through a top consulting team’s review. What this case reveals: ▪︎ Expose: AI can produce content that looks right but reads wrong. ▪︎ Demand: Every AI output needs human validation. ▪︎ Disclose: AI contributes to paid work must be stated clearly ▪︎ Validate: All references require verifiable sources. ▪︎ Reclaim: Integrity is cheaper than reputation repair. AI can assist accountability but it can’t replace it. Should firms be required to disclose when AI contributes to client work? 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03MvBqZ0 Reg CF offering via DealMaker Securities. Investing is risky. #artificialintelligence #aipolicy #data #generativeai #ainews

ChatGPT just learned how to use Spotify, Canva and Booking.com. It is a complete toolkit that connects AI with the real world apps like Spotify, Booking.com, Canva, Coursera and Figma. Cool features worth trying: → Agent Builder: Drag-and-drop workflows that connect reasoning, actions and tools → Connector Registry: Plug into APIs like Google Drive, Slack or Spotify in minutes → ChatKit: Drop an AI chat UI straight into your app or website → Evals & RFT: Measure and fine-tune how your agent performs, automatically How to start: → Pick a task (like creating a playlist, sending a Slack update, or designing in Canva) → Connect the tools in Agent Builder → Let your agent handle it end-to-end. This feels like the moment AI apps finally start building themselves. What’s the first tool you’d connect your agent to? 🚨 Bonus share window extended until Oct 17 🚨 We’re building the AI infrastructure for what comes next: community, education, tools, and agentic execution — all open and global. 13M+ in the community. 200+ companies on board. $3M ARR, bootstrapped. Believe in this future? Invest now and earn up to 25% more shares before the window closes. 👉 Own a piece of it ➜ https://hubs.li/Q03Mv5CQ0 Reg CF offering via DealMaker Securities. Investing is risky. #ai #openai #agentkit #developers #automation #chatgpt