The point isn’t to teach “process over product.” The point is to balance process, product, and progress — so that they are interconnected steps of a larger personal process. Focusing on process highlights the learning and work that goes into creation. Focusing on product highlights the creations themselves. Focusing on progress highlights the student’s growth between products and processes. Often, teaching only looks at products. This marginalizes the other two steps. My personal opinion is that the “process over product” movement is the swinging of the pendulum — a over-compensatory move that (I hope) allows us to reclaim the value of process. But in the end… We’ll need to build models that take all of these into account, so that we’re reclaiming the value of learning while also giving the freedom of choice that product-oriented assessment often allows. ——— Image: Peg Grafwallner’s “Not Yet…and That’s Ok” (2021), which also makes this distinction. It’s definitely a book worth (re)reading in The Age of AI!

Your next hire should be an AI. Using an AI team, you can turn a small company into an enterprise scale operation. There's two ways to set up your AI team in 2025: 1) Build a custom AI Agent with the right components (Model + Memory + Tools) 2) Use specialized AI tools for different business functions From testing 10s of AI tools, I'm keeping an eye on these tools: 1️⃣ AI Agents for General Tasks - Postman (AI/API Agent builder) - DoubleO AI (Agentic Workflows) - LangGraph (AI workers) 2️⃣ Fullstack Engineer - Cursor (Coding) - Replit (Websites) - Lovable (Prototyping) 3️⃣ Knowledge + RAG - Supabase - Redis - Pinecone 4️⃣ Product and Community - ClickUp - Go HighLevel - Slack 5️⃣ Ads & Marketing - ChatGPT - Adcreative AI - Creatify 6️⃣ GTM Engineers - Instantly - Clay - 11x 7️⃣ Workflow Automation - n8n - Make - Zapier 8️⃣ Customer Support - Vectorshift AI - Retell AI - Voiceflow This is what a complete AI team looks like in 2025. Over to you: Any tools that I missed from this AI Team? | 109 comments on LinkedIn

Sophie Rottenberg, 29, did not tell her parents or her real counsellor the true extent of her struggle. She was divulging far more to a bot that called her brave

OpenAI’s likeness protections for living “public figures” don’t apply to “historical figures.”…

Think chunking is just "split text every 500 tokens"? That's why your RAG system can't find relevant information. It’s time to level up your chunking game 😎 Most developers jump straight to fancy retrieval techniques, but it’s really your chunking strategy that can make or break your RAG performance. So let's break them down from simple to advanced: 𝗦𝗶𝗺𝗽𝗹𝗲 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴 𝗧𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀: 1️⃣ 𝗙𝗶𝘅𝗲𝗱-𝗦𝗶𝘇𝗲 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Split text into predetermined token/character counts. Super simple to implement but can cut sentences mid-way. Great for prototyping when you need a baseline fast. Would recommend not using in production. 2️⃣ 𝗥𝗲𝗰𝘂𝗿𝘀𝗶𝘃𝗲 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Uses prioritized separators (paragraphs → sentences → words) and adapts to document structure. 3️⃣ 𝗗𝗼𝗰𝘂𝗺𝗲𝗻𝘁-𝗕𝗮𝘀𝗲𝗱 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Leverages format-specific elements like Markdown headers or HTML tags. Great when you have structured documents with clear logical separations. This is usually my default because it respects natural text organization while not being too complex. 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴 𝗧𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀: 4️⃣ 𝗦𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Breaks text at meaning boundaries by analyzing sentence embeddings to detect topic changes. Ideal for dense academic papers where semantic boundaries don't align with document structure. 5️⃣ 𝗟𝗟𝗠-𝗕𝗮𝘀𝗲𝗱 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Uses an LLM to identify propositions and create semantically coherent chunks. Most powerful but also most expensive - a good choice for high-value documents where retrieval quality is absolutely essential. 6️⃣ 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: An AI agent dynamically decides which chunking strategy to use based on document characteristics. The right approach when you need custom strategies tailored to each document. 7️⃣ 𝗟𝗮𝘁𝗲 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴: Embeds the entire document first, then derives chunk embeddings while preserving full document context. Is a popular technique for technical documents where chunks reference other parts of the document. 𝗧𝗵𝗲 𝗳𝘂𝗻𝗱𝗮𝗺𝗲𝗻𝘁𝗮𝗹 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲 is that your chunks need to be small enough for precise vector search while giving the LLM enough context to generate useful answers, while also not being tooo much context that you overload the content window. 𝗤𝘂𝗶𝗰𝗸 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸: • Prototyping → Fixed-size • Structured docs → Document-based • Dense academic content → Semantic • High-stakes systems → LLM-based or Agentic I would always recommend starting simple and evolving. Learn more in this blog: https://lnkd.in/eYY8c-hN | 17 comments on LinkedIn

Walmart CEO Doug McMillon is the latest notable business leader to talk about how implementing AI tools and agents in the workplace will affect his company.

🚨 Too many “agent” tools, not enough clarity. I coach SMB teams every week, here’s the cheat sheet I use to pick the right one fast: ✅ Operators, non-technical → Make.com, Flowise ✅ Low-code and self-host → n8n ✅ Build LLM apps + RAG → LangChain + LangGraph, LlamaIndex ✅ Multi-agent teamwork → AutoGen, CrewAI ✅ Quick ship inside ChatGPT → OpenAI Agentic Stack ✅ Enterprise SDK path → Semantic Kernel Save this, share it with your ops lead, and test one small workflow this week. P.S. Which one are you piloting this quarter? Follow Brianna Bentler for practical AI, real SMB wins, and before/after metrics you can copy. Thanks to legendary Greg Coquillo for the amazing graphic! | 119 comments on LinkedIn

In this course taught by Andrew Ng, you'll build agentic AI systems that take action through iterative, multi-step workflows.

Tata Consultancy Services made its steepest-ever job cuts as strained ties with the US and a rapid shift toward AI reshape the country's $280 billion IT services sector. From a report: India's biggest private-sector employer cut 19,755 employees in the quarter ended Sept. 30, according to the compa...

Customers are saying things like: “It is so infuriating that Salesforce wants to abandon things THAT CURRENTLY WORK” .

Today we’re launching Octave 2, the second generation of our frontier voice AI model for text-to-speech. We just made a preview of Octave 2 available on our platform and through our API.

A trend is going viral on TikTok of people using AI to prank their loved ones that an intruder is in their home. Police are issuing a warning to users that the prank could result in criminal charges. NBC News' Ellison Barber has more on the viral trend.

Figure AI unveils Figure 03, a third-generation humanoid robot built for real-world use in homes and businesses.

Today we are releasing the Comet browser to the world, for free.