User Experience

“Loop engineering” is a hot buzzphrase after mentions of it by Boris Cherny (Claude Code’s creator) and Peter Steinberger (OpenClaw's creator) went viral on social media. Loops are now a key part of how we get AI agents to iterate at length to build software. In this letter, I’d like to share my 3 key loops, shown in the image below, for building 0-to-1 products. These loops guide not just how I build software, but also how I decide what software to build. Agentic coding loop: Given a product specification and optionally a set of evals (that is, a dataset against which to measure performance), we can have an AI agent write code, test its work, and keep iterating until the code is bug-free and meets its specification. This idea of closing the loop took off around the end of last year, and it has been a game changer in enabling coding agents to work longer productively without human intervention. For example, over the weekend, I was building an app for my daughter to practice typing, and my coding agent could easily work for around an hour, using a web browser to check what it had built multiple times before getting back to me, without needing my intervention. The engineering loop executes quickly. Every few minutes, the coding agent might build and test a new version of the software. I hear frequently from developers who are finding new ways to engineer more effective engineering loops. This is an active area of invention! Developer feedback loop: In this loop, a developer examines the current product and steers the coding agent to improve it. Last year, a lot of developers (including me) were acting as the QA (quality assurance) function for our coding agents, manually finding bugs and then asking the agent to fix them. But with coding agents much more able to test their own code, the amount of time we need to spend on this function has decreased significantly. This allows us to make higher-level product decisions, such as what key features to offer, where the UI needs improvement, and so on. The developer-feedback loop operates over time intervals between tens of minutes and hours — that's how frequently a developer might review a product and give feedback. In the case of the typing app, I changed my mind a few times about the visual design, what cat costumes she can unlock as she learns (she loves cats), and the user flow for a grown-up to log in and steer the child's learning experience. When a developer has a clear vision for what to build, it is still a lot of work to translate that vision into a specification for a coding agent to implement. Further, after the developer has seen an implementation, they might update (or perhaps clarify) the spec to steer it toward what they want. If you find that the system repeatedly runs into certain problems, building a set of evals for the agent becomes useful. [Truncated for length. Full text: https://lnkd.in/gKDQ6H9s]

𝗡𝗮𝗶𝘃𝗲 𝗥𝗔𝗚 𝘄𝗼𝗿𝗸𝘀 𝗶𝗻 𝗮 𝗱𝗲𝗺𝗼. 𝗜𝘁 𝗳𝗮𝗶𝗹𝘀 𝘁𝗵𝗲 𝗺𝗼𝗺𝗲𝗻𝘁 𝗿𝗲𝗮𝗹 𝘂𝘀𝗲𝗿𝘀 𝘀𝗵𝗼𝘄 𝘂𝗽. Embed → retrieve → generate looks clean in a notebook. Real requirements break it: → Questions whose answer is spread across many documents → Industry terms that embeddings get wrong → Bad chunks the pipeline never catches → Answers that live in how things connect, not in any single chunk → PDFs full of tables and images a text-only index cannot read These 5 architectures are how serious teams stay ahead in the agentic AI era: 𝟬𝟭 𝗛𝘆𝗯𝗿𝗶𝗱 𝗥𝗔𝗚 → Dense vectors find meaning. BM25 finds exact words. → Reciprocal Rank Fusion combines both ranked lists. → A safe baseline for almost every team. 𝟬𝟮 𝗚𝗿𝗮𝗽𝗵𝗥𝗔𝗚 → Pull entities and their relationships into a knowledge graph. → Retrieve subgraphs and community summaries, not chunks. → Best when the answer lives in how things connect. 𝟬𝟯 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗥𝗔𝗚 → A planner agent picks the right tool: vector, web, or SQL. → A reasoner agent keeps trying until the answer is solid. → Retrieval becomes a plan, not a single step. 𝟬𝟰 𝗖𝗼𝗿𝗿𝗲𝗰𝘁𝗶𝘃𝗲 𝗥𝗔𝗚 (𝗖𝗥𝗔𝗚) → Grade every retrieval before you trust it. → Correct → answer. Unclear → rewrite the query. Wrong → search the web. → This is what production RAG actually looks like. 𝟬𝟱 𝗠𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗥𝗔𝗚 → One embedding model (CLIP, ColPali) for text, images, and tables. → One vector index. One multimodal LLM. → No more separate pipelines for PDFs with charts. I built a runnable example for each of the five patterns. GitHub link in the first comment. The best teams in 2026 do not pick one. They combine them — hybrid retrieval inside an agentic loop, with a corrective grader, over a multimodal index. Naive RAG is a starting point, not a finish line. That is why most enterprise GenAI projects stall at the demo. Which of these five becomes the default RAG stack in the next 18 months — and which stays a specialized tool?

Don't need to comment, like, or connect. Download it. Read it. Use it. Learn with it. Over the last weeks I kept seeing the same pattern. Designers were curious about MCP, but stuck. The path was unclear. Setup felt intimidating. Real use cases were missing. So I built the Design MCP Adoption Toolkit. A practical guide for using MCP inside a real Figma workflow. No theory. No hype. Just execution. Inside you will find: → What MCP is in plain language. → The three MCPs that matter for design work. → The mental model for Anthropic Claude Code to Figma, OpenAI Codex to Figma, and Figma Console MCP by Southleft, LLC and TJ Pitre. → The full setup in nine clear steps. → Nine real workflows you can test this week: Accessibility audits. Ticket validation before handoff. Token migration. Multi platform component handoff. Component documentation generation and more. Our roles are evolving. We are moving closer to that old Webmaster model where design, systems, structure, and technology connect. The designer who understands systems and automation will have leverage. This toolkit is my contribution to that transition. Consume it. Test it. Break things. Ask questions. Explore your own use cases. These are exciting times, and we move faster when we learn together. If you build something interesting with it, share it. Concrete examples help the whole community level up.

🚛 WHEN TRANSPORT LEARNS TO THINK GREEN I came across a concept today that stopped me — an autonomous hydrogen truck-trailer drone designed for long-distance freight. At first, it looked like another futuristic vehicle. But then it hit me: this isn’t just transport evolving — it’s intent evolving. For decades, we’ve designed logistics around speed and scale. Now we’re finally designing around sustainability. This new concept merges autonomy, aerodynamics, and hydrogen power to do something radical: → Eliminate carbon emissions in heavy freight. → Cut operational energy costs through intelligent routing. → Reduce highway congestion with coordinated drone convoys. It’s not just engineering — it’s a shift in philosophy. A move from moving faster to moving responsibly. We often talk about “green tech” as a feature — but the real shift happens when sustainability becomes the invisible infrastructure behind innovation. It’s not an addition to progress. It is progress. What’s needed now isn’t more invention — it’s integration. We need to: ✅ Build networks where clean energy and automation reinforce each other. ✅ Redefine “efficiency” to include environmental balance. ✅ Shift from carbon offsetting to carbon prevention at design level. Because the next breakthrough won’t come from faster engines — but from systems that make waste impossible by design. That’s when technology stops being an experiment in innovation… and becomes an expression of intelligence. So here’s the question I keep returning to — 👉 Will the next era of transport be powered by fuel — or by foresight? #Innovation #Sustainability #Hydrogen #AutonomousVehicles #GreenTech #Logistics #FutureThinking

I once lived at distributor’s home in a small town because I had no choice... When Marico Limited was nascent, Bombay Oil Industries was still the family’s backbone. In those early days, I wanted our business to transform from a commodity trade into a branded consumer company. To do that, I had to understand the ground truth. There were no fancy hotels in the towns we visited. I stayed in dusty and small guest rooms. I sat with distributors over chai and samosas. I watched how coconut oil was stored, how shopkeepers priced it, how packaging changed hands. One day, a retailer told me matter-of-factly: “You always sell big tins. When people come back to buy, they carry a few kilos. If your packet is small, they will pick your brand at convenience.” That simple insight was a turning point. It nudged us to expand SKU ranges, introduce smaller packs, and think about how to become a “grab-and-go” brand, rather than just a bulk commodity supplier. If you ask me where innovation begins, it begins in the least glamorous places. In the musty shelves of neighbourhood stores, in conversations that feel insignificant, in paying attention to what people don’t say aloud. Takeaway for entrepreneurs: Your real research lab isn’t spreadsheets or agencies. It’s the ground. If you go build empathy for your customer at the shelf level, the brand strategy almost builds itself. #entrepreneurship #business #resilience #mindset #growth

Peter Thiel and Jeff Bezos just backed a $3B startup that destroys tumors with sound waves. It’s called HistoSonics, Inc., a US medtech company born out of the University of Michigan - now headquartered in Minnesota - developing a non-invasive cancer therapy powered by focused ultrasound. No scalpels. No radiation. No chemo. Here’s what makes it groundbreaking 👇 ▶ 1. How it works HistoSonics’ device, called Edison, emits pulsed ultrasound waves that create “bubble clouds” inside tissue. These bubbles collapse within microseconds - producing mechanical forces strong enough to destroy tumors at a cellular and subcellular level. ▶ 2. Proof it’s working The tech won FDA approval in late 2023 and is already used by 150+ hospitals, including the Cleveland Clinic. Each system costs about $1.5M, and the company expects $100M+ in revenue this year. ▶ 3. The investors betting big HistoSonics just raised $250 million, led by K5 Global, Wellington Management, and Bezos Expeditions, with Thiel Capital joining as a new investor - valuing the company at around $3 billion. The new funding will help expand treatments for tumors in the breast, prostate, pancreas, and brain. ▶ 4. The next frontier The company aims to scale globally, with European regulatory approval targeted for 2026. Founder Mike Blue calls it “a paradigm shift in how we treat cancer” - replacing invasive surgery with precision sound. And I agree.. it’s a bold vision: A future where tumors are treated with sound instead of surgery. But challenges remain - from long-term safety data to proving this can scale beyond high-end hospitals. Still, if HistoSonics succeeds, it could reshape how oncology looks a decade from now - not by extending life through suffering, but by removing that trade-off entirely. Do you think non-invasive sound therapies could become the new standard in cancer care? #entrepreneurship #healthtech #innovation

I flew to the US a few months back and attended Justin Welsh's talk on LinkedIn growth. The man has grown to 770K followers in 5 years. I'm still at 290K in 4 years. Here's what I am doing wrong/right: Let's talk about what I'm doing wrong first. → Value-add commenting - Justin uses the "CEA" formula (Compliment, Expand, Ask) on 5-10 posts daily with surgical precision. Meanwhile, I comment when something randomly catches my eye. → Micro-interviews - He actively reaches out to influencers with thoughtful, specific questions about their expertise, then shares their responses as posts (with permission). → Trend translation - He has Google Alerts set up for industry keywords and is consistently first to explain what breaking news means for his specific niche. → Early comment engagement - He replies to the first 6-8 comments within minutes of posting, which signals to LinkedIn's algorithm that his content is engaging and boosts reach. → Systematic repurposing - He turns one piece of research or insight into 5-6 different posts: the main insight, a contrarian take, a personal story illustrating it, a how-to guide, a mistake to avoid, and a future prediction. → Strategic collaborations - I used to create joint content, co-host discussions, and cross-promote with other creators in my space. But I've completely neglected this, missing out on audience overlap and compound growth opportunities. What I'm doing right: → Posting consistently - I publish 4 times a week without fail. This consistency has built trust with my audience and kept me top-of-mind. → Killer hooks - I spend genuine time crafting that first line to stop the scroll. Keep the hooks <280 characters → Behind-the-scenes content - I share failures, mistakes, and real struggles without sugar-coating them. → Contrarian takes - I'm not afraid to share unpopular opinions. → Strategic giveaways - I regularly offer valuable downloadable resources that actually solve problems. → Authentic storytelling - I write like I talk, sharing personal experiences that illustrate broader business principles. The biggest difference is that he's intentional about every single move. I post consistently and create good content, but I haven’t been strategic enough with profile optimization, early comment engagement, or systematic repurposing. Time to level up. Growth isn’t just about posting more. It’s about being smarter with how you use the platform.

Want to stay motivated every single day? Borrow a strategy from Harvard. Then borrow another from stand up comedy. Together, they’re a powerhouse for momentum, motivation, and mastery. Here’s how it works: Let’s start with Harvard. Researcher Teresa Amabile studied 12,000 daily work diaries across 8 companies. She wanted to know: What truly motivates people on a day to day basis? What she found changed how we understand drive. The #1 driver of daily motivation wasn’t: Money Praise Perks It was progress. The days people made progress on meaningful work were the days they felt the best. Progress isn’t a luxury. It’s a psychological necessity. So how do we make progress feel visible especially on days when it’s not? Use a “Progress Ritual.” → At the end of the day, pause. → Write down 3 small ways you moved forward. → That’s it. No fanfare. Just ritual. This works because we rarely notice our progress in real time. It gets buried under busyness, meetings, and mental noise. The act of looking back gives your brain the reward it needs to keep going. Momentum builds from meaning. Now let’s add some comedy. Young Jerry Seinfeld had one goal: write new material every day. To stay on track, he created a brilliant system. Each day he wrote, he put a big red X on his calendar. Soon, a chain of Xs formed. And here’s the key: Don’t break the chain. One red X becomes two. Two becomes ten. Ten becomes identity. Whether you’re writing, coding, or training Daily action + visual chain = long-term motivation. Summary: The Two-Part Motivation System From Harvard: Record 3 ways you made progress each day. From Seinfeld: Mark an X for each day you show up then don’t break the chain. Progress fuels purpose. Consistency fuels confidence. Apply both and you’ll stay on track especially on the tough days. Because when your days get better, your weeks get better. When your weeks get better, your months get better. When your months get better, your life gets better. It starts with one small win today.

I ignored Duolingo for 3 months. Here’s how they got me back - and why I (neeeeearly) stayed. A LOT of products just try to get churned users back into an app session without thinking about the onward journey. When I came to Duo - it was a full reset: ✨ Fun animation made it feel like a moment, not a mistake ✨ They asked if I wanted to switch goals not just carry on ✨ Made me commit to a new language (or recommit to my old one) ✨ Quick email follow-up 'welcome back' kept the momentum going ✨ CRM nudged me towards a 7-day streak (users who hit it are 2.4× more likely to stick) It’s easy to focus on getting users back. Much harder to design what happens after. Duolingo does that well. As a rule of thumb: 1) Focus on getting users to the magic moment again ASAP 2) Fix the elephant in the room by welcoming them back after their break 3) Aim for re-set by setting goals or showing them what's new Full breakdown in slides here ↓ Curious if you've seen any other reactivation sequences do particularly well? ---- Hiiiii, I'm Rosie 👋 I do a weekly deep dives on growth, product & UX. 📅 Last week: How LinkedIn reached $2bn in subscription revenue 📅 Previously: How to write surveys that actually work: lessons from Strava 📅 Next up: Sign up to find out Get them at 🕺 growthdives.com 🕺

Netflix Is Going Physical — And It Might Just Rewrite the Experiential Playbook At Cannes Lions, Netflix unveiled more details about its boldest move yet in fan engagement: Netflix House — permanent, immersive venues launching this fall in Philadelphia and Dallas. Think “Stranger Things” escape games, “Squid Game” obstacle courses, “Wednesday” carnivals, mini-golf through your favorite titles, themed cocktails, exclusive merch, and yes — a TUDUM Theater to host fan events and screenings. But this isn’t just a cool activation. It’s a strategic pivot that’s worth unpacking: ✅ Strategic Intent: Netflix isn’t trying to build a theme park empire. This is about deepening emotional ties with fans, amplifying buzz, and future-proofing its brand beyond the streaming wars. These venues aren’t just fun — they’re fan conversion engines. ✅ A New Content Loop: Every attraction is designed to be shared — built for UGC, influencer walkthroughs, cosplay, TikToks, and viral moments. Fans become marketers. Data becomes feedback for future IP development. The venue becomes a living R&D lab. ✅ Not Just Eyeballs — Wallets: With exclusive merch, themed dining, and potential collabs (think Netflix x Funko or Netflix Bites F&B), the monetization flywheel is in motion. Even modest visitor volume could generate $25–30M/year per location — and that’s before you count the uplift in brand love or viewership. ✅ Global Signals: This could be the first step toward regional pop-ups, international localization (imagine a “Lupin” heist experience in Paris or “Money Heist” in Madrid), and even a Netflix-con-branded event model. It’s fandom scaled offline. 💡 Big Picture? Netflix is building something Disney mastered decades ago — real-world storytelling at scale. And if this works, it unlocks a new dimension: streaming IP that lives, breathes, and sells in the physical world. 📊 Our modeled impact: ⌙ ~1M visitors in Year 1 ⌙ 100M+ earned impressions ⌙ 10–15% churn reduction among local superfans ⌙ $5–10 lift in ARPU among engaged segments ⌙ Payback in ~3–5 years — with marketing ROI baked in 🎯 This isn’t about “content” anymore. It’s about building culture. Kudos to Marian, Greg, Josh, Mitzi, Emily, Nidia, Lauren, Jessica, Nikki and team. #Netflix #Cannes #Media #Licensing #ConsumerProduct