Adam Watt

BREAKING Claude Code just shipped automated code review 🎉 when a PR opens, Claude dispatches a team of agents hunting for bugs in parallel, verifying findings, and ranking by severity. One high-signal summary comment plus inline flags. Anthropic ran this internally for months: ▫️ PRs with substantive review comments went from 16% to 54% ▫️ Under 1% of findings marked incorrect by engineers ▫️ On 1,000+ line PRs, 84% surface findings averaging 7.5 issues each Pricing is $15-25 per review, billed by token usage and PR complexity. One reply said it best: "Are you kidding me? That's a year of salary." Available now in beta for Team and Enterprise. How many hours per week does your team spend on code review?

19

5 Comments

🚨 ~$1B raised in Dev + Infra across Nov–Dec. The most unhinged seed rounds I’ve seen lately. 🚨   Seed is the new Series C guess.   Almost half a billion went into infra alone - someone finally opened the AWS bill.   AGI is taking a breather, small language models are everywhere. I guess size doesn't matter after all!   Here are some of the coolest rounds:   🔥 $475M Seed - Unconventional AI Yes, seed. Backed by a16z, Sequoia, Lightspeed, Databricks, Lux. This isn’t a startup - it’s a full-on rethink of the computing substrate for AI (hardware + software + energy).   🔥 $75M Seed - Tenzai Battery, Lux, Greylock. Builds an AI-powered penetration testing platform that continuously simulates real-world attacks against your systems. Designed to proactively surface vulnerabilities as software and AI-generated code rapidly expand the attack surface.   🔥 $70M Seed - Gradium Audio-native foundation models. Develops a unified audio architecture covering speech generation, transcription, voice transformation, and dialogue, optimized for real-time, low-latency interactions rather than text-first pipelines.   🔥 $50M Seed - Inception Diffusion-based language models with lower GPU requirements. Replaces traditional autoregressive generation with diffusion techniques to improve efficiency, controllability, and latency, targeting production use cases where compute cost is a bottleneck.   🔥 $41M Seed - Mirelo AI AI-generated sound for video. Provides tools and APIs that automatically generate synchronized sound effects and music from video content, enabling faster and more scalable post-production workflows. A lot of cool stuff being built in Europe!! Shout out to orq.ai Yasu Specific (YC F25) Albatross AI INTELLIGENT CORE NobodyWho Open Machine

71

4 Comments

These 10 fast-growing YC startups have recently closed their Series A 🔥🔥 Whether you’re job-hunting, investing, or just tracking the next wave of breakout tech, these companies are ones to watch: 1. Escape: AI-powered offensive security platform automating API and application testing. $18M Series A. 2. Depot: Software delivery and CI system built to radically accelerate Docker image builds and GitHub actions. $10M Series A. 3. Lio (formerly askLio): Agentic AI platform designed to automate enterprise procurement workflows from end to end. $30M Series A. 4. Harper (YC W25): AI-native commercial insurance brokerage connecting mid-sized businesses with over 160 insurance carriers. $46.8M Seed+Series A. 5. Dots: Global payouts API platform catering to gig workers, creators, and contractors. $8.9M Series A. 6. Tamarind Bio: No-code computational biology platform that accelerates AI-driven drug discovery. $13.6M Series A. 7. Humand: AI-powered operating system creating a centralized hub for managing the deskless workforce experience. $66M Series A. 8. Kombo: Unified API and global infrastructure layer connecting HR, payroll, and recruiting software. $25M Series A. 9. Alaan الآن : Leading corporate card and AI-powered spend-management platform operating in the Middle East. $48M Series A. 10. Pasito: AI-native workspace built to aggregate and process scattered employee benefits, health data, and insurance claims. $21M Series A. - - - - - - - - - Bonus: Looking to affordably bring over international talent 🌍? Talk to us at Gale (YC W25) and we'll get you sorted.

227

8 Comments

YC backs 5-10 versions of the same company every batch. For YC, this strategy works. They deploy $75M per batch across 150 companies anyway. One winner pays for five competitors when you're operating at that scale. If one becomes the breakout, the 7% they own covers the losses from backing the others. For VCs, the math is different. You’re concentrating capital in ~3-4 companies per quarter, not 150. We have to pick one, maybe two if the approaches are fundamentally different. The problem is beyond just portfolio construction. It's operational. You can't split attention between direct competitors. You can't share confidential information when both companies are asking for intros to the same customers. Most seed funds face this dilemma eventually. You back a company, then six months later a better team tackles the same problem with stronger traction. Some investors refuse to back competitors out of loyalty, then watch the better company rise and dominate the market. Others back multiple competitors and destroy trust with all of them when founders realize they're competing for the same investor's attention. Is it better to miss an opportunity than to create conflicts you can't manage? Concentration matters when you're trying to add real value beyond the check. YC can afford to spray and pray. Most seed investors can't.

32

7 Comments

Key takeaways: - The core shift is economic, not hype. LLMs turn “AI” into a recurring operating cost, so pilots quickly become capacity planning, governance, and ownership. - Power and timelines show up because production inference needs sustained compute capacity. When usage increases across the org, you run into procurement, security review, integration work, and infrastructure readiness challenges. - The winners are not the teams with the flashiest demos. The winners are the teams that convert compute into measurable outcomes inside a workflow with traceability and controls. - In 2026, “AI strategy” that is not tied to a specific workflow, KPI, and operating model is mostly theater. If you are building the application layer on top of the datacenter buildout, execution systems, workflow automation, data-layer tooling, compliance automation, I want to see what you are shipping. Founder intro: www.axion.ventures

6

How to get into YC (Advice from a 2x founder) 1. Everyone says YC invests in people. YC doesn't invest in people, it invests in narratives. Your application to YC has to tell the first chapter of a great founder's biography. If you're wondering why you got rejected from YC, it's probably because when the reviewers finished reading your application, they didn't feel compelled to turn the page. 2. Have an idea, make a logo, make a web landing page, make an MVP, and find someone willing to pay you to use it. Everything else is not essential. 3. Get referrals. It matters less how many referrals you have, and more what they're saying about you. Your referrals need to reflect the same narrative that you sell in the application: "This person is, for some reason, special." Referrals don't exist to back up the idea that you're accomplished (that's what your resume is for). Referrals support the idea that when your peers look at you, they see someone they believe can build a great company. 4. Do something substantial, and put it in your application. Hundreds of thousands of people went to top schools and then worked in big tech. This is not substantial. Your application has to reflect the idea of what you are capable of breaking this mold in some considerable way. DM me if this was helpful, and I'll send you a full guide.

87

8 Comments

YC and Tier-1 VCs prefer technical founders. They might be right. Vungle hit hundreds of millions in revenue but as a non-technical CEO/Founder, I often felt helpless. My co-founder and I were both business guys. We didn’t know how to hire engineers, evaluate architecture, or debug production. My hiring filter? “Seems smart, speaks well.” I promised features to customers without a clue how they’d get built. Over time: • Our SDKs crashed client apps • DevOps melted down on New Year’s Eve (we lost millions in one night) • Everyone pointed fingers while tech debt piled up I had outsourced product entirely to the early team. Whilst they lacked customer context. I lacked technical depth. That gap widened until it cracked everything. I thought we had technical debt but in reality we had something called trust debt. Here’s what I learned: If you’re a non-technical CEO, you can’t hand off product. You must own the translation layer between user pain and roadmap. You must ensure the team builds the right thing - don’t chase after shiny objects. After Vungle, I learned to code. It took years and it was painful. When the AI wave hit, I felt like something clicked. Like I could finally understand how tech works. If you’re a non-technical CEO, spend at least 100 hours learning to code. Not delegating. Not whiteboarding. Just learning. If you can, make it a hobby. Learn how engineers think. Watch tutorials. Break things. You don’t need to be the best engineer in the room (people like us never can be). But you need to earn their respect and that starts with understanding how things actually get built. AI has made it easier than ever to close that gap. So build. Break things. Ask dumb questions. Learn just enough to be dangerous. Then surround yourself with technical people you trust. But never ever step away from the product. That’s the part you can’t afford to outsource.

179

49 Comments

Stripe engineers are now minion babysitters. "A Stripe engineer might have half a dozen running at a time." The agentic coding is cool, but what's equally impressive is the speed at which the dev environments that make this possible are provisioned. The dev environment needs to: - clone gigantic git repositories - warm Bazel and type checking caches - start code generation services that continually run on devboxes ...and more. They get all this done in 10 seconds. So in ~10 seconds, a minion gets it's own dev environment to get to work. If you want to see how stripe is using ai agents to change the job of software engineering you should read the entire post: https://lnkd.in/gdhkfWwX

167

2 Comments

Running a Company on Markdown Files CompanyOS: a skills-only system that turns Claude Code into the operating layer for an entire company. No application code, no web UI - just markdown files that teach Claude Code how to run business operations. https://lnkd.in/gACFhTMk #AdventuresInClaude #ClaudeCode #CompanyOS

269

52 Comments

Last month, in Y Combinator's call for startups for their Spring 2026 cohort, they called for "AI Native Agencies." YCs frames this opportunity simply: "agencies of the future will look more like software companies, with software margins. And they'll scale far bigger than any agencies that exist in these fragmented markets today.” We believe YC is right. We differ, though, with YC in our belief in where the value will accrue. More on this here 👇👇 https://lnkd.in/g6iNtTUh

21

5 Comments

This YC founder does not exist. There is no DataFlow (YC S22). And yet, they are leaving a stellar five-star review for this product. This is an actual screenshot from an actual YC company's website. And this actually isn't a knock on that company. Obviously, this isn't a good thing. And if it was intentional, it's very bad. But I suspect it was just an accident: someone vibe-coded the website, which generated placeholder testimonials, they shipped it, and moved on to the next thing. Placeholder testimonials were a problem even pre-generative AI, but the problem is about to get 100x worse.

39

10 Comments

Asked Claude to uncover insights from this database of 407 startups from YC's 2025 cohorts. The results are quite interesting. While 94 companies are fighting over developer tools (that's a new competitor every 4 days!), entire billion-dollar markets have almost zero competition. Here's what I found: 🔴 Saturating sectors: - Developer tools: 94 companies - Productivity/workflow: 61 companies - Sales & marketing: 39 companies 🌊 Wide open opportunities: - Government tech: $600B market, only 1 company - Insurance AI: $1.4T market, virtually no players - Healthcare analytics: Only 3 companies total But here's the kicker: The companies winning in crowded markets aren't winning with better tech. They're winning by becoming AI workers that replace expensive humans entirely. 📊 What's in the full analysis: ✅ Complete breakdown of all sectors. ✅ Market saturation maps with competitor counts ✅ YC's exact positioning strategies that work ✅ Specific blue ocean opportunities with market sizing ✅ How to stand out in saturated domains. The gap isn't technical. It's positional. 💡 Bottom line: Stop building AI assistants. Start building AI workers. Want the complete breakdown? The full report includes detailed company mappings, competitive analysis, and actionable market entry strategies. Comment "REPORT" and I'll send you the full report with all the data, graphs, and insights. What surprised you most about these numbers? #YCombinator #AI #Startups #DataAnalytics #MarketAnalysis

59

44 Comments