Jon Dahl

How We Built GrantBite In 2 Months With #Supabase When we started GrantBite, we had one constraint: Validate fast. Not “fast for enterprise software.” Fast, fast. Two months from idea to production. That forced an early question: What’s the minimum infrastructure needed to ship something real, safely, and repeatedly? The answer was clear: • Not #AWS from scratch • Not microservices • Not rebuilding auth and APIs again This is the same philosophy we followed when building #Palou — using managed primitives instead of recreating foundational infrastructure. What tools have made you move faster lately?

26

🚨 ~$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

Most people see Claude Code and think: "Now everyone can vibe code." They're missing the bigger shift. The real insight isn't that non-technical people can build apps. It's that startups can now deliver complete platforms instead of point solutions. When software was expensive, you had to pick your battles. Build one feature well. Expand slowly. Stay narrow. That was the traditional VC advice: "Focus. Don't go multi-product until you've dominated one category." When we started Vessel, we ignored that advice. We went multi-product on day 1. Fundraising. Reporting. Co-investments. LP onboarding. Portfolio company reporting. All in one platform. In the old world, that would have been suicide. Too much surface area. Too many engineering resources required. Too slow to market. In the AI era, it's a massive advantage. Here's why: Clients don't want five point solutions that don't talk to each other. They want one platform they can trust. One unified data layer. One relationship. Every GP I talk to says the same thing: "We're drowning in tools that don't integrate." The fragmentation tax is real. Different logins. Different data formats. Different support teams. Different security reviews. AI coding tools collapse the cost of building breadth. What used to require a 50-person engineering team can now be done with 5. The startups that understand this are building platforms from day one. The question isn't whether you can code with AI. It's whether you're building something that stands alone — or something that becomes a feature.

15

2 Comments

To date, writing extensions for Envoy has been a laborious process that involves writing C++ and compiling a fully custom build of the entire proxy.  Can we make that easy and make Envoy more approachable by developers?. Can we create an ecosystem of reusable extensions around it - akin to how npm ecosystem fundamentally altered Node.js's trajectory. With that vision, we at Tetrate have been working with Envoy community to first make Envoy easy to extend by any developer. We added dynamic modules a year ago and then added a proper Go and Rust SDK around it. Today, we are making the developer experience much easier and announcing a home for Envoy extensions with an open source extension marketplace around it. Thanks Matt Klein Anna Berenberg for your support and thanks Rohit Agrawal Ben Plotnick Anubhav Aeron for early feedback and trials. You can read more about Built on Envoy here: https://lnkd.in/gbNz9-3v and see a cool demo video from our engineering lead Ignasi Barrera on how to write a simple Go extension and test it locally for Envoy to not allow AI calls to specific models. https://lnkd.in/gcpYJyjD

74

GTM goals for early stage technical companies: 1. Be referenced often in internal meetings of your prospects (exploratory phase of adoption) 2. Be referenced often in externally facing architecture diagrams of the top companies in your market (first path to production + PMF) 3. Be endorsed often in communities as experts/influencers see 1 + 2 appear without your promotion (your real ARR + NRR story happens here) 4. Be endorsed as the gold standard by communities (moat, hang on for dear life and keep shipping and making money) #2 is v hard as things shift often, but Founder time should be with customers to make this occur as often as possible. #3 + #4 become your roadshow focus over time so you can scale your company.

21

2 Comments

AI isn’t just “helping” us build software anymore - it’s changing who can build software. Real-world example from Chimani: We’ve run a custom CMS on Ruby on Rails for ~16 years (hosted on Heroku/AWS). Plot twist: I don’t know Ruby or Rails. Historically, when something broke, I’d hire help (often via Upwork). And every time Heroku announces an end-of-life stack upgrade…I brace for a crash. Late last week, while fighting off a cold, I decided to do something "mindless": migrate the repo from GitLab to GitHub so I could hook it up to OpenAI Codex and see how far I could get modernizing the codebase. Five days later: - running on the latest Ruby/Rails + updated gems - Postgres upgraded - deployed on a new platform (Railway) - working implementation end-to-end And I’m still the same “hack” programmer I was five days ago. What made it work: 1) Speed matters. The Codex experience in the right environment (vs. a slower workflow) dramatically improved iteration loops. 2) Grounding + guardrails are everything. Clear constraints, constant verification, and tight scope kept the model from wandering and kept changes shippable. The takeaway isn’t “AI replaces developers.” It’s that AI compresses the distance between intent and implementation - and that changes the game for founders and small teams.

20

i don't think most people have priced in the fact that codegen is now trivial. the bottlenecks are shifting upstream -- why build, what to build -- and, importantly for us, downstream -- whether what's built holds up in the real world. no software will ever be perfect. bugs happen in production. new corner cases uncovered. reality wins. so the differentiator going forward is designing systems that expect failure, then learn from it. when something breaks, it should be detected, root-caused, and fixed -- fast. and the system should remember the issue so that it's more resilient in the future. anti-fragility will be the new baseline.

9

2 Comments