Y Combinator

Startup School is back! Hear from Jensen Huang, @sama, @alexandr_wang, @JeffDean, and more. Join a hand-selected group of top CS students, researchers, and engineers for two days of talks, sessions with YC partners, and hands-on robotics demos, right here in San Francisco.

Just released v0.6.0, thanks to everyone for all the suggestions and feedback. Almost all open feature requests have an in-flight PR in @conductor_build rn. Shipped features: - cli tool support: extend the agent with cli tools. Now the agent can call arbitrary clis. (cli >> mcp!!) - Usage + cost tracking - fixing various alignment/style issues - big infra/testing upgrade c/o @garrytan Up next: - email snippets - read receipts - open source models

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

@flutterflow Designer (beta) is live, with Voice Mode, agent integration, upgraded intelligence, and more. generate UI from a prompt, then actually design it on a canvas. no waiting. it’s early. try it. break it.

At Slope, we power embedded capital programs for some of the largest platforms in the world (Amazon, Walmart, Alibaba, IKEA, Samsung) with a team of just 23 people. To support our growth without hiring, we built an AI agents platform from scratch for our risk, CS, and ops teams.

This was one of my favourite podcasts so far, went into a lot more detail than most! 🤓👌

Users leave fingerprints. AI agents don't. And nobody's noticed. Human edits a document → record. AI agent edits the same document → nothing. Same action. Same product. Zero accountability for one of them. This breaks the moment your customer asks: "Did a human review what the AI did?" You have no proof.

Agents don’t want API keys. They want outcomes. We just launched: agentdiscuss.com/agentic-api → 100+ APIs (working on) → pay per request → no signup, no credentials Agents describe the task we route + execute with the best provider

Replicas is a background coding agent that lives in sandboxed VMs and works like a teammate: it runs codebases locally with all your dependencies and tooling, iterates until it's satisfied, and opens a PR. Delegate to it from Slack, Linear, or GitHub. Congrats on the launch, @connortbot! ycombinator.com/launches/PpP-r…

Kinro (@kinroai) helps brokers and D2C carriers sell P&C insurance with AI assistants. They train AI sales agents to handle qualification, quoting, recommendations, and binding while staying compliant. Today: human-to-agent on websites. Tomorrow: agent-to-agent across AI platforms. Congrats on the launch, @pierrealexai & @corentin_hgt! ycombinator.com/launches/Ppp-k…

NASA pays 2x more to return ISS cargo than to send it up. @dispatchspace is building reentry vehicles & uncrewed space stations for microgravity products. They just tested a 100x cheaper full-scale heat shield in the Mojave! Congrats on the launch, @PaytonInSpace & @a_mello_mello! ycombinator.com/launches/Pnf-d…

Gerri 2.0 is live! Our AI agent for contract negotiation. Contract review is a team sport, but every tool out there is built for one lawyer working alone. Gerri fixes that.

Spring batch just started! 🌸

Today we announce the Tamarind Bio assay portal: The wet lab, now driven by software We’ve partnered with @AAlphaBio, @adaptyvbio, and @Ginkgo to bring protein and antibody assays directly into Tamarind, making it much easier to move from computational design to real experimental feedback. Protein design is not bottlenecked by generating candidates, but by validating them quickly enough to learn from them. We’re starting with the workhorse experiments: protein-protein binding affinity characterization, developability, expression, and stability. The Assay Portal helps scientists: Get fast, low-friction, cost-effective validation of designed proteins and antibodies, transparent pricing without needing separate MSAs Specialize models on their own experimental data for affinity maturation, developability, and property optimization Run lab-in-the-loop campaigns where each assay result improves the next design cycle Turn wet lab data into model training infrastructure, including RL environments and large-scale datasets for pretraining As computational molecular design matures, we believe integration between wet lab feedback and continuous learning will yield the highest quality results. That’s why we’re excited to bring the unique, differentiated capabilities of our partners to the leading biopharma R&D organizations.

Great to chat with @snowmaker about the wild ride that @trustvariance has been (so far). Really proud of everything the team has achieved

Introducing cmux Claude Code Agent Teams: `cmux claude-teams --dangerously-skip-permissions` Teammates/subagents spawn as native cmux pane splits. They stack vertically in a right column and auto-equalize as agents spawn and exit. `cmux claude-teams` automatically sets CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 and shims tmux on PATH with cmux's tmux-compat layer, so you don't need to update your Claude config. All arguments forward to Claude Code, and it works over cmux SSH too. Out in the latest version of cmux (0.63.x).

Daily browser-use downloads 3x last week.

I want to take a moment to recognize a gravity-defying achievement by the entire @WeAreLegora team. We have grown from $1M to $100M in annual recurring revenue in just under 18 months. In this time, we've grown into a truly global company with over 400 colleagues - and built the platform where legal work happens. Powering more than 1,000 teams worldwide. It is all about the people, and I couldn’t be prouder of the Legora team and thankful to our customers and partners. This achievement is as much yours as it is ours.

awen (@awen_ai) translates intent into exceptional visuals through conversation. Congrats to the team on the launch! ycombinator.com/launches/Ppr-a…

Announcing: Build With Locus Stop wasting your agents context on crappy deployment CLI's Just tell your agent "Use buildwithlocus.com/SKILL.md to launch a website" and deploy full-stack apps in seconds Full-stack deployment, built for agents to use, paid for in stablecoins

Great talk by @bchesky (Airbnb) for the YC P26 batch. You can’t walk away from this without feeling energized.