Building real-time semantic search for Conversational AI - @usemoss (YC F25)
San Francisco, CA
Joined April 2024
- Tweets 80
- Following 326
- Followers 390
- Likes 116
Pinned Tweet
I was telling my team yesterday that as we gear up for bottom-up adoption, hitting 100 developers would be a solid first milestone.
Today my CTO casually drops: “we’re already 100+.”
Stakes just went up 💪 next stop is 1,000. ⚡️
Attending my first @TechCrunch Disrupt! Excited to meet founders, and builders shaping the next wave of AI.
If you’re building in Conversational AI and facing retrieval latency challenges, little “r” or DM me. We’ve built a solution you’ll love.
Browser is the new compute layer. It’s inevitable. The next decade of software will be running natively in the browser!
Come join us at #HackHalloween
Proud to have @usemoss sponsoring #HackHalloween 🎃
We’re building real-time semantic search for conversational AI. Bringing search close to where your agents take decisions.
If you’re hacking on agents that earn, control, or evolve, come build with us⚡
#Moss #SemanticSearch
Was chatting with another founder today - it’s fascinating how LLM code generation tools nail TypeScript and Python, but still fumble with the basics in Rust. The gap between language popularity and model competence is still very real.
Every YC founder I meet teaches me something new.
Today I met an engineer building AI to automate insurance - he actually got a real insurance license just to embed himself in the field and feel the pain firsthand. That’s the kind of obsession that builds enduring companies!
You can’t defy physics.
Even the best cloud round-trip hovers around 100 ms P95.
Add multiple hops like vector DBs, model APIs, caches, and you’ve already broken the illusion of “instant thought.”
The only way forward is shorter distance between data and reasoning.
Engineering gives you control.
Sales reminds you you have none.
Doing both daily is how you learn to stay calm in chaos.
Data is the new code!
the rumor is openai drops “agent builder” tomorrow and wow, if that's true thats a BIG DEAL
for the 12 months, people have been stitching together tools like n8n, zapier, make, vapi, and claude workflows to simulate autonomy.
it worked but it was duct tape.
now IMAGINE that entire stack, native to openai, with one-click access to MCP, chatkit widgets, and every model they’ve trained (no API chaos. no patchwork. one smooth canvas)
this is what happens when ai moves from tool to infrastructure.
before: you needed 10 tabs, 5 plugins, and a weekend to build an agent.
after: you’ll drag a few blocks, add logic, hit “publish,” and deploy a production-ready workflow.
what app store did for software, agent builder COULD do for intelligence.
it’s the beginning of the “no-code ai economy,” where building an autonomous agent is as simple as building a notion template. developers get leverage.
non-technical founders get superpowers. businesses get workflows that run 24/7 without ops teams.
openai might launch the app store for intelligence tomorrow.
the DOWNSTREAM effects:
- zapier and n8n lose their monopoly on automation
- claude and perplexity become upstream research assistants for agent networks
- indie agents replace indie apps
– data becomes the new code
tomorrow's dev day should be INTERESTING.
We talk about moats and that is a useful concept but maybe less so right at the beginning of your startup journey
Moats are useful concepts after PMF because it gives you a language you can use to deepen your PMF
But when everything is brand new you could do almost any real thing users want and make a valuable business and that time is now
In the early days, the only moat that startups have is speed. Once you make something people want, the question becomes what deeper moats can you build on to defend against the competition.
On the @LightconePod, @garrytan, @harjtaggar, @sdianahu, and @snowmaker dive into Hamilton Helmer’s Seven Powers framework to find out how these moats show up in practice today in AI startups.
00:00 - The Moat Problem
01:30 - The Seven Powers Framework
04:20 - When to Think About Moats
08:40 - Forward Deployed Engineering
10:18 - Process Power
14:34 - Cornered Resources
19:30 - Switching Costs
24:54 - Counter Positioning
31:24 - The Workforce Displacement Reality
34:00 - Brand & Speed as Moats
37:30 - Network Economies
41:00 - Scale Economies
43:44 - Final Advice
Spicy thing on how nobody uses knowledge graphs in industry 👀👀
Really curious how memory folks are getting benefited with knowledge graphs.
Cursor CPO @sualehasif996
Turbopuffer CEO @Sirupsen
Notion AI Engineer @akm_io
Braintrust CEO @ankrgyl
Great discussion on all things Semantic Search and honestly the hard parts of AI Engineering
some spicy things said
- “in my entire career nobody uses knowledge graphs”
- “turbopuffer will run every sql query”
- some cool details on how Cursor uses turbopuffer
Apps can personalize on the fly. If you’re looking for super-fast search and retrieval to power your applications and agents, DM me or hit us up at - contact [at] inferedge [dot] dev
We take care of the hard parts of the search infrastructure like indexing, scaling, and performance, so developers can focus entirely on building great AI experiences. We believe our tech will unlock a new class of AI experiences.
InferEdge gives developers instant search and retrieval built for real-time conversational AI agents. Retrieval runs where the interaction happens, such as in browsers, apps, or enterprise infrastructure. The responses feel as fast and natural as human conversation.