We’re Aamir & Luis and we’re excited to launch our company’s new brand ahead of demo day this week: Orchestra (orchestra.ai)

Companies move their frontier inference spend to Orchestra, with a single URL swap. We relay their traffic through to the frontier labs, while building evals in the background automatically. These evals allows us to post-train specialist models that replace the frontier, at a fraction of the cost. When our model beats the frontier at your product’s work, we promote it and your bills go down. Then we repeat the loop continuously, forever; your AI inference gets smarter and cheaper the more you use it.

We’re going straight after the $104B inference market; Anthropic grew 14x YoY but we’ve seen time and time again that start-ups and enterprises are willing to move spend if you can prove the performance. That it. We don’t have FDEs. We don’t pre-train models. Most of the application layer lives in a CLI that coding agents use.

We just do one thing really well: building and routing your work to the most efficient model. We found the easiest way to build a fast growing business is to sell your product to people who have pain and who are already spending money with someone else. The more we focused on the fastest growing segments of our business, the faster our product got better at solving a simple, but incredibly painful problem: “You don’t need a PhD to work in the mailroom”.

It’s the “Age of the Unc”

I shared my experiences doing YC as a 38-year old (https://x.com/lluismanrique/status/2094846605925241259) but for any founders out there thinking about coming back to YC: do it. We worked 60-hour weeks, not 100-hour weeks. We had a network of angels, customers and advisors. We ran the marathon, instead of sprinting it. At 38, I’m old enough to understand that pedagogy is the point of YC. Learning how to work on your company is the point. You’ll get so much more out of YC when you’re doing it again with the hindsight of all your past successes and failures. Having done it before is a huge advantage, but having a long career in which I’ve encountered a lot of situations and had to make a lot of decisions and a few mistakes, and see others do the same, allows me to understand how YC’s advice impacts the real world. You will manage your agents better. You will close customers faster. You’ll build a better company.

Why Orchestra?

Branding is everything, especially outside of silicon valley. We launched as the ‘alternative’, the ‘backup’; but as time went on, we realized how open companies are to moving all their spend to a single place and making it work efficiently. Inference is the fastest growing market in the history of capitalism and we needed a brand that reflected our ambitions. Orchestra is going to win a huge chunk of that market; not the back-up, but the entire show.

Say hello to Orchestra. And if your bills are too damn high, send us a note. We’d love to help.