Yeah, this is just slop. No benchmarks, no concrete case studies, just some vibecoded "platform" to finetune models on your own traces.

Which is an idea that has some value, but also some weaknesses. And this implementation of it isn't forthcoming with that concept. You have to really dig in to understand what they're even talking about.

Valid criticism. Happy to answer any qs. We're still working on solidfying results.

Benchmarks are the ultimate consolidation of halnons razor.

[flagged]

That's cool, thanks for sharing!

Open source models.

wmo routes requests between frontier models and open source models that continuously train using Tinker. As the smaller models improve, more traffic gets routed to them.

Calculating cost is just tokens in/out.

Thanks :)

Fixed formatting which will help with readability. We do routing, distillation, and token compaction.

It's open source!

We do have a platform we'll be launching as well to manage training + serving for you which will require more diligent privacy guarantees.

Let me know if you have any questions!