Own your intelligence.
Anyone that opposes this has something to sell you (perhaps your own intelligence back to you.)
Local models need to continue proliferating. The “absolute unit” massive models may be out of reach for now locally, so your best option is to, indeed, keep your options open. Own your harness, data, connections, runtime: your agents. Embrace open source.
Prediction: Local model will be the biggest story of 2027
Claude Model is getting smarter and bigger
But smaller model is getting better.
I don’t need my own model to solve the hardest math problems. Why shouldn’t I distill a smaller one focusing on the things that matter to me and free?
Local model is free.
Local model is private.
Local model is fast.
Such future will disrupt the inference market and anthropic/openai’s current biz model as we speak. We still need inference but we also need an entire new generation of hardware that powers this type of inference.
Who is building this new future? Raise your hand!
The frontier labs are constantly flexing image and video generation capabilities. But does that mean AI can do the work of a designer?
We have SWE-bench for code. We have LAB for legal work.
Introducing Atelier: the benchmark for professional design.
https://t.co/U9sfP2o8lO