Super excited to share this work we've been doing at AC (major kudos to @AtharvNaphade)! It allows us to take a legitimate step towards taking our hands of the wheels during model training. Looking forward to sharing more about how we were able to use it for automating a full end to end training loop 👀

I implemented a system in @appliedcompute’s platform for automated failure mode clustering with Jev to surface errors at an even larger scale than before. RL training produces billions of tokens in traces. I always manually read many traces to understand model behavior, but finding agent failures (like reward hacking / hallucinations) at scale is easy to miss without automation. Here’s how it works:

Working with @Drewch and the shopify team is awesome. Honestly never seen a group so customer obsessed and willing find the best path to delight them no matter what. Plus an insane iteration speed by being so data + experiments driven!

Onboarded to instinct recently. I told it to figure out if I needed to do anything to be able to drive in italy. It found me a AAA within walking distance for an IDP and also found out my rental car was booked for the wrong airport Absolutely saved my ass.

Creating a platform that researchers actually want to use for training is quite hard, AC2 is our complete solution. Can't wait to see what people cook up! @benscharfstein already trained away em-dashes

Awesome to see more and more of these techniques we are using for customization make their way into the platform! I'm a big believer that hinting is the next generation of prompting, this gives us and our customers a way to easily see the impact.

At Applied Compute, we believe Training wins Inference. In the limit, inference alone is a commodity: providers serve the same models, compete on price, and race toward thinner margins. Training changes that. Once an AI workload reaches scale, serving a generic model stops making sense. Companies will fine-tune, distill, and continually optimize models around their own data, product, economics, and definition of quality. That is the real advantage of open-weight models: you have the weights. You can make a model better on your evals, faster for your workload, and cheaper per token. Over time, companies will choose the best model for their specific task - not the best generic model. We believe that, in the long run, more than 90% of inference on open-weight models will come from trained variants rather than untouched base models. Training also de-commoditizes inference! A provider serving the same off-the-shelf model as everyone else has little room to differentiate. But a company will pay more for a trained version of GLM-5.2 that materially outperforms the base model on the evals that they care about. The providers that help companies train differentiated models (Applied Compute) will be best positioned to serve them in production, too. Training and inference will converge into one continuous system. The long-term end state is continual learning: models that improve from proprietary data, real-world usage, and direct feedback. The companies that close this loop fastest will build products that get better with every interaction and compound their lead over time.

"You need to have a platform that is enabling every firm in the economy to create its own token capital" @satyanadella describes a vision @ypatil125 shares: a platform for enterprises to power the economy on their own terms. It's the goal I joined AC to build — and we're honored to collaborate with Microsoft to do it.

Specialized SotA models has a nice ring to it Proud of the team (@raymondmfeng + @jacob_dphillips) for a great achievement & being very principled in all the evals/hill-climbing (bring back error bars!)