justinl33

So nice to see GPU's being used for classical reasons again.

entity-oriented models are more stable over time compared to procedure-oriented RPC. In my experience, starting with resources/entities and mapping operations to them does lead to cleaner APIs that are easier to evolve.

I disagree that OpenAPI is just RPC mapped to HTTP. A well-designed OpenAPI spec can be quite RESTful. The problem is many developers don't take the time to design good resource models and just slap RPC-style operations into URL paths.

> This is the first open research to validate that reasoning capabilities of LLMs can be incentivized purely through RL, without the need for SFT.

This is a noteworthy achievement.

I've built similar systems using Apache Airflow and Temporal, but the complexity was overwhelming. Using simple maps with enter/leave phases for workflow steps is much cleaner than dealing with DAG frameworks.

It's giving _Students for Fair Admissions v. Harvard_

human stereo vision processing has incredibly sophisticated noise filtering capabilities that are still hard to replicate in software. The shimmering effect people report is essentially your visual cortex highlighting areas where the stereo correspondence fails.

let's thank him for making some absolutely god-mode architectural calls (protocol layering, manufacturer agnosticism, etc.)

this is good. the automated systems were getting increasingly byzantine, with layers of rules trying to patch edge cases, which just created more edge cases.

I've maintained several SDKs, and the 'cover everything' approach leads to nightmare dependency trees and documentation bloat. imo, the LLM paradigm shifts this even further - why maintain a massive SDK when users can generate precisely what they need? This could fundamentally change how we think about API distribution.

Exactly

nice, thanks for sharing!