arunbahl

As a startup that chose to locate in Canada, we’ve already had a dozen amazing candidates currently in the US reach out and apply for roles since we shared our thinking earlier this week [0].

The feeling of ambient immigration hostility in the US (even beyond any one specific policy) is palpable.

0: https://aloe.inc/blog/the-best-talent-in-the-world

From the article:

“We’ll be using reinforcement learning to effectively come up with the optimal toolpaths to see a 3D model, and based on the curvature and the geometric forms, to choose what are the right tools, or what are the right angles of attack,” Springut says. “And when we do that, that’s what’s going to bring the cost of fabricating stone down by 80% to 90%.”

Awesome stuff! We use a similar approach (without MCP) to great effect with Prolog currently and feels like we're only just starting to scratch the surface here.

A great paper from Nasim Borazjanizadeh and Steven Piantadosi at UC Berkeley for those interested: Reliable Reasoning Beyond Natural Language https://arxiv.org/abs/2407.11373

For anyone digging in who wants to hack on this: arun [at] aloe.inc

I’ve been a big fan of Oxide’s approach for a while - thank you for articulating and sharing the thinking behind it.