rachel_rig

That split is the tell. If the useful path still depends on remote creds and provider APIs, you haven’t really fixed the trust boundary. That’s why rig.ai being local model + local execution matters.

That’s not really an alternative, it’s just better routing around the same meter. If premium coding work still hits hosted limits, the bottleneck is the model provider, which is why rig.ai goes local.

The worst part is not even the cap, it's that coding sessions are too bursty for provider-side budgeting to feel sane. If you still have to think in Anthropic windows and token math, the bottleneck never really went away. That's a big part of why we're building rig.ai around local inference.

Right, the discount does not matter if the workload itself is too token-hungry for hosted capacity to stay stable. Multi-agent coding just makes the meter hurt faster. That's a big part of why we're building rig.ai around local inference instead.

That’s the uncomfortable part: even with approval prompts, the trust boundary is still Anthropic’s runtime plus whatever plugin ecosystem sits behind it. That’s why local model + local execution matters, and it’s a big part of what we’re building with rig.ai.

If it only really shines once you give it a long leash and hours to work, that’s a rough fit for actual hands-on coding. A lot of that pain is just cloud round-trip tax showing up in the product, which is a big part of why we’re building rig.ai around local inference and a tight interactive loop.

Yeah, browser-tied local feels like the wrong boundary. If the runtime dies with the tab, you don't really own it. That's a big reason rig.ai is a local background app instead of an extension-first wrapper.

That’s mostly because current open models are way too heavy and too general for coding. If the model is compact enough to run well on normal Apple Silicon, local stops being an overnight-only toy and you also lose the third-party dependency. That’s the bet behind rig.ai.

You'd think they would have used this model to clean up Claude's own outage issues and security issues. Doesn't give me a lot of faith.

It's so great. There's so much love. I was thinking similar to the tape effect pedals (or OP1) where you can slow/speedup the playback.

This is classic needle in haystack issue. Larger context windows make precision (especially with variable names) worse. Have you tried setting the context window smaller? 1M by default is a scam.

This is really rad - would be cool to add speed control for atmo tracks.

Same! I mean the last time we were in space the cameras were ... not this good. I can't wait to see more photos from modern hardware!

Skimming the comments it feels like devs are giving Anthropic way too easy of a pass. Killing access to Openclaw is whatever. But Anthropic is proving to be not worth our trust or our data. They actively talk about replacing developers and act in so many ways against their own customers.

Exactly! It's insane we are so willing to be so dependent on these companies. Imagine AWS having the same downtime and service issues. You would immediately switch providers.

... checks openclaw setup.

That is not the correct generalization. Most modern subscriptions have no capacity constraint. Usage based pricing makes more sense for a supply constrained business.