movedx01

Is it the first one priced reasonably and transparently, or am I missing something here?

Don't forget about the "ChatGPT 5 Pro" too :) which is a bit like Deep Research but not quite?

Anything for artifacts perhaps? ;) We use external runners(not blacksmith) and had to work around this manually. https://github.com/actions/download-artifact/issues/362#issu...

Is that true though? Every time i tried to get Sonnet to write out any actually complicated layout/component, I ended up with a novel of classNames with a ton of duplication and useless additions. Those are very hard to spot inside a soup of tailwind classes, whereas I can see them immediately in a nicely laid out stylesheet. This happened extremely often when I tried to iterate with Cursor/CC on the given design, instead of giving up immediately after it failed to one-shot it.

This is great, and I keep my fingers crossed for Hatchet!

One use case I imagine is key here is background/async agents, so OpenAI Codex/Jules style, so that's great if I can durably run them with Pickaxe(btw I belive I've read somewhere in temporal docs or some webinar that Codex was built on that ;), but how do I get that real-time and resumable message stream back to the client? The user might reload the page or return after 15 minutes, etc. I wasn't able to think of an elegant way to model this in a distributed system.

That's great! It would be even better if one of the features included in the table was whether given MCP supports OAuth Dynamic Client Registration, which optional in the MCP standard.

anything OAuth

But you can generate synthetic data using a 3.0 program to train a smaller, faster, cheaper-to-run 2.0 program.

I was almost exclusively using Claude Code for a couple of weeks, and after recently trying Cursor with Sonnet 4 in MAX mode, I think it now comes close. Those are requests paid on top of the sub price though.

It's always radio buttons, every single time

watching baron harkonnen verbally create me code for todo list in React was rather amusing, thanks

GCP does, at least in GKE.

From my experience, the embedded solution becomes a distributed orchestrator and, thus, is no longer simpler.

fireworks does: https://readme.fireworks.ai/reference/requirements-and-limit... but its limited to rank 64, CloudFlare workers too, but there rank is limited to as low as 8

For me it meant: these routine performative discomforts make me more likely to "take the slow route" when it really matters.