Nobody is going to mark their game as "Made with AI" unless customer sentiment shifts or there is a very strong incentive not to lie. Right now releasing a game with a ai label is just net negative. Meaning the filter would be utterly useless even if it was introduced.

This is fun. It would make sense to let agents submit proposals for adding new tools as well because right now the tool selection to review is very limited. It would be a great signal for OSS maintainers too I think

"Trivially" is relative. In the case of wow the standard way to do botting is to use the in-game lua scripting (normally used for UI addons) with more functionality unlocked. The difference between introspecting the whole game state and controlling the game in-process with good api vs running realtime visual model on the captured visual feed and using fake keyboard inputs is pretty large.

TBH a "global arena thing" can be very good for performance rather than a bunch of random allocations/deallocations

My understanding is that code mode is supposed to be implemented by the harness, not the MCP provider. You chain multiple MCP providers as well as other harness provided tools inside the sandbox.

My timeline might be wrong (I remember a cloudflare article mentionning the "in MCP" case), but anyway yes there's tools to do it in the harness now and it's the better idea.

Can we talk about the point the article makes that the MCP servers should implement JSON apis instead of text APIs? It makes sense if your harness supports code mode, but an embedded code sandbox is a pretty big piece of machinery and not everyone implements it. It seems like a massive change to the way one should build an MCP. Also why again are we using MCPs instead of OpenAPI when it's now supposed to be just stateless JSON APIs?

Teachers have said that about every generation since time immemorial.

So it's an electron app but you don't support editing or even viewing on the web? And no windows client?

windows is on our roadmap (fingers crossed launching tomorrow!). and as far as the web-based affordances, we are working on it! wanted to start with the local dev experience before trying to build multiplayer features.

windows is ready! https://install.dev.fast/windows

You cannot introspect model training by prompting the model. Whatever answer it hallucinated on your query "where do you know that from" has almost certainly very low bearing on reality.

True. It sort of doesn’t matter how it was trained. The fact that it knew the company based on an internal idiom is super surprising

Personally I've found the opposite useful - I mostly use OAI models with OMP harness and I have a rule instructing it to summarize if any work was omitted or if it made any surprising changes. Sometimes the agent will forget to do some piece of work or invent a new bizarre way of doing things, but quite often it is able to self reflect on this at the end of the session.