It's such a dumb discussion.
It's such a dumb discussion. MCP is an API with some description. It adds tools to your agent, along with some context. The (common) complaint is that the principle of progressive disclosure isn't working because all tools, with all their descriptions, are loaded into context right at the start. This is a somewhat reasonable complaint, as the structure makes it hard for the harness to progressively disclose the tools. This is a fundamental issue with anything that just adds a bunch of tools, whether it be via MCP or HTTP (still sad that MCP won over OpenAI's HTTP based approach). How might it be solved? Well, we could work with sets of tools. That's pretty much what the CLI approach does: Wait until you need it, then invoke the help command to discover what to do exactly. The caveat of the CLI being that it's a nightmare to secure. At the end of the day, every capability eats some amount of context because the LLM needs to know when to invoke it.
Posted by solarkraft on May 30, 2026, 3:30 PM
Tags
Related issues
- I wish there was something that could be done beyond feeling powerless.
- I don’t hate taxes per se but I hate that the government will take tax dollars earmarked for one purpose, which is how they got the tax a...
- I've never thought SF Symbols were beautiful.
- > conservatives don't like being told that their ideas, while legally allowed, are not welcome in polite company As long as there are a ...
- Yeah but this AI music is worse than any AI produced code.
- I wish there was a way I could get educational-only youtube as an adult.
Raise real-world problems, find people facing the same issue, and discuss solutions.2,000 issues0 comments
complaint (Other)
I wish there was something that could be done beyond feeling powerless.
NewI don’t hate taxes per se but I hate that the government will take tax dollars earmarked for one purpose, which is how they got the tax a...
NewI've never thought SF Symbols were beautiful.
New> conservatives don't like being told that their ideas, while legally allowed, are not welcome in polite company As long as there are a ...
NewYeah but this AI music is worse than any AI produced code.
I wish there was a way I could get educational-only youtube as an adult.
people say taste is the differentiator but I hate that because taste is personal, whereas culture is communal taste without curation for...
I've always enjoyed hacking on random projects and AI does that really well.
It's such a dumb discussion.
It's such a dumb discussion. MCP is an API with some description. It adds tools to your agent, along with some context. The (common) complaint is that the principle of progressive disclosure isn't working because all tools, with all their descriptions, are loaded into context right at the start. This is a somewhat reasonable complaint, as the structure makes it hard for the harness to progressively disclose the tools. This is a fundamental issue with anything that just adds a bunch of tools, whether it be via MCP or HTTP (still sad that MCP won over OpenAI's HTTP based approach). How might it be solved? Well, we could work with sets of tools. That's pretty much what the CLI approach does: Wait until you need it, then invoke the help command to discover what to do exactly. The caveat of the CLI being that it's a nightmare to secure. At the end of the day, every capability eats some amount of context because the LLM needs to know when to invoke it.
Comments (0)
No comments yet. Start the discussion.