Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A critical look at MCP

raz.sh

Post reply on HN
Print view View original

101–110 of 348 posts

Re: A critical look at MCP

#101

Post by solidasparagus »

wisemang wrote: post #39
immibis wrote: post #25

Earlier quoted context omitted.

I understand those with experience have found that XML works better because it's more redundant.

Is it the redundancy? Or is it because markup is a more natural way to annotate language, which obviously is what LLMs are all about? Genuinely curious, I don’t know the answer. But intuitively JSON is nice for easy to read payloads for transport but to be able to provide rich context around specific parts of text seems right up XML’s alley?

Or is is because most text in existence is XML - or more specifically HTML.

Re: A critical look at MCP

#102

Post by _raz »

traviscline wrote: post #98

Thanks for this, I’ve been feeling similarly. I’m working on some Go programs/tools with the explicit goal of describing existing servers in a language neutral manner to try to get some sanity into the mix. I was reenergized to pick this back up because Google is working on a version so I want to get these tools ready. Open to ideas and input, have been noodling on it for a bit now, lots not in form to share but figu…

I had to take a break from extending, our go LLM wrapper, https://github.com/modfin/bellman with mcp to write the blog entry. So some sort of server-like-thing will be added soon

Re: A critical look at MCP

#103

Post by traviscline »

traviscline wrote: post #98

Thanks for this, I’ve been feeling similarly. I’m working on some Go programs/tools with the explicit goal of describing existing servers in a language neutral manner to try to get some sanity into the mix. I was reenergized to pick this back up because Google is working on a version so I want to get these tools ready. Open to ideas and input, have been noodling on it for a bit now, lots not in form to share but figu…

In the current state you can insert “mcpspy” in front of a server and it intercepts and streams out a plain text format that’s nice for humans and machines. There’s also a replay tool that emulates previous traffic, including in mock client and server modes, and a diffing program that is mcp protocol aware.

Oh, and most importantly, a vim syntax plugin for the .mcp file format.

Re: A critical look at MCP

#104

Post by _raz »

baxtr wrote: post #99
hirsin wrote: post #13

In the same way that crypto folks speedran "why we have finance regulations and standards", LLM folks are now speedrunning "how to build software paradigms". The concept they're trying to accomplish (expose possibly remote functions to a caller in an interrogable manner) has plenty of existing examples in DLLs, gRPC, SOAP, IDL, dCOM, etc, but they don't seem to have learned from any of them, let alone be aware that t…

To this date I have not found a good explanation what an MCP is. What is it in old dev language?

A RPC standard that plays nicely with LLMs?

Re: A critical look at MCP

#105

Post by phillipcarter »

ethical_source wrote: post #56
phillipcarter wrote: post #51

Earlier quoted context omitted.

…what? Literally nothing you wrote is accurate.

You'll come around to my perspective in time. Don't take it personally. This generation isn't any worse than prior ones. We go through this shit every time the tech industry turns over.

The people who built MCP are seasoned software engineers, as are most folks who work for these labs. What are you even on about?

Re: A critical look at MCP

#106

Post by auggierose »

vmaurin wrote: post #57

It will probably never work. Companies have spend probably the last decade(s?) closing everything on Internet: * no more RSS feed * paywall * the need to have an "app" to access a service * killing open protocols And all of the sudden, everyone will expose their data through simple API calls ?

I think it is a huge opportunity to dethrone established players that don't want to open up that way. Users will want it, and whoever gives it to them, wins.

Re: A critical look at MCP

#107

Post by Scotrix »

keithwhor wrote: post #91
keithwhor wrote: post #90

On MCP's Streamable HTTP launch I posted a issue asking if we should just simplify everything for remote MCP servers to just be HTTP requests. https://github.com/modelcontextprotocol/modelcontextprotocol... MCP as a spec is really promising; a universal way to connect LLMs to tools. But in practice you hit a lot of edge cases really quickly. To name a few; auth, streaming of tool responses, custom instructions per to…

Quick follow up: I anticipate alignment issues as well. Anthropic is building MCP to make the Anthropic experience great. But Anthropic's traffic is fractional compared to ChatGPT - 20M monthly vs 400M weekly. Gemini claims 350M monthly. The incentive structure is all out of whack; how long are OpenAI and Google going to let an Anthropic team (or even a committee?) drive an integration spec? Consumers have barely int…

OpenAI takes the backseat and wait until something stable/usable comes out of it which gains traction and takes it over then. Old classic playbook to let others make the mistakes and profit from it…

Re: A critical look at MCP

#108

Post by kaoD »

baxtr wrote: post #99
hirsin wrote: post #13

In the same way that crypto folks speedran "why we have finance regulations and standards", LLM folks are now speedrunning "how to build software paradigms". The concept they're trying to accomplish (expose possibly remote functions to a caller in an interrogable manner) has plenty of existing examples in DLLs, gRPC, SOAP, IDL, dCOM, etc, but they don't seem to have learned from any of them, let alone be aware that t…

To this date I have not found a good explanation what an MCP is. What is it in old dev language?

In a nutshell: RPC with builtin discoverability for LLMs.

Re: A critical look at MCP

#109

Post by ethical_source »

phillipcarter wrote: post #105
ethical_source wrote: post #56

Earlier quoted context omitted.

You'll come around to my perspective in time. Don't take it personally. This generation isn't any worse than prior ones. We go through this shit every time the tech industry turns over.

The people who built MCP are seasoned software engineers, as are most folks who work for these labs. What are you even on about?

LOL

Re: A critical look at MCP

#110

Post by traviscline »

traviscline wrote: post #103
traviscline wrote: post #98

Thanks for this, I’ve been feeling similarly. I’m working on some Go programs/tools with the explicit goal of describing existing servers in a language neutral manner to try to get some sanity into the mix. I was reenergized to pick this back up because Google is working on a version so I want to get these tools ready. Open to ideas and input, have been noodling on it for a bit now, lots not in form to share but figu…

In the current state you can insert “mcpspy” in front of a server and it intercepts and streams out a plain text format that’s nice for humans and machines. There’s also a replay tool that emulates previous traffic, including in mock client and server modes, and a diffing program that is mcp protocol aware. Oh, and most importantly, a vim syntax plugin for the .mcp file format.

https://github.com/tmc/mcp/blob/next/cmd/mcpdiff/testdata/sc...

This is what the tests look like, for both the tools and to validate the servers.

Post reply on HN