Underwhelming MCP vs Hype
As I see the current state of hype, the experience is underwhelming:
Confusing targeting — developers and non devs both.
For devs — it’s straightforward coding agent basically just llm.txt , so why would I use MCP isn’t clear.
For non devs — It’s like tools that can be published by anyone and some setup to add config etc. But the same stuff has been tried by ChatGPT GPTs as well last year where anyone can publish their tools as GPTs, which in my experience didn’t work well.
There’s isn’t a good client so far and the clients UIs not being open source makes the experience limited as in our case, no client natively support video upload and playback.
Installing MCPs on local machines can have setup issues later with larger MCPs.
I feel the hype isn’t organic and fuelled by Anthropic. I was expecting MCP ( being a protocol ) to have deeper developer value for agentic workflows and communication standards then just a wrapper over docker and config files.
Let’s imagine a world with lots of MCPs — how would I choose which one to install and why, how would it rank similar servers? Are they imagining it like a ecosystem like App store where my main client doesn’t change but I am able to achieve any tasks that I do with a SaaS product.
We tried a simple task — "take the latest video on Gdrive and give me a summary" For this the steps were not easy:
Go through Gdrive MCP and setup documentation — Gdrive MCP has 11 step setup process.
VideoDB MCP has 1 step setup process.
Overall 12, 13 step to do a basic task.
"I got an alert for X, who can help me" surfaces a bunch of helpful stuff. Not groundbreaking in the sense that I could do the leg work myself, but fantastic from automation point of view.
What scares me is the "YOLO mode" and similar in cursor. It's not practical to review each and every tool execution after the first couple of manual tool executions... Prompt injection is still a thing, whether intentional or accidental. Little Johnny Tables is still around.
It's still early days, the complicated setup with OAuth and or API tokens will be improved imo.
I imagine a simple Bad Request could derail the convo since it's added to the context?
Fundamental things are keeping me away
- Me: "Do this"
- LLM: "Ok, let me try do this.. nope something's off. Let me try another way".
Occasionally, I do have to abandon a run but in general it seems successful.
I wonder why there is a protocol at all for end users since you are using natural language as the ACTUAL protocol.
Is MCP not for developers?
I thought of it as like a GraphQL in that it takes over the network comms layer but for LLM queries instead of payloads.
Idk
Check this out https://smithery.ai/
I see - fundamental misunderstanding on my part about what it is.
It's a chat thing - for the end user. The word "protocol" is very confusing here.
Makes sense why it didn't make sense, if that makes sense.
I invented something like this months ago to pull up a calculator, browser, etc. in the chat UI (try it chat.blacksand.cloud)
If you prompt what is "2 + 2" it will respond with "4" and render a calculator tool in the chat. It knows what tool to render based on this system.
I guess that is what they are formalizing.
I get it now, cool.
There is a video of what I am talking about here: https://www.blacksand.cloud
Prompt "What is 2 + 2" and it will not only answer, but also bring up a calculator tool. You can try this here https://chat.blacksand.cloud
I'm still confused on what MCP is.
Prompt "What is 2 + 2" and it will not only answer, but also bring up a calculator tool. You can try this here https://chat.blacksand.cloud
I'm still confused on what MCP is.
MCP is just hype. Hot air.