Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Can you please stop posting unsubstantive/fulminatey comments? We've already asked you this multiple times:

https://news.ycombinator.com/item?id=45022004 (Aug 2025)

https://news.ycombinator.com/item?id=44396920 (June 2025)

https://news.ycombinator.com/item?id=43028401 (Feb 2025)

https://news.ycombinator.com/item?id=39337678 (Feb 2024)

You've obviously got some substantive points to make here, which is great, but indignant putdown rhetoric has a destructive effect on the threads. If you could just make your substantive points thoughtfully, we'd appreciate it.

https://news.ycombinator.com/newsguidelines.html





OK, I will tone it down.

However, it IS also a description of the current state of affairs in the MCP land. The comment threads and proposals in the MCP projects are dominated by the LLM-generated text, so it is almost impossible to keep the full picture in one's mind. LLMs made it possible to create an overwhelming amount of activity with ease.

Moreover, a lot of _code_ for the MCP servers is also AI-generated and has never been used in practice. It's easy to verify. Here are Github search results for the ProxyOAuthServerProvider that is supposed to delegate the authentication to a third-party server: https://github.com/search?q=ProxyOAuthServerProvider&type=co...

There are 215 results at the time of writing, and all but 3 of them are either forks of or LLM-fueled rewrites of the same code from the `modelcontextprotocol` repo. And one of the 3 is mine, and it doesn't quite work.

So yes, "It's just more of AI slop". Sorry. That's just a neutral description of the current state of affairs in the MCP/AI world. And yes, it's absolutely horrifying.


I don't have any problem with this argument* and if you had posted this comment originally I wouldn't have replied. The "absolutely horrifying" bit at the end is a bit beyond the pale, but not, er, absolutely horrifying.

(Btw, that makes your comment a case of the so-called rebound effect (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...) - one of the more interesting phenomena we've noticed.)

* I mean qua moderator of course. Qua reader, I have no opinion.


Yes, I'm sorry about that. I really should not write one-line comments without taking time to detail them. I will just stay silent in future in these cases.

> bit beyond the pale

Uhm... Why? It's an honest question. I thought that "horrifying" (as in "inducing horror") is a normal descriptor, not racially/sexually loaded or anything. The AI situation certainly induces real dread in me.


I did say "a bit" :) - I guess because it pattern-matches to internet rhetoric for me. By 'internet rhetoric' I mean something where the largeness/sensationalness of the claim is high and the information content is low. (Oh also, the leading "And yes" is a bit of an internet trope)

This is quite minor though; if it distracts from the main point, I'd say forget it.


I've made some contributions to the typescript-sdk and I'm part of the MCP Contributors server and can confirm, it was all vibecoded and the direction it is going is really not promising.



Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: