Granted, it was so creepy seeing the posts asking questions about MCP before any information was available about it. And the continuous repetition of how it is like, not USB, but USB-C, which in my mind is a regression because USB really was one cable that did everything whereas USB-C is a maze of twisty little cables that all look alike... Except one of them will charge your phone to 100% overnight and the other one will charge it to 8%.
I'd be disparaging about what a terribly designed protocol MCP is but back in 2017 I designed a terrible protocol for an AI trainer too. People who are good at one thing (rockets) often are bad at other things (politics) -- that somebody is going to design a bad protocol as opposed to "just use websockets" is how it goes. Also AI confounds normal expectations in many ways, I never really found a right answer for a model trainer which can (1) train a foundation model in three weeks on a moderate-sized cluster and (2) small train 20 models based on that foundation model in 3 minutes and pick the best one out of those. Back then I knew the importance of evaluation and model selection and all that but didn't have the mastery of it I have now.
I wrote this essay years ago when I had similar feelings
It gets repetitive but daily visiting HN is too much living in a bubble (I open it several times per day, too). https://news.ycombinator.com/best is how I judge quality and diversity of stories. Relative diversity of course, HN is still a lot of tech news.
HN isn't your personal platform. If you're not interested in a post, don't upvote it (or even flag it), but I don't see what brings you to complain about the posts that other people are upvoting. It's not your lawn.
In theory, HN should be easily manipulatable. Why should I automatically trust it as a representative of what the crowd thinks?
You should open (at least) two threads in your mind. In one of them, it is a representative of what some crowd thinks. In the other, it is a representative of one or more third-parties performing to a smaller crowd.
It's a very simplistic example, but enough to get started.
If you feel this to be a problematic situation, do not accept it as the norm. If you wish to find others aligned with you, share these grievances and engage with the community, and within yourself, to identify a solution.
Honestly it's just all blurring together. The way I would put it I used to be excited to see what content got the most traction each day on HN. Now it just feels like I'm seeing the same thing over and over. But we haven't really had any sort of major breakthroughs that justify the level of saturation.
I count 5/30 (15%) LLM related posts right now.
Granted, it was so creepy seeing the posts asking questions about MCP before any information was available about it. And the continuous repetition of how it is like, not USB, but USB-C, which in my mind is a regression because USB really was one cable that did everything whereas USB-C is a maze of twisty little cables that all look alike... Except one of them will charge your phone to 100% overnight and the other one will charge it to 8%.
I'd be disparaging about what a terribly designed protocol MCP is but back in 2017 I designed a terrible protocol for an AI trainer too. People who are good at one thing (rockets) often are bad at other things (politics) -- that somebody is going to design a bad protocol as opposed to "just use websockets" is how it goes. Also AI confounds normal expectations in many ways, I never really found a right answer for a model trainer which can (1) train a foundation model in three weeks on a moderate-sized cluster and (2) small train 20 models based on that foundation model in 3 minutes and pick the best one out of those. Back then I knew the importance of evaluation and model selection and all that but didn't have the mastery of it I have now.
I wrote this essay years ago when I had similar feelings
https://ontology2.com/essays/HackerNewsForHackers/
but I can't really stand by it because things have changed.
It gets repetitive but daily visiting HN is too much living in a bubble (I open it several times per day, too). https://news.ycombinator.com/best is how I judge quality and diversity of stories. Relative diversity of course, HN is still a lot of tech news.
HN isn't your personal platform. If you're not interested in a post, don't upvote it (or even flag it), but I don't see what brings you to complain about the posts that other people are upvoting. It's not your lawn.
In theory, HN should be easily manipulatable. Why should I automatically trust it as a representative of what the crowd thinks?
You should open (at least) two threads in your mind. In one of them, it is a representative of what some crowd thinks. In the other, it is a representative of one or more third-parties performing to a smaller crowd.
It's a very simplistic example, but enough to get started.
TL;DR your boredom can be manufactured.
All great projects start with a grievance.
If you feel this to be a problematic situation, do not accept it as the norm. If you wish to find others aligned with you, share these grievances and engage with the community, and within yourself, to identify a solution.
> same repetitive posts
You can't just say that without telling us which posts exactly you mean, with links.
Honestly it's just all blurring together. The way I would put it I used to be excited to see what content got the most traction each day on HN. Now it just feels like I'm seeing the same thing over and over. But we haven't really had any sort of major breakthroughs that justify the level of saturation.
I don't have the same impression.