2026-03-16
I had (and to some extent, still have) a habit of checking out Github's Trending page every couple of days. The habit is objectively useless, but I like checking out the developers whose repositories made it to the page, and mostly seem like people who are worth following (in the tech space). Or that is the feeling that I used to have. For the past year, that page has been featuring more and more AI repositories, which is expected given the boom of AI agents and LLMs. 8 of today's (16th March) 11 trending repositories are focused on agents or language models in some way. Day by day, it feels like we are seeing the some different iteration of the same things again and again. There is only so much skill repositories that you can be tempted to check out. Don't get me wrong, programming has definitely gotten more interesting because of this, and I am one of those people who enjoy it more than the average user. But the saturation and public visibility of open source work concerned with these topics really make it hard to find stuff that are more organic, or even human.
4 years back, I did a pet project, Dev Sites. Each day, the website would pull in the profiles of developers made it into Github's trending tab who also had something on the "website" field. The list grew each day, compiling a list of developers who were mostly interesting and inspiring. Some of these websites were company or product pages, some just links to their socials. But most of them led to the personal websites/blogs of these people, where a lot of them expressed their thoughts, be it technical or not. My initial goal for the compilation was to have a central hub for these voices, that I could browse any time later. I have found a good chunk of people to follow this way.
The site collected data for about half a year before some issues with the Firebase configuration it was initialized with. I a) did not care so much and b) had such a big list already compiled that I never fixed that. So, the pulls came to a halt, and the site froze with a list of 1066 websites, which is still accessible.
These people had a lot of variety among their works in terms of the problems they were tackling, the tools they were using, and the medium they were putting their efforts in. This variety is something that I have been feeling is "missing" for the past year.
I had also stopped checking the site often. In the past two years, I have visited the site maybe three or four times. Yet, it sits among my pinned repositories, something that I find worth to show off. On my latest visit, what I noticed is that a good portion of these sites have either gone dark or simply gone inactive. A lot of the sites have had fewer and fewer activity over the past few years. A lot of the sites have pivoted towards writing about AI mostly. I ran a crawl on the websites and the number of posts mentioning AI topics cover up ~40% of the total posts in 2025-26. It is exciting, yet cooks up a burning feeling of discomfort inside my stomach. Have we stopped writing about the things that we like? Or have we simply stopped liking or writing about the thousand other topics that really deserve to be talked about even in this ridiculously fast paced time of agents and automation. I now see less posts going over why a language is beautiful to some people, or content talking about system optimizations, or people in large scale systems speaking about what cool things they do. It feels like we have evolved to the point where most of our approaches, or innovation, or tech stories are centralized only on some few topics. Ironically, I might also be contributing to that percentage with this piece.
These are the results I came up with for the links in the site. A detailed report is here in this repository.

Site type breakdown:
- personal: 778
- company: 47
- product: 44
Sites mentioning AI topics anywhere: 242 / 869 (28%)
**AI mentions per year** (among sites with posts that year)
- 2022: 42 / 169 sites with posts (25%)
- 2023: 54 / 162 sites with posts (33%)
- 2025: 88 / 200 sites with posts (44%)
- 2026: 54 / 127 sites with posts (43%)
**Sites that pivoted to AI** (wrote about AI in 2025/26 but not in 2022/23): 27
**Average posts per year** (across feed-tracked sites)
- 2022: 169 active sites, avg 5.7 posts, max 49
- 2023: 162 active sites, avg 5.8 posts, max 42
- 2025: 200 active sites, avg 7.8 posts, max 251
Sites with last detected activity before 2023: 144

2026-06-25
So this post came up the charts of HN today. Commenters in this thread have talked about this whole discourse (and then some more) from a lot of different point of views. I think this is the correct term I would use for what I am feeling. Highly recommend checking that out. And before you get any wrong idea, the tiredness also comes from my expectations and optimism about the future of AI. AI fatigue is not boredom with AI itself, it's the cultural cost of living in the gap between optimization promises and organic expectations.