Earlier this week, this post came up on HN: Death of ShowHN. It makes something clear: the amount of noise when trying to show your shiny new invention is enormous.
I’ve noticed it too (before Arthur’s article). The Renaissance era for software is afoot, and ShowHN is a proof of that.
I was intrigued, how far can this go? We see the AI slop invading every corner of our lives. For example, the job market is being affected by every job posting receiving an avalanche of AI resumes, then ATS AI systems fighting AI with more AI. Social media bots chatting with each other, dead internet theory, dogs and cats living together, massive hysteria!
So I went to the source; the code! first: get the data from ShowHN from August 2025 to February 2026. I picked this period because it appears to be a common agreement that is the turning point for some models (particularly Anthropic’s Opus).
The starting theory here was, if there’s an uptick in posts, there should be an uptick in GitHub repos. And if there is an uptick in repos, well… we can start analyzing them to understand the long-tail effects of this period.
From Aug 2025 – Feb 2026 I found 20,416 posts which published 5,714 GitHub repos. Understandable, since ShowHN sometimes has app sites and sometimes just repos.
The first glance reveals the following:
If we measure the age of the repo in days, we can see the time anybody waits to show their brand new creation. An interesting trend is surfaced here, the majority of the new repos have a short life. This could be coincidental, but it tells me the code was produced with haste. As if people were pumping tokens for source code like there’s no tomorrow.
How many repos actually go to live and thrive after their (supposed) HN hug of death you may ask?
31.7% are dead in the water instantly after their 3 minutes of fame on Show HN.
Time will tell for some of these of course, but more than a quarter of what hits ShowHN and has a repo, goes to die.
If we look deeper here ~50% of Show HN repos are less than a week old. What happens to them?
At this point, I realized I wanted to have more data, this might be a fluke, maybe this is how 2025 started?
So I decided to go for another period, let’s say a year before August 2025.
| Period | Posts | Github repos |
|---|---|---|
| Aug 2024 – Aug 2025: |
22,885 | 5,192 |
| Aug 2025 – Feb 2026 |
20,416 |
5,714 |
| Au 2024 - Feb 2026 |
43,301 | 10,906 |
Interesting data point here as well, the amount of posts/repos in 1 year prior doesn’t reach the amount of posts/repos of this short period August 2025 - February 2026.
We can clearly see this is not how last year was, the amount of side projects doubles in 2026.
Combined with repo longevity we also validate this trend, more code (slop?) is being pushed to Github also faster than ever.
The survival picture breakdown also paints a story; we used to have more “mature” things being presented. Now we have a deluge of very hot off the presses code thrown at us.
I think we sense there’s a shift here, we are able to ship things faster and in great volumes. Github alone seems to be getting full of side projects that might go nowhere.
I personally don’t think this is sustainable, this low signal to noise ratio is making probably great side projects sink.
What will make your project stand out in the future?
Most of those articles linked here are promoting something, which I imagine is a way to go around the avalanche of new “stuff”.