Thread by @yazijys

Hacker News turned 12 last February. Here's the traffic since the beginning — or a dozen-years brief internet history: dropbox.com/s/9f4m0bwnve51…

Disclaimer: except for tremendous love and respect, I have no affiliation with @ycombinator.

How did I get the data then?

PG (@paulg) used to publish HN traffic screenshots every now and then. His last was when HN turned 6: news.ycombinator.com/item?id=5253773

I had that 6-year traffic screenshot saved for years; no idea why.
Couple weeks ago I started annotating it to understand how HN grew and evolved in an attempt to apply some of that to Health News (news.nukleosome.com)

Was planning to share the annotated 6-year traffic (annotations look terrible now in retrospect) but emailed PG asking whether he doesn't mind me doing so.

He of course said he doesn't mind but I may want to ask Daniel Gackle, who's now in charge of HN.

Daniel said it's fine with him as well.
But also said this: "If you wanted to bring it up to date, we could send you the equivalent graphs to the present."
There are probably very few places (& people) on the internet who are that good, open and positive. Here's Dan in PG words:

I got 6 screenshots for the next 6 years' graphs (2013-2019) and from there started some laborious, painstaking pixel work to glue all 12 years together in one graph.

To do that, I used some complex AI/ML models running on supercomputers somewhere in Antarctica, until penguins destroyed them in a strike against climate change.

Now seriously, although it took some time, it all came down to basic principles of simply adjusting pixels coordinates to each graph's scale in order to extract the data points then regenerate everything in one graph.

Now few things I've learned;
1. HN is very unique and hard to replicate because the people behind it were so and worked hard to keep it this way up to this day.
If you'd like to have a glimpse of such incredible work, you can go over Daniel's comments: news.ycombinator.com/threads?id=dang

2. I didn't know that @paulg had started this as an experiment and repeatedly tried to make it less addictive that he even polled users to turn off HN few hours a day.

3. I also didn't know that, beside his elegant work on spam, PG had 3 patents, which sadly weren't in the timeframe of the graph and couldn't be highlighted.

4. Although the graph shows strong survivorship bias, you may notice a couple times throughout how competition did not kill some startups (i.e. @Dropbox) and how PG's sentiment about the issue is highly plausible.

5. If you ever see @ProductHunt and @Twitter having some fun tweets together, that might be one reason why:

6. I didn't know that @patrickc built, beside many things, an offline version of Wikipedia for the iPhone

7. One signal that may get lost in the noise is @balajis and @vijaypande's "Startup Engineering" MOOC on @coursera, which sadly is no longer available but was such a great feat, especially for people not in Silicon Valley.

8. And last for now, something I didn't get to highlight in the graph is that @ycombinator was copied multiple times throughout this 12-year period that @patrickc made this website:

9. Given the nature of venture investing, it was somehow surprising how fast @pmarca and @bhorowitz built @a16z and its reputation. Or how @naval did so for @AngelList.

10. In case you haven't already, give this point of the graph a quick @zoom_us