Late Night Writing About AI
I set myself the strange goal of writing about artificial intelligence (AI) every day for 500 days. It seems inherently weird to start thinking about the consequences or reflecting on this more thoroughly on day 129, today at the time of writing. Do not expect any revelations about AI in this article, this article is on writing about AI.
Saving Up or Publishing Late
I have said to people whom have asked that I try to save up articles on a Sunday to publish them day by day. This was true at some point until it became false. Especially when things stacks up in regular life with marriage, studies, job etc. Then it becomes an outright lie as several weeks go by, not that I have lied — I did tell people asking that it has degenerated into too many late nights. When those repeat it becomes an unfortunate situation. Only stubbornness keeps me going, or the streak? A Snapchat streak is when you send direct snaps back and forth with a friend for several consecutive days.
It is a bit like I am a Snapchat streak writer.
Let’s Be Real
Sometimes the mind can feels like a mess and there is a sense of bad conscience as I tap away senselessly on a keyboard — attempting of course to understand the topic of artificial intelligence. My wife calls it the ‘tap-tap-tap’. It is because my keyboard is loud and she can hear the sound, so she knows that I am writing.
My rule of thumb in the 500 days of AI is that as long as I publish before dawn it counts as within the day. However this has led to some peculiar situations of which I will mention a few.
- Coming back home from a wedding at 3am and staying up to 5am to finish an article while staying with my wife’s family (I’m a horrible person, but it was Sunday the day after).
- Writing articles on my mobile phone when I had no chance to access the Internet. Sitting on a bus writing about AI on the Medium app feels nice, but very awkward.
- Being drunk while writing an article after a party desperately trying to finish the draft to publish it while doing some edits the morning after.
- Writing until 1am one night just to find out that Medium deleted the whole article by some weird bug, then rewriting the article and posting it.
It is really a collection of great moments once you press that ‘ready to publish’ button. It does give a buzz, however
Who Do You Write For?
When people ask me I generally tend to say that I write to learn, which is true. However I generally tend to write for someone, and one of my favourite writers asked me to picture who I write for.
Again, being honest I have not done that too much yet. I could do it now. Who are you reading this? I imagine someone one of two people, either a friend reading this out of guilt (I share so many articles) or a stranger in a room trawling through Medium.
There might be something wrong with me, yet I cannot seem to find something as accurate. If I had imagined writing for my grandmother so she would stand artificial intelligence it would be very different. How would I described it?
My grandmother does not read Medium, I think.
On Medium I realised my most technical and in-depth articles were the most popular. My articles about dataflow or semi-supervised learning have more views as an example. My first 30 days I had about 900 views. These last 30 days I have had more than 12'000 views. Readership increased from 400 to 4000 — the difference being those who look at it or actually read it.
What I appreciate the most about having readers is having discussants. When I meet people who read an article I can discuss with that person what they thought about during or after. Harsh criticism is not something that I have received yet online, maybe I am not controversial enough, or most Medium members are very polite.
I discovered a feedback lesson in a seminar on International Security Policy was probably the harshest critique I have received thus far. I handed out my paper, it was read, notes were taken and it was badly received in the best way possible. They pointed out clear inconsistencies in both texts, numbers, words and paragraphs.
What I will do is hand more people my reading in an organised fashion and ask for their feedback. I can send it to a few people and ask for their criticism. Recently I printed 100 copies of a book draft that I was working on titled Fifty Shades of AI – The Climate Crisis and Artificial Intelligence Safety. I have sold or handed out almost all of these yet received feedback from two people so far.
Therefore I conclude that even if you ask for feedback if the setting is not right, or you don’t ask for it within a given time it may not be given. This could be a valuable lesson. Feedback in a few subtle ways is great, even if it is something simple as in someone stating that a specific chapter was great or did not work.
If you want to give me feedback on any of the 129 days feel free!
It is a bit silly to sit up late typing on a computer when I could be sleeping, and still it has to be said or typed. I really have enjoyed sharing the journey thus far with you my reader. There has been many days of not understanding AI and there is likely to be many more, 371 days if we count, but hey who’s counting? I did when I started this project and it made every day a milestone.
Lately I have been invited to debates and talks about artificial intelligence where I mostly talk about Climate Crisis and AI or AI Safety depending on the circumstance. Making these talks have forced me to condense the information that I have found to attempt communicating it in a summarised manner.
If this has been helpful or you enjoyed my article please consider responding here or sending me a message on a social media platform.
Thank you for reading and being patient with me.
I will try to be more diligent and cut down on my late night AI writing.
This is day 129 of #500daysofAI.