I have a theory that whatever popularity my blog has comes from writing to the level of the reader. My favorite example of the opposite behavior is how every blog post on a topic involving Bayesian inference seemingly must start with Bayes’ formula.
I really like Katex. I used to use MathJax for my blog, but the page reflow was annoying. Now I use Katex with server-side rendering, and even pages with a lot of math load quickly and without reflow.
I didn't implement a CNN from scratch, but a few years ago, I wrote a blog post on CNNs [1] because, like the other commenters, I could find almost no decent blog content on what exactly a CNN was. Maybe it will help in your efforts.
Ever since reading Bertrand Russell's "The Conquest of Happiness", I've thought a lot about the importance of being able to sit inside my own head. One quote that I wrote down:
> A life too full of excitement is an exhausting life, in which continually stronger stimuli are needed to give the thrill that has come to be thought an essential part of pleasure. A person accustomed to too much excitement is like a person with a morbid craving for pepper, who comes at last to be unable even to taste a quantity of pepper which would cause anyone else to choke. There is an element of boredom which is inseparable from the avoidance of too much excitement, and too much excitement not only undermines the health, but dulls the palate for every kind of pleasure, substituting titillations for profound organic satisfactions, cleverness for wisdom, and jagged surprises for beauty... A certain power of enduring boredom is therefore essential to a happy life, and is one of the things that ought to be taught to the young.
In graduate school, I worked for 2+ years before my first paper was published. In that time, I passed my PhD qualifying exam, took classes, wrote code, read papers, learned math, and so forth. Yet when I applied for internships, I received no interest from employers. I suspect this was because I had no concrete signal that I knew anything in my field.
While working on my second paper, I started blogging. In the language of this article, I started generating public intellectual capital for myself. I have definitely experienced the effects of this capital on subsequent job hunts. Now I can point people to my blog to demonstrate knowledge, technical skills, and communication skills beyond the scope of my peer-reviewed work. Furthermore, there is no question about who contributed to my blog, and when I learn something new, I can externalize that quickly.
This is great. I've written previously about why I write as a researcher [1], but none of my reasons were this author's reason: that narrative structure clarifies understanding. This resonates with me because I've noticed that I remember blog posts better, that I understand the topic deeper, if I really flesh it out with narrative writing: why this model, who developed it, what are the alternatives, etc. It often feels like a waste of time initially, but I almost always find that the process makes me realize there are details I initially missed.
This is good advice, and both parts rely on being aware of your own mind and even tricking it a bit. This reminded me of a Moxie story about nearly dying while sailing [1]:
> Strangely, I could feel myself drawn towards the temptation of giving up, even though I knew failure meant certain death. In hindsight, I think it's because the act of giving up feels so similar to the sensation of success, at least in a superficially immediate way.
Over the years, the mantra that "giving up feels like success" has really helped me not give up simply to access that feeling.
This sounds a lot like a tool I built for myself [1], sans the database. I agree that Markdown + Katex with a local server seems like the right move for most technical people. Lots of things like encryption, backups, and basic text search can be done via other Unix tools. I also agree that the big win is owning your data long-term, even if you get tired of maintaining the software.
I didn't want to make the post overly long by adding a section on "failure modes," but certainly one of them is: spend a long time building and tweaking a blog rather than writing. I've done this too and work to avoid it.
I write drafts in Markdown + MathJax and then build everything into a static site using Jekyll. I didn't use a theme or template, since I found most were too noisy. It's just custom HTML/CSS.