If we don't blog, what will the next big AI models be trained on?!

Good point. "Blog as much as you can, as otherwise the training set will be Twitter!"

I think there is a more important reason to blog besides the 3 reasons listed: to force yourself to slow down, organize your thoughts, fill in the holes, and articulate your points.

"Writing is understanding."

"Writing is nature’s way of letting you know how sloppy your thinking is" - Guindon

Though admittedly I first heard it from Leslie Lamport and that's who I associate it with.

That's what I was trying to cover with the "make your newly-acquired knowledge concrete" bit, and was my focus in the previous post. This time around I wanted to look into the aspects that might be impacted by AI (and why I didn't think they would be).

You don't really need to blog for that though? You could just write to a text file on your PC and not throw it online.

Socrates would like a word with you.

Third time! It’s good. I like reading these kind of stories so I resubmit if they don't get attention.

If I submit something that's recently been submitted, HN just loads the prior submission. How do you avoid that?

Wait at least 24 hours and only if it didn’t get many votes.

Are/were you a prof at UW?

I was 6 years ago but in math.

Did you make SageMath? I used it a bunch in college. It was/is such a cool tool. There was a LaTeX plugin or something and I used it to grade students papers. It's also how I learned about hardware and software precision. Along with a whole bunch of other stuff.

Also very much thought it was cool to know someone can nerd out on math and skateboard.

Carnegie Mellon ;)

You might be surprised how many people read your blog! If you hang out in techie places, you might even get some, "oh, I read your blog!", comments from strangers.

Very nice, I didn't know this. I used pyppeteer and selenium for this previously which seemed excessive.

See my article from January 2023, "Natural language is the lazy user interface".

https://austinhenley.com/blog/naturallanguageui.html

Lately I’ve made working on tools that make it easier to gather context for ChatGPT. For example, copy the relevant code files to my clipboard.

Read about a few of them here: https://austinhenley.com/blog/promptcontext.html

I ran a study at Microsoft years ago that’s relevant to this. Can I inject static analysis warnings into code reviews? How many? Which ones? What about nearby unchanged code?

https://austinhenley.com/pubs/Henley2018CHI_CFar.pdf