ai

An article in The New Yorker about feeling a little sad about the inevitable replacement of coding with AI really hit home for me. I’ve not used AI for coding yet professionally, but I’ve done a couple of experiments that indicate how useful it might be in replacing not only the boilerplate and drudgery but also some of the thinking that goes into it. I don’t think it will eliminate the role of people in creating software, but it will change it dramatically by potentially devaluing all of the deep techy things it was really fun to learn (e.

Amanda Mull has a nice piece in The Atlantic about retail self-checkout. Self-checkout is horrible. For the most part, it doesn’t make things faster and in some cases (like you want to buy a beer), it makes things much slower. The only use case that it works well is if you are picking up a very few items that have big and obvious bar codes and don’t require a bag. Even then, there is a chance things will go awry.

I like the idea of labeling things as written by humans or AI so you know the origins of what you are reading on the internet. Not By AI has some nice badges that I’ve added to my site because, it is written by my primitive human brain. I thought about adding it to each post that I write, but since I don’t use AI to write the articles (except some small parts when I’ve written about AI and want to show some example output) that seemed like too much overhead on each article.

AI is all over the news these days. It’s even inspired a pretty awesome stock market rally. I’ve been doing some exploration of what is out there and one of the more interesting things I ran across was MusicGen. It does what’s on the tin: generates music using a Language Model (LM). It’s interesting even if the music isn’t great, it does fit with the description used to generate it. I haven’t played around with it yet, but I plan to see if I can generate some cool theme songs for my dogs.

I was curious to see how well ChatGPT could solve a simple programming problem, so I asked: Write a JavaScript function to return the prime factors of a number in an array. A few seconds later, it came back with a pretty good answer: function primeFactors(num) { const factors = []; // check if number is prime function isPrime(n) { if(n<2) return false; for (let i = 2; i <= Math.

I don’t know why I bother. I’ve been following advances in AI with interest and it’s getting scary good. A big story recently is about how uncannily good ChatGPT is at writing text that is essentially indistiguishable from what a person might write. So, I tried it by generating some text on a subject like yesterday’s post: what is JavaScript spread syntax? The ChatGPT response? JavaScript spread syntax is a way to expand an array or an object into a list of items or key-value pairs, respectively.

The scale of the GPT-3 neural network is stunning. There are some models that are even larger, but GPT-3 is pretty much state of the art. It’s a very interesting thing to play around with. If you sign up and get a free account, you can access the Playground which has some examples of different kinds of problems (e.g., chat, explain to a second grader, translation, etc.). There is also an API that is straightforward.