jdilla.xyz

Friday threads

2024-04-19

• How school rankings work. I have so much suspicion of school rankings. First, I am skeptical that they actually measure the effectiveness of schools. Second, I suspect that they direct billions if not trillions of dollars of spending every year when parents go to shop for a home and check school ratings 8 and above.
• Snowball earth! Everytime I run into an article or youtube video like this I falll for it. Just go through wikipedia and then dramatize it a little bit for me.
• “Surely nothing could “survive,” whatever that means exactly, for 100 million years. But when brought back to the lab and offered nutrients, the microbes began to grow and multiply.” How deep does life go?
• The Founder Effect. The closer you are to the original source, the more variation you’ll find.
• Albert Hirschman on decision making as the scarce factor in economic development. One of my main observations from working at YouTube was the company was decision constrained, not resource constrained.
• Are LLMs reaching diminishing returns? On some level, I find this comforting — I feel confident that I’ll be able to work in a way that compliments the LLM and there will still be room for humans in the loop.
• “Alex died suddenly, when he was still relatively young. The evening before he died, Alex said to Pepperberg, “You be good. I love you.” The Great Silence by Ted Chiang.

Friday threads

2024-03-30

• Ezra Klein on parenting. Three things I took from this podcast:
  • “I don’t know of any policy intervention that reliably increases fertility” <- a strong argument that culture more than policy is what is leading people to have less children
  • The dueling tension between being the ideal worker — giving 100% to your job, being available, etc. — and the ideal parent as someone who is actively spending time giving their child attention
  • A reminder that opting in to these cultural values is at least to some degree a choice.
• War on the Rocks: “As I have seen in Ukraine and have observed in other theaters, the introduction of robotic and autonomous systems into the force is liable to increase both the number of people and the diversity of skills necessary within the force.”

The English coined "soccer"

2024-03-27

From my friends at Duolingo:

The people that affectionately call their Prince William "Wills" and £5 and £10 notes ~"fivers" and "tenners"~ are responsible for shortening "Association Football" to just "Assoc."—which, when written, looks like it might be pronounced "Assock." (This "Association Football" name is the same as the French Football Association in FIFA!) In late 1800s England, at Oxford, there was also a fad of adding -er to some words. And thus, "soccer" was born. In England. 🏴󠁧󠁢󠁥󠁮󠁧󠁿

Quantity precedes quality

2024-03-25

Found via Dynomight:

Quality over quantity. I often worry that I write too much on this blog. After all, the world has a lot of text. Does it need more? Shouldn’t I pick some small number of essays and really perfect them?
Arguably, no. You’ve perhaps heard of the pottery class where students graded on quantity produced more quality than those graded on quality. (It was actually a photography class.) For scientists, the best predictor of having a highly cited paper is just writing lots of papers. As I write these words, I have no idea if any of this is good and I try not to think about it.

I hadn’t heard this before, but I do find it to be true. Creativity is a habit. The way to quality is through quantity.

Claude Review

2024-03-25

You can read my review of Phind, another LLM tool, here.

For the past several weeks, I’ve been using Claude Opus (the paid version of Anthropic’s ChatGPT competitor).

For the first time, I think ChatGPT has been outclassed.

For context, I use Claude for help coding (you can read about my process for coding with GPTs here), for understanding new-to-me technical concepts, and for marketing, business, and technical writing.

Here’s what I like about Claude:

• It feels smarter. I’m not sure exactly how to quantify this, except to say it’s like when you sync slightly better with one coworker than another. On coding projects, I notice fewer suggestions that lead me down a dead end or that add unnecessary scope.
• It has less of a heavy tone on writing. I feel like ChatGPT gives me decent feedback on my writing, but when I ask it to write for me, it feels like it has been written by ChatGPT (some hallmarks: excessive excitement, over explaining). This is fine in some cases, but in others, I end up having to remove the ChatGPT-ness of it.

With that said, Claude has some pretty big limitations:

• Just 5 images per chat is a crime. Screenshots of error messages is the default now.
• I really miss the context I’ve been able to add to my own customGPT for my coding project. It’s not a lot of extra information, but having the LLM know my basic directory structure saves a lot of extra messages.
• If there is a way to stop a message, edit it, and resend it, I haven’t found it yet in ~3 weeks of daily usage.

Despite all these limitations, given the choice between which one to work with, I am consistently choosing Claude. I wonder what this says about the stickiness of these tools?

Now for the ultimate test: have I cancelled my ChatGPT subscription yet?

No, but for an unexpected reason: the ChatGPT mobile app. I’ve been having voice conversations with ChatGPT mobile, both personally when there’s a topic I want to dig into, and increasingly with my daughter. We pick a topic she’s interested in (e.g., where do dinosaurs come from) and just riff on it. It’s perfect for a curious toddler — she can keep asking questions over and over again. I think this is what Tyler Cowen means when he says creators are competing with LLMs for attention.

We’ll see how long this lasts — I don’t expect that I will continue keeping two LLM subscriptions forever. Hurry up and add a mobile app, Claude team!