@simon Just read your post about ChatGPT horoscopes. Spot on.

I did want to note, you don't even need to be using the memory feature!

It just gives a vague-enough, flattering-enough answer that seems plausible!

It says to me:

"I can infer that you value directness and critical thinking over flattery or vague responses. You seem to have a preference for thoughtful, efficient communication and appreciate when others challenge your ideas constructively. You probably aim to create an environment where both clarity and growth are prioritized.

This might suggest you have a strong sense of self-awareness and a distaste for superficiality, which could sometimes lead you to focus on cutting through excess detail to get to the heart of a matter. Maybe you already know this about yourself, but if not, it could be worth reflecting on how this approach influences your interactions and decision-making."

https://chatgpt.com/share/670ef75d-4a64-8000-8ff8-db86dfdbec1c

Great use for #Claude projects: I'm starting a new job and have had a ton of documents given to me about benefits, procedures, background info, etc.

I've dumped them all into a Claude project, and I can ask things like, "How much will my insurance cover for the crown I've got to get at the dentist?"

You can of course use any #LLM with a large enough context window...Claude projects just make it really convenient.

#ai #career #jobs

I was just having a flashback to when I learned about insertion sort in high school many, many years ago.

I was confused for quite some time because when I read about it, it seemed like the elements were being *moved* rather than *inserted*. I could, of course understand that insertion could be viewed as a component of moving, but I was afraid I was misunderstanding something because it seemed obvious to me that it should be called "move sort" or "shift sort".

I'm not sure that I wouldn't have hit another stumbling block if it had been called "shift sort", but to me it highlights the importance of naming things well!

#algorithm #education #programming

If you're wanting to learn #python Protocol, a useful way to think about them is how they are different from abstract base classes.

Difference 1: Your implementing classes don't have to inherit from the Protocol.

Difference 2: The Protocol isn't enforced at runtime, it's "enforced" by whatever you're using for static type checking.

(There's other more nuanced differences, but I've found this is a good way to get people started on the path to understanding)

When answers.microsoft.com appears in search results, this is the appropriate action to take.

(requires uBlacklist extension)

#search #google #chrome #microsoft

(Reason: https://w4t.pw/f4)

I'm working on #django application wherein the user makes some selections in a form to generate a document.

Given that the document generation leverages some ML models, it'll be 5-30 seconds before the document is ready.

The requirements are that the user is updated with progress of the generation which takes 6 steps.

I started to think about if I wanted to use websockets or polling an #API endpoint or what, and then I remembered something I made earlier this year: https://w4t.pw/eg

You can use #html Declarative Shadow Dom to do async out-of-order progress updates on the page with zero #javascript

#ux

Something I just read on Dwarkesh Patel's substack in a post where he shares his notes about Pinker's "The Language Instinct":

"Slave holders in some Caribbean slave societies imported slaves from different ethnicities, so that the slaves couldn’t organize a revolt. In order to communicate with each other, the adult slaves developed a combined pidgin - a half language which is a mishmash of the ones of offer with an incomplete grammar. But when children are brought up in these societies, every single time they spontaneously develop a creole out of that pidgin - a full language with complete grammar."

https://w4t.pw/f3

#language #linguistics

hot take: When I'm interviewing candidates for a job, using #ChatGPT or #Claude during coding interviews is not cheating. I expect them to use the tooling, environment, tricks, and techniques they'd use during their job.

I only care about their thought process and their answers.

I expect them to understand their code, but whether they come up with all of it themselves, or use something they got from an #llm or #google...I don't care as long as they understand it and can explain it to me.

One reasonable exception: If the job or company prohibits usage of such a tool during the course of their work.