I mostly do research to create new tooling, but explicitly not with ChatGPT, because I am convinced that that (and all hosted chatbots with it) is a massive trojan horse, like FB. So I only use local implementations and I since a month and a bit I have a little set-up where I queue up large LLM requests and run it on rented cloud hardware.

Under (early) development:

• A "dream" mode where idle GPU/NPU time is being used on a RAG knowledge base to re-prioritize knowledge and update importance of stored knowledge. I'm not sure if this is going to work. For now I'm re-tooling it completely because I started in a wrong direction with it initially.

In "testing":

• A case builder that judges relevance of additional information to be added to a case. I.e. I can feed it a link, I'm currently working on this article from this morning. I programmatically extract all links and fetch all these, convert videos, podcasts and pdfs to text, which are added to a RAG knowledge base powered by all-MiniLM-L6-v2 as a sentence transformer. Then 2nd layer links are each examined for relevance, and so on. This is currently eating most my NPU on my laptop, especially since I'm messing with PDF transposition with minerU which takes for. ev. er.

In "production":

• News article aggregation against clickbait titles (#1006868) I use this daily as an alternative RSS reader. I need to rework it though, I'm thinking to maybe use Airflow for orchestration, the prototype suxx because of NATS not being stable when ran locally (which was why I wanted that and not kafka).
• I noticed that I use Brave's AI-enhanced search results more often than the algo results.

Video calls with Lily from Duolingo )

My most frequent recent use cases- friend group has gotten me addicted to gambling in fiat stocks, I know they are shit coins but it's a lot of fun and better than betting on sports.

1. I have a project for "Technical Analysis" (upload a screenshot of the recent price action chart) and it gives me support levels for when to invest in tranches. As someone who only knows the basics of TA this helps me fight my internal instincts to YOLO buy tops and get scared when the price drops - basically it helps you identify and buy dips. My buddy is great at TA but I was tired of bothering him constantly for advice. Here's the sample output for BTC daily chart today (o3): https://chatgpt.com/share/687a8d6d-379c-8000-a4eb-f0629bff3561
2. I have a second project "Investment Analysis"- you feed it a stock ticker and it gives you a great overview of the stock, the company, recent price action, headwinds, tailwinds, recent news, P/E, etc. Probably a lot of this is duplicated in other apps but it's very thorough and easily customizable to what you're interested in. Here's the sample output for analyzing OSCR (Oscar Health) - thinking of nibbling in if it gets <$12.5: https://chatgpt.com/share/687a8e28-5d70-8000-a368-9e988c0c7096

I use a little bit of AI every once in a while, it's pretty useful to get quick explanations and solutions to common problems.

It has also been really helpful for language learning for me, as it can do quizzes, definitions, and so much more, and correct you as you go.

When it comes to programming, it's useful for some common issues, but this is where it can fail. I wouldn't trust AI to lead a whole programming project, I've frequently see it generate broken code but the AI will confidently claim that it's all fine and will only admit the mistake if you point it out, so it's crucial to know what you are dealing with, this is why Vibe coding without knowing any programming is often a bad idea.

I use AI a lot to accelerate coding. Writing code blocks that I know how to do, but might take me 5-10 minutes of thinking, can now be completed in a few seconds with a simple tab.

I also use it for research. It's a good springboard for helping you find relevant sources and then digging deeper from there. I always make sure to double check any claims that the AI makes and any sources it cites.

I've occasionally used it for shopping advice, like "find me a product that can do this".

I don't use it for writing, at all.

It's very good for processing a bunch of numbers. Often times I need to deal with numbers, that I need to use to break into other numbers, that I translate into some kind of other ratio of numbers... That and explaining a common concept.

Mostly for questions instead of asking Google I ask AI. Some artwork change photos into different styles. I use only PPQ.ai. I used to use unleashed.chat but found ppq.ai PWA much better.

To write my objectives and reviews at work

Legal advice, rewriting emails, memes, chapter art for my third book and help writing my third book, fix broken garage door, recipes

I use it instead of Google now