What’s your daily driver general purpose LLM, i.e. non-coding use cases?

For me ChatGPT o3. 

A bit of Gemini (PDFs), Claude (writing suggestions), Grok (contextual).

Crypto and Blockchain enthusiasts with no friend IRL to talk crypto with so here I am

Husband | Cat Dad | Civil Engineer | /plants - host | $Degen HODLer

What’s the benefit of project folders when search exists?

I'm not inputing historical data, just the current status and I ask it to suggest exercises or training plan. my current prompt is this:

When answering, please consider:
- I'm a runner (60 km per week)
- I'm 178 cm tall and weigh 65 kg
- I'm 47 years old
- My goal for this year is to increase my upper body muscle mass
- I work out at home in a home gym and have the following equipment available:
- Adjustable dumbbells
- 5 kg and 15 kg barebells
- 10 kg barebell curls
- Adjustable bench
- Squat rack + bench rack
- Various weight plates (0.5, 1, 1.5, 2, 2.5, 5, 10, 15)
- Resistance bands

How has your experience been with adding data like body measurements, exercises and such??

My typical method is to tell chatGPT to make a table to log it, but it tends to hallucinate or just f’up the table at some point lol

I have a folder for strength training. I added as a context to all prompts my running routine and body measurement so all the answers are already tailored on me. 

Folders are a killer feature if you have many question about a topic and all need the same context.

Sooo I make a project folder for each of my interest 

@0xdesigner gave me the idea 

I setup custom instructions for each folder so it know my general inputs (i.e. token chart sent in crypto folder) and my desired output (i.e. give me latest news on that token or log that chart as my entry piece for a trade)

Just time saving when it knows what to except and what I expect

Been seeing you vibing with MCPs… I will add it to the to do list to improve my flow

do it in claude desktop and connect mcps!

Better memory for all chats in the folder + document upload