i looove this. i have been working on something similar. are you using AI to help with balance forecasting? would you be willing to share how it has been working out for you? how much set up is necessary for you to set it and forget it. i am currently using monarch and only ever look at account balances and only sometimes look at transactions.
Balance forecasting is part of the app! The app has been working well for me, I have since unsubscribed from Copilot. I think the amount of setup will depend on your spending pattern. Obviously, the more transaction data, the better the forecast. Otherwise a fair amount of tweaking might be needed for it to match your expectation. The MCP makes all the projections available to your agent though, so you can certainly analyze and make those tweaks through an AI chat.
one idea i was thinking for anonymizing data when sent to models is to hash or modulo with salt some values before sending it to the model and then unhash/unmodulo on the result. in other words, a way to keep the users data their own. just an idea.
> Most of the “breathtaking” innovations of the early pioneers of the field are obvious solutions that anyone would have thought of when faced with the problems they encountered.
This is just straight up arrogance without any proof. Every breakthrough builds off another's work. Then to claim that 'AGI has been correctly predicted', I honestly don't even know what you are talking about.
Yeah, probably not the best strategy but it is a strategy. I just think this is generally how most wars in history won. Biggest army to just pummel the enemy.
isn't it more like: i have 1000 bits, i transfer 1000 bits but depending on the person, it might be lossy, so they only understand about 700. they then come up with the 300+- on their own, potentially putting them over 1000 or they come back and ask questions to fill in the blank. the bits don't ever have to be bit identical.
They're lossy. I might make an esoteric reference that no one gets, or an analogy that doesn't quite land, and then have the LLM help me come up with something a bit more understandable to a general audience. There's a difference between dumping the output that you spent 5 minutes with, and taking an hour to craft that perfect analogy.
Sure, but that seems orthogonal to what I'm saying, no? We could continue clarifying the definitions of information in this context, but my fundamental argument is LLMs don't add information to my writing that couldn't be added by the reader themself.
This of course has the potential to change with personal LLMs that can have shared private context with me. However, that isn't a defense for sending people AI slop, it just turns it from "LLMs don't add value" to "LLMs may add value when used judiciously."
How do people mentally price things. I'm on the fringe but I'm willing to pay $60 for a AAA 60 hour game that I'm positive I won't finish but am find if I get 6-10 hours out of it.
