Nice app! Feedback since you asked: The most obvious must-have feature IMO is to paste automatically. Don't require me to hit a shortcut (or at least make it configurable)
The next most critical thing I think is speed and in my tests it's just a little bit slower than other solutions. That matters a lot when it comes to these tools.
The third thing, more of a nice to have is controlling formatting. By this I mean - say a few sentences, then "new line" and the model interprets "new line" as formatting, not as literal text.
> where you have control over your estate via the will, maximally until some time after the death of a specific named person who was alive at the time of your death. (which I find a strange stipulation)
This is a compromise of allowing some of your wishes to be fulfilled after your death while preventing long term 'dead hand' control over property and resources. You can set up a foundation dedicated to "thing" but can't control what that foundation does until the end of time.
We probably don't want to live in a world that is controlled by the wishes of someone gone for 500 years who had a view of the world completely at odds with the present. It centralizes power and ties the hands of those living in the present.
I've run into something similar twice now in the last month. A candidate pauses or says 'let me think about that' on a relatively simple question as if to give an LLM time to respond. After the pause they give an overly long detailed answer - again like an LLM response.
One candidate was absolutely stumped and could not answer why and when they became interested in technology. They couldn't say anything about themselves personally. It was baffling.
Manhattan is a better example - there is no fast alternative route between the city and airport, and that's why you see a lot of helicopter transfers. I can see this working in a lot of US cities where public transport is poor and the airport is far away - Seattle, LA, Bay Area etc.
I keep reading this, but I haven't been able to find much data on intentional vs. accidental consumption with the exception of people who thought they were taking heroin. Do you have any sources on how many deaths can be attributed to laced or misrepresented non-opiates (e.g. marijuana, adderall, mdma)?
Some of them are, some not as much. There are certainly more obvious poetic devices like a "murder of crows" yet there are still many other collective nouns that are more "legitimately" part of English - pack, herd, pod, flock, swarm, and school for example. You don't have to know these terms as using 'group' would be understood - but they are common enough in books / documentaries.
Still growing and the main reason worldwide population continues to increase. That said it is projected to slow down over the next several generations like other areas.
This is why you see a number of hotels linking the A/C (and electricity etc) to keycards. If the keycard isn't in the slot, then the A/C isn't running.
This is something I've been wanting for a while now, and I even considered doing exactly what you're doing and trying to figure out OCR myself. The reason I wanted it is because I would like to keep track of my bank balance over time, but I am really not comfortable giving access directly (like Mint required). But I then have the same issue when it comes to using an OCR SaaS tool because you'd have data. How are you thinking about privacy in this regard?
I'd be open to testing and providing feedback if you're interested.
I believe that in many cases judges have discretion to reverse a guilty verdict returned by a jury if there are problems with evidence. They can't do this later though - once the verdict is handed down they can't revise their verdict and sentence, as far as I understand. They also cannot do the reverse and unilaterally find the accused guilty over a jury that has voted to acquit.
Nice looking site and design - kudos. I don't think there is anything wrong with the idea but you do have to compete with easier alternatives (e.g. credit cards, gofundme, craigslist). You can't just be different than the alternatives to effectively compete, you have to be better - especially for the creator who in your scenario needs money to make ends meet.
As others have mentioned your biggest challenge is the two-sided marketplace, it is unfortunately just about the hardest thing to pick for a startup. You won't get creators until there is a large enough volume of supporters for it to be worth it to them (why set up a campaign if no one responds). You won't get supporters until you have enough creators to match services they might want.
Many marketplaces like this start with at least one of two things - they either already control one side of the market via relationships (they have access to a large amount of supply, for example), or they have a lot of capital they can use to fund the concept until it gains traction (by funding creators for example). Ideally you have both.
My advice, given the time you have already invested, is to figure out how you can niche down very narrowly to start. Pick one geographic area, and perhaps one service type, and then grind it out manually to match supply to demand. Maybe you're matching creators who tutor on the side, to students in Seattle (or whatever). You want to be narrow because it's too hard to build supply and demand if you're broad, especially by yourself without funding. You'll have to try lots of hacks to see what works to get initial users - it will be manual, you will have to talk to a lot of people, it won't "scale" but that is ok to start. If it works you can grow into other markets later (example: Facebook started only for Harvard, Uber was only black limos in SF).
Also consider looking for a charitable benefactor who will fund creators (even if they don't use the services) to get a bit of energy into the marketplace. If you can find someone, or a company that wants to do this for a bit of good will and press you'll have some capital to fund the initial creators.
> Maybe, just maybe, if I set things up right, this effort will be worth it.
Worth it how? For your own curiosity or in some meaningfully productive way for your future self?
> Maybe in the future I will want to know exactly what the thought process that made me choose one font over another was
Maybe, but if it's trivial it's trivial so you don't need to remember it. If it was something you spent a lot of time deciding you probably are going to remember it anyway - if it was important. This kind of historical decision documentation becomes more useful when you have teams and want to onboard new people (and even then it has a limit)
> Do you have any suggestions? How do you stop yourself from overthinking and over-engineering task/project management?
For single-person projects you can go very simple. Most people are worried they're going to forget important tasks, but important tasks have a way of coming up again and again. I think storing list of items in a simple todo app or on paper is about all you need - this list can be as long as you want, but the longer it is the more diminishing the returns. Choose some small set of priorities you will do today - maybe 4-6 items. Do those then move down the list. Feel free to re-prioritize frequently.
> How do you determine what is worth specifying/documenting/archiving, and what should be ephemeral and only exist for the sake of facilitating taking the next step?
Specs are most useful in making you write, and thus helping you think. If you can't write it down you probably don't understand it. Small tasks don't need a lot of writing because you understand what you need and why. Specs are worth writing for the big items if you need to force yourself to think deeply. They're not important for historical purposes - once the feature, project, whatever is done, the spec is no longer needed. And in fact it's not very useful to review again as it's always better to review the thing you actually built instead of the spec (which in most cases will differ). Documentation (how to use the thing) can be useful if you want your future self to remember or others to use it - so document APIs, Libraries, things like that. Don't spend time documenting self-explanatory functionality.
Just FYI - Pihole won't block video ads on Youtube on your TV because it is limited to DNS blocking. Youtube is structured such that blocking the necessary domains would block all of Youtube.
The next most critical thing I think is speed and in my tests it's just a little bit slower than other solutions. That matters a lot when it comes to these tools.
The third thing, more of a nice to have is controlling formatting. By this I mean - say a few sentences, then "new line" and the model interprets "new line" as formatting, not as literal text.