Digital Tools I Wish Existed (2019) | Hacker News Reader

Digital Tools I Wish Existed (2019)

jon.bo

75 pointstheshrike7918 comments

For the queue management bit Fabric[0] seems like a decent idea, it's kind of trying to be a personal, private, local AI assistant. You feed it the stuff you'd normally read and it summarises it for you.

Then you can decide which content you should spend more of your time on.

Not nearly complete, but an idea I had even before I found out about it. And kinda similar what Apple is attempting with iOS18 and Apple AI.

[0] https://github.com/danielmiessler/fabric

While I understand the motivation and wish for similar tools myself, I think the healthier solution is to just put less information into the top of the funnel.

But how do you know which bits of information to ignore at the top of the funnel?

There is an old joke.

A boss looks at a thick stack of résumés he had received for a job listing. It's too thick to go through it in a day. He takes a random bunch of papers from the middle of the stack, and throws the papers away. His subordinates are surprised, so he comments: "We don't need any unlucky losers anyway".

Doesn't that question apply regardless of the amount?

"goodreads clone but with relational db functionality exposed" has been on my wishlist forever too

https://thestorygraph.com isn't perfect but for discovery it's better than Goodreads in my opinion

In order for this to be even remotely doable, you need explicit and transparent data flows between devices and applications which is currently a no starter for tech companies and app developers.

Apple letting you out of the walled garden? Forget about it. Google letting anyone else use their data? How about 3 more soon-to-be-killed products instead of a proper API.

Not to mention you need a rewrite of the DMCA in order to complete this.

There's no way for official APIs outside of GDPR data request exports.

But you should be able to create custom parsers for the relevant bits.

I would like to see Evernote's AI evolve to a point where it would process my ~4500 item knowledge base (KB) of personal notes, clipping and documents, and then allow me to ask questions that would be answered based on that volume of information.

Notion's AI is decent for this, but not very good at finding things

What you're asking for is more or less already a solved problem with a number of different AI tools. Whether it will work with minimal effort on an Evernote account, I'm not sure.

I just want a tool that collects and organizes everything I've ever written online.

https://indieweb.org/POSSE

Never have your content only on someone else's site.

I usually write every comment / opinion in Obsidian if it's either long or seems like something I need to explain multiple times.

That way I can just copy-paste the same comment again, read through and edit either the original or the pasted version to fit the situation and I save a lot of time and research =)

Previous discussion: https://news.ycombinator.com/item?id=25228089

Even more previouser discussion (2019): https://news.ycombinator.com/item?id=21659876

A multi-threaded emacs ?

putting this into my archivebox to read later /s

Reply on news.ycombinator.com