My entire product of doing in-image advertising[1] was created in order to build a new concept website for my uncle's nautical antique store[2]. He has a store built out of an old lighthouse in Old Town San Diego, and it's basically a museum (where you can buy the stuff inside it). When he came to me, he asked me if I could build a website that has the feel of his museum-like store, so I said, "Why don't we just take pictures inside of the store itself, and let your customers browse the pictures by hovering their mouse over the products and getting more info on the products they're interested in!".
The punchline: I spent months building a pretty amazing site, to have him say that his audience was probably too old to understand how to use the site I built. So far, that technology has led to two granted patents and my startup![3]
I've posted it here several times before, but I made this stream-of-consciousness journalling tool for myself: http://write.itskrish.co
The differentiator is that you choose a fixed duration, and if you stop typing within that duration, all your writing is erased. Such a constraint has been effective at getting me to pour out everything on my mind, without hesitation.
I only check my email 3x a day. But, I had such a hard time maintaining the will power to NOT check email, I had to build a tool to help. The service holds my email outside of my inbox and only delivers it at the three times a day I specify.
It has helped me focus on my real to do list, and not on email.
I am also working on an AI/ML bot that plays Out of the Park Baseball -- a baseball sim where you play as the team manager. The AI manages the roster and lineups and makes the big in game decisions. It does not handle pitches or batting (that is not part of the game).
When searching for an apartment in Switzerland a few years ago I found no good listing aggregator. So I built https://www.immomapper.ch/map which crawls the largest providers and currently shows over 100'000 listings on a big map with filtering, email alerts etc
Unwaffle (https://unwaffle.com) started as an admin feature of S3stat. I use it on a daily basis, and it has actually added enough revenue that I'm no longer worried that it would have been a waste of time to build it, should it turn out not to be commercially viable.
Like I said, it started internal, but it was so apparent after a few days that it could be useful for other businesses that I separated it out with an API and stuck it at a different URL, in case it turned out (which it did) that I wanted to polish it into a product of its own.
Public transit schedule checker that took into account wait times for connecting links from light rail (Silicon Valley) to CalTrain. So it would tell me stuff like:
- leave in 3 minutes for 32 minute wait connecting to bullet train
- leave in 3 minutes for 10 minute wait connecting to slow train
- leave in 33 minutes for 2 minute wait connecting to bullet train
- leave in 10 minutes for 3 minute wait connecting to slow train
Then I could see clearly my options in the way I preferred. I cared less about what time I would get home, and more about how much time I would spend waiting and sitting on a train, so this let me optimize for that.
I wanted to be able to quickly & easily resize images to particular sizes, so I made Bulk Resize Photos (https://bulkresizephotos.com); image resizing in your web browser without uploading the images, and without downloading/installing anything. Works best in Chrome.
I've built/building Orchestra (http://orchestrahq.com) to help me build better professional relationships. It helps me keep track of details of my interactions and also reminds me when I've not spoken to someone for a while.
I used to spend a bunch of time bouncing between HN and Reddit looking at headlines. So I built myself a tiny aggregator with React and socket.io that auto refreshes so I can glance at it throughout the day.
Some of these are just hacked-together tools, but I can think of several:
Turn on/off Dropcam inside home when security system is armed/disarmed.
Quick way to launch baby announcement website from within delivery room, depending on sex/name of my child, plus birth stats – my wife was thrilled ;)
My best "personal" product that I have also made available for others is http://footballpickempool.net. There are now great mobile apps for a pool like this, which are preferable for scoring and season-long rankings, but some still need paper. I was putting these together for our office each week, so I automated it.
Many webapps used internally for my job, so not personal, but a few that are hosted on their own domains and see public traffic.
Having to work a lot across timezones, I created a web app to track time and weather around the world and quickly find a good time to meet across multiple zones:
We had a large number of cron jobs at my regular employer that ran across multiple servers and applications. Things started to get a bit unwieldy and insight into any issues was negligible, so I built a cron job monitoring tool for internal use. Eventually I figured that if we found it useful maybe others would too. So I looked around and realized there were a few others already in existence that seemed to be making money. Mine offered some features the others didn't so I figured I'd go for it. It's been a lot of fun.
I wanted to better discover nearby architecture when walking around the city (Chicago) on weekends, so I made https://chicagoarchitecturedata.com/
Over 13,000 Chicago buildings with architecture style, architect (sometimes), year built, and other data points. Uses GeoDjango and browser location api to show stuff around you when you use on a phone or you computer. Also serves as a guide to learning about architecture styles and neighborhoods.
I write tiny programs in the Unix tradition: (mostly) stream based primitives that fit together with other tools, like format conversion (json-yaml) or command repeater rept(1), which is great together with lam(1).
I also enjoy writing the man pages, porting to very old compilers, creating packages, etc. It's a refreshing distraction from working with enterprise CMSes.
Ctrl-F Video (https://video.loud.red/) to look for keywords in YouTube videos. Needed it to quickly jump to relevant places in videos. Works well, even with YouTube's auto-generated CC, which surprised me. For some reason it is really popular in China.
I made an app that notifies family when a QR code is scanned. It also provides information for ems providers and allows for several levels of interaction between the two. It works well for the few people that use it. https://emsinfo.me.
I want to start podcasting so I built myself a podcast CMS/RSS generator https://github.com/gmemstr/whiterabbit it's my first go at Golang so I'm still learning slowly.
I made Cachet (https://cachethq.io) as I didn't want to pay for alternatives. In the end, the system I built for this to sit behind didn't get any love.
I created https://www.HireLoop.io because I wanted something to manage my hiring which _always_ closes the loop with my applicants regardless of them being good or bad.
Im building a personal notes classifier, a website where I can write down links, bookmarks, ideas, thoughts, blog posts, projects, articles, etc. and it classifies them based on the text. It also knows whether something should be public or private - so that just I can see it. I want to get it to the point where I use it as part of my daily workflow as a way to organize my thoughts. My hope is that by process of needing/using it, I will understand well what is needed to make it better and the feedback cycle will make it a valuable tool quickly, as opposed to me making something that is intended to be used by someone other than me.
hey, I am really interested in learning more. I am working on something very similar. I have already developed some apps in this direction but I am going for a new iteration. I am just curious about your approach. The "it classifies" means the app does it automatically? How do you go about that? do you have something online?
Im still learning/building. I will link you when I have something more substantial to show. For the first iteration, I am making it with d3.js - something like this: https://bl.ocks.org/mbostock/3127661b6f13f9316be745e77fdfb08...
where each note will be a point. The placement of the notes on screen will be generated by the classifier model. So all bookmarks should be in one cluster, all hobby aspirations in another cluster. I may even use k-means to cluster them.
Then when a classification is incorrect, the user can drag the item to another location so that it is where it should be which will trigger a callback to re-train the classifier model so that it can continually improve.
nice! So, you are going more for the "mind map" approach right? I am a bit confused now. So links, notes, blog posts, and other 'items' you named on your first answer would be nodes here? I see how it would be very easy to find the category you are looking for, but how would that visualization help at the time to find a specific node? When you said "blog posts", "articles", "projects" where you mostly meaning a way to link to them, or to embed them into the app itself? In other words, is this a bookmarking tool or also a publishing/blog tool?
My idea is quite in a different direction. My best solution so far is http://multikanban.com . It really is just a personally customized kanban board. But it lets me navigate through multiple boards, each board is a project/goal/collection of tasks (maybe like your 'categories'?). Inside each board I store several things, notes, bookmarks, etc. And due to it being a kanban it helps me in the progress of getting it done too. Getting it done might mean several things, for example I have a Kanban board for books I want to read, so getting something done there means I just read a book. It serves both the purpose to see what I want to do, and what I've done, ad it helps me prioritise it too. However, it has lots of limitations.
Even though I'm quite happy with that solution, I want to complement it with something more along the lines of blog/publishing. I do have a blog, but I barely write there. I want to find something in between the small notes and the long blog posts. Which serves the purpose of storing thoughts, links, and categorizing it with tags, but without each entry needing to be a post, as we are used to in common blogs. I'm still conceiving it...if that wasn't obvious yet haha
Yep, every item would be a node. Similar items like bookmarks would be in 1 cluster while blog posts would be in another.
As far as publishing, its mainly for my own personal use, for me to organize my thoughts, bookmarks, aspirations, etc. Im not sure how to integrate that part (what a visitor might see)
I started on this road because I have a lot of notes in a variety of apps/files on different devices and every three months or so, I write them all down on a big whiteboard to see any common threads/themes, and to try and reorganize/prioritize my life. Its very helpful but every time I do it I have re-erase the whiteboard. This is a more permanent tool to replace that need.
However, finding a specific node might be tricky. Maybe something like a search bar up top with a zoom to node function (like google maps). Or a sort by feature? I haven't thought that out yet.
Ill try out your multikanban site and get some feedback to you. My initial impression is that setting up a kanban board is work - work that could be spent doing other things. To use it in the first place is a time investment. I wonder if it could provide intermediary benefits before the long term benefit of getting your work done more efficiently? So for instance, if you were using it to blog somehow, and a kanban board was created just by happenstance, the user could kill two birds with 1 stone! Anyways, Ill give it a go and let you know what I think.
as opposed to what? An app? A website is accessible from the largest number of places - my phone, any computer, etc.
Also, 1 intention of this is to streamline my browser use. So that instead of just opening up a new tab and getting distracted with social media/article aggregators, I can organize my time online. I open a new tab to this website and see my todos, things Im interested in, projects Im working on, etc. And thus I have a starting point to go from.
Im also considering adding Kanban board type functionality to it so that I can prioritize my personal tasks/notes/todos and then move them to in progress and done. This needs a lot of screen real estate.
A daily todo that results in long-term stats and habit formation.
Using my own 'product' every day has been a good exercise in thinking about UX and maintenance (I've had the same DB for several years now with little breaking changes). It's also a responsible-way to do resume-driven development and use something weird or unconventional just because.
Aaaah, I thought your username was familiar! I was talking to 'hsribei the other day and opined that I'd never thought of putting it up as a SaaS product because I didn't think anyone would buy it.
I'm glad someone is putting it out there because the concept has been very helpful to me. Hope it continues to do well.
Well, not done, but I am working on basically an information hub for myself. Unfortunately I am not a software engineer, so I am slowly learning python to get it set up.
Basically it will pull a variety of information and display it on a simple website. Things like notifications for social media, my to do list, sports scores/times, weather, etc.
It's a larger task than I realized. I'll probably never finish it tbh haha.
I made a python module / cli utility to collect all the log events that would usually be emailed to me throughout the day, aggregate them, and send digests on schedules based on their importance: https://github.com/asteriske/translogrify
I made https://sysplatter.com. I don't know how much other people's use cases it accomplishes but for me, I like a lot of customization of my statistics so it helps with my other stuff.
Not really used that much by other people; didn't click at all with anyone on HN, sadly.
Deduplicated File Storage CLI with compression and encryption. Still a little buggy, but I needed some way to slim down my data archives while still being able to access them rather quickly.
The punchline: I spent months building a pretty amazing site, to have him say that his audience was probably too old to understand how to use the site I built. So far, that technology has led to two granted patents and my startup![3]
[1] https://www.youtube.com/watch?v=tf4pE0xtYTo
[2] http://westsea.com
[3] http://pleenq.com