Pill tracker with BLE connectivity. This could be a very simple stick on device with motion tracking chip. You attach it to a pill bottle and pair it with a phone or local base station. The idea is to remind elderly people to take their prescriptions on time, and alert their relatives if something is wrong. First Steps: look to see if this is on the market already in this form, and if there are any parents out. Make a prototype.
Motion activated necklace with BLE. For similar reasons, a necklace like device that uses motion tracking and proximity tracking to monitor elderly people with cognitive diseases. The BLE device could alert their relative if the person wearing the necklace leaves their house, stops moving for a prolonged period of time, or falls down. First Steps: check what's currently available on the market, if nothing, make a prototype.
Robotic submarine with no tether and AI vision system. This would be ideal for salvage operations, particularly for the search portion of the operation. The device I am picturing would be a small battery powered autonomous submarine, about 2 feet in length, with cameras on board, magnetometer, infrared sensors, sonar, and a powerful processor. The subs would be designed to dive all the way to the bottom of the ocean autonomously, do a grid pattern search, using cameras and sonar for location data, than rise to the top and record and broadcast their GPS location. The idea is that the boat would come around, retrieve the sub, download the video, and use automated video review combined with people reviewing the video to find any interesting art effects that might be hidden beneath. Idea is to drop a several of these and then come back in 2 hours to collect them. First Steps: see what's already on the market, make a better one as a prototype, see if anyone want's to buy more.
Drone navigation that uses vision only, no GPS. The other day there was a company that showed off a system that can bring down a drone by emitting a highly directional RF beam that jams GPS signal and standard WiFi frequencies. This causes drones like the DJI to go into failure mode and land. The idea is to start a research company to research and sell technology solutions for drones that would allow them to fly strictly line of sight using the camera feed alone. This is pretty straight forward, technologically speaking. First Steps: build a simple drone prototype using off the shelf components. Get notoriety by claiming to have a tamper resistant drone, sell company to DJI :)
Online SaaS software for moderating video debates. Video debates between all kinds of people are popular right now, especially as we are becoming more and more divided as a nation. Lean into that dynamic and make a SaaS platform that allows people to broadcast either live or prerecorded debates with automated and human moderators. The idea would be to have the participants in the debate be in different locations, each has a window on the screen where you can see them. The moderator can silence either of the participants, there is a hard clock on how long they can talk, comments viewers make during the debate can appear as part of the video, etc. I got the idea while watching the technical difficulty Potholer54 had while trying to set up his debate with Crowder. First Steps: see what solutions are available, and if there are some, why are people not using them.
Sorting file manager for quickly saving files in the right directory. This is a simple idea that I might write in Python and QT one of these days, just to practice. The idea is you save a file from any program using the normal windows or mac "save as" dialog inside a special in-box folder. Afterwards, you run the sorting program that shows you the files in the inbox folder one at a time (some files can be shown with preview, others would just show the file name), and you can use the keyboard to type a few letters of the name of the folder where the file should end up and the program will try to autocomplete the path for you. The file will be moved to the desired folder when you hit enter. Simple. First Steps: pay around with QT and build this as a project to get to know Python and QT, or maybe someone already made this, do some searching.
Kreepy idea: E-Reader like the kindle with eye tracking IR sensor for schools. Schools can now be "big brother." The idea is you give children these e-readers, and assign chapters to read at home. The e-reader will know the chapters you are supposed to read, and would have algorithms built in to make sure you are actually reading the pages by tracking how fast you are flipping through the pages, and if you are looking at the e-reader when doing so (IR eye sensor). Next day, in class, if you haven't been reading, it would rat you out to the teacher, and he/she will know who to pick on. This would be easy to defeat with a bit of coding and robotics, but let's be honest if you can do that, you probably wouldn't learn much from the assigned reading as is. First Steps: check with a few school teachers I know about what they think about this idea. Make an Android prototype. Figure out how to live with my self when this becomes popular and brings misery to so many school children (millions of dollars I would make from this should help ;) ).
Open source book initiative for standard educational books. If we are going to have creepy eye tracking e-readers, we will need books for them. Create an organization to promote an open source book initiative that would be tasked with creating open source, public domain, up to date books for K-12 on various subjects. Make them public domain so any company can print and sell them to school districts, which would drive down the price to basically just above printing cost through competition. Make them also available on digital platforms. Save schools money so they can concentrate on spending more money on kids. First Steps: contact some people who know how to get something like that rolling. I have no clue. Just being honest.
Stuffed toy that reads stories to children. While we are on a reading roll (couldn't resist the alliteration), how about an idea for a stuffed toy that reads classic books to little children with a pleasant computer generated voice (but with internations). The idea is to give your 2-5 year old kid a stuffed bear that has a WiFi connection and a speaker inside and is designed to sense when the child is playing with it, and read that child classic stories from the public domain. There are studies that show that children's performance in school, and life, can be correlated to the number of words they hear per day while growing up. So many parents are so busy nowadays, maybe a stuffed bear could do the trick. First Steps: I don't have kids, so who knows what's available right now. Do some research, build a prototype with a Raspberry Pi or Android Stick and a stuffed toy.
Electric surfboard for assistive surfing. One of the difficulties in surfing is being able to paddle fast enough with the wave to get enough momentum to popup and actually catch the wave. An electric surfboard can solve this problem by helping you paddle with the wave. It can also help you surf for longer on smaller waves, by giving you a little power boost. Basically, like an electric surfboard, but for surfing. Use turbine electric motors and sealed lithium batteries. First Steps: learn how to surf, build a prototype, have people who know how to surf test it out.