Here's something I think ought to be more or less straightforward but don't yet know any of the relevant machine learning theory for and so I've been putting it off. I could find a variety of handy uses for it if implemented the way I'm imagining. If it exists already, I don't know how to search for it.
It should ideally be a simple/minimal-dependencies CLI program that can take a command and arguments for that command, go ahead and run the command with given arguments, and record those arguments in relation to each other. When called with any of the command arguments missing, it will run the command filling in the blanks with a prediction for the best values for the missing arguments, based on the historical usage patterns.
My example use case is that I currently manually run a terminal command to adjust my external monitor brightness to a comfortable level at various times of day, and I would love to have a simple cron task scheduled to have it auto adjust throughout the day based on time of day and possibly other variables I can retrieve such as ambient brightness or whatever else. The dumb version would just be a predefined lighting schedule hardcoded in cron, but I specifically want to continue making adjustments as needed based on comfort and have the automation trend towards optimum over time, minimizing the frequency that I need to adjust without replacing my option to do so temporarily. The ML part I don't have the education for yet and I've been putting it off, but I can imagine a lot of other use cases for something like this implemented in a very generic way. The way I first imagined this working, I'm not entirely sure any sophisticated algorithm is even necessary, it might be doable with some high school level math run against the stored data. I presume that with such a generic approach it won't perform impressively well right out of the gate but with an accumulation of training data can be coerced into remembering what the user wants/expects gradually. As I said, I'm not educated in machine learning but I suspect this should be straightforward to do at least a crude version of given some of the libraries available.
It would make sense for me to be able to preheat my oven or stovetop remotely. Going all the way into the kitchen and back just to flip a switch and to come back and wait for 5 minutes seems insane these days. Things that could be dangerous like these would need additional sensors and safeguards against forgetfulness or deliberate exploitation, of course, but with this topic we're talking about a next-gen stack of technology instead of a Ben Franklin coil in a box with a knob. I believe the reason these things have not already been tackled (at least not for the mass-market) is that the production costs are disproportionately high compared with consumer demand for those small conveniences, considering most of these types of devices already do not need frequent replacement or upgrade, and are prohibitively expensive for a lot of households in the first place.
My laundry machines and dishwasher are relatively smart appliances, but again, it's needlessly wasteful to have to remember to walk all the way across my house to check on their progress instead of just receiving a notification when they are done, and being able to view their usage and performance stats on a dashboard somewhere to make better decisions about how and when to use them. It'd be nice to be able to load up each machine and let it decide when to begin its cycles based on how full it is and with delayed timing to optimize for the lowest cost electricity.
A refrigerator that can track the weight and volume of its contents and help maintain a shopping list is something we've talked about having for decades, but none has ever been successful yet. I'm not suggesting this is easy functionality to solve for, but perhaps thinking about it in different ways would help. What if each compartment were operated independently and could do more than just maintain a cool temp? Maybe a box that controls temp all the way from freezing to roasting as well as humidity. It could take your dishes prepped for cooking the night before and go ahead and heat them up to be ready to eat at the appropriate time, with no second manual process needed. It could double as a dehydrator, and maybe even keep track of when things will expire and pre-emptively dry them to avoid waste. If this were a stackable thing big enough to fit a big turkey or a pizza inside, I could imagine having several of them stacked or spread out instead of the traditional fridge/freezer combo box that can only be installed in one place in my house, and is hard to access under/behind for maintenance.
Speaking of maintenance, if we're already adding some small general-purpose computer to each of these things, why not have them all self-contain operation and maintenance manuals, and politely remind when they are due for some cleaning or repairs? Most of these devices don't need to have a built in display, but instead should be managed via a web page or connected software on a phone, tablet, or other PC. I think it'd be pretty nice to open a general-purpose smartphone app and within a few taps get a step-by-step illustrated guide on how to properly replace some part or do routine cleaning on any of my dozens of household devices. The next step beyond that is to even be able to order the replacement parts or 3D-print them locally from the same app.
What I'm especially interested in at the moment is aquaponic/aeroponic household gardening/farming and how to assemble a reasonably small and self-contained low-to-zero maintenance system that just tells me when to add food or chemicals and when individual plants/fish are ideal for harvest. I've been playing around with a Back to the Roots Aqua Farm, which is just a simple Betta fish tank with an integrated herb garden system on top, where the fish feeds the plants and the plants keep the water clean. This is really just a lower-hassle way of keeping a small fish, but the same principle works for larger setups with edible fish and plants and there are a lot of "everything you need to get started" kits out there but I have yet to see a true "open box, add water, insert fish and seeds" package other than this one, and of course it's extremely dumb as far as self-maintenance goes, the only real benefit being that you have a fish tank that doesn't need to be manually cleaned as often. A more sophisticated setup would involve timed pumping, grow lights, and autonomously administered food and maintenance chemicals, where really all you have to do is harvest and replace the plants and fish, and keep re-filling the stockpiles of whatever food and other additives are needed to keep everything in optimal health. I'd be really interested in buying such a setup in some type of one-box form-factor, and I think that such a thing may really be the future of sustainable agriculture. Something where I install the thing and simply extract food from it, and the pool guy can handle the periodic cleaning and re-supply that it requires.
Ubuntu has gotten me started on Linux thanks to the simple/low-hassle install process, and enough polish to be a bearable alternative to OSX or Windows. Now that I'm getting the hang of things, I'm beginning to prefer it in spite of the issues. I also like distros using Gnome 3+ but Ubuntu is ubiquitous enough that I defaulted to it.
It should ideally be a simple/minimal-dependencies CLI program that can take a command and arguments for that command, go ahead and run the command with given arguments, and record those arguments in relation to each other. When called with any of the command arguments missing, it will run the command filling in the blanks with a prediction for the best values for the missing arguments, based on the historical usage patterns.
My example use case is that I currently manually run a terminal command to adjust my external monitor brightness to a comfortable level at various times of day, and I would love to have a simple cron task scheduled to have it auto adjust throughout the day based on time of day and possibly other variables I can retrieve such as ambient brightness or whatever else. The dumb version would just be a predefined lighting schedule hardcoded in cron, but I specifically want to continue making adjustments as needed based on comfort and have the automation trend towards optimum over time, minimizing the frequency that I need to adjust without replacing my option to do so temporarily. The ML part I don't have the education for yet and I've been putting it off, but I can imagine a lot of other use cases for something like this implemented in a very generic way. The way I first imagined this working, I'm not entirely sure any sophisticated algorithm is even necessary, it might be doable with some high school level math run against the stored data. I presume that with such a generic approach it won't perform impressively well right out of the gate but with an accumulation of training data can be coerced into remembering what the user wants/expects gradually. As I said, I'm not educated in machine learning but I suspect this should be straightforward to do at least a crude version of given some of the libraries available.