I think the answer to this question, and broadly to AI in general, is to focus on building yourself.
By that I mean focus less on creating some external thing, and focus more on “sculpting” your self.
That can mean anything from learning how to repair cars, how to draw, a foreign language, getting in amazing physical shape, etc.
The point is that as people rely more and more on some external tools to create things, you’ll be less dependent on them. Such a person will stand out more and more as time goes by.
Someone that speaks five languages and can draw well is immensely more impressive and capable that someone using Google Translate and an AI image generator.
Some day, you'll get old. Then you'll discover that real needs of older folks (and you can throw in kids, and some other groups) are almost entirely unmet.
So, DJT gets an exoskeleton, intensive medical care, a chauffeur, prepared meals (not just Big Macs and KFC) and millions struggle first, to pay the rent and utilities, but also struggle with 2-point type food labels, walking around, lifting heavy things, driving long distances for medical care (because a conglomerate bought and destroyed the local medical center) and other goods, repairs, travel and even simple nearby fishing and photo excursions. I've been a techie all my life, and know what's going on, but many seniors are phioshed, and scammed mercilessly and don't know the score. They end up with annuities, with astronomical overhead and fees (and very generous sales commissions!) or end up becoming Amway franchises, or buying time-shares. Information is scattered all over the web (ask me how I know) and generally confusing, contradictory and hard to find. Search is COMPLETELY broken. I use DDG and get brutally useless results most of the time. Pity the non-techie. LLM'S instead? They have led to suicides with their obsequious answers drawn from every byte in the universe rather than a bottom up and local-first data store. Social media? Advice from trusted friends? Don't make me laugh. It's bot-ridden misinformation galore.
So, if you want a compelling project, know that search is broken, LLM's are dangerous and designed not to be objective, and that accurate and useful information is harder than ever to find.
The issue is that I just turn to ChatGPT, Muse, or some other AI app and it just does it for me.
Most of my issues, such as keeping a running task list, managing my groceries list or pantry inventory, and even fun/leisure things are handled by these apps.
Assuming you are thinking to build something that would make money - whatever you build someone will clone it with AI shortly thereafter. without a serious moat (which is insanely hard to get) I don’t think there is much to build anymore. “Good” thing is that nowadays you can build things fast and fail/succeed fast unlike “before” where building something took serious time
Hardware and specialized Android appliances (Android TV / Google TV / Digital Signage).
A lot of focus is currently on browser-based LLM wrappers, but there is huge untouched ground in headless device management, local orchestration, and edge utility tools for living-room/kiosk hardware.
For instance, managing TV boxes, local ADB orchestration, clean custom launcher provisioning, and lightweight local automation on constrained living-room devices without relying on bloatware or invasive accessibility hacks. Building tools that solve practical, physical-device problems will always have solid utility.
Hardware for an open source Android TV distribution like LineageOS would be a good idea. Needs to be modern, hackable, and powerful enough for smooth 4K playback, light gaming, etc.
I think the answer to this question, and broadly to AI in general, is to focus on building yourself.
By that I mean focus less on creating some external thing, and focus more on “sculpting” your self.
That can mean anything from learning how to repair cars, how to draw, a foreign language, getting in amazing physical shape, etc.
The point is that as people rely more and more on some external tools to create things, you’ll be less dependent on them. Such a person will stand out more and more as time goes by.
Someone that speaks five languages and can draw well is immensely more impressive and capable that someone using Google Translate and an AI image generator.
Your humanity.
Make this but open source: https://getcoldturkey.com/
What's the most ambitious thing you figure you could build?
We need to know what you're capable of.
I enjoy building real time multiplayer systems.
Some day, you'll get old. Then you'll discover that real needs of older folks (and you can throw in kids, and some other groups) are almost entirely unmet.
So, DJT gets an exoskeleton, intensive medical care, a chauffeur, prepared meals (not just Big Macs and KFC) and millions struggle first, to pay the rent and utilities, but also struggle with 2-point type food labels, walking around, lifting heavy things, driving long distances for medical care (because a conglomerate bought and destroyed the local medical center) and other goods, repairs, travel and even simple nearby fishing and photo excursions. I've been a techie all my life, and know what's going on, but many seniors are phioshed, and scammed mercilessly and don't know the score. They end up with annuities, with astronomical overhead and fees (and very generous sales commissions!) or end up becoming Amway franchises, or buying time-shares. Information is scattered all over the web (ask me how I know) and generally confusing, contradictory and hard to find. Search is COMPLETELY broken. I use DDG and get brutally useless results most of the time. Pity the non-techie. LLM'S instead? They have led to suicides with their obsequious answers drawn from every byte in the universe rather than a bottom up and local-first data store. Social media? Advice from trusted friends? Don't make me laugh. It's bot-ridden misinformation galore.
So, if you want a compelling project, know that search is broken, LLM's are dangerous and designed not to be objective, and that accurate and useful information is harder than ever to find.
Deshittify the universe, but one bit at a time.
in your own life, ask yourself, what do I need? what could make my life, or the lives of those I love, better?
The issue is that I just turn to ChatGPT, Muse, or some other AI app and it just does it for me.
Most of my issues, such as keeping a running task list, managing my groceries list or pantry inventory, and even fun/leisure things are handled by these apps.
Look deeper. Or more niche/esoteric
If you ask the machine you will get the median answer and do what everyone else is doing.
Assuming you are thinking to build something that would make money - whatever you build someone will clone it with AI shortly thereafter. without a serious moat (which is insanely hard to get) I don’t think there is much to build anymore. “Good” thing is that nowadays you can build things fast and fail/succeed fast unlike “before” where building something took serious time
Hardware and specialized Android appliances (Android TV / Google TV / Digital Signage).
A lot of focus is currently on browser-based LLM wrappers, but there is huge untouched ground in headless device management, local orchestration, and edge utility tools for living-room/kiosk hardware.
For instance, managing TV boxes, local ADB orchestration, clean custom launcher provisioning, and lightweight local automation on constrained living-room devices without relying on bloatware or invasive accessibility hacks. Building tools that solve practical, physical-device problems will always have solid utility.
Hardware for an open source Android TV distribution like LineageOS would be a good idea. Needs to be modern, hackable, and powerful enough for smooth 4K playback, light gaming, etc.