I have "new genius" ideas very often. After doing quick search I discover that any idea worth thinking of implementation is either implemented already or what seems to be low barrier to entry clashes with some legal obstacles.
I have the opposite problem. I have a genius idea, and I start to research it.
I find a company that actually built a solid product, dangit this is really good. They appear to have executed well, but they failed, or went nowhere, heck the app is still out there. Maybe they are even chugging along but its a smaller business even with a better product than I would have been able to build. Had I been a founder of the product, I would be questioning staying.
Then I also find sometimes I was doing it all wrong and the world has moved past my notions of products. I think there's a market opportunity because I don't realize that the rest of the world is already cool using a $15 plant hygrometer bluetooth device which can also keep track of your medicine or food in your cooler, my notion of the value of something is skewed by western costs
Interestingly that sort of research is actually what I've used Claude/Chatgpt deep research and openclaw for. If I have an idea, I get an agent to go and do some product research for me and see if there is a market, if anyone has tried it, and if there is anyone doing it.
It has unironically saved me a lot of time I would have otherwise spent going down rabbit holes.
Of the models I've found that claude doesn't gas you up as much as GPT, so for stuff like this where the answer can be "no, that's not a good idea" I usually use claude.
Yup. I do have a 4-step process for this (just for prompts and some bash scripts that call CC). 1. Breadth first 2. Compress 3. Per player deep research 4. Per player compression. Then I just merge all the markdown files, fit it into 250k tokens, load any model that supports that much and you can pretty much "tall to market".
The biggest limitation here is data access though. A lot of market data is gated behind registration or anti-bot captchas, so the project that my CC is working on now is a playwright clone that is not easily detectable + can be used with CLI same as playwright itself.
I find ChatGPT so infuriating the way it always agrees with everything you say. The product is optimised for engagement so it wants its users to be delighted
Jim Rohn said one time “just pick a direction and go, you will find out sooner”, if its good or bad.
That was adjusted for 80s. In todays world you can know whether something is worth pursuing in minutes. Tip - in 99.9% of cases its not, but you will still learn along the way. Maybe you find something new.
Hmm I often have ideas that I don't see anywhere else but I'm just in it for rent curiosity and learning. I absolutely hate the business side and usually I do stuff for free just so I don't have to complicate my taxes.
I find a company that actually built a solid product, dangit this is really good. They appear to have executed well, but they failed, or went nowhere, heck the app is still out there. Maybe they are even chugging along but its a smaller business even with a better product than I would have been able to build. Had I been a founder of the product, I would be questioning staying.
Then I also find sometimes I was doing it all wrong and the world has moved past my notions of products. I think there's a market opportunity because I don't realize that the rest of the world is already cool using a $15 plant hygrometer bluetooth device which can also keep track of your medicine or food in your cooler, my notion of the value of something is skewed by western costs