Finding a high-signal, low-noise source of problems to think about
Darko SavicAug 16, 2021
While trying to come up with one new idea per day for a year I realize that finding interesting problems to think about is the most difficult part.
Going for original ideas I often end up thinking of something that has already been done. I then try to come up with significant improvements (innovation) of already existing solutions. If there's nothing new I can contribute the only remaining option is to switch fields and go for a different problem.
Serendipitous discovery of interesting problems
How would one find a steady stream of diverse, interesting problems? I have a few, but they are mostly high noise-to-signal ratio. It takes time to run into an odd good problem in the sea of repeating ones. The effort that gets put in the search and extraction seems to eat away at the mental energy that would be better used for problem-solving. I often get tired and have to take a break without finding a worthy problem.
High noise-to-signal sources
There is a thin line between the benefits of content consumption and wasting time. Especially when the content source was not designed to let you go once you've had enough.
Pretty much any online platform is designed to capture you in the attention vortex and keep you consuming/creating content there. It doesn't take long before you run into random interesting material that has nothing to do with your goals. Chasing interesting tangents exhausts your mental resources and prevents you from working on your ideas.
Ideally, there would be an ever-updating list of well-defined problems, summarized into one sentence each. You would browse them like a catalog until you find something that you want to work on. From there, the interesting problems would be loaded into such a though process:
The best problem-finding tool I have found to date is the list view of brainstorming session on this platform:
Darko SavicAug 16, 2021
On Quora, you can often tell what problem someone is having just by looking at their question.
The downsides of using Quora as a source of problems
Quora recommendation algorithm serves you questions based on what it "thinks" is interesting to you. This means the questions tend to be repeating and revolving around the same areas of interest you previously reacted to.
Just like any social media platform, Quora is designed to capture you in the attention vortex and keep you consuming/creating content there. Soon enough you run into random content you want to consume that otherwise has nothing to do with what you set out to accomplish. This exhausts your mental resources and prevents you from working on your ideas.
I have yet to find a way to filter Quora questions in a way that bypasses the above two points.
Using Twiter's advanced search function to extract worthy problems
Darko SavicAug 16, 2021
Twitter has a pretty advanced search function. I have yet to come up with a good filter that outputs tweets that are diverse in nature but contain an identifiable problem.
There are some ideator accounts that are known for posting original ideas. But even those often stray into entertainment, motivation, and agenda promotion. Examples:
who else comes to mind?
Let's try to come up with some Twitter search filters that output diverse problem/idea containing tweets. Here are a few copy/paste strings that return mediocre results (still high noise to signal ratio) but some useful stuff in there:
"solve the problem of" (would OR could OR should OR we OR how OR help OR let's) min_faves:3
"solution to" min_faves:3
this idea (great OR brilliant OR love OR wow OR genius) min_faves:3
this solution (great OR brilliant OR love OR wow OR genius) min_faves:3
"how can we solve" min_faves:3
"is there a way to" min_faves:3
"a challenge to" min_faves:3
List of ideas on this platform
Darko SavicAug 16, 2021
Every idea solves a problem. It takes little effort to identify a problem in a well-defined idea. So browsing through ideas on this platform is a fairly good source of worthy problems. You can then come up with different ways of solving them.
People feeding their problems in through a dedicated platform
Povilas SAug 16, 2021
Differently from quora, this could be a platform dedicated more to inspire ideators rather than help people solve their problems. I mean it would be both, but the ideators would get a constant feed of problems that are truly relevant to someone on the "other side" of the platform.
The people who feed in their problems would have a chance for their problem to get picked up by some top ideator and solved for free. For them it might be worth trying - simply describing your problem (anonymously if preferred) doesn't take much effort. You could feed in any type of problem, from the most stupid and mundane ones to the most complex and unique ones.
AI could help filter the repeating ones and pick the unique ones for the ideators. The ideators on the platform could be divided into groups of different expertise levels - the same would be done with the problems that people feed in, the top ideators would get the most unique problems whether those starting from the very beginning would get the average, repetitive ones to warm up with. The better you got at solving the problems the more unique problems you'd see, this would motivate the ideators to raise their expertise level.
Such a platform could have a marketing motto: "Do you have a truly unique problem? Share it and get our top ideators solve it for you for free".
AI-generated lists of problems that people often write about
Darko SavicAug 17, 2021
Inspired by this tweet - why don't we tap into GPT3 AI-powered language prediction model?
Is GPT3 still getting new data from people worldwide? I imagine that any system such as Wordtune or Grammarly tap into GPT3 via their API and constantly feed it new data live, as people type it in. This would be the equivalent of having keyloggers installed on many systems worldwide and feeding everything into GPT3.
It would be amazing if we could get a constant stream of concisely described problems, summarized into one line each. Reading such lists of problems would prime ideators and get them going.
Update: this seems to work:)
Using AI21 studio's Jurasic-1 language model, I primed the AI with a few examples, then it continued to produce a list of random problems. One of the suggestions inspired this idea.
Study existing solutions and their occasional lapses
Samuel BelloSep 10, 2021
How to come up with new problems
Generating problems is easy. You just need to find a way to improve everything you come across. If you try to make things better, faster, cheaper, or clearer. I think the most abundant sources of problems are existing solutions. People who have the most difficulty in coming up with new ideas are the ones who respect existing solutions too much. Another advantage of studying existing solutions is that one's mind is familiarized with other people's problem-solving techniques.
High noise low signal sources
I believe the internet and life, in general, give information in a high noise low signal way. Sifting through the noise is a major part of logical thinking and information processing. It is more important to master thinking processes that quickly differentiate the noise from signal than to have access to a platform that performs the same function.
I also believe that moderate noise levels in the information a person takes in can help to improve one's creativity. The reason for this belief is that almost everything a person knows seems to be noise (when solving a problem) and is perceived to be irrelevant to the problem at hand until a logical connection is made. The ideator's mind should be able to comfortably fiddle with the 'noise' part of information since they frequently give hints about past and future problems.