Without knowing your background, it is difficult to suggest any specific idea. You mentioned hosting costs and 3 months build time, so I assume you want to build a hosted solution that can be built in 3 months most probably a MVP.
I will suggest you build a blog. For first 2 months read books on how to discover and develop innovative ideas, customer discovery, identifying customer needs, user flow and stories, everything a good product manager does before a product specification is written.
For first two months, Write blog posts summarizing what you learnt from each book, and implementation worksheets/mindmaps that can be used to execute what you learnt in that book. By the 3rd month, you might have good idea how you may want to build something and a bunch of mindmaps and worksheets to use for any idea that comes to your mind.
In third month, go to Show HN tab or product hunt, take the first idea that you find attractive and apply your worksheets and mindmaps to that idea. Contact the original poster and offer to collaborate and if you think you can do better alone, go for it.
...It answers the complete opposite question the author has asked
...Tells the author to not build a developer tool and still manages to be a top comment on an HN post
...Currently remains the top comment on a post entitled "I'll Build the Top Comment"
...and still manages to provide some useful suggestions despite all the above.
https://stackingthebricks.com/developers-dont-pay-for-things...
Any suggestions for books?
Dan Olsen, The Lean Product Playbook
Mart Cagan, Inspired
Ryan Singer, Shape Up
Jake Knapp, Sprint
Jeff Patton, User Story Mapping
Stephen Wunker, Costovation
Need: people have niche needs and are willing to pay about $10 to $15 per hour for it.
Students would probably like it a lot to be able to work different interesting jobs. Examples: being a tutor, helping someone write a letter, help an old person with computer issues [1] or whatever.
Strategy: ".edu" addresses only, have templates to kickstart student's brainstorming on what tasks they could do, let them price it. Marketing to students: go to student coordinators and let them send out mass emails to students. Marketing to consumers: look for Reddit topics where it's clear that people say "I like to learn xyz" and then reply that you found a good site for it to learn that xyz skill.
Revenue model: either self-advertizing (educational courses), normal advertizing (wouldn't recommend). Do not take a cut of the students, you want to give them every incentive to not go behind your back (students normally do that by teaching in private).
Something like that, pure napkin sketch on what you could build. It's something I'd like to see.
[1] How to reach digitally illiterate people? Their grandchildren could find you, or a student posts a business phone number on there that could be called. You could create a routing system where they get a business phone number that will be connected to their personal cellphone when needed via some app or sth.
I wasn't able to find any easily available solution that could do this.
Everything on Google now leads to a sponsored blog. Every major site has had its algorithm gamed, and only comes out with mediocre recommendations.
Maybe something like Product Hunt, but also for cars, recipe books, private hospitals near you.
I often consume podcasts and videos where a speaker will reference source material, suggest recommended reading, quote an individual, or tell a story. I find these 'embellishments' incredibly compelling but I frequently encounter a lot of friction when trying to remember what and when something was referenced. Given that I, and likely many others, primarily consume these media while otherwise occupied (getting ready, working out, cleaning, commuting, eating, etc) I see a distinct need for functionality that allows some type of community driven, timestamped tag cloud.
A clear example of this in the wild is any top YouTube comment composed of timestamps and topics. Google is certainly aware of the potential utility [1].
A technical implementation with similar crowdsourced tagging is SponsorBlock [2] and I think there are potentially awesome second and third order effects unlocked with a broader feature set. Imagine making API calls to learn the duration and frequency of a sponsor segment, by creator, ordered by time. Now extend that functionality only slightly, and you could query all instances where a book, podcast, video, person, topic or anything was tagged. Run transcription and NLP to infer sentiment about a tag, set up monitoring to be alerted of new tags meeting certain criteria, create customized recommendation feeds based off of tag filters... could there be endless possibilities?
Content with a production budget/team often solve this in advance. Podcasts have episode notes, YouTube videos have annotations and chapters, courses have syllabi, wikis have indexes. This is asymmetrical, from creator to consumer. What if, when someone shares a recommendation, you could engage with the content and also see the sections + tags that they found most valuable?
Presenting a standardized method for tabulated tagging, designing interactions to avoid interruption (I can't believe it took till iOS 14 to minimize Siri), building cross platform integrations (when feasible) and nurturing a community of tag contributors, are all significant elements of this product roadmap.
I've long considered launching this as a commercial project, but I believe with conviction this must be open source. My email is in my bio should anyone share interest.
[1] https://9to5google.com/2020/11/25/youtube-testing-easier-tim... https://9to5google.com/2020/11/24/youtube-begins-testing-aut...
My workflow for such things is so complicated it hurts. I have to remember whether or not I found the <audio> file in the inspector...then download it, then upload it to a separate transcription service I bought LTD from AppSumo. I have found I learn best by reading (retention etc) so I miss out on alot if I don't follow this workflow. Most Podcasts don't automatically provide transcripts etc.
So this would be aces.
