I wanted to work on something really interesting, such as problem statements that don't have good solutions yet.<p>But I fail to find things that could be done that haven't been implemented yet.<p>It seems like all the low (and even medium, high) hanging fruit has been covered. [Of course not calling Redis and such low hanging fruits, but just to give examples of something fundamental.]<p>How would you explore what to work on? Especially something non-AI.
I get the idea you want to work on generic tools. And create a new one unlike anything that has come before? Can you name one of these "problem statements that don't have good solutions yet"?<p>You need to reduce your problem space. I suggest working on any real-world challenge, like climate change. You will quickly run into problems. Work on the biggest ones you can solve.
Then you're not looking deep enough at problems that people have or you haven't been in the field long enough. There are tons of issues out there, and I wish I had a department full of me's who could work on all of the problems I want solutions to.
A high quality speech-to-text and/or text-to-speech program that just works out of the box<p>Challenge: make it require at most (1) opening a website (2) a single packet manager installation (3) a single app store installation<p>a good speech-to-text that integrates with programs as we use them would help my hearing-impaired friend understand the world better<p>and a nice sounding text-to-speech would help me listen to Wikipedia articles<p>I believe there are already good AI models there that can serve as the core, but they aren't entirely developed (missing features such as "really real time"/instant transcription <a href="https://github.com/openai/whisper/discussions/608">https://github.com/openai/whisper/discussions/608</a> and speaker separation: <a href="https://github.com/openai/whisper/discussions/264">https://github.com/openai/whisper/discussions/264</a>) and I haven't encountered a single good interface for them yet.
I'm also someone who are driven and get excited about working on something fundamental. My recent endeavor is the result of years to figuring out what can be the fundamentals for the web -- Building Blocks for the Internet.<p>It would be helpful to understand why you feel like version control (git, not github), a programming language (javascript), a browser, a database are all fundamentals? They are not the first in their respective categories, just that they all became popular? I feel like your examples also lean heavily towards developer only tools?
Look at the products from startups accepted into YC. Few (if any) haven't been done before and the same goes for most open source. Your goal should be to identify something that the others missed.
I think continuous integration is pretty terrible compared to what it could be. The problems I see, at least with GutHub Actions and Azure DevOps:<p>* Generally rebuilds everything from scratch, which get slow on large projects. Reducing overbuilding requires lots of manual work.
* Hard to test, often you have to push a commit.
* No flakey test detection and management.<p>There are some companies working on productionizing Bazel-based CI systems, like BuildBuddy, but there might be other ways of approaching this problem.
There is bleeding edge, and there is filling in holes.<p>Bleeding edge in software is going to be all ML now, mainly in the research of new architecture, and low level stuff to make compute faster.<p>As far as holes go, this is where things like Redis fall. Nothing about Redis is really that novel, its just a nice package of pretty standard functionality, which has a very common use case.<p>And there are A LOT of holes to be filled.
I’m assuming your post is related to open-ish source developer tools.<p>> It seems like all the low (and even medium, high) hanging fruit has been covered<p>Not at all. In software, fruit grows on top of fruit.<p>My personal wishlist: a desktop DB client with simple UX, tooling for DB functions, a tool for performance testing in various geographies, and a browser with symmetric encryption.