If you had no concern about market fit and funding, what would you work on?

[goksankobe]

Let's assume that finding funding is not and issue and you don't need to worry about the current state of the market in terms of existence of a stable customer base or availability of certain goods (e.g. if you'd need thousands of H100 GPUs, you'd find it). The only limitation (and this is probably open to interpretation) is that it should significantly enhance the human civilization in one or multiple domains (so things like rendering Mandelbrot set infinitely faster does not cut it IMO :)

[solardev]

I'd want to build some a generalized community-making/matchmaking system that can bring supply and demand together, but optimized for user satisfaction instead of external profit. For everything like:

* Nonprofit online dating that emphasizes quality matches and relationship satisfaction, not Match Group profits

* Municipal Uber/Lyft that connects drivers and passengers and takes a minimum cut just to maintain the service, with an increased emphasis on carpooling/ridesharing to decrease traffic

* A better Craigslist for apartment finding, with good filters, integrated background/application scoring, etc. that's run as a community service for both renters and landlords, cutting out the profit-driven middlemen and helping both sides quickly and easily find matches

* A better scheduling system for college courses, with integrated calendaring/schedule solving and a much better UX

* Community solar, where different neighbors can come together to buy a small solar farm (like a community garden) and then all share in the power produced without having to install small systems on each home. Much better economies of scale, but higher overhead since you're essentially running a small biz and responsible to multiple stakeholders

------------

In my mind all of these are essentially the same problems, having a robust graph in the backend, building exceptional UX on the frontend, earning and increasing trust with users through reputation systems and humans-in-the-loop (for things like background checks for rents and drivers)... and most importantly, being able to do so with user-aligned incentives or free money (funding isn't a problem).

All of these are relatively solved problems from a technical angle, but not really from a societal angle, because they're often predatory middlemen services. I wish we could rebuild their whole stacks in a reusable way and federate the services in a FOSS model across municipalities/states/nations. To me they could just be basic infrastructure services that governments provide and maintain for the betterment of their citizens.

[vegetablepotpie]

This seems similar to the types of projects that Code for America works on.

https://codeforamerica.org/about-us/

[solardev]

Hmm, interesting, thanks for sharing! Their open engineering positions are a bit old-fashioned, but nothing too bad. And their ask for "Experience with the justice system, social safety net, or other mission relevant experience" really speaks to me. I'd be very interested in applying if I didn't have a commitment to my current employer. I will have to keep them in mind for the future. Thank you for the tip!

[goksankobe]

Sometimes we just need a single variant of such fundamental services that just "works", but today we, in a very wasteful manner, get hundreds of variants each of which is trying to steal some percentage of the pie. Yeah sure thats how capitalism works and competition fuels improvement, but I think its also ridding us on bringing much more users under a single umbrella to maximize the interactions.

[solardev]

Yeah, it's a delicate balance, isn't it? On one hand having too many similar choices just leads to decision paralysis (and in the case of Match Group's dating services, they are all so similar these days it's kinda pointless anyway... a bunch of false choices with similar algorithms). On the other hand, governments also tend not to be very innovative or user-friendly (tried to handle DMV stuff or taxes lately?).

But I would love to see a model where the private sector can continue to innovate, but the government at some point steps in to buy/nationalize some operations (like a national fiber rollout, or basic digital infrastructure) the same way they build roads or other civil infrastructure.

[tikkun]

> it should significantly enhance the human civilization in one or multiple domains (so things like rendering Mandelbrot set infinitely faster does not cut it IMO

I feel compelled to note that "greatness cannot be planned" [1] and often things that significant enhance human civilization didn't seem like that at first.

[1]: https://www.youtube.com/watch?v=dKazBM3b74I

[qlk1123]

Thanks for sharing.

This presentation is pretty inspiring to me, but at the same time there is just no obvious way to leverage the claim. How can any management allow subordinates to do things without any objectives to justify?

Generally I buy the reasoning, but maybe that's just I cannot identify any fundamental flaws now.

[goksankobe]

Ah man, I knew someone was gonna point to Kenneth Stanley here:)

[giantg2]

This isn't tech related at all, but restoring old garden tractors. They're generally simple, but there are tons of details between the various manufacturers and models. They're mostly built to be very repairable and durable. Very satisfying to repair something that's 50 years old and have it be just as capable, if not more so, as the cheap ones they sell today.

[goksankobe]

There's certainly a lot of grace, especially in today's consumption mindset, in repairing a physical device. I hope you can monetize it someday (or maybe not if it will take the fun out of it)

[vegetablepotpie]

I would like to make an actually good open source UML/SysML tool.

There is doxygen, and it’s a very good tool, but it goes one way, from code to diagrams. There is Umbrello, it is bidirectional, in theory, but it’s really half baked. It feels like something put together in semester. It’s very buggy.

Would this significantly enhance human civilization? It’s hard to say. It would being together programming, systems engineering, and management on projects. A good outcome would be that if everyone is using the same models, then quirks in design of buildings, cars, software etc. might get revealed at the design stage and corrected before they are baked in during manufacturing and users have to deal with them. This would make the world an easier place to live in.

[mikewarot]

Probably not meeting your criteria, but anyway, I'd build a BitGrid ASIC[1]. I think it's possible to democratize access to Petaflops, while vastly reducing the power requirements.

So... yea, faster Mandelbrots, but for way less power?

[1] https://github.com/mikewarot/Bitgrid

[mikewarot]

I'd love to work on a Von Neumann probe. A machine that could set down on Mars and make a supply chain complete enough to replace itself, AND have enough capacity to allow the creation of supplies sufficient for human flourishing.

I'd start with one that works in my back yard, and extracts metals from the soil in small quantities.

[goksankobe]

Where would you start though? On software level, I can imagine how this could work, based on genetic algorithms where offsprings evolve by sharing the weights of their neural networks from parents. But self-replicating hardware is quite challenging isn't it?

[mikewarot]

Machine shops make parts for machine shops, with a lot of human labor and control.

So... Just start with machines that can mine and refine metals.

[imvetri]

Make the machine work for me than me working for the machine, is my vision.

https://www.linkedin.com/pulse/leaky-pipes-black-mass-initia...

[mikewarot]

I would document our supply chain completely by organizing and setting up a complete, but less economic, and smaller scale backup for it.

Everything you need, all open sourced, well documented, and ready to use to patch over any issues in our primary supply chain.

[goksankobe]

When I first read this, didn't sound like much. But after just thinking about it for 30 seconds, it would be a HUGE benefit for human kind IMO. It's the realization of first principles thinking. We could tap into fundamentals of how ANYTHING we create works. And furthermore, once you represent this chain in a graph data structure, it opens open immense (AI) possibilities for further optimization and prediction.

[tomcam]

A portable native code environment roughly like Visual Basic

[rozenmd]

Exactly what I'm doing now - building a sustainable business while also full-time employed.

[bwb]

Shepherd.com, trying to help people find books they love. I am a big believer that reading builds better humans.

Very hard to fund :)

[p1esk]

Music generation. Unfortunately that’s where I need those H100s.

[datascienced]

A CRM for personal use. To help organize life things.

[AnimalMuppet]

At the moment? Probably some kind of affordable housing.

[mooreds]

A catalog of perennial vegetables.