Makeabetter.computer – Microgrants for projects that make computers better | Hacker News Reader

Makeabetter.computer – Microgrants for projects that make computers better

makeabetter.computer

179 pointshirrolot86 comments

Dunno if author is around here, but I'd like to point out a group who are working to make low-spec, low-powerusage software.

https://100r.co/site/mission.html

They're a couple living in a boat (all by choice) making games, music and books/comics. Having encountered challenges like being off the grid for a prolonged amount of time and having to budget electricity for months at a time, they built a whole suite of creative apps that run on an incredibly limited amount of specs, including building a platform akin to PICO8 that can run such creative software on many embedded devices like the Gameboy Advance (https://github.com/hundredrabbits/awesome-uxn)

Orca is a delightful piece of software. A livecoding music creating environment, that’s both simple and very complex. Well worth checking out if you’re interested in any kind of algorithmic music making!

One thing I was a little surprised by was that such ideologically minded developers would shy away from a GPL style license for their work. It seems like for a lot of people “viral” open source licenses have fallen out of favour - but I’m not sure why.

They use an MIT license, which is equally viral. And is copyleft really more ideological than permissive licenses? Copyleft still leverages copyright to its advantage, which can be seen as validating the usefulness of copyright

MIT is not viral and it's what makes it so popular, you trade a copyright notice somewhere for the right to re-license to any license of your choosing, that makes it the opposite of viral if anything.

> And is copyleft really more ideological than permissive licenses?

They are equally as ideological, but copyleft is definitely "louder" in a sense since it actually forces people to play along.

> Copyleft still leverages copyright to its advantage, which can be seen as validating the usefulness of copyright

Copyleft is an unfortunate name but the fact that it validates copyright doesn't invalidate itself. In a way, copyleft leverages copyright to fight copyright, though that's not really accurate. It's fighting a popular usage of copyright.

Ah I thought you meant "viral" as in "popular". My mistake

> They are equally as ideological, but copyleft is definitely "louder" in a sense since it actually forces people to play along.

That's fair. I guess the ideology behind permissive licenses is to not force anybody to do anything

I get what you're saying, but the idea that a license forces you to do something seems a bit of a skewed description. Linux being GPL licensed doesn't force me to do anything. If I want to use linux, I can. If I want to modify linux, then some rules kick in. But it's still opt-in, because you're not being "forced" to leverage the fruits of someone else's labour in the first place. And if you want to leverage others labour, all that's asked is that you make your labour available for others to do the same with.

Which is, I guess, what I meant by "ideological". It seems like they are not just sharing some work they've done - they're sharing work that was inspired by a philosophy, and values they're trying to get others to adopt. Which is why it seems like a prime candidate for GPL licensing.

These two are some of the most interesting people I've come across. Both 100r.co and the xxiivv wiki are rabbit holes (ha) I go down in for hours.

I wrote over 300 ideas down

https://github.com/samsquire/ideas https://github.com/samsquire/ideas2 https://github.com/samsquire/ideas3 https://github.com/samsquire/ideas4 (incomplete)

I've emailed them and I am interested in donating to see what interesting projects people shall come up with.

I am interested in distributed systems, parallelism, data structures, algorithms, database architecture. I like the idea of a compiler that exposes its pipeline as a website that people can write materialized views over. Static analysis and optimisation crowdsourced. Like compiler explorer. If you represent your mappings as algebra then you can find equivalent plans like a cost based optimiser in a database.

Ideas4 has data structure ideas.

I would edit that list down if I were you. I found some reasonable ideas but 90% of them are obviously bad or vague ideas or already exist (unsurprising since there are 300 of them!).

Quality over quantity!

I am obviously sad that you don't think my ideas are good.

If you posted your ideas I would happily read them and share my reaction to them and my ideas in reaction to them. For free no less.

Nothing that exists that I have described works how I want. So that's why I have my GitHub projects to try bring them into being.

If you like idea sharing you might enjoy halfbakery.com and http://o2oo.io/

Yeah I remember half bakery! Great site.

Here's my best idea: phone cameras should all have square sensors and record the entire sensor, but display just landscape or portrait. That way you can choose the aspect ratio after shooting.

Kind of blows me away how much people complain about portrait videos but nobody has implemented this obvious solution.

There have been cameras that take square photos - e.g. the Snapchat glasses, but they restrict you to square aspect ratios and I'm pretty sure they actually use a cropped rectangular image.

That's a good idea.

When it comes to photos it's so difficult to find photos. When I witness someone trying to find a photo on their phone they scroll down the list of photos for what seems forever. I am sure there's a way of searching photos and indexing them so they're easier to find.

Thanks for sharing. I read some of the ideas and I had some similar ones. Which is a nice feeling, as in OK I am not the only one thinking that!

I will have a good read of more if yours when you get time, and maybe make my own public list. I have a private one but it is pretty short at the moment due to bad discipline in writing them down.

They are private more because of inarticulate-ness than secrecy.

If you published your ideas today and posted a reply to them I would read them all and share my ideas in reaction to your ideas and my thoughts on them.

So I would really find that exciting.

If you like idea sharing you might enjoy halfbakery.com and http://o2oo.io/

Our project was a recipient of a similar microgrant - https://shkspr.mobi/blog/2020/01/openbenches-is-a-recipient-...

It was fantastic! It let me buy some computing resources which I wouldn't have had access to. And, more than that, it was tangible proof that other people liked my idea - which is emotionally satisfying.

There was no complex application process and no auditing. That certainly opened it up to me as I hate filling in forms.

I hope one day I'm in a position to help others in a similar way.

> And, more than that, it was tangible proof that other people liked my idea - which is emotionally satisfying.

Not only emotionally satisfying. People forget how valuable such validation is for getting further funding.

> (1) Building a self-hosted compiler or interpreter.

> (2) Prototype an experimental browser.

> (3) Make your computer express emotions.

$100-$500 is a drop in the bucket for projects of this scale.

> (3) Make your computer express emotions

But I thought that already comes as standard on Windows...

https://i.imgur.com/lk57d.png

i think you miss the mark,Clippy was there in the 97.

Facebook changed name to Meta. Imagine if Microsoft had renamed itself Clippy.

Not if your time is free (as in: a hobby project).

The best effort I've seen at making a computer express emotions is Anki's Cozmo or Vector [1]. You can get one for less than $500 and hack it, or you can do your own hardware project for much less. Or make a conversational chatbot, APIs for GPT-J make that easier than ever.

Similarly, an innovative browser-chrome that tries novel UI ideas while using webkit for rendering is very achievable on a tiny budget.

1: https://www.vectorrobot.shop/

Either you are trying to encourage people with the money tag, or you are not doing anything and it is hard to imagine for this amount of money to be enough to move the needle for the type of person you need to make any progress on these types of big problems.

Improving something that has been optimized for years is obviously not going to be easy (which it might still be, non-obviously, but that's very unlikely). Offering a very small amount of money evaluates the required work, and by doing so pretty poorly actually devalues it. That seems bad.

I imagine the idea is "there are plenty of students with more time than money, who are held back from doing really interesting projects by comparatively small amounts of money".

This program isn't interesting to a software engineer working in Silicon Valley, but I know some brilliant 16-24 year olds who have time on their hands but would have problems investing $200 into software tooling or a little hardware.

Sure, those people won't build the next sentient robot, or the next gcc. But innovation often happens by exploring lots of ideas a little, to even figure out which ideas are worth exploring further. An army of hobby projects can be very effective at that.

> $100-$500 is a drop in the bucket

My fully-funded doctorate's federal research grant (Brazil) pays around 450 USD per month (cf. Valores das bolsas no país [1]).

[1] https://www.gov.br/capes/pt-br/acesso-a-informacao/acoes-e-p...

So you're saying it's a drop in the bucket?

No, he's saying that not everyone lives in America.

technically he lives in _America_ too

Of course? That's why they are "micro"grants.

Many projects that improve peoples' computers are done mostly for love - but that doesn't mean they don't get done a lot faster with a little money for expenses or to supplement income. A few hundred might let someone go to a conference to promote the project and get support, buy testing equipment, etc.

It's not going to fund anything startup-style, but it very likely could help lots of small useful projects run more smoothly and reliably.

> $100-$500 is a drop in the bucket for projects of this scale.

Agree, but also depends on the location. But seems even the cheapest places to live in the world comes close to $500 a month for living costs.

$1000-$5000 would have been nicer, then people in those locations would have at least a chance at a 10 month runway for the highest grant amount.

I don't think the author is expecting people to finish any of the funded projects, just to kickstart them or even give some inspiration.

>But seems even the cheapest places to live in the world comes close to $500 a month for living costs.

I live in outskirts of a second-tier city in southern India. Spacious rented room and I employ a maid for washing needs. My average monthly cost is $150.

What's the situation with power and internet...?

No issues with internet (via mobile). Power is an issue whenever there's heavy wind/rain (ex: power off for about 2 hours yesterday). Many houses have battery based backup. I could afford that, but haven't got yet (one of the reasons being I'm in front of computer almost all day).

I'm not sure the cost of living is the right comparison here. I would rather use a professional salary as a basis.

Somebody who can develop a self-hosting programming language that is useful for anything can probably earn at least $5000 regardless of where they live. I wouldn't expect them to live on 1/10th of that even if this is technically possible.

> Somebody who can develop a self-hosting programming language that is useful for anything can probably earn at least $5000 regardless of where they live.

I know quite some highly brillant people who have difficulties getting a decent job perhaps because they are really bad at self-promotion.

So being able to earn such an amount of money depends on having some further traits that don't have anything to do with intelligence or programming skills (rather perhaps a little bit with hustling skills).

I'm not sure what self-promotion has to do with getting a developer job. You send your resume, pass the interviews and you are in. If you live in a poorer country, it might take you going through a couple of jobs until you get to a decent salary, but you'll get there if you are either smart or hard-working.

Maybe some people require a translation for this project:

  It's free money to do cool shit.

I think i know the play here. It is funded by angel investors / founders who then get to connect with talented people they may want to hire. It is kind of a paid interview :-). I might be wrong but not a bad idea.

My government can give a approx $15000-$20000 grant but it has to be MVP for a business kind of thing.

Back to this… I would apply for the grant mainly to test if the idea was worthy of a grant for my own motivation. I probably don’t need money as there is enough free tier stuff to run it on.

Great idea. If anyone here wants a microgrant and has POSIX distro skills, I can offer $100 to you or your favorite charity for help getting the `num` command added to the POSIX standard track so it becomes freely available for everyone.

Num is free open source software that uses `awk` to calculate many typical statistics on the command line.

https://github.com/numcommand/num

That's a nice little utility. A poor man's version would be to use Perl, which is mostly everywhere and comes bundled with List::Util. Missing a fair amount of what you're providing, but would work in a pinch, especially with some aliases to scrub the boilerplate out.

  echo "10 9 8" | perl -aMList::Util=/./ -E 'say min @F' # min
  echo "10 9 8" | perl -aMList::Util=/./ -E 'say max @F' # max
  echo "10 9 8" | perl -aMList::Util=/./ -E 'say sum @F' # sum
  echo "10 9 8" | perl -aMList::Util=/./ -E 'say sum(@F)/@F' # average
  echo "10 9 8" | perl -aMList::Util=/./ -E 'say reduce {"$a,$b"} @F' # arbitrary ops

Good thank you, these are very helpful!

Correct URL: https://github.com/numcommand/num

I will donate $1k for GNU/Linux systems to handle cdate. /joke

What would that imply actually?

Ideally someone saying e.g. "I'm experienced with asking the POSIX standards folks to consider adding the `num` command to POSIX, and I'll do a few hours of volunteering to help toward this goal".

Or "I know how to package the `num` command for Linux (e.g. Ubuntu apt, Fedora yum) and BSD (e.g. FreeBSD pkg, macOS brew), and I'll do a few hours of volunteering to set up packaging, in order for users to be able to install more easily".

There's nothing wrong with the size of the grant, in the US a hundred dollar bill is a nice bag of groceries, this is one to five of them for showing off something cool which someone was already working on.

Something like this which had a clear path from a few groceries to full sponsorship at a comfortable salary could really take off, I think.

As is, it's in a sort of uncanny valley where you'd probably attract more high-quality work in total if the prize was "get on a list of cool new stuff" instead of money. It's easy for the friction involved in accepting money to exceed the value of a hundred bucks, which might prove a bit of a market for lemons.

Or it might not, it's a worthy experiment.

Props to the sponsors for doing this, I look forward to seeing what comes of it!

I was going to be mean, but the numbers actually work.

If we take the medium salary for Software Dev in Denver at $149,000[1] (I know, I know some are paid vastly more), which is $105,077 after tax, and divide that by work-hours in a year, 1,768. Then we see they're paid about $60 an hour. So $500 means about 8 hours work. A day to play around with some project and add something is kinda do-able, especially if the dev throws some extra hours in because they're having fun.

I would absolutely say however that their advised projects are toned down slightly.

1. https://www.levels.fyi/Salaries/Software-Engineer/Greater-De...

I dunno, if your goal is to change computing as we know it, what is a middling dev in Denver going to do about it in 8 hours? Not to say anything about devs in Denver, but it’s more the problem is so large that I can’t see what $500 could do.

It would probably be a better use of money just randomly handing it out to middle school science students to see what they can do. Get them on a trajectory to change the world rather than soliciting any concrete ideas; no one with an actual idea to make computers better is going to even blink for $500.

If someone is paying $149k/y to work on ones own passion projects for free and keep the IP rights, let me know where to apply.

The amounts are actually not that bad for someone who is working on something in POC stage.

The last few days I'm trying to integrate a JavaScript engine with Vapor(an OS backend framework written in Swift) to create a standalone, dead easy to run(no Docker, package managers or anything like that, easy to install and run as installing Firefox) multiplatform low code backend tool which roughy falls in the "Building a self-hosted compiler or interpreter." category. The idea is that it will allow developers focus on the product instead of dealing with server and data management at reasonable cost, control and performance. My app will take care of authentication, data storage and allow custom logic to be written using JS. Firebase kind of does it but it cannot be self hosted and there are horror stories of mistakes that can cost tens of thousands of dollars in a blink of an eye. The custom logic with functions is also too slow, it takes forever to wake up and run a function(though they introduced some solutions for the issue).

As it turns out the JavaScript engine integration part is not as easy if you are not well versed with C/C++ and its tooling because builds will fail, platform ports are missing etc. JavaScriptCore is very nice on macOS but it appears that compiling it on Linux is not just running the two lines of code as described in the documentation.

Maybe I should apply for this and use the money to pay someone proficient in the topic to compile JavaScriptCore and help me make it talk with my Swift codebase on Linux.

Given the small sums (and the relatively limited scope of the examples), these seem more like "nano" grants to me.

Well… the median of the range quoted in title could cover a month of living in my city (if you don't need to pay rent). Although you're probably not going to prototype a browser or build a compiler in that time, and anyone capable of doing so would move the hell out of here to make at least 20 times that much anyway.

Yeah, I've reconsidered my comment.

These grants are definitely in line with the amounts of what are often called "micro-loans", which in some contexts are often as little as $200 (eg. Kiva). In other contexts a micro-loan is up to $50k (eg. SBA).

Isn't housing most people's single biggest outlay?

Idk, the amounts are small, not really enough to spur any great new thinking, but maybe what's going on here is that there are a lot of people with ideas and opinions already, who might just share their thoughts if someone bought them a free lunch or two.

Seems to me VCs do a lot of work in scouting out potential new companies and trends. This would be a cheap way of collecting those in a “pull” manner rather than “push”. They’ll end up getting a lot of early-stage ideas from people desperate for cash, but maybe that’s exactly what they want.

This is an awesome idea. When the web was young, getting people to read and comment on your ideas was a great way to get feedback and encouragement for good ideas.

Advertising has distorted and rendered the open web far less useful in this regard.

Funding small one time projects without strings is something that DARPA used to do. It's been missed.

In their application, they offer an option for the funded projects to write up their project as a blog post. I think it would also be good for them to offer funded projects an additional option to to be included in list/post from the funder's side. It would be a nice way to add perspective to outside observers.

Does anyone know of a similar program with grants that are about 10x this size ($1k to $5k)? That's what I would need to get some of my ideas demonstrated.

Funny you should ask. My employer is starting something along those lines:

https://www.futo.org/grants

I'm shameless, so I applied with my LLC ( https://www.adama-platform.com/ ). At this point, it's less about the grant and more validation and marketing.

I love the idea, and this could have helped a younger version of myself when I was poor.

I want to build a screen-less computing system that has a high-degree of affordances for people with different abilities. I want to be able to write software specifications in maths on an e-ink tablet, or a white-board, and have them interpreted and sent to an interactive theorem prover that displays the result back to me (or dictates it).

A system akin to https://screenl.es or https://dynamicland.org/ that can be assembled from existing parts as much as possible and extended to the needs of the user.

But the focus is to move away from keyboards and glowing screens and this fixed notion of computing that we've been stuck in since the 90s.

Nice to see projects like this. It makes me wonder if there are others out there with bigger budgets. As much as I enjoy hacking my reMarkable on the odd weekend there's a lot of work to do to get good gesture recognition capable of recognizing maths as well as hand-writing that is fast and can run on low-power devices. Plus the software stack to run the computing system on needs a lot of work to adapt to a screen-less paradigm. If I could dedicate myself to it full time I'd probably get much further than I can right now which has been... several years and mostly just have some hacky stuff running on my old RM1 and some email scripts.

Just one of those projects that probably won't see the light of day because it's not capitalist enough.

Round Investment Service is a Reputed and Licensed Money Lender, Offering Loans At Low Interest Rates. The Company is Registered with the Ministry of Law. Our Mission is to help People Who are in Need of Cash by Lending them Money. We do not need any credit score, Collateral and Deposits to Process your Loan Request. Why Not Apply for a loan today? Some of the Services Provided by the Company are, 1. Unsecured Private Loans 2. Business Loans & Funding 3. Payday Loans 4. Debt Consolidation 5. Personal Loans Email: roundinvestmentservice@gmail.com

I’m working on an educational programming language (pickcode.io) and having something like this to pay off whatever AWS/Netlify stuff isn’t in the free tier would be great.

https://paul.totterman.name/posts/free-clouds/

> (1) Building a self-hosted compiler or interpreter.

Aren't every compiler or interpreter self-hosted, because they run on your own machine?

> Aren't every compiler or interpreter self-hosted, because they run on your own machine?

Self-hosting in this context means that the compiler is its own toolchain, it doesn't rely on another compiler to create its executable.

That said, nearly all compilers need some other compiler to bootstrap to self-hosting status, though there are projects that aim to minimize that chicken-and-egg issue. More pointers here: https://en.m.wikipedia.org/wiki/Bootstrapping_(compilers)

No. A self-hosted language means the language's compiler or interpreter is written in that language.

Go is written in Go. rustc is written in Rust. clang is written in C++. Etc.

FYI this domain is getting blocked via some OpenDNS block list on the public wifi I’m using at the moment.

Love the idea, hate the website. Something about these 'clever' console animations in a web browser seriously puts me off. We spent last 40 years moving away from consoles only to be sucked into fakes ones via this retro design approach. Ugh.

Taking the target audience into account, I think the design is right on spot and successful.

Food for thought: why do people still use consoles inspite of all the progress made so far?

If you answer is inertia, think again.

- It's fast, always was fast, I hate waiting for computer to do what should be instant

- copy paste always works, good luck selecting text from button or menu of a gui app

- font and colors same everywhere

- if for some reason GUI is temporarily not working it is not downgrade for me

- takes less resources, leaves more for actual work

- fast and transparent network access through SSH

- strangely, i feel that TUI is more aesthetically pleasing that GUI, perhaps related to uniformity of monospace text and everything fitting to grid.

If you mean why do devs still use the terminal over GUI programs, I'd say a big reason why is because its kind of like its own universal GUI.

Take git for example. If I use VSCode and you use Kraken for example, and you have a problem with git and ask for my help with it, depending on the complexity of the problem I might not be able to figure out how Kraken works or even what the issue is. We could fix that by settling on one GUI git program, but you say we should settle on Kraken and I say we settle on VSCode.

The terminal meanwhile is the default choice, because its on every machine by default and is compatible with any other tool you use.

Nah, it is pretty much inertia. You can see it whenever PowerShell comes up, with all its vast improvements over the bash + coreutils ways of doing things, and people are still all "well it sucks because it doesn't work exactly the way I'm used to".

Frankly, software developers gave up on GUI in the early 2000s despite huge discoverability and UX advancements because the money was suddenly in shoving ads down people's throats via the web. Fortunately for them, when users are your product instead of your customer UX is relatively unimportant.

$500 to AnIdiotOnTheNet to revive the GUI!

Right, they should have used a tailwind template right from the first google result :/

It's exactly the type of site many people are crying for here on HN.

They just used a console font and animated in the first 4 paragraphs over maybe 1000ms. It is not the most obnoxious site I have used by a long shot.

Is this a parody?

These amounts are so trivial, it's almost not worth applying. They need to increase by about 20x to start being worth it.

If they are trivial to you that probably just means you aren't in the target group. $200 is nothing to a full-time software engineer, but for a student $200 might be much more than they are willing (or able) to spend on a hobby project.

Exactly.

The upper end looks to be about the average monthly developer salary of a poor country with many highly educated programmers (such as Belarus). There will also be a pool with talented individuals who are not yet on an average monthly developer salary, and if either your individual project or the scheme as a whole is somewhat successful there will also be a CV benefit in having been sponsored. And of course many developers all over the world would jump at an opportunity to get a month's funding to work on their pet project and keep the IP.

…getting you back to square zero, where tightly-scoped micro projects never get investment. I mean if one of these tiny things is actually a killer idea, it’ll make a big enough splash to get a couple orders of magnitude more funding, anyway, right?

I think the point is that you’re not going to get any killer ideas from this. More power if they do, but the people working in the “make a better computer” space have been doing so deliberately for years, spending much more than $500 in the process.

Maybe you get a better computer by sprinkling $500 checks around to random devs in the developing world, but it’s not likely to work as a legitimate strategy to improve computing. If you really want to do that, take this entire fund and invest it in one of the many future of computing projects out there already well under way.

I'd guess the places where this is relevant will have plenty of rich-enough people who can self-fund. If if they can't, then the amount is definitely not trivial. $200 that you don't have is still $200 you don't have.

$500 mean little to me, [senior software engineer from a rich country], but it's higher than the average salary in many places.

Reply on news.ycombinator.com