This is great. I really hope that efforts like this will spread around more. If smaller companies like Tarsnap can do it, then what stops others?
> If you're at a startup which relies heavily on open source software, please take a moment to ask yourself: How much does your company contribute back?
Quite a few years ago, Gabriel Weinberg from DuckDuckGo launched a "FOSS Tithe" website which encouraged companies to donate some percentage of their profits to the FOSS projects they rely on. I've tried to encourage my former employers to do this when possible, or to pay people to contribute features to FOSS tools we were using. With varying degrees of success.
Now that I'm self-employed again, I can simply donate some of my profits, and am more able to contribute patches.
Sponsoring contributors to build features for your org is probably the best way to go. Your org will get an invoice, you'll get a feature. The project will be happy, and the developer will be happy.
Not a public company but Bloomberg's FOSS giving falls under their overall corporate philanthropy program, for example. They're involved in FOSSFunders.com behind the scenes but lots of bureaucracy to actually adding their logo.
In practice I doubt this is much of a barrier. Shareholders don't inspect every line item of a company's budget. For a decent-sized company, a few hundred thousand or even a million or so $$ per year would probably go more or less unnoticed by shareholders.
Remember that it's a common -- incorrect -- trope that public companies are legally required to seek maximum profit, etc. It's a lot more complicated and nuanced than that.
Is it really that black-and-white? It certainly has at least some characteristics of charity, such as being non-compulsory, and often done out of a sense of goodwill.
To me it is more similar to giving money to a busker.
The busker is there doing his thing, he's not asking for money to play. He's working though, and payment acknowledges the value of his work.
Charity, on the other hand, is supposed to expect no work in return. You help people whether they did some work or not.
The fact that the OSS dev didn't charge for the software in the first place doesn't mean paying him for that software is charity. It's acknowledgement, and a possible investment if you depend on the software he provides.
That might be an issue for smaller projects but at larger scales there is often an associated charity. For example https://apache.org/ and https://numfocus.org/community/mission are both 501(c)(3) registered charitable organizations.
In the sense of being "charitable", sure; but the legal sense of being "a charity" is quite a different matter. I record most of this as "sponsorship" and get Tarsnap's logo put in various places; if the Canada Revenue Agency asks any questions I can absolutely point to evidence that this brings in customers to justify it as a business expense.
Tangentially related, apparently Sentry's own post mentioned in the article was removed from HN according this tweet https://x.com/bentlegen/status/1716884717394670015?s=20. Does anyone know the circumstances regarding why?
> Does anyone know the circumstances regarding why?
As someone working at Sentry my running theory of this is that we're not a YC funded startup and the posts are removed for marketing. Removal in this sense is that the post is still there, but it's pushed down in rankings artificially to be removed from the first two pages.
> The guidelines define “interesting” as “anything that gratifies one's intellectual curiosity.” Surely a real contribution to the OSS sustainability crisis counts?
Arguing about whether the submission meets the guidelines is pointless when it ignores that users downvote what isn't interesting in their own opinion. The moderators are never going to artificially keep a submission active after users have downvoted it just because it meets a definition in the guidelines.
Uninteresting is off-topic, per the guidelines. “On-Topic: Anything that good hackers would find interesting. That includes more than hacking and startups. If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity.” (emphasis added)
It is curious that users would downvote/flag our post yesterday but not Colin's post today, which begins, "Yesterday I read a great article from Sentry". I agree it's kinda shouting into the wind to complain. Oh well. :)
It's well known that to get to the front page you have to submit then get two dozen friends (with good rep) to upvote you within a short period of time.
Since the moderation log is hidden I can only guess what happened. But last year we got the confirmation of the removal via email and the explanation was that the post wasn't "that interesting". https://twitter.com/chadwhitacre_/status/1716947994338021885
I expected this again and thus kept monitoring the index page actively. The post stayed on there for quite a while and suddenly dropped off to the second page and then a while later completely.
Crazy I missed that. That certainly damages the trust I have in this site - it already benefits YC to have this community, they don't need to play favorites.
It feels like we're being treated as a sales funnel when they pull tricks like that.
Speaking of Sentry's program... I recently received a notification that as an open source developer, I am eligible for a payout from this program. Excited, I logged in to eventually find out that the payout amounted to... $15.
Don't get me wrong, I really appreciate the thought and effort that went into this, and I really hope more companies that use open source software follow suit, but I wish they would adjust their formula to not bother with the payout if it's less than a certain amount. It's just more hassle than it's worth, and the nice gesture can easily be mistaken for something else when it gets that low.
It's $15 per month so $180 allocated overall. Would it be more worth it for you if it were paid out in one lump sum for the year vs. doled out monthly?
Thanks for clarifying. I missed the part about this being a recurring monthly payout - $180 does clear the threshold of what I'd consider a significant donation, so I appreciate that.
I don't mind it being spread out monthly. I think it would help if the process made it a bit more clear what to expect (in terms of the total payout and timeline) in the initial message and/or the UI.
Again, I really appreciate the thought and work that goes not only into this program but also into getting other companies on board.
Is there a particular reason for the monthly payment when you already know (and have?) the total for the year? If it were me, I'd rather have money in my hand now instead of later.
Mostly a design choice on Thanks.dev and GitHub Sponsors parts. I would be fine to give it all at once, especially since conceptually for me it's payment for the past year.
Where is the evidence it was removed by a moderator? I know that's the first thought, but HN doesn't have many moderators, and usually the moderators leave a comment on a submission after taking a moderator action.
I regularly come across a submission on the front page that looks interesting, read the article, then come back and notice it's been pushed to the 2nd or 3rd page. My assumption is that it was downvoted by a bunch of people, but it's still annoying. I wonder if there are any sites that track these movements?
> ”Tarsnap has spent 100% of its December operating profits [$274,482 over 14-years] on supporting open source software”
Off topic: if you extrapolated that math, it should be hugely encouraging to aspiring indie/solo developers that you can earn a fantastic income as a 1-2 person business.
Nothing unique about December in terms of revenue. Even if people were prepaying large amounts in December (which isn't happening as far as I can tell) that would go into the "unearned income" reserve; profits are only realized when the service is actually delivered.
Usually its fourth quarter where many businesses move into the black for the year based on how they budget for things, hence things like "Black Friday".
Yep, but actually many companies unrelated to Retail use this or similar measures as they gain fourth quarter accounting success, pretty common all over the place.
is it really a fantastic income? tarsnap is well-known and we're seeing the value after 14 years of work. $250k/y is much less than what you would be making at a FAANG.
Just because it's less than a FAANG doesn't mean it's not a fantastic income. I make less than this, in a HCOL area, and I think I have a pretty fantastic income. $250k/yr is 3.3x the median household income in the US ($74,580 according to the 2022 Census).
"fantastic" is always going to depend on context -- you could also say $250k/yr is 80x the median global income. In this case, I'm sure the owner could easily make a FAANG income if he so desired, which is why I used it as a comparison point.
Your math only makes sense if you’re assuming they got all their customers year one and have had no growth since. Who’s to say the donations last year weren’t $100,000 putting the estimated profits at $1.2m (assuming all months are roughly equal)?
Who’s to say the donations last year weren’t $100,000 putting the estimated profits at $1.2m
To answer the rhetorical question: I can tell you that. I wish, though. ;-)
Reality is somewhere between those two hypotheticals. The past few years, my now-2.5-year-old has been taking a lot of my attention and I haven't really been focused on Tarsnap's growth.
Doesn't that makes the calculus worse for the startup, not better, due to time value of money? The counterfactual FANG worker chugging along at 250k/year has had health insurance, deductible retirement contributions, and years of stock market growth to appreciate their nest egg.
If you’re making $1m in profit after 14 years it doesn’t take long to catch up even if your first few years were slow. You also are sitting on a business you could likely sell for $5-10m on top of that.
If Apple, another big player in the BSD ecosystem, were to donate their December operating profit to open source, that'd be a game-changing $10 billion a year!
Shoutout to Tarsnap, who's already doing awesome stuff like this.
You know GitHub sponsors is a broken experience when you find out GitHub only supports 7 people on their own platform https://github.com/github companies really need to fund their dependency tree…
Yes, perfect training grounds for your future “solo-pilot coding agent”.
Let’s not kid ourselves, by now we know that data is gold and we hand it to these corporations without ever getting paid for it.
One of the main reasons they can “incur losses” on co-pilot is not only future monopolization by market share but also by being provided endless amounts of golden data through usage.
So yes, hosting “social git” with some fancy UI and CI and spinning up GPUs (they will constantly need to ramp up on anyways) is peanuts in comparison to the upside.
Yes, open source developers as well as co-pilot users should get paid for their labour. Directly and by MSFT.
>The folly here is you believe MS does this because they want to support open source...In reality they get a TON of value from this "free" service
What is the difference? Nobody supports open source for the sake of it, we support it because we get tons of value from the "free" service. It sounds like incentives are properly aligned.
I see nowhere that GP said anything about what they believe. They might believe that, but until they say so, you're reading more into the comment than is there and making an assumption
another common folly is thinking that paying puts you in a special tier. free stuff is solely funded by the value of data, ofc there are exceptions funded by donated money and hard work. The things that are made to make money though, don't suddenly stop using the value of your data because you pay for them.
I recently learned that Kubernetes is running on a lot of free labour. The main draw and ever increasing underpinning for lots of services provided by the behemoth infrastructure entities…
Software labor is so funny. Here we are, with software engineering salaries some of the highest of any profession in the US. While at the same time much of the most critical and commonly used software was written completely for free. It’s an interesting juxtaposition.
I wonder if there's a license pegged to some meaningful metric (a hard problem unto itself), that prevents usage if the dev fund/devs are not compensated an X amount in a specified interval.
Something like an SOS License.
Support Open Source
...
All development on this project will cease to exist if dev fund does not meet "DOL Software Dev CPI adjusted Blah Blah Salary" * Number of active/approved devs.
Bugfixes will stop if dev fund falls below Y amount or something
...
Just some half baked thoughts, many conditions can be added or removed depending on the project structure/goals, but the underlying thought would be that Devs Get Paid is a key requirement in there somewhere.
Lots to consider here, but it's weird that no popular licenses have this as a clause in there - almost like it's frowned upon to ask for money?
> I recently learned that Kubernetes is running on a lot of free labour.
Not sure what is your source for that? At a quick glance most contributions come from Red Hat, Google, VmWare, Microsoft, Amazon. Lot of independent contributions are done independently but as part of employment, which I guess is not free labour?
It is opensource so there is still a lot of volunteer work going on there but the infrastructure (CI/CD, image hosting, etc) is also paid for by corporations and not by the community. Its millions of dollars a year in contributions.
i wouldn't really call $400k/year in total "heavily subsidized", that's around the all in cost of a single developer at one of these sponsoring companies
Depends. Giving $1m to free open source is “bad” but hiring 4 engineers and have them write core code and then open source that is good. Microsoft basically does this (more than 4 of course)
> Giving $1m to free open source is “bad” but hiring 4 engineers and have them write core code and then open source that is good. Microsoft basically does this (more than 4 of course)
Apple bought CUPS (printing) and now owns it and hired the original developer (who left at some point):
They ought to hire him back. When I upgrade my Macbook Pro to Monterey, it no longer prints to my Canon D530 printer. I have to copy files to a 10.6.8 Snow Leopard Mac Mini and then send them to the printer. Yay progress!
How can you say something like this with a straight face? There are millions of examples of "good" that have come from the products and services for-profit companies produce.
They have tens of thousands of workers. They ensure the sustainability of what they need and I bet they give back the bare minimum to abide by the laws. Apple is famous for benefiting much more than contributing to FOSS.
Apple are also infamous for making terrible software (iTunes put of off even looking at Mac for a decade and a half) so it might actually be a good thing..
>> Shareholders will boot Tim Cook if he even thinks of that..
> Which is why seldom good ever happens in the land of profit-driven business.
Many responses disagree with this comment. My knee-jerk reaction is the same.
OTOH, there is truth to the comment in that many examples of "good" are examples in which a business is driven by vision or value, not by profit first.
The history of computing is intimately tied with government investment and academic research. Hence ARPAnet and ALOHA. The impetus for world changing innovation tends to come from governments trying to improve things that are structurally inefficient.
Let's take a look at where the exponential growth of technology kicked off, and when the first actually almost-free-market began. Hmm. Quite a strong correlation.
That is just an outright lie. Even most of the open source stuff I'm going to assume you are thinking of could never have existed without adjecent profit-driven ecosystems.
Profit drives innovation, and more importantly makes technology accessible at scale. Do you think most of the world would have access to washing machines (which are of ridiculous utility, especially when you factor in their price with profit margins and all!) if it were not for some schmuck wanting to line his or hers of pocket? Because it would not.
At its heart, capitalism is a (greedy) solution to a planet-scale resource allocation problem, and there are no viable alternatives (give or take some modifications/variations).
>like US cable and telco companies like to remind you every time they do an anticompetitive move.
And despite that, who is helping move the telco world forward? Pointing to one mediocre example is certainly not proof that profit doesn't drive innovation.
No red tape needed, when comcast temporarely lowers their rates anywhere a new provider is trying to gain ground. Even the threat of doing that is enough to stop most attempts dead in their tracks.
It's not impossible to enter a market offering better services for bigger price than you competitors really. Or having better tech it's possible to attract enough investments to overstay lowered prices. Or use other ways which definitely work outside of US telco industry.
This is like the nursery rhyme version of economics we tell to freshmen. The truth is pretty far from this.
My personal take is that in practice most of the innovation we see in tech comes from the universities. E.g. the tech in search engines, smart phones, the internet, LLMs and AI etc.
The tech is then, in practice, given to the private sector. There the profit motive is used to commercialize rather than innovate.
Commercialization does often require some innovation around the edges, which companies often get by hiring talent from academia.
I realize this is unlikely to be a super popular opinion for some HN readers. But to me it's the one most defensible by evidence.
>>The tech is then, in practice, given to the private sector.
100% false, normally the people that created the tech while in university are the ones that create the business using the test (see facebook, google, etc), it is their tech to begin with it was not given to them
In most cases the University is also either paid a fee, or is part owner in the company as a startup, cashing out when funding starts.
Nothing is "given away"
now I personally think anything developed using Government Grants should be Public domain, non-patentable, non-copyright, 100% public domain, but ....
The technologies I mentioned were all developed by dozens or hundreds of people over the course of decades. Statistically, P(you become a founder | you helped invent the tech) is basically 0. Although you are probably right that P(you helped invent the tech | you are a founder) may be nonzero.
"(...) nursery rhyme (...)" is a nice way of not addressing what was said. Good for you.
Yes, a lot of discoveries (not "most innovation", but besides the point!) happens at Universities. But so what? Universities are either for-profit themselves, or they are financed through taxes which depends on the profitability of businesses in the economy which they exist in and contribute to. And like a sibling comment pointed out: Researchers are often the ones commercializing a discovery. Moreover, I specifically stated that "[the commercializing actor] makes technology accessible at scale."
Your statement that "the profit motive is used to commercialize rather than innovate" is a half-truth, because the profit motive is very much present once you consider more than the first-order effects.
I 100% agree that commercialization of new techology (again, innovation is a lot more than bleeding-edge research) e.g., search engines, smart phones, the internet, LLMs and AI, is optimally done through some kind of interaction between univiersities and commercial actors – but to the mutual benefit of a spiderweb of stakeholders affiliated with both academia and commercial actors, and not some quasi-altruistic act of sacrifice on the part of the university.
You should realize that the second part of your claim is just a postmodernist mantra (written to sound fine, links to reality skipped). In observable world one needs better economy to contribute more time to not-for-profit activities. While each, and every attempt to destroy evolutional capitalism to replace it with revolutional whatever "non-gready" model consistently lead to notable worsening of economy.
Worsening of economy by what measure? The argument is circular if you're saying "less profit is made in economic setups which encourage not-for-profit activity". It hardly seems like a reasonable criticism.
More interesting is what people would achieve if their time wasn't dedicated to profit seeking. The measure of success then may be the subjective usefulness of someone's contributions. It is a big subject but waving it away as you have isn't a satisfactory answer at all
You don't understand. You can't just e.g. sleep late, fish a little, play with your children, take siesta with your spouse, and spend evenings playing guitar and sipping wine with your friends; you have to spend more time fishing and with the proceeds, buy a bigger boat, then with the proceeds from the bigger boat buy several boats, eventually you would have a fleet of fishing boats, then you open your own cannery, controlling the product, processing and distribution. You would expand, branch out, acquire and merge, do an IPO and then, in twenty or so years if everything goes well, you could retire with millions in your bank account and then, I don't know, move to live somewhere on a nice coast, where you would sleep late, fish a little, play with your children, take siesta with your spouse, and spend evenings playing guitar and sipping wine with your friends.
Linux and git, probably the two most successful free software programs out there were both created by one person (the same person) WITHOUT relying on a "profit-driven ecosystem". His code made trillions of dollars but that wasn't what motivated him nor he got some % as those new kids (and their VC parents) joining FOSS wants to.
He could not have made Linux into what it is today without additional contributors, and many of those contributors have been in the employ of companies that use Linux.
Yeah so he helped create a system where contribution beats direct monetary benefit. Microsoft tried hard with hundreds of billions to prevent Linux success, the fact that it spends so much money on it nowadays proves that no amount of money could have stopped Linux success.
Linux and git were examples I was thinking of. He could still have made them, but most of their value relies on interactions with commercial offerings. Not to mention technology financed on the back of taxed profits. And their availability to regular people (notice how I talked about "making technology available at scale") would be diminished beyond belief.
But yes, people make the most incerdible stuff without being paid to do so.
> At its heart, capitalism is a (greedy) solution to a planet-scale resource allocation problem, and there are no viable alternatives (give or take some modifications/variations).
I am pretty certain we had economics for millenia before capitalism has emerged about three hundred years ago.
Do you mean greedy, as in the sense of a greedy algorithm? Because I don’t have a formal mathematical description for this, but it feels more ‘optimal’ than a purely greedy algorithm.
Local agent behaviour may be greedy, but this could still result in something approaching a global optimum - the economy could work like Ant Colony Optimisation in aggregate.
Put the FOSS grants under the Marketing budget, call them "sponsorships" or "comarketing" or whatever. Corporations have a budgets for warm fuzzy stuff like that.
You realize that people like retirees have parts of their savings invested in things like Apple stock?
So this is literally taking money that would otherwise go to grandma and giving it to open source.
Money that would go partially to hedge funds, sure. But money that's also partially going to grandmas.
How about you let individual shareholders themselves decide if they want to donate to open source personally, since the profit belongs to them? They can take their dividend and they can transfer it to open source. But not have Apple force them to.
I mean, I personally just don't like other people donating my things for me. And if I want to donate my money, maybe I want to donate it to something else like fighting poverty or disease?
It's awesome for a privately-owned two-person company to choose to donate to open source. It would not be cool for a publicly traded corporation to take a twelfth of the dividends I receive and send them to a charity that is not of my choosing.
Great! 55 year old+ have an average net worth of about a million dollars[1]. They can spare a few bucks for open source which they use to generate a huge pile of cash anyway.
First of all, why don't you take a look at the median rather than the average.
Second -- if you're 65 your savings have got to last you for the next 20-30 years. Including often expensive nursing home care. Suddenly even a million bucks... isn't that much.
So no -- in general, they can't spare an entire twelfth of an investment. But more importantly, even if they want to, it should be up to them. Not somebody deciding for them.
Because the moment Apple announced that they were going to implement a policy like that, the value of the stock would drop by a twelfth. Instantly.
So by the time you found out about it, the damage would be done.
But you're missing the bigger picture anyways: business is business and charity is charity. The job of publicly traded corporations is to make money, not to give to charity. Then individual people can take the money the business made, and give to the charities of their individual choosing.
Having to choose which stocks to invest in based on their charitable giving portfolios just mixes everything up. The profits are all the same in the end, but charitable giving decisions around those profits should be in the hands of individual people, not corporations.
considering he's been doing it consistently since 14 years ago, he's a responsible businessman and smart guy (i.e., putnam awardee). good for you, mate! hope your business reaches the spot where you want it to be.
A trick for easily estimating powers of two is to remember that 2^10 is about a thousand. Thus 2^20 is about a million, and dividing by four, 2^18 is about 250,000.
Yeah. Just remember that each multiple of 2^10 is approximately 10^3 (1000) and then memorize the single-digit powers of 2: 1, 2, 4, 8, 16, 32, 64, 128, 256, 512. It's useful for doing order of magnitude estimates of multiplications in your head.
If your company uses open source software, please contribute back, financially. For large projects, there are foundations, for smaller projects there are other ways (like Github Sponsors).
There are some ecosystems where this is becoming commonplace (I mostly live in the Clojure world), which is a great thing!
This matters, even if the amount of money isn't huge. What is important is for it to be recurring: a subscription. If you use the software and rely on it, keep contributing to ensure continued maintenance and development. Far too many people associate "open source" with "free".
I have been trying to do this for the last two years, and while the amounts are not huge, I try to increase them along with my SaaS revenue growth.
Thanks for all the contributions @cperciva. FreeBSD update doesn’t get talked about much but going from having to rebuild FreeBSD for updates from source to binary updates has been incredibly useful.
No, that's USD. Most of the amounts were in USD to begin with; the rest (e.g. the $1000 award at my alma mater) I converted at the relevant exchange rate for the month, which is why they're not round numbers.
But am I reading this right? 274k is 1/12th of tarsnap profit over 14 years, so tarsnap has made 274 * 12 / 14 = 234k profit annually? That's only 234,000,000,000,000,000 picodollars!
On average. Factoring in growth 2023 annual profits is probably ~$500k. Probably $1m/yr is achievable, great success for such a simple (in a good way) and cool project, and very inspiring for the rest of us!
As the sibling pointed out, that would be the per year amount, not per month.
But also it's safe to assume Tarsnap's December profits have been steadily growing over the past 14 years (or at least for some of those years), so the number is probably a bit higher than $275k/year. $275k/year is more like the average profit for the past 14 years. (Even then, that assumes that the profit numbers are roughly the same every month, which may not be the case.)
people always complain that companies don't "do good", but a little-known fact is that this is often down to this being actively discouraged by governments.
in most countries, it is basically impossible to write off donations as expenses. this means the company have to pay taxes for money that isn't even there anymore.
if people want companies to support good causes, the first item on the agenda should be to lobby for changed tax laws.
It makes sense to me. We tax companies on the money they make in excess of what they needed to spend to operate, and deciding not to make a donation would have no impact on their income. I would not expect it to be considered an operating expense.
In personal income taxes, I have always seen the deductibility of charitable donations and (in the US) mortgage interest as being a form of social engineering -- donating to charity and home ownership are individual behaviors that the government wants to promote, so it effectively subsidizes them for most people.
Businesses are generally run to maximize profit so you could assume a deduction for charitable giving would be immediately weaponized by creating a charitable arm of the company and donating most profits to yourself, or something equally repulsive.
I see the deduction of mortgage interest on owner-occupied property as putting prospective owners on closer to equal footing as prospective landlords when competing for housing. (Business interest is a deductible expense.)
I don’t want the net financing costs for a rented property to be significantly lower, which would serve to help a landlord outbid an owner-occupant.
(Yes, there is also a social engineering component, but there’s a practical economics-based argument for it as well.)
I feel open source donations could be seen as an expense. They are sort of paying for the development of software they presumably find useful. It's at least a start on fixing the problem of the commons type issue of open source benefiting multiple companies so they all want the other one to pay for it.
> Businesses are generally run to maximize profit so you could assume a deduction for charitable giving would be immediately weaponized by creating a charitable arm of the company and donating most profits to yourself, or something equally repulsive.
Aren't there a bunch of groups like for example Mozilla which have both company and non-profit arms? And yes it could be abused, but is a scheme that just donates money to a charity that has the company owner for a CEO and does nothing but pay high salaries is already forbidden?
It might also be an issue too, but it’s not actually a hard blocker.
Just ask the author for a proper onetime invoice. Maybe not all OSS authors will do it, but I woud be happy to make a quick PDF for a nice buck. :)
Now the real reason that doesn’t happen more often: Our financial departments and/or bureaucracy inside the org are actively hindering frivolous acts as being nice to others.
> If you're at a startup which relies heavily on open source software, please take a moment to ask yourself: How much does your company contribute back?
And this deserves emphasizing.