Serious question, what would the Linux foundation do with 5bn/year?
[1] https://www.coreinfrastructure.org/ [2] https://lwn.net/Articles/600571/
How about exploring the frontiers of open hardware (RISC-V)?
The back-of-the-envelope calculation from an LWN comment[+] (on the discussion "Is it time for open processors?") by Jon Masters, jcm, of Red Hat (but speaking for himself):
[quote] The problem with an "open" processor is the cost. It costs over $1.2 Billion to do a ground up 4 year OoO core of the kind of competitive performance that others have built. And even then, once silicon is deployed, security issues can still be found. In short, it will never happen. You'll get RISC-V cores like BOOM, you'll see some wonderful IoT designs, but you'll never see a high end Xeon-class core unless some billionaire funds it as a pet project, and keeps investing year after year for the greater good. And no, none of the major search/cloud vendors are going to go fund this - they want to make commercial vendors compete. [/quote]
I think it's a really "long shot", but we can dream. :-)
at which point they proceeded to shit on those devs for being so awful at maintaining a critical project. It was nasty and unwarranted.
GL driver, openGL Lib, etc all open source for all GPU.
Manage contributions from Intel, AMD, ARM, NVidia and other GPU silicon IP vendors.
The simple goal of opensource, advance the GPU driver, libraries for the benefits of everyone in the field.
After that, maybe define, manage the evolution of "AI" driver, library layers between the TensorFlow (or other AI framework) and the CPU, GPU, TPU.
Alternatively, its also possible the administration can go corrupt when there is so much money flying around. Non-profits can do mismanagement their funds.
The average person deserves to know how badly charity, foreign aid, and government programs can often be at improving human welfare. Not always, but often.
1: https://www.npr.org/2015/06/03/411524156/in-search-of-the-re...
The reason? No, at least one major reason is they need to diversify their revenue streams. Google for example has been desperately trying to find something to offset their lopsided dependence on ad revenue for ages, because their party will come to a grinding halt real quick if they lose a large share of the eyeballs. Why do you think Google has given so much money to Apple to stay on their platform as the search provider? It's not a good position to be in.
I mean, I get what you're saying... but honestly making 15b profit per quarter is really a pretty good position to be in.
A long while back, my city paid for a billing system. Was something like 30, 40 million bucks. They did not own the code, or if they did, were not in a position to make that the benefit it should have been.
Long story short, they ended up paying a lot more for support, updates, bug fixes and the like. And then did it again to finally arrive at something workable!
A group of people, myself included took this to government as an example where open code could be a much better investment.
For that same sum, an entity could be created to develop that code, making for local jobs, local revenue and a very likely savings over the long life possible.
Maintenance fees could provide the entity with the staff needed to maintain that code, until it pays for itself.
Additionally, that entity could make it available to other municipalities, and or consult on that problem domain. That turns what us currently a cost center into revenue that could, at a minimum, lower the cost of the utility currently paying way too much for software not really adding that much value.
We got bipartisan support. Industry lobbyists killed it off back then. Our bad. At the time, the thrust was to get OSS written into law as a must consider. The industry arguments and ability to lobby won the day, and were compelling to legislstors who were not well equipped to really grok the potential benefits.
What we shouod have done was push for and more fully develop the plans and non profit model needed to sell it as a project, or investment.
This all happened in the early 00's too. Its a bit different and even more plausible today, and legislators are more likely to understand it all better too.
Open code would mean other projects could build on that effort, and open data would mean the business of the people would be available where and how needed.
A project of this kind would end up sustainable and would yield consistent savings to every municipality using it, whether they developed it or not.
Once done, its a model for similar needs and costs and could be repeated anywhere. An umbrella foundation could unify these things, or several sprout up, each owning a needed domain, each making revenue and ending up self sustaining.
Over time, there would be various entities of this kind solving not sexy, but quite expensive and painful problems that really do need solving.
This kind of thing is not well matched to investors and founders looking to make big returns. The primary value is making public works, utilities, etc. lean, efficient.
It's also not well matched to proprietary software houses, who may well deliver great solutions, but at high margins and with closed and often expensive dependencies. Their solutions could be a bust too.
It's high risk.
Oregon vs Oracle comes to mind.
This kind of thing is well matched to a non profit type effort, or foundation and could easily be seen as a public work, of benefit to everyone, in that taxes, the cost of running the basics in society would be reduced for everyone.
Cheaper, efficient government basically.
Redirect the savings into things like potholes, schools, health care maybe.
With these kinds of resources, or even a modest fraction, a whole lot of good can be done.
Pair developers with domain experts and legislators, city council people and solve some problems. Each solution woukd likely displace closed, expensive software and doing that would be a public good where it's most needed.
While doing this would not make the Microsofts, Oracles of the world happy, it would mean a lot of people get to make a great living, and we all see the benefit of a more lean, efficient use of taxes actually adding value instead of mostly cost to our lives.
Thanks was a great story. Thanks for sharing.
I think personally, mandating open data is probably easier, and could be successful.
Industry lobbies will beat that open code mandate back hard. And I'm not sure we can win on that basis even today. Open data however, is a no-brainer.
I think the real answer is to just compete. If one of these projects were started, using the model that I put in my other comment, it would compete very well. I don't think it's even really a contest. And from there the savings and the economic forces would push it where it needs to go.
That would set the market opposition to the side, where it really should be. It may well be that closed proprietary solutions really are the best, and this would be one way to find that out, and find out where it's appropriate and where it's not. That's a big win for everyone.
Not to mention the numerous opportunities to take what that energy is learned in built and apply it all over the place Consulting, handling special cases and what not.
Ended up doing it all twice.