I LOVE the term “grass computing!”
This slide deck is from a talk I gave at a conference called border:none in Nuremberg, Germany last year:
It goes into a bit of the history of how we got here and why everything’s become so centralized. I think economics is only part of the story.
As far as funding goes, three of the projects you list are funded to some level by angel or venture capital: BitTorrent, ZeroTier, and Sandstorm. I think OwnCloud, which you didn’t mention, is funded too.
Some capital is going into this stuff, but it’s a very tiny trickle compared to what gets invested in centralized systems. That’s not because of any ideological agenda. It’s because centralized systems usually get more users (due to better user experience mostly) and make more money (for multiple reasons).
The only alternative would be to have government fund all this stuff. I’m not sure how folks elsewhere feel, but I don’t trust the U.S. Federal Government more than I trust VC firms.
I do wonder if the tide is turning. Centralized/decentralized is basically the old mainframe/micro cycle of reincarnation. First we used dumb terms to access mainframes. Then we used PCs. Then the Internet has made it easy to access really big mainframes in “the cloud,” which is really just marketroid rebranding of mainframe computing. (All the tech, like containers and VMs, is old mainframe tech reimagined/rebooted.) Now we’re starting to see some efforts to push things back toward personal computing again, albeit with a different more networked model from the old grey box PC.
There are aspects of the cloud that aren’t going away though. At central data centers it’s possible to achieve economies of scale that can make things like storage and compute cheaper in bulk there than they are in a distributed system. They can also be more reliable. I host many things in the cloud because it almost never goes offline, while my home Internet connection is much more flaky.
What we really need — and many many people including you have hit on this — is to find ways to change the technical power dynamic. The cloud should be the slave and the PC should be the master. The cloud should be encrypted “zero knowledge” backup, storage, and support infrastructure for personal computing.
The way Apple uses cloud is close to what we ought to be trying to build in the more free/open software community. Apple uses cloud exactly this way— as dumb storage and support for rich endpoint devices. They just don’t do the open or encrypted / zero-knowledge part, as they have no incentive to do so.
As far as economics go, I’ve started to believe increasingly that free “as in beer” is in fact the enemy of free “as in freedom.” This will be true until and unless we actually make it to some kind of post-scarcity society where production can truly be decoupled from powerful economic forces.
I’m not saying we need toll booths everywhere, but we need them somewhere. There must be some mechanism for open and free “as in freedom” projects to finance themselves. Otherwise they’re at a permanent disadvantage.
Free “as in beer” also encourages the development of dishonest pseudo-free business models, such as the use of surveillance to monetize free services. In the late 90s, many services and online media outlets tried a paid or freemium model. People mostly refused to pay. Meanwhile the services that opted for pseudo-free models built around stealth monetization strategies like surveillance grew exponentially and took over the market.
This is nothing all that new. Media has always been a deflationary race to the bottom, giving advertisers and astroturfers/propagandists the most power over what gets made and shown. The difference now is that now the media is bidirectional: it watches you as you consume it. In the past the cost of free was ads/propaganda slipped into your news and entertainment. Now the cost of free is total surveillance.
Thanks for posting! I enjoyed the link to Telekommunism, "venture communism" is a fun phrase.
You focus on funding, it's easier to monetize centralized services so more venture capital is poured into creating them. The other post mentions technical difficulties, distributed systems are hard and getting them to work even when you trust each piece to act in good faith is difficult. Writing a distributed system which is also a good product is harder than simply writing a good product; centralization helps you outcompete.
How might we get around these pretty substantial market forces?