> Google hopes to “standardize machine learning on a single framework and API,” namely TensorFlow, then supplement it “with a service that can [manage] it all for you more efficiently and with less operational overhead,” namely Google Cloud.
"vendor on-ramps" is the corporate equivalent of the open-source/pro split commonly seen in non-corporate projects like sidekiq: A bunch of features are available in the open source domain with liberal licensing, and beyond that there are additional paid enhancements. Nothing new or surprising here
What's different is that the corporate offerings used to be extra features or monitoring or support--things that were sometimes nice, sometimes must haves.
Now the corporate offering is the ability to run the code in a way that makes it useful. Setting up tensorflow is easy (at least, when I looked at the docs it seemed to be), but running tensorflow so you get useful answers out of it in a performant manner--that requires knowledge and data that can't be encoded in code.
So the corporate offering that is tied to open source goes from being a nice to have (or perhaps a requirement) to a hard requirement.
Software is complex, costly, and time-consuming. Perhaps some individuals are altruistic enough to take it on themselves without some sort of monitization, but in a capitalistic society, most open-source will tend towards profitability (in some avenue). And that’s just fine with me. We all have benefited greatly from the model.
This is a problem because it results in vendors with huge budget to dominate the open source space. This in itself is fine, but the side effect is that it discourages all the grass roots innovations that could have happened, and effectively kills off all the small time competition with no budget.
normally these large vendors have every interest to stay away from implementing anything that undermines their power in the ecosystem.
When that happens, people can “fork” out of the project since it’s open, you may say.
But imagine even thinking about forming something like react.js or tensorflow. Unless you have very strong motivation it would be very discouraging to even start.
So I think the problem with these large vendors dominating open source is that when things come to this state, we don’t have an alternative and everyone is locked into the “open source” projects given to us by those small number of companies. Not so much different from their own services dominating consumer market share.
>But imagine even thinking about forming something like react.js or tensorflow. Unless you have very strong motivation it would be very discouraging to even start.
Altruistic individuals still need to eat and pay rent. And if they have full time job, then they have few hours max for open source.
Larger project being sponsored by a company or having some kind of monetization seems inevitability to me. In the long term, open source developer is hungry too.
I don't know. All of the stuff that they are talking about in the article are things that are of interest to business... message queueing and iot and databases. I can see having the source code as an insurance policy against the vendor going bust, i don't see how these projects are of interest to people.
The number of open source projects that I think actually matter in the cyberpunk dystopian future we all live in is really small. I stopped believing the hype about open source a long time ago. All that change the world BS makes everyone feel great...you are a revolutionary because you use VI or Debian...without having to do anything that might cost them something. Witness all the helplessness and rage expressed when distros went to systemd. Literally source available, more choice than ever, and Linux running on billions of devices...but it's never quite good enough. How dare you oppress me!
I don't know which freedom we are talking about anymore. It seems to be available either in abundance or not at all depending on who you ask.
This article is of low quality. False dilemmas, appeal to authority and reasoning by analogy.
Even the on-ramp Kubernetes is now preferred container orchestration tool across cloud providers and on-premise. A more apt article would be how the big three cloud providers co-opt Open Source into managed services.
Isn't that a good thing, too, though? I personally love that I can spin up Redis or Postgres or whatever as a managed service on AWS. I care about using the software, not staying on top of patching the server running the software.
Adobe Experience Manager (AEM/CQ) is at its core open source. But you're right in that manner, that they don't promote these projects or help them to grow.
Some orgs go a little bit too far with this. Plotly for instance puts a publish button by default which is a direct tie in to their cloud data hosting. You can turn it off, but being opt-out instead of opt-in is a little disingenuous for an open source project in my opinion.
It its being open-source matters to you, you can go and alter the source to your taste and remove the button. (I'm not even talking about contributing something generally useful upstream.)
If someone just wants some piece of software for free, but does not care to have a clue, then maybe they're better served by a paid subscription.
Exploited = taken advantage of, to their detriment. How is this the case with a link placed somewhere?
People without both a clue and a desire to spend time on acquiring it are best served by some hand-holding and solving their specific problems for them. Unlike the code, this can be a paid service, on which they can spend money instead.
The link by itself doesn't come with a clear description of the security, privacy, legal, or licensing implications of its use, even accidental use, and could lead to unintended public data disclosure.
>Spanner depends on TrueTime, a powerful way of coordinating resources in disparate places. No one but Google has this technology, although AWS and Microsoft could conceivably build it. Pretty much no one else on Earth could.
This is what I thought when the Spanner paper first came out. As I've worked in industry, I've discovered that far more companies than you think have access to the hardware necessary to support such behavior. If the interest continues, I'd expect this hardware to become fairly standard on enterprise gear.
Kubernetes is also developed with the goal of preventing vendor lock-in... It is an on-ramp for "the cloud" in general, but there are many details that will decide where you run your cluster.
Another perspective on this topic by the CTO of influxDB:
My talk, titled "The Open Source Business Model is Under Siege," is about the existential threat that Open Source Database companies, or any open source software infrastructure, are facing from cloud vendors like Amazon Web Services, Google Cloud, and Microsoft Azure. The talk is based on my own experience building a business around open source over the last three and a half years with InfluxDB and our other projects.
> Google could open-source everything tomorrow without any damage to its revenue, but the code itself would provide other providers and enterprises only limited ability to increase their revenue unless Google did all the necessary prep work to make it useful to mere mortals not running superhuman Google infrastructure.
That's a poor excuse. Free software has more benefits than availability.
> AWS, Microsoft, and Google are all racing to figure out how to turn their innovations into open source on-ramps to their proprietary services
I'm sorry but these companies neither started the Open-Source movement nor do they define it today. Microsoft, in particular, was openly hostel towards the Open-Source community for at least a decade. Today's Microsoft has embraced Open-Source in some fantastic ways but still it is hard to forget comments like these:
> "Linux is a cancer that attaches itself in an intellectual property sense to everything it touches," former Microsoft CEO Steve Ballmer
> "A common trait of many of the companies that failed is that they gave away for free or at a loss the very thing they produced that was of greatest value--in the hope that somehow they'd make money selling something else," Microsoft Senior Vice President Craig Mundie.
"vendor on-ramps" is the corporate equivalent of the open-source/pro split commonly seen in non-corporate projects like sidekiq: A bunch of features are available in the open source domain with liberal licensing, and beyond that there are additional paid enhancements. Nothing new or surprising here