> Germany and France attacked Twitter Inc. and Facebook Inc. after U.S. President Donald Trump was shut off from the social media platforms, in an extension of Europe’s battle with big tech. German Chancellor Angela Merkel objected to the decisions, saying on Monday that lawmakers should set the rules governing free speech and not private technology companies.
> Rights like the freedom of speech “can be interfered with, but by law and within the framework defined by the legislature — not according to a corporate decision.”
As much as I support this pure democratic view of Angela Merkel, and as much as I hope Ben Thompson is right with his Internet 3.0 "Return of technology" and "open protocols" idea to have a counterweight to big corp, I think it's really difficult to escape Internet 2.0 economics.
e.g.
EU Cookie law - Good intention and poor implementation.
Whatsapp vs. X - We all on HN know, we want better, but network effects are really strong.
Merkel's view might be interpreted as pro-free speech at a glance, but the corollary of what she's saying is that no single entity should wield as much power as Facebook, Twitter, and Google at all. It's to be interpreted in the sense that additional legislative weight should be put behind disrupting the quasi-monopolistic dominance of these entities.
Having followed Angela Merkel's comments in the past I did not read her comments as "pro free speech". I read them as a concern about power. What happened this week, in the long term, will likely be looked on poorly by history.
> What happened this week, in the long term, will likely be looked on poorly by history.
It might well be, but it's worth keeping in mind that this isn't mutually exclusive with what happened last week. I don't just mean the armed insurrection in the Capitol with the tacit support of a sitting president who lost his re-election bid; I mean that, for a full week, no government agency even made an official statement about that insurrection.
The real story of January 2021 may well be that private companies have stepped in to take action not merely because they could, but because the government refused to do so. While I share Ben Thompson's discomfort at private entities having this kind of power over the public sphere, the even more uncomfortable truth is that we -- both (primarily American) citizenry and (primarily American) government -- have ceded that power to them.
This is just narrative. Immediately after the Capitol riots people were trying to create narratives that the FBI and Capitol Hill police intentionally didn't do their job. That turned out not to be true, as the piece by Brian Stelter showed. This reasoning seems to be a further manipulated form of that. I would not expect the American government to be making public statements. Generally it's the president that addresses the public and obviously in this case he didn't.
I didn't mention the FBI or the Capitol Police. There was an armed attack on Congress while it was in session and the federal government has not made an official public statement about it. Maybe you think there's nothing remotely weird about that. I do. Maybe you think no other administration would have had multiple briefings from law enforcement by now. I don't.
> Maybe you think there's nothing remotely weird about that. I do. Maybe you think no other administration would have had multiple briefings from law enforcement by now. I don't.
I do not. I would expect that in any terrorist incident, like 9/11 or others, that the President would be orchestrating a response. Clearly the President has been implicated in these things so that's not going to happen. You have Congress preparing impeachment documents and the FBI has responded to journalists, many of which created immediately hostile narratives about law enforcement. I do not know what else you're expecting them to do at this point.
If I were the FBI I would not be saying anything either. I'd have agents out in the field collecting evidence and arresting people, getting the story and having them turn in their coconspirators. If you make some sort of statement it prompts them to destroy evidence.
All right, with the longer explanation I see where you're coming from on this. I would still stand by my observations (the part you quoted), though; there hasn't been even the most anodyne public statement expressing sympathy, calling for unity, vowing to make a full investigation, or the like. That this administration may be resisting making such statements because they are implicated in the attacks is pretty extraordinary.
> It's to be interpreted in the sense that additional legislative weight should be put behind disrupting the quasi-monopolistic dominance of these entities.
I tend to agree with your read, i.e. this has nothing to do with freedom-of-speech type issues and everything to do with Germany (and Europe more generally) positioning themselves against Big Tech; my only question is why now.
I'd love nothing more than to see Facebook/Twitter take a beating, but in this particular case there isn't really any strong argument that the government should have intervened and prevented Twitter from blocking Trump. Is this just an extreme case of carpe diem?
Honestly I really struggle to see why this would even be 'democratic'. There's a pretty strong and convincing argument that it would be better if Twitter hadn't blocked Trump, but the notion that the government should be allowed to force at gunpoint a private entity to amplify speech that such entity disagrees with doesn't strike me as particularly democratic.
I get it, Twitter bad, I agree. But the implications of this idea are frankly much scarier than any "corporate decision" will ever be.
With constitutionality free speech the government shouldn't be allowed to force a private entity to amplify or censor speech. And don't get me wrong. I was happy about the ban in this moment.
On the other hand I wouldn't like to give all moderation power to private entities alone. If not opportune with current business model, company ethics are quickly changed (e.g. don't be evil). As long as you have small decentralized shops and platforms that's ok. With concentration of power a private company nearly acts like a utility. Maybe some kind of neutral and elected ethics committee could help large private platforms to maintain transparent and democratic standards. Would they have blocked him even earlier?
> With concentration of power a private company nearly acts like a utility.
I agree this is a problem. I believe the more rational way to solve it is to break the monopoly, i.e. using antitrust powers more aggressively and letting the market decide, rather than having some committee decide what's kosher.
> elected
Holy cow please no. I'm willing to believe you have the best of intentions, but anything elected would 100% become a stupid political game from day one.
And even if it didn't, popular votes on issues that potentially impact individual rights are a terrible idea: if 51% of the public votes $VERY_BAD_THING, do we have to go along with it? We enshrine fundamental rights in constitutions precisely because we don't want them to be endangered by the current political wind.
> the notion that the government should be allowed to force at gunpoint a private entity to amplify speech that such entity disagrees with doesn't strike me as particularly democratic.
"Private entity" is a very broad category that encompasses everything from individual citizens/entrepreneurs to trillion-dollar multinational corporations with armies of shareholders, lobbyists, and lawyers.
I think there are plenty of scenarios where the law should discriminate between the latter and the former, and this is one of them.
> I think there are plenty of scenarios where the law should discriminate between the latter and the former,
I don't see why this is the case. Private entities are made of people. If Twitter vehemently disagrees with something, I don't see any reason why the government should force them to go against their wishes.
> very broad category that encompasses everything
This is exactly the problem. While there is an argument that Twitter was wrong in the specific case, the implications of having the government force Twitter to say/amplify things they don't believe are __chilling__. Restricting speech is bad enough, but often understandable, this is frankly several steps beyond what I'm comfortable with.
From a strictly utilitarian perspective, Twitter's actions generated backlash that was probably avoidable had they continued with their previous policy of placing a label that basically said "this guy is an idiot" on every tweet. This obviously has to be balanced against the damage caused by letting him break very rule without (apparent) consequences. I'm not sure where I stand on this issue, but I can see an argument for both sides.
This is however a completely disjoint topic. "Should they have done X" and "Should they be able to do X if they so choose" are very different questions.
From my perspective, we're still talking about the actual events of Jan. 6, rather than whatever inane thing Trump would have tweeted this morning to deflect that conversation. In my mind, that's a HUGE win that far outweighs any backlash. I also, personally, wonder how overstated that backlash actually is. I don't know anyone IRL that is lamenting the fact that Trump lost his Twitter account.
If you're deemed a common carrier, you can no longer exercise full editorial control over the content you're carrying. We've gone back and forth on whether ISPs are common carriers or not. Twitter is a step even further, but possible.
Sure, but does the comparison really hold? ISPs provide a service that's strictly tied to very expensive and hard to duplicate infrastructure (often with strategic significance, even). To make an even more extreme example, if the companies controlling the North Atlantic TAT cables suddenly decided to arbitrarily deny service we'd have a huge problem, nobody is denying that.
But social media are literally a database and some javascript, that's not even remotely in the same league.
Again, I do agree we have a quasi-monopoly problem; but if we do, the logical solution is to break the monopoly. Imposing political control creates more problems than it solves.
Agreed that social networks are not natural monopolies due to physical characteristics, but the network effects are still significant barriers to entry. Another angle is mandating interoperability of some kind, such as mandating that mobile carriers had to support porting of phone numbers. None of these ideas seem like a slam dunk, though.
This is the usual hypocrisy of european politicians.
Things are ok and not problematic when it is them that do it because they have the power. But when they are the subjects of similar things, then they don't like and want to have this power for them.
There is a very good and ironic example of that in France:
The former president Nicolas Sarkozy created and pushed a lot of nefarious 'security' laws when he was president. For example, the possibility for police to monitor phone calls without a warrant and things like that.
To critics, he was replying that the state is 'trustful' and that only bad people could fear for their privacy.
Back now, a few years later, police wiretapped a phone line that he opened under a fake name to secretly discuss about another police investigation that is currently targeting him and he allegedly used this line to abuse of authority and try to get insider knowledge from law officials in exchange for a special position.
At his trial and in medias, he cried everywhere that it is unfair and abused that his phone lines could have been wiretapped like that to provide evidences against him. Like if he is a victim and not the person that pushed these bad security laws against the population despite a lot of critics of people concerned by freedom and privacy topics.
I don’t know that I’d call it hypocrisy. Most European politicians who’ve commented on this seem to be expressing a strong, principled view that the state is on top and nobody should be allowed to exercise power over it. This sounds strange to a lot of us in the US, where we don’t generally believe the government deserves special respect, but it’s our attitude that’s atypical from a global perspective.
I can understand that from far you could have this impression that they have good will and for them "the state is on top and nobody should be allowed to exercise power over it".
But make no mistake, this is just communication/propaganda and what makes me say that it is hypocrisy. Politics here are champions of double talk. Despite pretending to be democracies, a lot of leaders are now trying to grab the maximum power and undermine citizen decision power.
For example, in France, normally the President and government is just here to execute the laws decided by the national assembly. But in the past decade, majority members of the national assembly are now in a party whose purpose is to "support the president" and so, you could be excluded if you would not vote like the president want you to.
Also, more and more the government decide new laws unilaterally, sometimes in secret or after secret negotiations with lobbies, and will do everything needed to force the assembly to approve it.
Sometimes it is just pressure, sometimes it is manipulations like presenting multiple time the same law, even if it is rejected, until it will pass. Or a present it at a specific time, like at night when there are other events, so that opposition will not have time to come to vote.
They are also more frequently using anti democratic tools when they can't manage to have their law to pass, like something call 49.1 that enact a law without vote of the national assembly.
And lastly, we have seen the case a lot with "fake news" and "hate speech", where governement or governement member will spread "fake news" or send bad "hate" messages.
But when you have breaking news of bad behavior of them, then they will pretend that it is "fake news"/"hate" message, and that the state should be able to censor that.
In this regard, they are very similar to Trump.
To give one last example, during the first part of the covid crisis, the government knew that they did not have enough mask, because of bad management, and instead of telling the truth, they said that pharmacy were not allowed to sell them, because people would not know how to use them and that they are useless to deal with the covid.
Later the proof was given that they were voluntarily lying.
We get to vote for who represents us in government, do we get to vote for which laws Twitter , Google, FB choose to enforce? Can we vote with our wallets out of this one? Why does it feel like advertisers crowned themselves arbiters of truth and a lot are ok with it because they currently don't like the same bad things.
Govts are of the people for the people, corporations are for making profit over anything, do we hate people (ourselves) this much?
> This is the usual hypocrisy of european politicians.
If a chimp somehow learned that 2 + 2 = 4, would you point out that given he is a chimp and "chimps don't not math", that the statement about the subject is wrong, even that is clear it has a merit in itself and who says what basically don't matter when we think about something being right or wrong?
This is not as simple as 2+2 of course, but i rather prefer that the merit of what being discussed is taken the proper focus while who says what, only have more prominence when
hidden intentions that can actually cause harm cant be neglected.
And i think this is clearly not the case unless your name is Mark Zuckerberg, Jack Dorsey or anyone who will profit from this digital neo-feudalism power grab.
People seen to forget easily how and why sovereign states with the rule of the law were built, and how, if we forget the lessons of the past, it will be very hard to get out of a state where we all have no recourse against our new lords once we go through this path of powerless, anemic sovereign states.
I don't see a way out of this that doesn't include government-funded technology services.
If Twitter is a free speech platform to be used by the people, there is no way to exercise that consistently without violating Twitter's rights.
I know the US built a public postal service because we saw mail delivery as a requirement for a functioning country. Is the infrastructure necessary to run an Internet now falling into the same category?
And not just the cables, but the routing technology, hosting technology, etc. More like AWS than Comcast. Where is the line, other than for protected classes?
I would much rather see something more decentralized. It seems like politics is increasingly looking like whichever party is in power uses their power to get more power.
There is definitely an argument to be made that the Internet is so much ingrained in society that denying people Internet access is like denying people access to grocery stores and electricity.
Open protocols, or non profit open source projects will never compete with true entrepreneurship and hardcore entrepreneurs creating products. They just can't move as quickly, or deliver good enough products.
What I think internet 3.0 will be is websites and services segregated politically, just like the media is. Just like we have CNN and Fox, we will have Twitter and [Parler?] Facebook and Rightbook.
It will take time to build. But the assault on half the country will result in the market taking care of this. Just like parler almost did.
(It's not clear exactly what Ben means by his internet 3.0, his article could be interpreted this way as he says "decentralised".)
> non profit open source projects will never compete with true entrepreneurship and hardcore entrepreneurs creating products. They just can't move as quickly, or deliver good enough products.
How do you reconcile that statement with the absolute dominance of Linux? I don't think they this is universally true. Maybe with the qualification "... for consumer products".
I think the rest of your comment is probably, sadly, accurate.
Seems to me Linux success is largely due to commercial vendors developing many features as aggressively as they would do for closed source products. If it were just engineers working gratis on saturday afternoon Linux wouldn't be where it is today.
Isn't that an argument for it being free and open? Whether contributors get paid is orthogonal to freedom and openness.
It's hard to imagine what would have happened if there was no GNU Linux, Apache, SQLite... [insert gigantic list of open source technology that almost everyone in our field relies on].
I think the events of the last week will give a boost to things like ActivityPub. It has over a million active users, and two of the big projects (Mastodon and PeerTube) came out of EU countries. They have their problems, but if I were an official looking to reduce dependence on the US, that's where I would start.
You are presenting true entrepreneurship and open protocols as a false dichotomy
This has been changing for some time now: MongoDB switched to an open source model a decade ago, and a number of companies have followed suit. The strategy is hard to pull off, but even in the case where Docker the company very much failed, Docker the technology did not.
The next wave of these are blockchain companies. I think the jury is still out on these open source companies becoming breakout successes, but they did succeed so far in one critical aspect: the funding model. These companies have tens of millions of dollars at their disposal to build open protocols, and a plan for how to make more once they ship usable products.
Yes, building centralized tech is easier. No, that doesn’t mean it will inevitably win. These things take time.
>
MongoDB is free and the source is available.
Versions released prior to October 16, 2018 are published under the AGPL.
All versions released after October 16, 2018, including patch fixes for prior versions, are published
under the Server Side Public License (SSPL) v1.
You're not wrong, but the reason is more fundamental: open source and open protocols do not have an economic model sufficient to finance the immense amount of brutal work required to make products that are easy to use and appealing for the general public. FOSS and open systems will always lose (outside nerds and certain power user audiences) because closed systems have an economic model in the form of either paid SaaS, commercial licenses, advertising, or surveillance capitalism (or some mix of these things).
... or in the case of Parler perhaps paid propaganda financing from far-right wealthy financiers...? Of course political propaganda is a form of advertising so it fits under the adware / surveillanceware category.
In software, making something work in the algorithmic and technical sense is usually less than a quarter of the work. In the case of social networks and chat apps and similar stuff it's probably only a single digit percentage of the work. The vast majority of the work is in the domain of user interface, user experience, and product design.
Even worse, this is usually the type of work that programmers don't enjoy. It involves endless tweaking, endless rewrites of user interfaces to hone in on something people like or to chase trends, and the soul crushing drudgery of bug fixes and workarounds to edge cases. Because this work is not fun, people generally must be paid to do it. Because it's the majority of the work, the cost of paying people to do it dwarfs the cost of building the technical underpinnings of the product.
I used to think that the failure of decentralization was a technical problem. I wasn't totally wrong there either. There are major technical challenges around decentralization, but I've realized since then that the lack of an economic driver to finance the UI/UX aspects are the real issue. This issue is not solvable because nobody is willing to actually pay for software.
Does anyone else think building product is not that hard? You make or break on the Alpha release when the direction of the product is set. The hard work comes after you know you’ve got a good bet.
Aside: Some people pay tons of money for software due to addictive game design, social and UX patterns.
But I agree. Building good UX is hard, iterative work and requires a connection of right and left brain thinking via communication or generalists.
Graphic, UI, UX and game designers are typically not as engaged with open source as programmers and engineers. Is this a cultural issue or is it inherent? Ultimately any given UI is a concretion, but I've noticed a strong trend to abstraction in these fields. Design Systems/Language, Design Thinking and so on. I mean the term "UX" is in of itself an abstraction. And the digital tools that designers use today are much more powerful and varied than just a decade ago.
Maybe it is a matter of time until there is more engagement in open source from the design world.
> Open protocols, or non profit open source projects will
> never compete with true entrepreneurship and hardcore
> entrepreneurs creating products.
Nginx and Apache httpd are, in that order, the most used web servers. Both are FOSS, and both serve HTTP, an open protocol. There were definitely attempts to build proprietary protocols, none of which were successful, and in the long run the proprietary web servers ended up second place to the open source projects.
Here in the US, I get reminders via SMS from my bank or my doctor telling me that payments went through or that I have an upcoming appointment. More thorough messages are sent via Email. Every form I've filled out since the turn of the millennium asks for my email address, none of them have asked for my MySpace, Twitter, or Facebook handles.
I edit my code in either Emacs, Vim, or Visual Studio Code -- one of those is an open-source product by a very much for-profit company, the other two have survived decades as powerful editors/IDEs.
If I had to make a prediction, I'd say that over time open protocols will end up winning -- not because of some kind of ideological purity, but because they permit people to build platforms that interoperate with one another. The Fediverse is young yet, and Twitter/Instagram/FB are making adversarial interoperability hard, but the entire history of the Internet and the World Wide Web is one where open has won over the long haul.
Entrepreneurship is not about putting a price tag on everything and making closed proprietary stuff. It's about making a profitable business using the tools at your disposal, which might even be open source (like in the "open core" business model). It may be part of your strategy to open source your product, as many do.
> If I had to make a prediction, I'd say that over time open protocols will end up winning
I think Alan Kay made a prediction to that effect a few years ago, so you're in good company.
I certainly hope you're both right. Patience is required, though.
I imagine it's something of a cycle. Companies try to lock users into their ecosystem, but there are opposing forces that want the protocols to be free.
I certainly hope that the internet of the future looks a lot different than its current form. We regularly hear of data breaches; something that people take in their stride.
When will people wake up that the current model is not a good one? I don't know. It's not certain that they will, but my sense is that current practises are not sustainable long term.
We now have "IoT" devices that are beholden to company servers. That's utter insanity, in my books. Will we wake up to the fact that it is a deeply flawed idea?
> What I think internet 3.0 will be is websites and services segregated politically, just like the media is. Just like we have CNN and Fox, we will have Twitter and [Parler?] Facebook and Rightbook.
That defeats the connection desire of user-generated content. I'm on FB mainly to learn what my friends/family are up to (and see cute animal pix and some maths jokes). If a bunch of them were on {right,left}book and I were on the other I'd miss out as would they.
If the political claptrap self-segregated into verticals that would be fine by me. FB would still work. Of course such arrangements are not stable as terms like "left" and "right" aren't really meaningful (the 1950s "right" and "left" enthusiasts would not recognize how those terms are used today).
>Open protocols, or non profit open source projects will never compete with true entrepreneurship and hardcore entrepreneurs creating products. They just can't move as quickly, or deliver good enough products.
The best products don't win (Windows's monopoly is proof of that, Linux not being based on Plan 9 is another), and then you have the game theoretic threat (lock-in, insecurity, existential jeopardy) of proprietary products thrown in to boot. Those are usually more important for fundamental, backbone kind of things though (OSes, browsers, servers, crypto, etc. -- as opposed to your music player or a game).
Just a nitpick but not "half of the US population" follows Trump.
Not even getting into the fallacy of attributing more than can be to participants of a binary choice dilemma, the American Republic presidential elections are not "1 person 1 vote" like most other democratic Republics.
> Open protocols, or non profit open source projects will never compete with true entrepreneurship and hardcore entrepreneurs creating products.
Meditate on the fact you're posting this on the Internet, as opposed to CompuServe's or AOL's own networks, and say that again.
In addition to the fact you're drawing a false dichotomy between Open Source and open protocols and business, no business has ever created a fundamental communications network of the scale or success of the Internet, and none have even come close even though they tried. AOL, CompuServe, Tymshare... they're all dead and buried or absorbed into the Internet to the point they no longer have an independent existence on the technical level.
You can say this about programming languages: Compare the success of Python to the success of AutoLISP or even the success of Lisp Machine Lisp to the success of Emacs Lisp. I can't think of a single widely-used programming language which is only available in a proprietary implementation. (And by widely-used I mean beyond a single OS and beyond a single industry.)
> The place we compute shifted from a central location to anywhere; the time in which we compute shifted from batch processes to continuous computing.
An interesting case of a true statement that is (imo) actually turning facts upside down.
Yes, it is true, "the place" where we access information and computing services is now "anywhere", since we carry a jazzed-up dumb terminal in our pockets.
All computation is now actually happening in central systems called "clouds". This was Larry Ellision's vision as articulated in late 90s.
The kicker of the analysis is this:
What is notable is that the current environment appears to be the logical endpoint of all of these changes: from batch-processing to continuous computing, from a terminal in a different room to a phone in your pocket, from a tape drive to data centers all over the globe. In this view the personal computer/on-premises server era was simply a stepping stone between two ends of a clearly defined range.
Centralized clouds are not the "logical" conclusion of what happened in the 80s and 90s. Personal computing was precisely that: Personal. Your machine, your code.
My personal speculative understanding of this transitional period in history of computing is that PC caught the policy makers by surprise. What we have now with clouds was the vision of IBM in 60s and early 70s.
The 'black swan' of PC, which was coincidental with public cryptographic breakthroughs (PGP), placed serious obstacles in terms of projected policy goals, which certainly included considerations pertaining to sensitive national security capabilites, such as dissemination of information, and ad-hoc network communication. (Raise your hand if you remember clipper chip efforts, for example).
So yes, the "range" was "clearly defined" per policy recommendations coming out of think-tanks, then PC happened and we spent 2 decades watching the co-option of the PC champions by Big Tech and SV money, and now we're treated to convenient "this is where we were supposed to be folks!" narratives.
But the point I wanted to make was : making web requests is batch-processing. These are short-lived executions of code and it does not matter if the process dies at the end. It can be restarted for every web/batch-request.
Which is something fundamentally different than an application ( video-editor, game etc ).
> Companies, meanwhile, will note the fate of Parler.
This is one of the most significant things that I've already seen. Many discussions in the past few days about how this represents a business risk. E-mails from providers about TOS updates, some even saying that if a client gets caught up in something or indicted, we get kicked. Many similar stories from others; companies starting to invest in at least making their tech multi-cloud or multi-cloud-ready. This may drive a shift from cloud computing as well: it's more difficult to shut off a company that runs its own data center.
Thing is, cloud computing is actually quite expensive. There's a reason why Amazon, Microsoft, Google, IBM are all lining up to grab a piece of the pie.
Yes there are economies of scale which means overall the cost should be lower, but once you factor in profit margins, I'm not sure.
The main benefit is definitely not requiring as much technical acumen or teams to build your infrastructure, but as more skilled workers enter the labour force, I think that too will start to get marginalised to some degree and teams will shift back to having their own IT departments. Just a guess.
I personally hate the cloud, way too expensive for me and the design choices they do are geared towards profit maximisation and not building a partnership with customers where both win. E.g. so many different modules, each with limited functionality...when I can just run my own virtual server and do everything myself much better
> but as more skilled workers enter the labour force, I think that too will start to get marginalised to some degree and teams will shift back to having their own IT departments
I am not sure if this would the case though. Why would an employer use skilled workers to do something mundane like setting up a kubernetes cluster vs something that actually differentiate their own business from others.
Large and established businesses hire high-paid high-skilled workers to coordinate their legal affairs, maintain their accounting, manage their public image, and control their finances.
None of those things are part of the core business or product differentiation. Yet large corporate orgs still expend enormous amounts of resources on those activities. In fact in most companies the second highest-paid executive is the COO, which pretty much just means the guy in charge of all non-core activities.
I think the simple answer is that most corporates are highly risk averse. Being in startup culture blinds you to this. Startups are comfortable with existential threats in the background, because they're constantly on the verge of failure anyway. They might neglect their liability, public image, or cloud TOS risk. Fast innovation in the core product is a far higher priority for a hypergrowth startup.
In contrast, large and mature organizations are going to be extremely sensitive even to small risks. The typical large company is slow-growing but highly profitable. The number one priority is to avoid killing the goose that lays the golden egg.
Think of the most conservative, bean-counting , button-down loan officer in suburban Topeka. Once "Cloud TOS Risk" enters the business lexicon, said beancounter is likely going to insist that the company in question has a plan, before releasing the tens of millions in loan refinancing that modern corporate America relies on in their leverage capital structures. It doesn't matter how big of an actual risk is. These processes are driven by "check-the-box" CYA theatre.
Partly due to cost-benefit. I am hopeful that there is such a large mass of skilled labour or labour that can be skilled (so many graduates these days), that we can give them jobs and avoid having everything tech infrastructure related being pushed off to the big cloud vendors. As another comment replying to you mentioned, firms hire in-house accountants, lawyers, etc. --> If the cost of tech employees becomes favourable vs. outsourcing to the cloud, then why not internalise?
Its a big risk to put all your eggs in somebody else's hands. With in-house tech, you have somewhat more control. My 2c.
The fact that all the top-line Cloud hosts are really (really!) expensive just adds insult to injury; so perhaps this will start to accelerate the plans of those looking for contingencies. The good news is that we now are in the era of K8s and cheap(er) CPUs (AMD), so there are a lot more options now than just "betting on AWS".
Funnily enough, I'm now starting to think that the only reason AWS proliferated as much as they did was to only fuel the VC capital boom of the last decade and not much else. Startups we able to promise huge valuations and returns based on the belief that AWS would "scale for us" and they didn't really need the inhouse expertise needed to grow. Just sell the growth.
That's an interesting angle and makes a lot of sense - lot of these tech startups are not run by tech people, so having AWS helped them bridge the gap when it came to securing funding.
You have to have a really weird business model or otherwise borderline content to risk getting kicked off from AWS. Sure, a risk for businesses to track, but surely a very low risk (with admittedly high impact).
I think the test is pretty straight forward and is basically an issue of risk and is not really directly related to the internet.
If having anything to do with X creates the perception of Y being liable for any legal and civil improprieties, then Y will dissolve any of its relationships with X.
<my take>
For the parler case, Amazon didn't want to be held responsible for hosting content that (maybe) caused the problems in that nations capital, and certainly doesn't want to be held responsible for any future content/actions.
I think this can be justified by preservation of shareholder value.
</my take>
We don't know what the line is going to be going forward. If it is refusal to take down calls for violence or other illegal content--well that seems pretty fair to me.
But we should be skeptical of what exactly big tech will do this power going forward.
> Megalothymia is “the desire to be recognized as superior to other people”, and “can be manifest both in the tyrant who invades and enslaves a neighboring people so that they will recognize his authority, as well as in the concert pianist who wants to be recognized as the foremost interpreter of Beethoven”; successful liberal democracies channel this desire into fields like entrepreneurship or competition, including electoral politics.
Trump can be described as someone who has megalothymia, wielding his money and more recently politics to assert his perception of himself as 'superior'. This, along with what I opine is blindness of the heart, leveraging this megalothymia as a tool to worship his ego.
These recent days, a number of tech companies, of course made up of people, flexed back when they felt their reasoning outweighed the risks of such a decision. Is this tyranny from these powerful people (Companies et al.)? And thus we are seeing folks (Trump & Companies et al.) taste each others tyranny, taking turns? Or are companies et al. simply claiming a piece of that same echelon Trump claims for himself, as an act of fighting back against a tyrant? Or some third or fourth option... etc.
I cannot help but suspect that such a move would go down in history books as the reigning champion of Reasonable-Sounding Laws With Terrible Consequences.
That is an impossibly broad category. Anything could be construed as political in one way or another. This would result in no one being allowed to remove anyone at all from their property or platform.
This brings up the question - What is the role of government?
Ensure contracts are upheld, to harm bad actors (via prison, fines, etc.) and to protect us from foreign invasion.
so, where did that fall apart in our current American system and what can we do to save our democracy? As the article has hinted at, regulation has gotten too esoteric for a bipartisan split in congress. When you have almost 500 people arguing about the use of SPACS for going public, limits on PPE loan, rights of citizens and business in regard to the pandemic, etc. etc.
John adams said 'God forbid we be 20 years from [..] a rebellion.'
We as Americans need a bit more of a centralized system and move a bit more towards an oligarchy. Same things that plagued the roman republic (Bipartisan politics) are plaguing our system. Augustus' answer? bring it down to him and 30 other people who where the most important people from each 'area' and the only thing that gave him more power than the rest was the fact the he was the only member who didn't rotate. Otherwise, it still came down to voting albeit with a forum that allowed for open discussion without rambling into nonsense and listing to 500 people say the same thing over and over.
I for one have been a huge fan of platforms like Kubernetes, because at least they abstract the cloud away from specific providers to some degree. Had they additionally used IPFS for storage, switching cloud providers wouldn't be much of a burden at all. Most enterprises building redundancy also have to think of what happens if one of their cloud providers go offline.
The problem is that many companies are so invested in AWS that they don't know anything else. Many companies are building their entire IT around AWS. In that regard, unless we create legislation that requires unified APIs and an open standard for cloud services; then we are going to have to start treating cloud providers like utilities.
I supported Twitter kicking Trump off its platform. I think AWS kicking Parler off is practically crossing the line, and that legislation needs to be made to tell cloud providers that they need to create an open standard for their services or be considered an essential utility.
Centralization works for a reason: It makes prices go down.
For a company like Google, Twitter or Facebook it matters very little if they have 100, 1.000 million users or 2.000, or 3.000. Most of the cost is in salaries and the salaries are the same because the technology is scalable.
Once they have the market of hundreds or billions of users, almost nobody can compete with them because even paying enormous salaries the price per user is very small.
The problem with decentralization is that they don't have scale, so it is a serious problem just paying a single engineer salary.
They were built in the early days of computing but now our smartphones are far more powerful than desktops of 5 to 10 years ago.
Moore's law suggests that this trend will continue (unless purposefully downgrading performance to increase repurchases) and with 6G coming, we shouldn't need to be so dependent on the classic datacenter-client architecture wich is the very source of FANG's power.
The new Internet and decentralization completely shifts this power-balance towards the masses. It is democratization of our data instead of being arbitraged by FANG who pay pennies on the dollar to resell it to others.
Is there any significance that he chose "Internet 3.0" rather than "Web 3.0".
I think it is important to distinguish between websites/broadcast SMS (Facebook, Twitter), hosting providers (AWS, GCP) and "app stores" (Apple, Google). I hope readers understand the differences.
Here is a thought experiment: How large of an audience, i.e., how many daily visitors, does a website need to have before an act of banning a contributor becomes a free speech issue. What is that magic number, the threshhold over which the website comes to be perceived as a home for "free speech". Most websites fall below this threshhold. For example, journalists, bloggers, politicians, etc. do not debate hellbanning on HN as a free speech issue.
> For example, we do not debate hellbanning on HN as a free speech issue.
You clearly don't have showdead marked. There are persistent (and, I should say, entirely incoherent) complaints about speech suppression in the gray comments at the bottom of many threads (often still riddled with the slurs that actually got the user banned).
Point taken. I will have to enable showdead and see what I'm missing. However that is not what was meant by "we". I have edited the comment to be more clear.
> Is there any significance that he chose "Internet 3.0" rather than "Web 3.0".
I think so, the closing point of the article is that we are due for a return to decentralization, even if that is spurred on in the short-term by "competing centralized solutions." The primitives required to run distributed/federated services that can compete with the big web giants probably aren't there just yet, but they are starting to emerge.
With cloud hosting providers, K8s is a big step forward, because you can deploy an open-source system end-to-end an arbitrary cloud k8s platform. Part of the pain of switching to more open alternatives in the past has been managing operational deployments which this begins to address. Blockchain technologies, for all their misuse and hype have brought us some interesting innovations such as the InterPlanetary File System (which probably isn't going anywhere but is a start). None of this is better than the current crop of closed system, compare Diaspora to Facebook for example, but we have the start. But you can only have the "Web" applications he envisions if you have the "distributed internet substrate."
To be fair, concerning Western civilisations it's the North Americans who have become the most crazy; always have been quirky, but in the past decade ever more so.
Turbocapitalism might have something to do with it...
Always worth remembering, that sane societies with sane institutions and conduct exist. One has to consider the "human element" if one wants to keep it that way.
I would like to see a rise in decentralized services, but it's not possible while Apple and Google's curated stores are the only way to install apps on a smart phone.
As soon as you can install an app from a file like you can on desktop, then we can have some real progress.
I'm thinking of Apple saying "we won't let you download that app" for my macbook. How did we let them have that power on mobile computing?
It's still perfectly possible (for now) to use an alternative app store like F-Droid on Android phones. It requires sideloading an .apk (literally "install an app from a file") and jumping through a couple hoops (basically: changing your security settings to allow installing apps from a file) to get started, but it's not an insurmountable barrier.
Apple is of course a totally different story, as their devices are 100% locked down walled gardens where Apple is the sole arbiter of what software you're allowed to install. It continues to astound me that a substantial number of HN users see absolutely nothing wrong with this.
> it's not possible while Apple and Google's curated stores are the only way to install apps
I disagree. We have DNS, email, and HTTP. That's a pretty capable set of distributed protocols that have been on those phones since the beginning.
The stores are not a problem here (although they are in other ways); rather, there is a lack of apps that people want to use, or perhaps a lack of people using those apps.
I use Riot (now Element), which uses the Matrix protocol across all my iDevices, and it runs just as well as on my Mac. but hardly anyone I know is on it. It's much more private of a solution that WhatsApp, and although not as refined, it works pretty darn well for my use cases.
If we're talking about technical feasibility, then you're right, it's possible to compete with the app store with the web browser. However, the masses don't care about technology and protocols. It's about convenience and motivation. Any extra step will decrease adoption and engagement on the whole.
All things being equal, if it takes any more energy to do something then, statistically, it will be done less.
84 comments
[ 9.1 ms ] story [ 168 ms ] thread> Rights like the freedom of speech “can be interfered with, but by law and within the framework defined by the legislature — not according to a corporate decision.”
As much as I support this pure democratic view of Angela Merkel, and as much as I hope Ben Thompson is right with his Internet 3.0 "Return of technology" and "open protocols" idea to have a counterweight to big corp, I think it's really difficult to escape Internet 2.0 economics.
e.g. EU Cookie law - Good intention and poor implementation. Whatsapp vs. X - We all on HN know, we want better, but network effects are really strong.
It might well be, but it's worth keeping in mind that this isn't mutually exclusive with what happened last week. I don't just mean the armed insurrection in the Capitol with the tacit support of a sitting president who lost his re-election bid; I mean that, for a full week, no government agency even made an official statement about that insurrection.
The real story of January 2021 may well be that private companies have stepped in to take action not merely because they could, but because the government refused to do so. While I share Ben Thompson's discomfort at private entities having this kind of power over the public sphere, the even more uncomfortable truth is that we -- both (primarily American) citizenry and (primarily American) government -- have ceded that power to them.
This is just narrative. Immediately after the Capitol riots people were trying to create narratives that the FBI and Capitol Hill police intentionally didn't do their job. That turned out not to be true, as the piece by Brian Stelter showed. This reasoning seems to be a further manipulated form of that. I would not expect the American government to be making public statements. Generally it's the president that addresses the public and obviously in this case he didn't.
I didn't mention the FBI or the Capitol Police. There was an armed attack on Congress while it was in session and the federal government has not made an official public statement about it. Maybe you think there's nothing remotely weird about that. I do. Maybe you think no other administration would have had multiple briefings from law enforcement by now. I don't.
I do not. I would expect that in any terrorist incident, like 9/11 or others, that the President would be orchestrating a response. Clearly the President has been implicated in these things so that's not going to happen. You have Congress preparing impeachment documents and the FBI has responded to journalists, many of which created immediately hostile narratives about law enforcement. I do not know what else you're expecting them to do at this point.
If I were the FBI I would not be saying anything either. I'd have agents out in the field collecting evidence and arresting people, getting the story and having them turn in their coconspirators. If you make some sort of statement it prompts them to destroy evidence.
Indeed, but I have no doubt that part will be taken care of.
I tend to agree with your read, i.e. this has nothing to do with freedom-of-speech type issues and everything to do with Germany (and Europe more generally) positioning themselves against Big Tech; my only question is why now.
I'd love nothing more than to see Facebook/Twitter take a beating, but in this particular case there isn't really any strong argument that the government should have intervened and prevented Twitter from blocking Trump. Is this just an extreme case of carpe diem?
Honestly I really struggle to see why this would even be 'democratic'. There's a pretty strong and convincing argument that it would be better if Twitter hadn't blocked Trump, but the notion that the government should be allowed to force at gunpoint a private entity to amplify speech that such entity disagrees with doesn't strike me as particularly democratic.
I get it, Twitter bad, I agree. But the implications of this idea are frankly much scarier than any "corporate decision" will ever be.
On the other hand I wouldn't like to give all moderation power to private entities alone. If not opportune with current business model, company ethics are quickly changed (e.g. don't be evil). As long as you have small decentralized shops and platforms that's ok. With concentration of power a private company nearly acts like a utility. Maybe some kind of neutral and elected ethics committee could help large private platforms to maintain transparent and democratic standards. Would they have blocked him even earlier?
I agree this is a problem. I believe the more rational way to solve it is to break the monopoly, i.e. using antitrust powers more aggressively and letting the market decide, rather than having some committee decide what's kosher.
> elected
Holy cow please no. I'm willing to believe you have the best of intentions, but anything elected would 100% become a stupid political game from day one. And even if it didn't, popular votes on issues that potentially impact individual rights are a terrible idea: if 51% of the public votes $VERY_BAD_THING, do we have to go along with it? We enshrine fundamental rights in constitutions precisely because we don't want them to be endangered by the current political wind.
"Private entity" is a very broad category that encompasses everything from individual citizens/entrepreneurs to trillion-dollar multinational corporations with armies of shareholders, lobbyists, and lawyers.
I think there are plenty of scenarios where the law should discriminate between the latter and the former, and this is one of them.
I don't see why this is the case. Private entities are made of people. If Twitter vehemently disagrees with something, I don't see any reason why the government should force them to go against their wishes.
> very broad category that encompasses everything
This is exactly the problem. While there is an argument that Twitter was wrong in the specific case, the implications of having the government force Twitter to say/amplify things they don't believe are __chilling__. Restricting speech is bad enough, but often understandable, this is frankly several steps beyond what I'm comfortable with.
There is? I haven't heard it.
This is however a completely disjoint topic. "Should they have done X" and "Should they be able to do X if they so choose" are very different questions.
Again, I do agree we have a quasi-monopoly problem; but if we do, the logical solution is to break the monopoly. Imposing political control creates more problems than it solves.
Things are ok and not problematic when it is them that do it because they have the power. But when they are the subjects of similar things, then they don't like and want to have this power for them.
There is a very good and ironic example of that in France:
The former president Nicolas Sarkozy created and pushed a lot of nefarious 'security' laws when he was president. For example, the possibility for police to monitor phone calls without a warrant and things like that.
To critics, he was replying that the state is 'trustful' and that only bad people could fear for their privacy.
Back now, a few years later, police wiretapped a phone line that he opened under a fake name to secretly discuss about another police investigation that is currently targeting him and he allegedly used this line to abuse of authority and try to get insider knowledge from law officials in exchange for a special position.
At his trial and in medias, he cried everywhere that it is unfair and abused that his phone lines could have been wiretapped like that to provide evidences against him. Like if he is a victim and not the person that pushed these bad security laws against the population despite a lot of critics of people concerned by freedom and privacy topics.
But make no mistake, this is just communication/propaganda and what makes me say that it is hypocrisy. Politics here are champions of double talk. Despite pretending to be democracies, a lot of leaders are now trying to grab the maximum power and undermine citizen decision power.
For example, in France, normally the President and government is just here to execute the laws decided by the national assembly. But in the past decade, majority members of the national assembly are now in a party whose purpose is to "support the president" and so, you could be excluded if you would not vote like the president want you to.
Also, more and more the government decide new laws unilaterally, sometimes in secret or after secret negotiations with lobbies, and will do everything needed to force the assembly to approve it.
Sometimes it is just pressure, sometimes it is manipulations like presenting multiple time the same law, even if it is rejected, until it will pass. Or a present it at a specific time, like at night when there are other events, so that opposition will not have time to come to vote.
They are also more frequently using anti democratic tools when they can't manage to have their law to pass, like something call 49.1 that enact a law without vote of the national assembly.
And lastly, we have seen the case a lot with "fake news" and "hate speech", where governement or governement member will spread "fake news" or send bad "hate" messages.
But when you have breaking news of bad behavior of them, then they will pretend that it is "fake news"/"hate" message, and that the state should be able to censor that. In this regard, they are very similar to Trump.
To give one last example, during the first part of the covid crisis, the government knew that they did not have enough mask, because of bad management, and instead of telling the truth, they said that pharmacy were not allowed to sell them, because people would not know how to use them and that they are useless to deal with the covid.
Later the proof was given that they were voluntarily lying.
Govts are of the people for the people, corporations are for making profit over anything, do we hate people (ourselves) this much?
If a chimp somehow learned that 2 + 2 = 4, would you point out that given he is a chimp and "chimps don't not math", that the statement about the subject is wrong, even that is clear it has a merit in itself and who says what basically don't matter when we think about something being right or wrong?
This is not as simple as 2+2 of course, but i rather prefer that the merit of what being discussed is taken the proper focus while who says what, only have more prominence when hidden intentions that can actually cause harm cant be neglected.
And i think this is clearly not the case unless your name is Mark Zuckerberg, Jack Dorsey or anyone who will profit from this digital neo-feudalism power grab.
People seen to forget easily how and why sovereign states with the rule of the law were built, and how, if we forget the lessons of the past, it will be very hard to get out of a state where we all have no recourse against our new lords once we go through this path of powerless, anemic sovereign states.
If Twitter is a free speech platform to be used by the people, there is no way to exercise that consistently without violating Twitter's rights.
I know the US built a public postal service because we saw mail delivery as a requirement for a functioning country. Is the infrastructure necessary to run an Internet now falling into the same category?
And not just the cables, but the routing technology, hosting technology, etc. More like AWS than Comcast. Where is the line, other than for protected classes?
There is definitely an argument to be made that the Internet is so much ingrained in society that denying people Internet access is like denying people access to grocery stores and electricity.
The rule of the law prevents we get back to the rule of the jungle which resorts to the more powerful actors forcing their will against all others.
you are against anti trust, anti labor, anti discrimination laws that were passed by "lawmakers" .
I pray this thinking doesn't go mainstream.
What I think internet 3.0 will be is websites and services segregated politically, just like the media is. Just like we have CNN and Fox, we will have Twitter and [Parler?] Facebook and Rightbook.
It will take time to build. But the assault on half the country will result in the market taking care of this. Just like parler almost did.
(It's not clear exactly what Ben means by his internet 3.0, his article could be interpreted this way as he says "decentralised".)
How do you reconcile that statement with the absolute dominance of Linux? I don't think they this is universally true. Maybe with the qualification "... for consumer products".
I think the rest of your comment is probably, sadly, accurate.
It's hard to imagine what would have happened if there was no GNU Linux, Apache, SQLite... [insert gigantic list of open source technology that almost everyone in our field relies on].
This has been changing for some time now: MongoDB switched to an open source model a decade ago, and a number of companies have followed suit. The strategy is hard to pull off, but even in the case where Docker the company very much failed, Docker the technology did not.
The next wave of these are blockchain companies. I think the jury is still out on these open source companies becoming breakout successes, but they did succeed so far in one critical aspect: the funding model. These companies have tens of millions of dollars at their disposal to build open protocols, and a plan for how to make more once they ship usable products.
Yes, building centralized tech is easier. No, that doesn’t mean it will inevitably win. These things take time.
...and then moved back 2 years ago.
From the https://github.com/mongodb/mongo/blob/master/README :
> MongoDB is free and the source is available. Versions released prior to October 16, 2018 are published under the AGPL. All versions released after October 16, 2018, including patch fixes for prior versions, are published under the Server Side Public License (SSPL) v1.
AGPL is open source: https://opensource.org/licenses/AGPL-3.0 SSPLv1 is not listed, and SSPLv2 does not seem to have been approved either at last mention https://opensource.org/LicenseReview122018
... or in the case of Parler perhaps paid propaganda financing from far-right wealthy financiers...? Of course political propaganda is a form of advertising so it fits under the adware / surveillanceware category.
In software, making something work in the algorithmic and technical sense is usually less than a quarter of the work. In the case of social networks and chat apps and similar stuff it's probably only a single digit percentage of the work. The vast majority of the work is in the domain of user interface, user experience, and product design.
Even worse, this is usually the type of work that programmers don't enjoy. It involves endless tweaking, endless rewrites of user interfaces to hone in on something people like or to chase trends, and the soul crushing drudgery of bug fixes and workarounds to edge cases. Because this work is not fun, people generally must be paid to do it. Because it's the majority of the work, the cost of paying people to do it dwarfs the cost of building the technical underpinnings of the product.
I used to think that the failure of decentralization was a technical problem. I wasn't totally wrong there either. There are major technical challenges around decentralization, but I've realized since then that the lack of an economic driver to finance the UI/UX aspects are the real issue. This issue is not solvable because nobody is willing to actually pay for software.
But I agree. Building good UX is hard, iterative work and requires a connection of right and left brain thinking via communication or generalists.
Graphic, UI, UX and game designers are typically not as engaged with open source as programmers and engineers. Is this a cultural issue or is it inherent? Ultimately any given UI is a concretion, but I've noticed a strong trend to abstraction in these fields. Design Systems/Language, Design Thinking and so on. I mean the term "UX" is in of itself an abstraction. And the digital tools that designers use today are much more powerful and varied than just a decade ago.
Maybe it is a matter of time until there is more engagement in open source from the design world.
Here in the US, I get reminders via SMS from my bank or my doctor telling me that payments went through or that I have an upcoming appointment. More thorough messages are sent via Email. Every form I've filled out since the turn of the millennium asks for my email address, none of them have asked for my MySpace, Twitter, or Facebook handles.
I edit my code in either Emacs, Vim, or Visual Studio Code -- one of those is an open-source product by a very much for-profit company, the other two have survived decades as powerful editors/IDEs.
If I had to make a prediction, I'd say that over time open protocols will end up winning -- not because of some kind of ideological purity, but because they permit people to build platforms that interoperate with one another. The Fediverse is young yet, and Twitter/Instagram/FB are making adversarial interoperability hard, but the entire history of the Internet and the World Wide Web is one where open has won over the long haul.
I think Alan Kay made a prediction to that effect a few years ago, so you're in good company.
I certainly hope you're both right. Patience is required, though.
I imagine it's something of a cycle. Companies try to lock users into their ecosystem, but there are opposing forces that want the protocols to be free.
I certainly hope that the internet of the future looks a lot different than its current form. We regularly hear of data breaches; something that people take in their stride.
When will people wake up that the current model is not a good one? I don't know. It's not certain that they will, but my sense is that current practises are not sustainable long term.
We now have "IoT" devices that are beholden to company servers. That's utter insanity, in my books. Will we wake up to the fact that it is a deeply flawed idea?
That defeats the connection desire of user-generated content. I'm on FB mainly to learn what my friends/family are up to (and see cute animal pix and some maths jokes). If a bunch of them were on {right,left}book and I were on the other I'd miss out as would they.
If the political claptrap self-segregated into verticals that would be fine by me. FB would still work. Of course such arrangements are not stable as terms like "left" and "right" aren't really meaningful (the 1950s "right" and "left" enthusiasts would not recognize how those terms are used today).
The best products don't win (Windows's monopoly is proof of that, Linux not being based on Plan 9 is another), and then you have the game theoretic threat (lock-in, insecurity, existential jeopardy) of proprietary products thrown in to boot. Those are usually more important for fundamental, backbone kind of things though (OSes, browsers, servers, crypto, etc. -- as opposed to your music player or a game).
Not even getting into the fallacy of attributing more than can be to participants of a binary choice dilemma, the American Republic presidential elections are not "1 person 1 vote" like most other democratic Republics.
Meditate on the fact you're posting this on the Internet, as opposed to CompuServe's or AOL's own networks, and say that again.
In addition to the fact you're drawing a false dichotomy between Open Source and open protocols and business, no business has ever created a fundamental communications network of the scale or success of the Internet, and none have even come close even though they tried. AOL, CompuServe, Tymshare... they're all dead and buried or absorbed into the Internet to the point they no longer have an independent existence on the technical level.
You can say this about programming languages: Compare the success of Python to the success of AutoLISP or even the success of Lisp Machine Lisp to the success of Emacs Lisp. I can't think of a single widely-used programming language which is only available in a proprietary implementation. (And by widely-used I mean beyond a single OS and beyond a single industry.)
An interesting case of a true statement that is (imo) actually turning facts upside down.
Yes, it is true, "the place" where we access information and computing services is now "anywhere", since we carry a jazzed-up dumb terminal in our pockets.
All computation is now actually happening in central systems called "clouds". This was Larry Ellision's vision as articulated in late 90s.
The kicker of the analysis is this:
What is notable is that the current environment appears to be the logical endpoint of all of these changes: from batch-processing to continuous computing, from a terminal in a different room to a phone in your pocket, from a tape drive to data centers all over the globe. In this view the personal computer/on-premises server era was simply a stepping stone between two ends of a clearly defined range.
Centralized clouds are not the "logical" conclusion of what happened in the 80s and 90s. Personal computing was precisely that: Personal. Your machine, your code.
My personal speculative understanding of this transitional period in history of computing is that PC caught the policy makers by surprise. What we have now with clouds was the vision of IBM in 60s and early 70s.
The 'black swan' of PC, which was coincidental with public cryptographic breakthroughs (PGP), placed serious obstacles in terms of projected policy goals, which certainly included considerations pertaining to sensitive national security capabilites, such as dissemination of information, and ad-hoc network communication. (Raise your hand if you remember clipper chip efforts, for example).
So yes, the "range" was "clearly defined" per policy recommendations coming out of think-tanks, then PC happened and we spent 2 decades watching the co-option of the PC champions by Big Tech and SV money, and now we're treated to convenient "this is where we were supposed to be folks!" narratives.
Try again.
But the point I wanted to make was : making web requests is batch-processing. These are short-lived executions of code and it does not matter if the process dies at the end. It can be restarted for every web/batch-request.
Which is something fundamentally different than an application ( video-editor, game etc ).
This is one of the most significant things that I've already seen. Many discussions in the past few days about how this represents a business risk. E-mails from providers about TOS updates, some even saying that if a client gets caught up in something or indicted, we get kicked. Many similar stories from others; companies starting to invest in at least making their tech multi-cloud or multi-cloud-ready. This may drive a shift from cloud computing as well: it's more difficult to shut off a company that runs its own data center.
Yes there are economies of scale which means overall the cost should be lower, but once you factor in profit margins, I'm not sure.
The main benefit is definitely not requiring as much technical acumen or teams to build your infrastructure, but as more skilled workers enter the labour force, I think that too will start to get marginalised to some degree and teams will shift back to having their own IT departments. Just a guess.
I personally hate the cloud, way too expensive for me and the design choices they do are geared towards profit maximisation and not building a partnership with customers where both win. E.g. so many different modules, each with limited functionality...when I can just run my own virtual server and do everything myself much better
I am not sure if this would the case though. Why would an employer use skilled workers to do something mundane like setting up a kubernetes cluster vs something that actually differentiate their own business from others.
None of those things are part of the core business or product differentiation. Yet large corporate orgs still expend enormous amounts of resources on those activities. In fact in most companies the second highest-paid executive is the COO, which pretty much just means the guy in charge of all non-core activities.
I think the simple answer is that most corporates are highly risk averse. Being in startup culture blinds you to this. Startups are comfortable with existential threats in the background, because they're constantly on the verge of failure anyway. They might neglect their liability, public image, or cloud TOS risk. Fast innovation in the core product is a far higher priority for a hypergrowth startup.
In contrast, large and mature organizations are going to be extremely sensitive even to small risks. The typical large company is slow-growing but highly profitable. The number one priority is to avoid killing the goose that lays the golden egg.
Think of the most conservative, bean-counting , button-down loan officer in suburban Topeka. Once "Cloud TOS Risk" enters the business lexicon, said beancounter is likely going to insist that the company in question has a plan, before releasing the tens of millions in loan refinancing that modern corporate America relies on in their leverage capital structures. It doesn't matter how big of an actual risk is. These processes are driven by "check-the-box" CYA theatre.
Its a big risk to put all your eggs in somebody else's hands. With in-house tech, you have somewhat more control. My 2c.
Funnily enough, I'm now starting to think that the only reason AWS proliferated as much as they did was to only fuel the VC capital boom of the last decade and not much else. Startups we able to promise huge valuations and returns based on the belief that AWS would "scale for us" and they didn't really need the inhouse expertise needed to grow. Just sell the growth.
If having anything to do with X creates the perception of Y being liable for any legal and civil improprieties, then Y will dissolve any of its relationships with X.
<my take> For the parler case, Amazon didn't want to be held responsible for hosting content that (maybe) caused the problems in that nations capital, and certainly doesn't want to be held responsible for any future content/actions.
I think this can be justified by preservation of shareholder value. </my take>
But we should be skeptical of what exactly big tech will do this power going forward.
> Megalothymia is “the desire to be recognized as superior to other people”, and “can be manifest both in the tyrant who invades and enslaves a neighboring people so that they will recognize his authority, as well as in the concert pianist who wants to be recognized as the foremost interpreter of Beethoven”; successful liberal democracies channel this desire into fields like entrepreneurship or competition, including electoral politics.
Trump can be described as someone who has megalothymia, wielding his money and more recently politics to assert his perception of himself as 'superior'. This, along with what I opine is blindness of the heart, leveraging this megalothymia as a tool to worship his ego.
These recent days, a number of tech companies, of course made up of people, flexed back when they felt their reasoning outweighed the risks of such a decision. Is this tyranny from these powerful people (Companies et al.)? And thus we are seeing folks (Trump & Companies et al.) taste each others tyranny, taking turns? Or are companies et al. simply claiming a piece of that same echelon Trump claims for himself, as an act of fighting back against a tyrant? Or some third or fourth option... etc.
Ensure contracts are upheld, to harm bad actors (via prison, fines, etc.) and to protect us from foreign invasion.
so, where did that fall apart in our current American system and what can we do to save our democracy? As the article has hinted at, regulation has gotten too esoteric for a bipartisan split in congress. When you have almost 500 people arguing about the use of SPACS for going public, limits on PPE loan, rights of citizens and business in regard to the pandemic, etc. etc.
John adams said 'God forbid we be 20 years from [..] a rebellion.'
We as Americans need a bit more of a centralized system and move a bit more towards an oligarchy. Same things that plagued the roman republic (Bipartisan politics) are plaguing our system. Augustus' answer? bring it down to him and 30 other people who where the most important people from each 'area' and the only thing that gave him more power than the rest was the fact the he was the only member who didn't rotate. Otherwise, it still came down to voting albeit with a forum that allowed for open discussion without rambling into nonsense and listing to 500 people say the same thing over and over.
Eth can uphold contracts
The problem is that many companies are so invested in AWS that they don't know anything else. Many companies are building their entire IT around AWS. In that regard, unless we create legislation that requires unified APIs and an open standard for cloud services; then we are going to have to start treating cloud providers like utilities.
I supported Twitter kicking Trump off its platform. I think AWS kicking Parler off is practically crossing the line, and that legislation needs to be made to tell cloud providers that they need to create an open standard for their services or be considered an essential utility.
For a company like Google, Twitter or Facebook it matters very little if they have 100, 1.000 million users or 2.000, or 3.000. Most of the cost is in salaries and the salaries are the same because the technology is scalable.
Once they have the market of hundreds or billions of users, almost nobody can compete with them because even paying enormous salaries the price per user is very small.
The problem with decentralization is that they don't have scale, so it is a serious problem just paying a single engineer salary.
Moore's law suggests that this trend will continue (unless purposefully downgrading performance to increase repurchases) and with 6G coming, we shouldn't need to be so dependent on the classic datacenter-client architecture wich is the very source of FANG's power.
The new Internet and decentralization completely shifts this power-balance towards the masses. It is democratization of our data instead of being arbitraged by FANG who pay pennies on the dollar to resell it to others.
I think it is important to distinguish between websites/broadcast SMS (Facebook, Twitter), hosting providers (AWS, GCP) and "app stores" (Apple, Google). I hope readers understand the differences.
Here is a thought experiment: How large of an audience, i.e., how many daily visitors, does a website need to have before an act of banning a contributor becomes a free speech issue. What is that magic number, the threshhold over which the website comes to be perceived as a home for "free speech". Most websites fall below this threshhold. For example, journalists, bloggers, politicians, etc. do not debate hellbanning on HN as a free speech issue.
What is this number. Or is it something else.
You clearly don't have showdead marked. There are persistent (and, I should say, entirely incoherent) complaints about speech suppression in the gray comments at the bottom of many threads (often still riddled with the slurs that actually got the user banned).
I think so, the closing point of the article is that we are due for a return to decentralization, even if that is spurred on in the short-term by "competing centralized solutions." The primitives required to run distributed/federated services that can compete with the big web giants probably aren't there just yet, but they are starting to emerge.
With cloud hosting providers, K8s is a big step forward, because you can deploy an open-source system end-to-end an arbitrary cloud k8s platform. Part of the pain of switching to more open alternatives in the past has been managing operational deployments which this begins to address. Blockchain technologies, for all their misuse and hype have brought us some interesting innovations such as the InterPlanetary File System (which probably isn't going anywhere but is a start). None of this is better than the current crop of closed system, compare Diaspora to Facebook for example, but we have the start. But you can only have the "Web" applications he envisions if you have the "distributed internet substrate."
Always worth remembering, that sane societies with sane institutions and conduct exist. One has to consider the "human element" if one wants to keep it that way.
As soon as you can install an app from a file like you can on desktop, then we can have some real progress.
I'm thinking of Apple saying "we won't let you download that app" for my macbook. How did we let them have that power on mobile computing?
Apple is of course a totally different story, as their devices are 100% locked down walled gardens where Apple is the sole arbiter of what software you're allowed to install. It continues to astound me that a substantial number of HN users see absolutely nothing wrong with this.
I disagree. We have DNS, email, and HTTP. That's a pretty capable set of distributed protocols that have been on those phones since the beginning.
The stores are not a problem here (although they are in other ways); rather, there is a lack of apps that people want to use, or perhaps a lack of people using those apps.
I use Riot (now Element), which uses the Matrix protocol across all my iDevices, and it runs just as well as on my Mac. but hardly anyone I know is on it. It's much more private of a solution that WhatsApp, and although not as refined, it works pretty darn well for my use cases.
All things being equal, if it takes any more energy to do something then, statistically, it will be done less.