A Guest post!

by Michael Cain

Today the Senate Judiciary Committee approved antitrust legislation sponsored by Amy Klobuchar (D-MN) and Chuck Grassley (R-IA) that would require Apple, Meta (Facebook), Amazon, and Alphabet (Google) to effectively separate the platform and application portions of their businesses. I will be fascinated to see how far up the software stack they try to push "platform". Eg, is Facebook a platform, or only the very large compute and storage farms it runs on?

Since then I've had a chance to read through S.B.2992. What does it take to be designated "a platform", and
what do you have to do if you get designated? The qualifications seem to be
(a) a web or mobile-app front end,
(b) sufficiently many users (millions), and
(c) sufficiently large capital value (billions).
If you reach that point, as I read it, you have to open up your content management system (CMS) in some fashion.
The easiest way to do that and avoid any questions of whether you have or not
is to split the business. One part sells usage of a particular hardware/software set.

The other is in the content business: sign up users to post and make comments, sign up advertisers, enforce terms of service on content. The "platform" company charges everyone the same rate.

Whatever it is that Klobuchar and Grassley think they will accomplish, I predict they will fail. Why?

In the economics of telephony and related services, there's a thing called "the network effect." It says that the value of the service provided by a communications network increases more than linearly with the number of customers using the service. Early telephone companies figured out that they needed to interoperate and have a single consistent numbering plan.

Enterprise-only e-mail, despite having a greater range of capabilities, lost out to [email protected] internet e-mail because being able to communicate with everyone was more valuable than those other capabilities. The value of the Facebook service is that almost everyone is on. A few people might move exclusively to a Parler -- which could, under S.B.2992, be running on the Facebook platform -- but only a subset of their friends and families will. Any initial surge will tend to die out.

I'm old enough, and was doing R&D in the appropriate line of business, that I lived through the breakup of the Bell System in the 1980s. Based on that experience, access to the platform will not be what the Facebook competitors want. What they will want is to ability to "sell" a rebranded Facebook service with Facebook doing all of the heavy lifting. Access to the customers -- just go to your settings and say you want the "Brand X" version of the service. Access to the advertisers -- no need to do selling or billing or accounting.

Oh, and if someone subscribes to the "Brand X" version, no one will screen their content and there will be no banning. Something for nothing, and your clicks for free.

Comments

[wj]

I wonder if another parallel might be the time (early 1970s IIRC) when IBM was forced by the government to allow other hardware manufacturers to run its proprietary operating system on their gear. The (one?) competitor which arose never grew very big. But the anti-trust effort was definitely there.

[CharlesWT]

"Doing something about the big tech companies would be politically advantageous. This is something. Therefore it must be done."

[liberal japonicus]

I'm curious if anything was taken from Warren's ideas

https://www.newyorker.com/business/currency/how-elizabeth-warren-came-up-with-a-plan-to-break-up-big-tech

[wj]

Doing something about the big tech companies would be politically advantageous.

It is possible to oppose doing something solely because it would be politically advantageous. But still not consider the characteristic of being politically advantageous sufficient grounds to not do something. And that's what you seem to be suggesting -- that just because something might be politically advantageous is sufficient reason to denounce it.

The right question is: Political considerations aside, does something need to be done? Once we answer that, we can have a meaningful discussion about whether the particular proposal is the right thing to do.

[ral]

wj, I worked for Amdahl Corp. back in the late 1970s. They were the principal beneficiary of running IBM's software on compatible hardware (Dr. Gene Amdahl, founder, was the chief architect of IBM's System/360).

The company was successful for a while but everyone knew the days were numbered because eventually IBM would adjust to the new environment. It took them a while, though.

[Michael Cain]

I wonder if another parallel might be the time (early 1970s IIRC) when IBM was forced by the government to allow other hardware manufacturers to run its proprietary operating system on their gear.

Another case worth considering is the Microsoft one at the end of the 1990s. Both IBM and Microsoft were (at least initially) found to have a "monopoly" under the traditional antitrust definition. The big thing about SB 2992 is that it creates a new definition that has nothing to do with defining a market or a company's share. It's just about being big.

If the bill were to pass and be signed, I would be willing to make a small wager that Office 365 is a "platform".

[CharlesWT]

From about 1973 to 1975, I worked with a Digital Scientific IBM 1800 emulator. It, among other things, collected data from IC chip testing equipment.

One day it started crashing and restarting over and over. We finally determined that my new shoes were generating enough static electricity to crash it whenever I walked behind it on my way to the keypunch.

[Pro Bono]

Who's this Michael Caine fellow?

[liberal japonicus]

That extra e was my squeal of joy at getting the guest post! Will fix it now so I appear more sober and staid

[Girl from the North Country]

Something for nothing, and your clicks for free.

:-)

[wj]

In addition, even back then we exclusively used recycled electrons for all our work.

[Pro Bono]

You're only supposed to blow the bloody 'e' off.

[JanieM]

CharlesWT: ...enough static electricity...

A few weeks ago an odd thing started happening. Every now and then my computer (a tower that sits under my desk) would make a noise as if a peripheral had been plugged in and then unplugged. The first couple of times I thought it was just some anomaly and ignored it. But it kept happening, and eventually I realized that it only happened when I got up from my chair and left the desk. I'm convinced it's a static electricity effect.

Weird, and I don't know why it has only started this year, since although I've rearranged the geometry, I'm using the same furniture I've always used.

At least the machine isn't crashing. Small favors.

[jack lecou]

I wonder what they *do* hope to accomplish with this.

As you say, it doesn't seem like this does much to solve what many people perceive to be the real problem with respect to Facebook and the like -- the nasty one-two-three punch of network externalities forcing everyone to use them, the lack of transparency/straight up malignancy of the "algorithm" which decides what posts you see, and being too large and all-encompassing to be able to moderate effectively (even if they wanted to).

And proposals to address those problems exist. The best, IMO, revolve around forcing the large social media companies to participate in some kind of open federation protocol. I.e., make the services interoperate like email or long distance companies, where you can pick your provider, perhaps plug in your own "algorithm" etc.

On the other hand, *this* proposal looks more like it would just force Facebook, Youtube, etc. to calve off AWS competitors.

Which might be helpful too, all things considered. But is very weird if it's intended as a remedy for the first problem.

[Michael Cain]

At least the machine isn't crashing.

Shielding and overvoltage protection are both enormously better these days. Some of that is because contemporary integrated circuits run at (what used to be considered) ridiculously low voltages and have to be protected, plus the need to contain the RF emissions of the electronics.

From long ago, the first rule of trouble-shooting a problem involving motion is "Check the physical layer" :^) That is, make sure all the connectors are firmly plugged in to the appropriate sockets. I discovered how much I harped on that when I was running a bunch of field demos. We were setting up and a newbie said that such-and-such demo wasn't working. Three more experienced people in the area asked, in unison, "Did you check the physical layer?"

[CharlesWT]

In the early days of DIY home computing, there was a company that manufactured motherboards or something similar. They were having a severe problem with customers receiving boards that didn't work even though they carefully tested and packaged the boards. They finally determined that the boards were being zapped by the static electricity generated when customers ripped off the tape on the envelope containing the board.

[Michael Cain]

The best, IMO, revolve around forcing the large social media companies to participate in some kind of open federation protocol. I.e., make the services interoperate like email or long distance companies, where you can pick your provider, perhaps plug in your own "algorithm" etc.

That's my reaction as well. I acknowledge that the usual drawbacks are likely. Innovation takes a hit, because it has to be approved by whoever is setting the standard. Performance takes a hit, because of the probable need to move potentially large amounts of data across an external interface rather than an internal one. And assuming that the spec is big enough, little guys have to buy "platform" service from someone else because they can't afford to build it from scratch.

I may just sound discouraged because many years ago I had to be involved in writing standards for interoperability between telephone networks. Decades ago there was a Dilbert strip where the sales rep was telling Dilbert, "The product complies with all pertinent international standards." To which Dilbert responded, "You mean it doesn't work and it's not your fault."

[jack lecou]

"The product complies with all pertinent international standards." To which Dilbert responded, "You mean it doesn't work and it's not your fault."

Well, the irony is most of the big guys, and a not an insignificant number of little guys, are already running almost entirely on open standards and open source under the hood. HTTP, HTML, JSON, browser APIs. You name it.

The standard making processes for that stuff seem to be, for the most part, fairly robust and innovative. Maybe it's because web standards have always had a "make it work, *then* make it standard" philosophy.

There's no particular reason a federated social media standard couldn't work the same way, though, admittedly, capturing that same energy in a federal regulatory context might be a challenge...

[Michael Cain]

Well, the irony is most of the big guys, and a not an insignificant number of little guys, are already running almost entirely on open standards and open source under the hood. HTTP, HTML, JSON, browser APIs. You name it.

I'm thinking at least a level above that. How do you implement the Facebook concept of "share this video clip but only with my friends and family" when friends and family are spread across multiple services? Or make Alexa aware that I subscribe to three music services, operated by different companies, and am entitled to listen to songs from any of them?

[wj]

when friends and family are spread across multiple services?

Who counts as "friends" (or even family) can differ between services. For a reason. (And not just for bigamists.)

[ral]

Good afternoon, ladies and gentlemen... and others.

Of course I know all our members are ladies and gentlemen. What I mean is that some of you have brought friends.

[Pete]

Regarding the open federation protocol thing:

To use Office 365 as an example, I don't need to use it to access the content. Meaning, OpenOffice and LibreOffice are free and support Office formats, so I can read/write Word/Excel/Access files without having to subscribe to any Microsoft services, but share fairly seamlessly with those who do. I may not have all of the fancy bells and whistles and ribbons and templates that 365 users have but the lion's share of the functionality is there. Granted, I depends on the kindness of Mozilla and Oracle. But Microsoft doesn't appear to be appreciably affected.

I guess that's file-based as opposed to platform, but I kinda think it might be a model of sorts?

Simply setting up a smart tv is an afternoon project these days with all the a la carte content providers - Primes and Gos and Plusses, oh boy! I imagine a new system would work similarly: opt-in services one would need to set up and curate. So, frex, I could subscribe to ObWi's Facebook content without having to log into Facebook directly. Functionality could be limited - maybe I can't "like" or otherwise proprietarily "react" to things or even comment on them, and maybe I'd have to use a 3rd party means of communicating. But I wouldn't be completely locked out. Personal content would work the same way - I opt-in to the platforms that can utilize it: Facebook but not Instagram, Reddit but not Twitter.

Whether that would utilize new browser functionality or an application is to be determined. Maybe your library can provide email addresses or whatever login authentication would be necessary? Seems like that's not too onerous as they would only have to manage the authentication itself without the content overhead? Better yet, peer-to-peer handled locally? I dunno what the pitfalls and benefits might be there and I'm probably missing something glaringly obvious.

[Michael Cain]

But Microsoft doesn't appear to be appreciably affected.

I included Office 365 intentionally for two reasons. It appears (to me, before a bill has been passed, and long before such a law has been to court) to satisfy the definition of "platform". Browser or app front end, millions of users, billions in market cap. As a minimum that would seem to require MS to actually document the formats -- LibreOffice wouldn't be reverse engineering any more, they would be implementing from a spec. At an extreme, it would seem to require that any changes MS wants to make have to go through an external standardization process.

For example, the Linux crowd gets a spec for OneDrive that lets them write the driver to mount a OneDrive file system. Ditto iCloud, or Google Drive. And potentially, a standardization body where they can propose all sorts of potential changes to the standard(s) to make the filesystems friendly to Linux.

[jack lecou]

I'm thinking at least a level above that. How do you implement the Facebook concept of "share this video clip but only with my friends and family" when friends and family are spread across multiple services?

What wj said: concepts like "who are my friends and family" would be one of the things specific to the particular platform(s) you're using, and how you have them configured.

I'd imagine the basic elements of the system would be a lot more basic than that. Maybe just some kind of universal identifier ("[email protected]") and then a flexible message format that could carry all kinds of content (posts, messages, videos, comments, emoji reactions, etc.). The messages would be syndicated to other services, depending on various outgoing/incoming filters.

For example, your service ("Service A") might let you set up contact lists. And then if you share a post with your "friends" list (to: "[email protected]", "[email protected]"), it would encrypt the message using our public keys, and syndicate your post to B and C's servers.

Then on the other end, our services might let us choose from different algorithms and filters, and if we're configured to see your post, our own services would take care of decryption and putting it in our "feed" (or however our service presents things).

At least that's one idea. Smarter people than me have presumably worked out some real federation protocols more thoroughly. There are a lot of little details (security, privacy, efficiency) to work out and get right. But it all seems *theoretically* tractable to me.

Or make Alexa aware that I subscribe to three music services, operated by different companies, and am entitled to listen to songs from any of them?

That seems like a slightly different problem -- more home automation than social media. And there's no doubt a lot of standards and regulatory work to do there. For example, Apple/Google/Amazon have been rather regrettably stubborn about really letting each other interoperate, so we still have situations where Apple won't allow Amazon to publish an Amazon Music plugin to the Siri Store or, or whatever.

But those problems aren't really about any technical difficulties. They're political ones. Assuming Alexa actually had plugins available for all three services, it should just be a matter of entering all the credentials, and Alexa can take it from there.

(Maybe it'd also be nice to have some meta-service that could tell to copy your Google Home configuration over to Alexa or vice versa too. But, again, that's very technically tractable, if you've ordered the FAANGs to actually play nice with each other.)

[JanieM]

Most of this is above my pay grade and/or out of my competencies, such as they are.

But as to Office 365 -- I'm assuming you wouldn't say this about a standalone copy of Office? (That's what I still have, luddite that I am. Don't know when they'll get too old to still be useful.)

Pete, hey! Nice to see you! I tried LibreOffice and got frustrated with the word processor, though I could have used it if I'd had to. The spreadsheet wasn't all that functional from my POV. But that was a few years ago, maybe they've beefed it up since then.

[Pete]

@MC

I was using 365/Libre as a possible blueprint to address the "one-two" of jack lecou's 4:32 as I understood it. In other words, how to open a platform while preserving the original. Maybe a required proprietary import/export module written to a spec?

Just thinking out loud. And none of that addresses the three-punch, but I haven't the faintest idea as to how to get that crazy horse back in the barn.

Apologies if I took your inclusion down a completely wrong path.

[Pete]

Hi Janie!

I originally used OpenOffice for clients who needed to open attached files but otherwise didn't feel the need to spring for MS. Back in the day MS Office was notoriously buggy on the Mac OS and OO was not only more stable, it was more reliable at opening problematic email attachments. I do most of my writing in a text editor and only bother with Word if I need the fancy stuff, so open source is fine for me.

Full disclosure, I'm still running High Sierra on my daily driver as it's still compatible with my standalone Adobe CS suite. As a member of the Luddites, NY chapter, I kinda despise SaaS. Not just the managing of yet another subscription, but it feels like "Why buy tools when you can rent them from Home Depot?"

Great. Now I feel old and stupid. This is why I need to remember to keep my trap shut and stick to lurking. Cheers!

[JanieM]

Pete -- you gave me a good chuckle.

[CaseyL]

My only viewpoint on this is that, having used 365 Outlook, I hate it with a holy passion. It's buggy, doesn't sync well with desktop Outlook (meetings scheduled on the 365 calendar don't necessarily appear on the desktop calendar), and is missing some tools I like.

About the only good thing about 365 Outlook is the big icon it uses for email addressees, making it easier to avoid address autofill errors.

[wj]

I am fascinated (not to say amused) at how many of us here are
a) career computer folks
b) self-described Luddites.

Perhaps it's a matter of us knowing enough about computers that we don't share a lot of popular faith in them...?

[Michael Cain]

That seems like a slightly different problem -- more home automation than social media.

Neither "social" nor "media" appear in the bill's text. And Alexa runs on my phone now, which checks off the mobile app front end. Once they -- the FTC and Dept of Justice, jointly -- declare you to be a "platform" then you can't (quoting from the bill text):

- unfairly preference the covered platform operator's own products, service or lines of business over those of another business user on the covered platform...

- unfairly limit the ability of another business user's products, services, or lines of business to compete on the covered platform relative to the covered platform operator's own products, services, or lines of business...

- materially restrict or impede the capacity of a business user to access or interoperate with the same platform, operating system, hardware or software features that are available to the covered platform operator's own products, services, or lines of business...

There's going to be a ton of pressure from little guys for the FTC/DOJ to interpret platform broadly. Is Twitter as we know it a platform? Does "interoperate" mean there have to be interfaces so that someone on Mike's Message Service can follow a Twitter user? If Twitter bans a user, can they refuse messages to or from them after they sign up on Mike's Message Service?

It's going to be a nightmare.

[Michael Cain]

Perhaps it's a matter of us knowing enough about computers that we don't share a lot of popular faith in them...?

I'm kinder. We understand some of the underlying problems, and how hard those problems can be. This is why, when the topic of STEM versus liberal arts studies at the college level comes up, I'm always an advocate of requiring STEM students to take some liberal arts (or at least social sciences), and requiring liberal arts students to take some software classes. Not necessarily learn to code classes, but to gain an appreciation for how complex the systems are.

[Pete]

I ran across this personally-relatable comment recently:

"As a GenXer, I adopt new technology like a Millennial but I still curse at it like a Boomer."

[JanieM]

Pete -- great quote.

Michael -- We understand some of the underlying problems, and how hard those problems can be.

I think that's true of us here, and of a lot of people with a certain kind of sensibility. On the other hand, there seem to be a lot of people working in STEM fields who are utterly sure (in defiance of the entire history of the human race, never mind the past 200 years) that technology is going to solve all our problems within the next five years, if not tomorrow.

I've been listening to this guy at the rate of a few minutes a day (can't remember if I got the link here or at BJ). I thought it might be nice to see if my eye-rolling skepticism about the mysterious beast called an NFT was justified. If this guy -- who talks, and talks, and talks -- is to be believed, my gut instincts are correct. He is not kind to techbros.

All I can say is #notalltechbros(or gals)

[wj]

there seem to be a lot of people working in STEM fields who are utterly sure (in defiance of the entire history of the human race, never mind the past 200 years) that technology is going to solve all our problems within the next five years, if not tomorrow.

It's a view that has been common for quite a while. Certainly in the middle of the last century, when I was in school. And earlier, quite visibly during the Golden Age of Science Fiction. Even before that, the same view was widespread among the Victorians. Different technologies, of course. But the same expectation that it would deal with all the world's problems.

[CharlesWT]

It's going to be a nightmare.

This is often the case when a tech clueless Congress tries to reshape the tech world. We're lucky that technological innovation tends to outpace politicians and bureaucrats.

Congress may come up with some kind of all-encompassing regulation of the Internet giants just in time for some version of Web 3.0 to sweep in and make it all irrelevant.

[Michael Cain]

This is often the case when a tech clueless Congress tries to reshape the tech world. We're lucky that technological innovation tends to outpace politicians and bureaucrats.

Executive branch agencies as well. Some of our problems with the internet are rooted in decisions Congress and the FCC made in the mid-1990s.

[jack lecou]

Neither "social" nor "media" appear in the bill's text. And Alexa runs on my phone now, which checks off the mobile app front end. Once they -- the FTC and Dept of Justice, jointly -- declare you to be a "platform" then you can't (quoting from the bill text):

Oh -- I thought that question was in the context of "how would a social media federation protocol work".

In the context of this bill, I think what's relevant is Amazon would be barred from favoring it's own music service over Apple's or Spotify's or Mike's.

They'd have to allow other services to write and publish Alexa music plugins with the same APIs available to Amazon, and then wouldn't be allowed to make Amazon's own music plugin easier to access through certain keywords, etc.

(Still doesn't really have any bearing on *setting up* your Alexa though.)

- unfairly preference the covered platform operator's own products, service or lines of business over those of another business user on the covered platform...

- unfairly limit the ability of another business user's products, services, or lines of business to compete on the covered platform relative to the covered platform operator's own products, services, or lines of business...

- materially restrict or impede the capacity of a business user to access or interoperate with the same platform, operating system, hardware or software features that are available to the covered platform operator's own products, services, or lines of business...

That all seems fine to me.

Having had a chance to skim over the bill text, it sounds like the plain *intent* isn't to break up tech services, or legislate new interop layers or APIs.

The intent is to make access to the existing interface points fair and non-preferential:

Apple devices run apps. They shouldn't discriminate against some other company's apps for anti-competitive reasons.

Google crawls web pages and serves search results. They shouldn't preferentially favor Alphabet products in those results.

Amazon lists products for sale. They shouldn't use information available only to them to unfairly push their own versions of products.

Microsoft makes operating systems. They shouldn't have special APIs that only Microsoft apps are allowed to use.

Etc.

I'll leave it to the lawyers to figure out if the bill *text* actually serves that intent, but so far as what that intent is, it seems pretty clear.

There's going to be a ton of pressure from little guys for the FTC/DOJ to interpret platform broadly.

Well, see above. I'm not sure there's scope for that. The access boundaries aren't really negotiable or arbitrary -- in many ways, they're intrinsic to the service, following the contours of whatever points of access the service already offers, or what services like it customarily do. The bill just wants access to those points to be fair and non-discriminatory.

If there is still some wiggle room there though, more power to the little guys.

And in any case, some ambiguity in interpretation is always inherent in regulating things like this, so I don't think that represents any fundamental problem with this bill. That's largely what we have agencies like the FTC/DOJ (and the judicial branch) *for*.

Is Twitter as we know it a platform? Does "interoperate" mean there have to be interfaces so that someone on Mike's Message Service can follow a Twitter user? If Twitter bans a user, can they refuse messages to or from them after they sign up on Mike's Message Service?

Yes, no, yes.

Again, IANAL, but AFAICT, this legislation isn't saying that Mike can make his own *platform* and force Twitter to interoperate in arbitrary ways.

It's saying that access to the services Twitter already provides (subscribing -- to other Twitter users, posting tweets -- for other Twitter users, etc.) just needs to be provided in a non-preferential way.

In this case, maybe this would mean Twitter would need to provide access to twitter's APIs (including internal APIs) to allow Mike to make his own app. But it would still be an app for interacting with Twitter's platform: posting tweets to Twitter's platform, subscribing one user of Twitter's platform to another. Etc.

[wj]

They shouldn't have special APIs that only Microsoft apps are allowed to use.

To be really effective, the requirements would need to include making the specs for using said APIs publicly available as well.

[nous]

Two quick additions to the conversation. I agree wholeheartedly with the notion that a college education needs more meaningful depth in its commitment to breadth. I ended up where I am largely because I had spent some time between colleges working in IT (learning SQL through trial-and-error from the customer service side of a horribly run, high-concept dot com boom company). Those habits of logic and thoughts about how meaning is shaped by the information with which it gets associated were laced all thorough my intertextual readings of literary works and got me an unlikely toe in the door at a top Ph.D. program in my mid-30s.

If I had my druthers, no one would be allowed to declare a major (assuming we kept such a thing) until after they had two years of intensive, interdisciplinary breadth under their belts.

On another topic, I went to Open Office twice in an effort to get off of the MS Office merry-go-round before finally ditching it for good a couple years back for LibreOffice's version of Open Office. It started out usable, and despite moments of complete look-feel alienness, has kept me satisfied even with the option of MS Office right there via institutional access. It works just fine for most things.

My wife will not switch because the publishers she works with use MS Office for the editorial phase, but if it weren't for the need to get comments and track changes between her and her editors, LibreOffice could totally handle her 130K word manuscripts.

[Snarki, child of Loki]

"They shouldn't have special APIs that only Microsoft apps are allowed to use."

I'm not sure that opening up the API for Microsoft's Virus Transmission Protocol is a good idea.

[hairshirthedonist]

Maybe someone mentioned this earlier, but this sounds like it parallels net neutrality in a lot of ways to me, especially after reading jack lecou's latest comment.

[hairshirthedonist]

And now that I'm on that train of thought, it's also like the record company can't also own the radio station, or the movie company can't also own the theater chain, or the oil company can't own the railroad that transports the oil.

(This is me thinking maybe I'm beginning to understand this after only having a vague notion of what I've been reading until a few minutes ago.)

[wj]

this sounds like it parallels net neutrality in a lot of ways

The main difference I see is that net neutrality is, in effect, what we have currently and have had historically. There are arguments about removing that requirement, but for the moment, it's the standard.

In contrast, this is about changing an existing situation. Which is a far more difficult thing to do -- technically, as well as philosophically.

[jack lecou]

In contrast, this is about changing an existing situation. Which is a far more difficult thing to do -- technically, as well as philosophically.

No. Certainly technically there is no change. Or at most, minor adjustments.

Take Amazon for example. It offers APIs for everything already: listing products, adjusting prices, checking inventory. AFAICT, this law would not mandate any changes to any of that whatsoever.

The only difference is that Amazon's *own* use of their services would be explicitly constrained. They'd be barred from using the private data in ways that others can't: mining sales data in order to find profitable products to re-introduce under its own brands, duplicitously boosting its own brands in searches and recommendations, etc.

No technical change is mandated there. (I suppose Amazon could *choose* to comply by offering new APIs to mine data and boost products in the same way they do internally, but that'd be up to them.)

Nor am I convinced this could regarded as a change from the status quo "philosophically". I suspect most people's ethical senses would regard Amazon's manifestly unfair double dealing as out of bounds already.

It's a change from the status quo *legally*, of course, but that's sort of what new laws do. I think it'd be hard to argue that it's a negative change.

[wj]

Certainly technically there is no change. Or at most, minor adjustments.

Take Amazon for example. It offers APIs for everything already: listing products, adjusting prices, checking inventory. AFAICT, this law would not mandate any changes to any of that whatsoever.

In the case of Amazon that you cite, sure minimal technical changes. But if you want to have multiple social media systems interacting (e.g. Facebook and LinkedIn)? That would, I think, be a very different thing. That was the case I was thinking of.

[jack lecou]

In the case of Amazon that you cite, sure minimal technical changes. But if you want to have multiple social media systems interacting (e.g. Facebook and LinkedIn)? That would, I think, be a very different thing. That was the case I was thinking of.

Sure. But that sort of thing is not mandated by this law.

Facebook would be banned from arbitrarily blocking LinkedIn access to APIs it already provides, and vice versa. But no new APIs or interaction protocols would be required.

[wj]

I guess I see actually implementing a different social media system's APIs as a significant technical challenge. There is a bit of a difference between "allowed to use another's APIs" and "actually feasible to use another's APIs."

[jack lecou]

I guess I see actually implementing a different social media system's APIs as a significant technical challenge.

???

Facebook doesn't have to implement LinkedIn's APIs or anything.

Facebook implements Facebook's APIs. As per usual. It's just that once they do, they're not allowed say to LinkedIn: "nope, we're not giving you an account, you're a competitor".

[hairshirthedonist]

The main difference I see is that net neutrality is...

But that doesn't mean it doesn't parallel net neutrality in a lot of ways. What they (whoever "they" are) are trying to do here now is similar to what they (whether the same or another "they") have already done with net neutrality.

If it wasn't new, we probably wouldn't be talking about it in the first place. And if net neutrality weren't already a thing, I probably wouldn't think to compare this to it.

[jack lecou]

Net neutrality really is a pretty good parallel.

(Good enough that this bill might effectively legislate net neutrality if it were written just slightly more broadly. Some of the "have an app" qualifying language probably prevents that interpretation, but otherwise...)

[Michael Cain]

Sure. But that sort of thing is not mandated by this law.

Facebook would be banned from arbitrarily blocking LinkedIn access to APIs it already provides, and vice versa. But no new APIs or interaction protocols would be required.

We disagree -- maybe -- on what the courts are going to say about what's an API and what's not. Will it be "The API Facebook uses to query their database is SQL, an already open standard," or will it be "The API includes the database that Facebook queries using SQL." In the first case, you're right that Facebook doesn't have to do anything. In the second case, Facebook has to do something new: at least so far as I know, there's no Facebook API for a non-Facebook service to access Facebook's internal data at that level. SQL is easy; things like security are hard. Decades ago when they busted up the Bell System, the security answer was "distinct physical networks."

At some point (I believe) Congress will figure out that what they really want is not for Facebook and the rest to provide access to low-level resources (hardware, OS, SQL, load balancers, etc) on which a company can build a new service that starts from zero customers. They want the Facebook user content databases to be a regulated utility.

[wj]

They want the Facebook user content databases to be a regulated utility.

And perhaps someone with a little knowledge of history will point to the parallels with when electricity (then a new technology) became significant enough in people's lives that it merited status as a public utility. Then again, Congress may be ignorant enough to insist on reinventing the wheel.

[jack lecou]

We disagree -- maybe -- on what the courts are going to say about what's an API and what's not.

I mean, that's a wildcard. I'm not even going to pretend to know what the courts will say. I don't even know how a lawyer would read the text of the law.

What I do know is that there appears to be a very straight line between a number of actual instances of abuse by Big Tech and the solutions gestured at by this law, so what it's intended to do seems pretty easy to work out.

Will it be "The API Facebook uses to query their database is SQL, an already open standard," or will it be "The API includes the database that Facebook queries using SQL." In the first case, you're right that Facebook doesn't have to do anything. In the second case, Facebook has to do something new: at least so far as I know, there's no Facebook API for a non-Facebook service to access Facebook's internal data at that level.

The API is the API. There's no need to hash out what it is.

There is literally already an API (or, really, a collection of them) being used today by the various Facebook client frontends to communicate with the Facebook backend. They'll have calls like "getTodaysPosts()" and "getMyListOfFriends()" and "putNewPost()", all communicating with the Facebook backend on the user's behalf. Also, APIs for advertisers to place ads and specify targeting.

Unlike raw SQL, access permission is already sorted out for those APIs. (Or should be, Facebook's awful security culture notwithstanding.) For example, the client APIs should be gated to see only data visible to the logged in user. You can see *your* list friends, e.g., but not their friends. (If the API can leak more data than you're allowed to see, that would already be a security hole.)

These would be trivial to document and stick a meter on. And then anyone can write a third party client for FB's backend "platform" (or LinkedIn can post to FB on behalf of a user, etc.). Indeed, IIRC, large parts of this functionality are already publicly available.

The exception might be advertising special sauce. I'd assume there's some functionality equivalent to a "getTargetedAd()" call that fetches one of those creepy "congratulations on the baby" ad for diapers, leveraging targeting data trawled from information not visible to a single user, or through any advertising API -- correlations made based on posts by friends of friends of friends and so forth.

AFAICT, it's only this latter bit that would come into conflict with this law. And that doesn't require a new API, it just requires Facebook to play fair and target ads based only on information available through documented APIs, no secret spyware sauce.

Which seems like a good thing, all around.

[jack lecou]

From the bill:

(h)(8) ONLINE PLATFORM.—The term “online platform” means a website, online or mobile application, operating system, digital assistant, or online service that—

(A) enables a user to generate content that can be viewed by other users on the platform or to interact with other content on the platform;
(B) ... [non-FB relevant parts snipped

I think that pretty clearly means Facebook as a whole is the "platform". Or at any rate, some substantial chunk of the Facebook backend.

A plain reading of that definition implies a "platform" is some functional whole that lets users do those listed things. Some individual MySQL database ZZP-24Q in a data center in Nebraska doesn't qualify, because that doesn't "enable a user to interact with other content on the platform".

[liberal japonicus]

In contrast, this is about changing an existing situation. Which is a far more difficult thing to do -- technically, as well as philosophically.

well, for you, maybe...

I keeed, I keeeeeed.

I'm not up enough on the details to argue how the bill should work, my life is basically on the margins of these things, using them for an audience that it wasn't really intended for. I'd love to get away from Word, but I have to use it in it's Japanese incarnation so I can quickly answer student questions about how I do X or why won't Y work. With the pandemic, Japanese universities have basically grapped whatever looked vaguely hammer-like and plugged it in to provide distance learning and companies have abetted that, either by upselling what they have or by putting a coat of paint on it and selling it as a whole new product. And all of the foreign teachers here (and any Japanese teachers who are interested in this, which is an exceedingly small number) are in basically the same boat. Using discord/LINE/FB secret group to run classes (I haven't heard of anyone doing it on Instagram or Tik Tok, but I wouldn't be surprised) So my view is probably not at all similar to anyone's here.

But it does seem to me that things are broken so the argument that 'ohh, it's too hard to change things' seems a bit off. I think these platforms bear a big chunk of responsibility for the way things are, they need to fix it. Or be fixed.

[wj]

But it does seem to me that things are broken so the argument that 'ohh, it's too hard to change things' seems a bit off.

But pretending that it is simple and straightforward is the best way to assure that whatever is done by way of a fix will be as bad or worse.

P.S. In contrast, this is about changing an existing situation. Which is a far more difficult thing to do -- technically, as well as philosophically.

well, for you, maybe...

Which is why I am a Luddite as well as a conservative. ;-)

[jack lecou]

But pretending that it is simple and straightforward is the best way to assure that whatever is done by way of a fix will be as bad or worse.

I don't think anyone's doing that.

Nor do I think that such a thing is assured in any case. There isn't even a well defined notion of "bad" we can use here.

Things will be *different*, certainly. And there will still be problems -- even a solution that mostly works won't result in perfection. Nor can we possibly predict all the consequences.

But all of those things are always true with any legislation. To insist on perfect legislation or perfect knowledge is to be doomed to inaction.

(As an aside: we would have much, much better policy in general if both legislators and regulators had more leeway to experiment and get things wrong -- but do it rapidly and revise -- rather than our current insistence on perfection out of the gate, brutally punishing any experimentation. Being terrified of any change is, again, nothing but a recipe for stasis and corruption.)

[bobbyp]

Being terrified of any change is, again, nothing but a recipe for stasis and corruption.)

As everybody knows, the overarching metaphysical mission of "the Left" is to make the Perfect the enemy of the Good, but it would appear the field is crowded.

Well put, Jack.

[nooneithinkisinmytree]

"Being terrified of any change is, again, nothing but a recipe for stasis and corruption.)"

The dead hand of death cult conservatism, practiced to perfection by the Left.

[Michael Cain]

As an aside: we would have much, much better policy in general if both legislators and regulators had more leeway to experiment and get things wrong -- but do it rapidly and revise -- rather than our current insistence on perfection out of the gate, brutally punishing any experimentation. Being terrified of any change is, again, nothing but a recipe for stasis and corruption.

Consider an example. In the early 1990s, Congress redid telecommunications law for the US. In their wisdom, they decreed that there were two kinds of service: communications, and information. And that the FCC had to regulate on that basis. And that there were rules about established communications providers having to share their physical facilities.

Along came the internet. There were a small number of us in the industry that said the internet is a communications service: ISPs should hand out an IP address, then accept and deliver packets per the protocol spec. But also, that the "share physical facilities" makes no sense, and in the case of cable systems, renders them unusable for internet access. We lost. The FCC declared internet access to be an information service, which has caused all sorts of mischief over the years. Technology decisions were made, and hundreds of billions of dollars invested, on the assumption that internet access was an information service.

The Obama FCC attempted to reverse the problem. They said they would reclassify internet access to be a communication service (which would let them order net neutrality, the thing they were concerned about), but not enforce the portions of the statutes that made no sense. Everyone sued, for all sorts of different reasons. Eventually the FCC reversed itself, and internet access continues to be an information service with all the problems that entails.

Revamping the foundation of telecommunications regulation is a once-a-generation (or perhaps even less often) thing. At the bottom, that's because changes to the foundation cost people who are on the "wrong" side hundreds of billions of dollars. It takes years for Congress to work up their nerve. I'd guess we're on pace for Congress to act on internet access -- making it a third class of service and addressing things like shared physical facilities and net neutrality and possibly services like Twitter and Facebook -- along about 2028. SB 2992 is part of working up their nerve.

This is true for all kinds of regulation, not just telecommunications. Look at all the fudging that's gone on because Congress got nuclear power wrong. Anything where companies invest hundreds of billions of dollars is going to be problematic.

[jack lecou]

It's like the opposite of tried and true.

It's do something, and then stubbornly pretend for 30 years -- against all countervailing evidence -- that it must be true because you did it. (And now it is the way it is so it shouldn't ever be changed, because reasons.)

Anything where companies invest hundreds of billions of dollars is going to be problematic.

It certainly is, but it really shouldn't be.

The scraping deference to making sure past winners remain future winners, and that all the piles of money should remain in precisely the same pockets, is a big part of the problem. All the problems.

From an economy-wide perspective, that's precisely backwards. There were winners, there were losers. Now we need to adjust something, and there might be new winners and new losers. So what? Winning isn't guaranteed in business or investing. And shouldn't be.

The only important part is that, if the change is getting us closer to the right policy answer, the new mix of winners and losers should represent something closer to a genuine win for society as a whole.

[wj]

It certainly is, but it really shouldn't be.

It's just another example of the Sunk Cost Fallacy. And, bad as it is, not even the most egregious.

[CharlesWT]

Over time bad laws and regulations develop a constituency who will fight tooth and nail to keep them.

[Nooneithinkisinmytree]

Over time, good effective laws and regulations knee-jerk develop an enemies list who fight tooth and nail to abolish them.

In America, the mail-drop anonymous enemies are far richer than the aforementioned upfront constituencies, otherwise known as Soros/BLM Jewish conspiracies to the anonymous fascist white conservative extraordinarily overfed conservative minority.

[jack lecou]

Over time bad laws and regulations develop a constituency who will fight tooth and nail to keep them.

And they have a lot of allies.

[bobbyp]

Over time bad laws and regulations develop a constituency who will fight tooth and nail to keep them.

"Bad laws and regulations", similar to all those other laws and regulations, always have a constituency to begin with or they would not have been implemented in the first place.

[CharlesWT]

Sometimes bad laws and regulations get passed because politicians think they're a fix for a problem and/or will play well with voters. Sometimes they're presented as temporary fixes for transitory problems. But once passed, any special interest groups who benefit from them will lobby for their continued existence. If the groups are well-heeled, they will pay people a lot of money to sit in the national and state capitols to make sure nothing gets repealed.

[bobbyp]

Yes, Charles. "Politicians" just pull ideas out of their ass and pass them on sheer whim, and voters don't constitute a "constituency".

But have a nice weekend, a concept now shockingly in vogue and widely adopted as public policy obviously because a couple of politicians thought it would play well with the voters.

[wj]

bobbyp, amazingly enough, not everything that happens is a result of planning, or lobbying, or nefarious intent. Sometimes people, even politicians, do things on their own hook because they actually believe those things will accomplish something useful.

It is not completely unknown for them to be correct. As the adage has it: "Even a blind pig gets an acorn now and then." Alternatively: "Do not atribute to malice what can be adequately explained by stupidity."

[bobbyp]

bobbyp, amazingly enough, not everything that happens is a result of planning, or lobbying, or nefarious intent. Sometimes people, even politicians, do things on their own hook because they actually believe those things will accomplish something useful.

Really? Now I know how Bernie Sanders has passed so much legislation. Just wear red shoes and click your heels! Politics is easy. Who coulda' known?

[wj]

Didn't say it was a trivial effort. Or that just anything could get passed. Just that it does happen . . . so you can't assume it never does.

I would be interested in knowing what secret cabel you think responsible for getting the Civil Rights Act of 1964 passed. (And what their nefarious purpose was.) Just for one obvious example.

[bobbyp]

This....bobbyp, amazingly enough, not everything that happens is a result of planning, or lobbying, or nefarious intent.

and this...I would be interested in knowing what secret cabel you think responsible for getting the Civil Rights Act of 1964 passed.

wj,

When you get around to responding to the point I made rather than the one you dreamed up, I would be more than happy to respond.

Thanks.

[wj]

Since I obviously missed the point you were trying to make, perhaps you could have another go at it.

[bobbyp]

wj,

See my initial reply to Charles: 2/5/22 @ 3:55 PM

There is certainly nothing there regarding "secret cabals". What is your point?

My point is a rather general truism: All laws (assuming a representative government of some kind) have constituencies/supporters or they would not have been adopted to begin with. Even good laws will have "special interest groups" who would probably fight tooth and nail to keep them in place, no? The corollary is "bad" laws are also instigated at the behest of some constituency to begin with.

I merely point this out since politicians generally find parades to jump in front of. They typically don't start them.

Nothing nefarious about it.

Enjoy your Sunday.

[wj]

bobbyp, that obviously isn't the way read your initial post. Thank you for the clarification.