I hope we doesn't seriously believe that Facebook is maintaining open-source software out of the goodness of its heart.
Every open-source contribution that's sponsored by a FAANG company is done for a reason. Whether it's for branding and recruiting (largely the case for PyTorch and React) or to make something free while charging for a scarce complement (see Github or Kubernetes), is largely irrelevant to this argument. Take a look at the Open Compute Project, started by Facebook in 2011 to make switches that were differentiated by software so it could create a monopsony consortium and not pay Cisco. Along the same lines, look at this article (https://thebaffler.com/salvos/the-meme-hustler), about how O'Reilly created open-source as a brand in order to sell conference tickets and create a monopolistic publisher.
On the other hand, I'd dispute the characterization that "engineering innovations" are inherently less valuable. A massive amount of the unseen work at these companies is in the areas of reducing fraud, spam, and other abuse vectors. That these are still very much problems is testament to how difficult they are, not how little innovation there has been. If you want to see a company that doesn't invest in that innovation, look no further than Twitter and its nonexistent trust and safety feature set. Those "really interesting computational social science papers" are published to attract people who care about computational social science, who then can help build trust and safety innovations as well as engagement hacking innovations.
As another example, look at Waymo's self-driving project versus the projects at unicorns (Lyft Level 5, Uber ATG, Tesla Autopilot) or startups (Cruise, Zoox). Waymo is the closest to having a functional product while also having the strongest commitment to doing so safely.
However, all of this largely-invisible work is even harder to replicate than product-visible work and acts to further reinforce the market power of the large tech firms. That there's an exhaust of projects that become public goods shouldn't distract from that.
Github post-acquisition has been more active in product development but has been making more features available cheaper/for free. It's now a loss leader for Azure and the pricing drop harms its smaller and more featureful competitor Gitlab.
It is tough to think of Github as innovation. Github is an example of an innovation grab. Github lured developers in to post their code and then sold out to Microsoft. Microsoft had nothing to do with Github pre-acquisition -- and then boom, all of these developers had where they hosted code transferred to Microsoft, a company with a history of anti-competitive behavior and IP theft. Which ones? Windows was backward engineered Apple; Office backward engineered from Wordperfect and Lotus, SQL Server backward engineered from Sybase. (later purchased by SAP and significantly degraded). That acquisition should have been disallowed.
So the idea that Github is an example of innovation by Microsoft is absurd. Secondly, Microsoft could also be reviewing the various code uploaded to see what they can backward engineer. I can't know if this is happening, but it is entirely in line with how Microsoft functions.
The lack of criticism of this dangerous (for developers) acquisition by the media entities illustrates how much they defer to industry. The acquisition should have immediately been called out as foul play. It would be like selling a chicken coop to a group of wolves.
That is transferred to Microsoft as their host. And with Microsoft's full ability to search through every single project without regulation. This is not to say that the code transferred to Microsoft as ownership. However -- as I pointed out above, Microsoft tends to take a very flexible interpretation of other people's IP.
"On the other hand, I'd dispute the characterization that "engineering innovations" are inherently less valuable. A massive amount of the unseen work at these companies is in the areas of reducing fraud, spam, and other abuse vectors. That these are still very much problems is testament to how difficult they are, not how little innovation there has been."
AI to allow FB to discriminate so it can sell ads while not allowing others to discriminate in ways that annoy political actors is a hard problem. But FB doesn't get any credit for solving this problem any more than an engineer at Volkswagen gets credit for disguising car emissions to fool regulators. Twitter isn't great, but the fact that FB spends billions on 'safety' and Twitter doesn't, and both are dumpster fires, should show that the business model is the underlying problem.
Fundamentally FB is trying to engineer around a conflict of interest in the business model, a conflict of interest we should just bar via policy.
Self-driving is a great example of the fake promises of big tech innovation. After ten years of press releases they aren't even close to self-driving cars. They should be embarrassed.
Tech companies can continually promise whatever it is in the future, and be sure to not be held accountable to deadlines. Musk drives through deadlines like they are tissue paper, and media entities report each new announcement without including his terrible track record of his previous projections.
Great stuff Matt. Many of us are so burned out with media coverage that simply presumes that because a large company says something it must be true.
The fact that big tech is, further along, is not contradictory to Matt's point that you still don't see self-driving cars, that was promised to have been here by now. And that if the media had a member and were willing to challenge big tech, this failure to meet their deadlines would be considered embarrassing.
Elon promised that all Teslas could be robotaxis by now -- of course, he made that statement right before he had to raise a few billion. Then after he raised it he said "there are complications."
But that is not what he said when he wanted the $2.3 billion. At the time he said Tesla may stop selling cars, and just make cars to be robotaxis that would be making money 24/7.
MSM maybe, but in Elon's case I don't think he's truly been given a free pass. Tesla has the most active short community of any company on the planet and a notable group of people who think he should be in prison for securities fraud. That the SEC and FTC haven't taken more aggressive action isn't a media problem or a market power problem.
But, taking a step back: Matt's argument is that the US is less competitive because market power precludes competition and innovation. I don't think Tesla, as many promises as they have broken, are an example of that.
For now, sure. But the integration of Musk's companies is making them a player that will price-out competitors from entering the market in the future.
One of the major reasons Tesla could exist in the first place is that the battery technology used by them (Li-ion) was invented by public-sector or energy-company researchers with the consequence technology open to a variety of applications companies.
I don't know what cross-licensing deal exists between Panasonic and Tesla, but there's a chance it would preclude other automotive companies from using Tesla's battery technology. And unless anti-monopoly enforces mandate Tesla license and sell this technology to competitors, those competitors will be locked out on price.
Tesla (and so much of "tech") would not be where it is now if venture capital didn't know that they could purchase monopolies that wouldn't be broken by anti-trust enforcers.
Elon Musk's strategy of tech development squeezes out risk on the assumption he can feed it with enough cash for a long enough time. I'd rather mine hydrocarbons, oxygen and nitrogen in 24hr high noon sunlight from a near earth asteroid than go to Mars, but, his Starship plan is well thought out, makes the right compromises between practicality and performance and will succeed -- unless the money runs out!
That's why I hate his vulgar behavior -- it endangers the good things he has done and can do. My wife wouldn't remember if it was a Chevy or a Buick that crashed into her but she is attracted to Tesla vehicles, showing really wide currency for that brand.
He ought to practice real leadership, go straight and model good behavior by example. None of this "Pedo Guy" nonsense!
UPMC epitomizes much of what's wrong with American healthcare, and with large health systems. It got into a battle with the state when it decided to stop accepting patient health insurance from ONE Blue Cross provider - Highmark, insisting that if they wanted to be treated at UPMC facilities that the patients pre-pay. UPMC continued to accept patients with other Blue Cross plans, just not Highmark's. This is because, after years of each staying in their own area of expertise, UPMC started its own competing health insurance insurance program and Highmark retaliated, by moving into the hospital business. The state didn't really care about this regional spat until UPMC expanded to the more populated east and brought its predatory practices with it. Once it reached as far east as Harrisburg the state took action. UPMC caved to the state attorney general's valid complaints and UPMC again accepts patients with Highmark insurance again. At least for the next ten years.
For UPMC, it's all about revenue. They've bought up medical practices all over and their CEO Romoff was very upfront in saying that they wanted to be the Amazon of healthcare. Many physicians who still put patients first are quite frustrated.
UPMC, like Highmark, has abused its non-profit status for decades. It's hard to tell which is worse; they're both evil in their own way. But UPMC tried to take evil to a whole new level because it thought it could.
BTW, great book on monopolies, Matt. I bought it and enjoyed it thoroughly.
Plumbers are planners. There are design decisions that go into plumbing that have real impacts on the people who will use the plumbing. Why doesn't the Fed pledge to "buy up" (i.e. refinance onto the Fed's balance sheets) consumer credit card debt? Is this any more special than corporate debt? Consumers, like corporations, generally have assets that can be seized in the event of a default. With the significant drop in credit utilization this is obviously a "choked market" that needs to be "revived" ( https://www.foxbusiness.com/money/credit-card-use-drops-coronavirus ). Heck, the reduction in APR would have a significant stimulatory effect on the primary drivers of GDP (consumer purchases), without any effect on money supply like the stimulus payments.
----------
“really interesting computational social science papers” coming out of Facebook - I'm curious to what extent these are the equivalent of aerodynamics and failure testing by automotive and aerospace companies. I sincerely doubt Facebook's "social science" is seriously investigating the social science of people who don't use Facebook (30% of adults in the US). And to the extent they are it's highly doubtful they can accurately model them any near as well as they can Facebook users.
This isn't "innovation", it's basic "understand the product you're selling, and make a better product". Okay, I guess that later is "innovation", but it's not revolutionary innovation. It's a basic expectation of any company that designs products.
Love this issue, thank you for your work. What definitely stood out to me were the stories involving the DOJ. It is so important to point out when companies end up paying fines and settlements that are far less than the revenue obtained by corruption. There is no incentive for them to curb their practices in this case, they can see their corruption and those fines as an investment with a specific ROI which is, in this and many other cases, positive. I do have a degree in Business and its very clear that in those classes, future executives are taught to pursue endeavors with a positive ROI, why wouldn't these companies stay corrupt if they come out with a net profit of $850 million, especially if they have an obligation to keep growing quarter after quarter? Fascinating to see how inhuman some exective decisions can be.
I also truly appreciate the scrutiny of Facebook and the point you make about specific types of innovation. These tech monopolies have diversified their portfolios enough that they can now focus on these digital solutions like Machine learning and AI; they have the most resources so they can hire the best engineers and like you say, they have control over this data if only because they moved/consolidated quickly to dominate user market share. Google also can't be ignored in this respect (I appreciate that you touched on their inroads in banking software and education platforms). With respect to those crucial technologies that can determine a lot about the future of commerce, computing and even the makeup of the job market as more jobs are automated, we have an oligopoly and no one seems to realize it, its all just magic behind our apps made by smart people who probably know better than us.
I hope we doesn't seriously believe that Facebook is maintaining open-source software out of the goodness of its heart.
Every open-source contribution that's sponsored by a FAANG company is done for a reason. Whether it's for branding and recruiting (largely the case for PyTorch and React) or to make something free while charging for a scarce complement (see Github or Kubernetes), is largely irrelevant to this argument. Take a look at the Open Compute Project, started by Facebook in 2011 to make switches that were differentiated by software so it could create a monopsony consortium and not pay Cisco. Along the same lines, look at this article (https://thebaffler.com/salvos/the-meme-hustler), about how O'Reilly created open-source as a brand in order to sell conference tickets and create a monopolistic publisher.
On the other hand, I'd dispute the characterization that "engineering innovations" are inherently less valuable. A massive amount of the unseen work at these companies is in the areas of reducing fraud, spam, and other abuse vectors. That these are still very much problems is testament to how difficult they are, not how little innovation there has been. If you want to see a company that doesn't invest in that innovation, look no further than Twitter and its nonexistent trust and safety feature set. Those "really interesting computational social science papers" are published to attract people who care about computational social science, who then can help build trust and safety innovations as well as engagement hacking innovations.
As another example, look at Waymo's self-driving project versus the projects at unicorns (Lyft Level 5, Uber ATG, Tesla Autopilot) or startups (Cruise, Zoox). Waymo is the closest to having a functional product while also having the strongest commitment to doing so safely.
However, all of this largely-invisible work is even harder to replicate than product-visible work and acts to further reinforce the market power of the large tech firms. That there's an exhaust of projects that become public goods shouldn't distract from that.
On another topic... " (see Github or Kubernetes), is largely irrelevant to this argument." --> Can you explain the Github point? I'm curious.
Github post-acquisition has been more active in product development but has been making more features available cheaper/for free. It's now a loss leader for Azure and the pricing drop harms its smaller and more featureful competitor Gitlab.
It is tough to think of Github as innovation. Github is an example of an innovation grab. Github lured developers in to post their code and then sold out to Microsoft. Microsoft had nothing to do with Github pre-acquisition -- and then boom, all of these developers had where they hosted code transferred to Microsoft, a company with a history of anti-competitive behavior and IP theft. Which ones? Windows was backward engineered Apple; Office backward engineered from Wordperfect and Lotus, SQL Server backward engineered from Sybase. (later purchased by SAP and significantly degraded). That acquisition should have been disallowed.
So the idea that Github is an example of innovation by Microsoft is absurd. Secondly, Microsoft could also be reviewing the various code uploaded to see what they can backward engineer. I can't know if this is happening, but it is entirely in line with how Microsoft functions.
The lack of criticism of this dangerous (for developers) acquisition by the media entities illustrates how much they defer to industry. The acquisition should have immediately been called out as foul play. It would be like selling a chicken coop to a group of wolves.
That is transferred to Microsoft as their host. And with Microsoft's full ability to search through every single project without regulation. This is not to say that the code transferred to Microsoft as ownership. However -- as I pointed out above, Microsoft tends to take a very flexible interpretation of other people's IP.
"On the other hand, I'd dispute the characterization that "engineering innovations" are inherently less valuable. A massive amount of the unseen work at these companies is in the areas of reducing fraud, spam, and other abuse vectors. That these are still very much problems is testament to how difficult they are, not how little innovation there has been."
AI to allow FB to discriminate so it can sell ads while not allowing others to discriminate in ways that annoy political actors is a hard problem. But FB doesn't get any credit for solving this problem any more than an engineer at Volkswagen gets credit for disguising car emissions to fool regulators. Twitter isn't great, but the fact that FB spends billions on 'safety' and Twitter doesn't, and both are dumpster fires, should show that the business model is the underlying problem.
Fundamentally FB is trying to engineer around a conflict of interest in the business model, a conflict of interest we should just bar via policy.
Self-driving is a great example of the fake promises of big tech innovation. After ten years of press releases they aren't even close to self-driving cars. They should be embarrassed.
Tech companies can continually promise whatever it is in the future, and be sure to not be held accountable to deadlines. Musk drives through deadlines like they are tissue paper, and media entities report each new announcement without including his terrible track record of his previous projections.
Great stuff Matt. Many of us are so burned out with media coverage that simply presumes that because a large company says something it must be true.
I agree on FB's conflict of interest but disagree on self-driving. It's a genuinely hard problem and big tech is further along than anybody else.
Ben,
The fact that big tech is, further along, is not contradictory to Matt's point that you still don't see self-driving cars, that was promised to have been here by now. And that if the media had a member and were willing to challenge big tech, this failure to meet their deadlines would be considered embarrassing.
Elon promised that all Teslas could be robotaxis by now -- of course, he made that statement right before he had to raise a few billion. Then after he raised it he said "there are complications."
But that is not what he said when he wanted the $2.3 billion. At the time he said Tesla may stop selling cars, and just make cars to be robotaxis that would be making money 24/7.
Excuse me...I meant to write "if the media had a memory" not "a member" above. I was not questioning the manhood of the media in specific terms.
MSM maybe, but in Elon's case I don't think he's truly been given a free pass. Tesla has the most active short community of any company on the planet and a notable group of people who think he should be in prison for securities fraud. That the SEC and FTC haven't taken more aggressive action isn't a media problem or a market power problem.
But, taking a step back: Matt's argument is that the US is less competitive because market power precludes competition and innovation. I don't think Tesla, as many promises as they have broken, are an example of that.
For now, sure. But the integration of Musk's companies is making them a player that will price-out competitors from entering the market in the future.
One of the major reasons Tesla could exist in the first place is that the battery technology used by them (Li-ion) was invented by public-sector or energy-company researchers with the consequence technology open to a variety of applications companies.
I don't know what cross-licensing deal exists between Panasonic and Tesla, but there's a chance it would preclude other automotive companies from using Tesla's battery technology. And unless anti-monopoly enforces mandate Tesla license and sell this technology to competitors, those competitors will be locked out on price.
Tesla (and so much of "tech") would not be where it is now if venture capital didn't know that they could purchase monopolies that wouldn't be broken by anti-trust enforcers.
consequence = consequent
enforces = enforcers
Elon Musk's strategy of tech development squeezes out risk on the assumption he can feed it with enough cash for a long enough time. I'd rather mine hydrocarbons, oxygen and nitrogen in 24hr high noon sunlight from a near earth asteroid than go to Mars, but, his Starship plan is well thought out, makes the right compromises between practicality and performance and will succeed -- unless the money runs out!
That's why I hate his vulgar behavior -- it endangers the good things he has done and can do. My wife wouldn't remember if it was a Chevy or a Buick that crashed into her but she is attracted to Tesla vehicles, showing really wide currency for that brand.
He ought to practice real leadership, go straight and model good behavior by example. None of this "Pedo Guy" nonsense!
UPMC epitomizes much of what's wrong with American healthcare, and with large health systems. It got into a battle with the state when it decided to stop accepting patient health insurance from ONE Blue Cross provider - Highmark, insisting that if they wanted to be treated at UPMC facilities that the patients pre-pay. UPMC continued to accept patients with other Blue Cross plans, just not Highmark's. This is because, after years of each staying in their own area of expertise, UPMC started its own competing health insurance insurance program and Highmark retaliated, by moving into the hospital business. The state didn't really care about this regional spat until UPMC expanded to the more populated east and brought its predatory practices with it. Once it reached as far east as Harrisburg the state took action. UPMC caved to the state attorney general's valid complaints and UPMC again accepts patients with Highmark insurance again. At least for the next ten years.
For UPMC, it's all about revenue. They've bought up medical practices all over and their CEO Romoff was very upfront in saying that they wanted to be the Amazon of healthcare. Many physicians who still put patients first are quite frustrated.
UPMC, like Highmark, has abused its non-profit status for decades. It's hard to tell which is worse; they're both evil in their own way. But UPMC tried to take evil to a whole new level because it thought it could.
BTW, great book on monopolies, Matt. I bought it and enjoyed it thoroughly.
I like the format.
Plumbers are planners. There are design decisions that go into plumbing that have real impacts on the people who will use the plumbing. Why doesn't the Fed pledge to "buy up" (i.e. refinance onto the Fed's balance sheets) consumer credit card debt? Is this any more special than corporate debt? Consumers, like corporations, generally have assets that can be seized in the event of a default. With the significant drop in credit utilization this is obviously a "choked market" that needs to be "revived" ( https://www.foxbusiness.com/money/credit-card-use-drops-coronavirus ). Heck, the reduction in APR would have a significant stimulatory effect on the primary drivers of GDP (consumer purchases), without any effect on money supply like the stimulus payments.
----------
“really interesting computational social science papers” coming out of Facebook - I'm curious to what extent these are the equivalent of aerodynamics and failure testing by automotive and aerospace companies. I sincerely doubt Facebook's "social science" is seriously investigating the social science of people who don't use Facebook (30% of adults in the US). And to the extent they are it's highly doubtful they can accurately model them any near as well as they can Facebook users.
This isn't "innovation", it's basic "understand the product you're selling, and make a better product". Okay, I guess that later is "innovation", but it's not revolutionary innovation. It's a basic expectation of any company that designs products.
Love this issue, thank you for your work. What definitely stood out to me were the stories involving the DOJ. It is so important to point out when companies end up paying fines and settlements that are far less than the revenue obtained by corruption. There is no incentive for them to curb their practices in this case, they can see their corruption and those fines as an investment with a specific ROI which is, in this and many other cases, positive. I do have a degree in Business and its very clear that in those classes, future executives are taught to pursue endeavors with a positive ROI, why wouldn't these companies stay corrupt if they come out with a net profit of $850 million, especially if they have an obligation to keep growing quarter after quarter? Fascinating to see how inhuman some exective decisions can be.
I also truly appreciate the scrutiny of Facebook and the point you make about specific types of innovation. These tech monopolies have diversified their portfolios enough that they can now focus on these digital solutions like Machine learning and AI; they have the most resources so they can hire the best engineers and like you say, they have control over this data if only because they moved/consolidated quickly to dominate user market share. Google also can't be ignored in this respect (I appreciate that you touched on their inroads in banking software and education platforms). With respect to those crucial technologies that can determine a lot about the future of commerce, computing and even the makeup of the job market as more jobs are automated, we have an oligopoly and no one seems to realize it, its all just magic behind our apps made by smart people who probably know better than us.