Attempts at governance are futile until we recognize that we have built a machine and set it on its course. Instead, we pretend that the market is a natural phenomenon best left alone, and we fail to hold its mechanism designers to account. We need to tear down and rebuild that machine, reprogramming it so that human flourishing, not corporate profit, becomes its goal.
This article has useful insights. It points out that capitalism is an optimization engine for something. Shareholder value and CEO compensation, mostly. It reminds us of something that many people would prefer be forgotten - capitalism operates under rules set by governments, and those shape the market. Tax law matters - what's deductible warps spending patterns. Bankruptcy law matters - who gets left holding the bag? Corporation law matters - how immune are executives from the consequences of their decisions? Labor law matters - do people go to jail for wage and hour law violations?
All of those are changeable by legislation. Most have become much more favorable to companies since 1970. That's not necessarily permanent. We had a New Deal in the 1930s, following the Gilded Age of the 1920s and the crash of 1929.
There is a popular health blogger, Dr. Jessica Knurick, who makes similar arguments regarding the generally very poor health outcomes in the US, and I think her arguments are very persuasive. That is, our poor health outcomes are the logical consequence (not some anomaly) of a system that is designed to prioritize corporate profits over all other considerations.
The thing I find most convincing about these arguments is that I often hear many "free market boosters" say that any intervention in markets is a foolhardy idea, and that individuals are best equipped to make their own decisions. But IMO this is a BS argument because it pretends that we don't already have tons of rules built into the system that prioritize specific values and outcomes (e.g. how corporate formation often protects individuals from some disastrous consequences of their decisions).
I think that's the most important idea in these discussions: we do not live in some "default" state of fair, open competition. The system we built is a consequence of the rules we laid down to bring it into creation, and we can change those rules if we decide the current outcome is not what we desire.
I don’t think capitalism is an optimization machine for CEO compensation. I think it is just for shareholder value; if shareholders thought they could pay a CEO less they would in a heartbeat.
Why don't you say what you, yourself, think people mean when they use that very common word, and ask this person if they understand it the same way, or if they do not, what the differences are?
Asking people to put a bunch of work in that you're just going to dismiss with a no true scotsman is rude. I'm not saying that you're doing that, but it sure looks like you're doing that.
I was thinking about the industrialists of the late 1800s early 1900s.
Had we explained to them what modern capitalism is doing, they would be probably quite impressed by the productivity gains, but utterly perplexed at the offshoring and outsourcing to other countries.
Yes, it makes things cheaper, but they also knew about manufacturing bases required for defence, and innovation.
Fraud and lemon law strictures aren't counterproductive. The FDA isn't counterproductive. "Voluntary trade" reminds me of the idea that "all natural" products are inherently good. Nature (and other people) will commonly engage in toxic, exploitative behavior without intervention.
I don't think that's accurate as a matter of history. My understanding is that noncommercialized societies don't really use barter systems. They use credit systems, where if you need something you ask for it as a favor and you'll have to repay the favor later in a form determined then, not now.
> The market is a natural phenomenon of voluntary trade.
But the market requires oversight to ensure that you have meaningful choices. Otherwise, monopolies and corruption will work to remove your preferred options, forcing you to pay rents.
Depends on how low the barriers of competition are. Restaurants have been capitalist markets for hundreds of years and have never devolved into monopolies, even though they exist under relatively little oversight. Because the level of competition is so high.
This is also true of a wide variety of small businesses.
More often than we think, monopolies arise out of well-meaning legislation that ends up restricting competition, resulting in oligopoly/monopoly
I don't know about the restaurants having little oversight. Laws regulating restaurants are also as old as time, they are literally in the Code of Hammurabi!
"If a tavern-keeper does not accept grain according to gross weight in payment of drink, but takes money, and the price of the drink is less than that of the grain, she shall be convicted and thrown into the water."
Sure, but it sounds like you would acknowledge natural monopolies do also exist, which gets us back to the problem of creating rules for the marketplace
Counterproductive doesn't mean worthless. Better distribution of fewer resources is, at a certain point, desirable contra massive inequality with higher overall wealth. Regulation can also create wealth, it's not impossible. There's no natural law that requires unregulated barter to be even a local optimum.
That is simplistic. Without property rights established by laws and monopoly of violence to police and military there is not much market. There is nothing "natural phenomenon" of the formation of modern shareholder companies operating as separate legal entities but modern capitalism is built around it. You can't sell poisonous food without consequences.
Any of those you would call counterproductive?
A market is a human invention and construction, unless you are talking about a kindergarten versio of it that is so simplified it borders to parody.
I don't disagree with anything O'Reilly is saying but lets also call out the proximate cause which is that ZIRP era allowed these companies to grow to grotesque proportions and incentivized upstarts to shoot big in green fields rather than compete in known markets. Now interest rates are back and tech companies are cannibalizing their core competencies on the back of AI. Will markets begin slapping wrists for reckless spending and will upstarts see weakness and stab at exposed bellies? No guarantee they will but it is more possible than analysts are realizing.
Startups are about the least leveraged businesses out there and software, pre AI, was one of the least capital intense. This is not a matter of interest rates. The reason we don't see as many shots at greenfields now is because the simple ones are already taken.
If you want to claim that people are going into investment grade instead of high yield credit because of this, then that would be believable, but capital allocated to VC funds is chasing bigger returns and is not going to be distracted by 5% on the 10Y. VCs have not tightened up and are throwing money at anything AI related just like they were at anything "blockchain" a few years ago.
No, monetary policy isn't responsible for enshittification. Cheap money reduces pressure to enshittify (investors more willing to delay returns) and increases pressure to not-enshittify (competition is easier to capitalize), if anything, though it's certainly not a sustainable lever to achieve that.
No, at its core enshittification is just capitalism in the absence of consumer bargaining power. Every year, companies alter the deal and it gets a bit worse for consumers and a bit better for shareholders. Consumer protection is asleep at the wheel, antitrust is shot dead at the wheel (thanks Ronald Reagan and Robert Bork), and the hands-off approach enters its Nth year of doing what you expect it to do if you are smart and the opposite of what you expect if you are the opposite of smart.
The yuppies who dominated wallstreet in the 80s started streaming to SV. Now we have “VC influencers” - they sit around in a podcast style and talk endless nonsense about the “future” in a very self serving way. I don’t know if SV ever had integrity but these days it seems thinner than ever.
Not really. Most management, VC, and leadership in tech are ex-Engineers. It takes an engineering mindset to successfully optimize.
The world for those us who started in the 2010s is much more competitive than it was for the anachronistic minority of hippie types you'd see in SV when growing up (PS - Dario Amodei's parents were similar).
Since the early 2010s, getting into a target or semi-target CS/ECE/EECS/EE/CE program required having a 1500/2200+ SAT; 12-14 APs; honors societies like NHS & CSF; multiple semi-competitive ECs like Speech&Debate, Sports, Orchestra; and being competitive in Olympiads. Forget CS@Stanford - you won't get into CS@SJSU or UCR without a similar profile now.
To succeed this requires a deeply competitive, vicious, and tactical personality and mindset.
Those of us who came out of this background are now leadership (it only takes 10-20 years out of college to reach that position). Either you should be able to compete or you will go the way of the dinosaurs or the Aztec.
Go on LinkedIn and see the career history of most C-Suites, VCs, PMs, EMs, and VPs in any major tech company HQed in the US.
Heck, your undergrad alma mater's CS and ECE program has been a target program for VC analysts and PMs for over a decade now. And take a look at the trajectory of most of your school's alma maters now in tech leadership, founder, and VC positions.
Also perplexed by the phenomenon of those AI/"tech" talk shows where the apparent premise is a host interviews their own friends/coinvestors/roommates/etc, so they can agree about how awesome and correct they all are for 60 minutes.
Great breakdown of the problem, but we get no viable path to a solution, and IMO where are some misguided comparisons.
1) When Unix was made the hardware gap between Bell Labs and a uni comp-sci dept was a single minicomputer. Bell Labs was also legally barred from commercializing it due to the 1956 AT&T Antitrust Consent Decree. Today the frontier AI requires gigawatt-scale data centers costing a couple of tens of billions of dollars. No open software protocols or MCP servers can decentralize a market when the underlying compute substrate in itself is an industrial monopoly.
2) The author argues that governance fails unless we reprogram the corporate incentive machine, yet he expects AI labs to voluntarily build open ecosystems inside this capitalist market he just proved is inescapable with the shareholder arguments. Even if every lab agrees that displacing consumers ruins the macroeconomy, any lab that sacrifices margins for "human flourishing" will just be out-competed.
---
Like, he is against government regulations, but at the same time fails to see that Bell Labs would have probably commercialized UNIX if it was not forced to not do so by the antitrust law. So the previous solution is ignored, and the new solution 'reprogram the machine' has no plan or suggestion on how to do it.
Bill Gross came up with pay-per-click advertising, and then Google perfected it. And that totally was this new, again, once again, mechanism design, pay-per-click advertising, totally transformative economic mechanism.
Tim implies this is great, but...
Advertising on the commercial web led to adtech, which "perfected" surveillance capitalism, which is one of the things he is complaining about.
Perhaps he felt it was acceptable until "don't be evil" (rather inevitably) started limiting Google's growth (and was therefore abandoned.)
Greed, quasi-monopolies, megalomaniacs, fascists: Basically what you get through out history if a society allows for too much power getting concentrated in the hands of a few. And it never ends well for society.
> As Herbert Simon noted in 1971, an abundance of information leads to a scarcity of attention.
We are now about 15 years past when most Americans could be expected to have a smartphone, and nearly every kind of long form content is having a little trouble breathing like dinosaurs after a giant meteor. The entire movie industry appears broken and is basically taking its content from youtubers (a refreshing update to a long suffering dinosaur), only a very few (ie: none) musicians can make a living from selling records, and with the exception of perhaps you folks on hacker news, few people read books.
Depending on what you mean by "AI company" (which is an extremely ambiguous term that deserves a pretty rigorous definition before we talk about economics), it absolutely can be both.
Hell, it can be both within a single company. A commercial airline is both a famously terribly unprofitable venture and one that employs insane rent-seeking tactics (premium tiers, seat shrinkage) routinely. Hospitals and other healthcare industry participants are the same way: massive cost collapse leads to businesses closing and PE buyouts, at the same time as parasitic rent-seeking ("Office Visit" ICD codes, anyone?).
The recipe is usually: extremely inelastic demand/captive audiences for a service means that competition-based pricing assumptions pretty much immediately lead to rent-seeking and/or monopolistic behavior, and at the same time the margin pressure on the business might be intense.
Or, even more briefly: slim or negative margins don't always indicate a pareto optimum or the presence of competition. Sometimes, that just means that a competition/market-based pricing model isn't a good fit for a given industry.
But in reality they are not. Tech companies have a more favorable ownership split to founders and employees vs investors than any other major industry.
I think you can make a base thermodynamic argument. Value creation means climbing hard up against entropy. It means searching large solution spaces. It means risk with unknown payoff.
A rough analogy would be drilling a gas or oil well versus replicating what plants do with solar panels or harnessing atomic energy directly. Those are way harder than burning what was already deposited at high concentration. An imperfect but not totally wrong comparison I think.
Value extraction just requires that you identify value already there and make a plan to get and store or monetize it.
A lot of the debate over the merit of capitalism would be clearer if we made a hard distinction between value creating capitalism and value extracting capitalism. The latter is closer to imperial conquest or raiding the next village but with many more steps and in the modern world often without direct force.
The most common way to sack the populace today is trickery and exploitation of the human dopamine system. Financial con games rank highly too.
Value creating capitalism gave us modernity and a higher standard of living than Pharaohs. Raiding and conquest was what humans were jacking off doing for all the preceding 100k+ years when standard of living barely moved for most people.
> We got there the slow way, through steady organic growth. That is quite anathema to the rush to the top of whatever appears to be the latest hill that seems to characterize Silicon Valley today
It went from an all-round aspirational place to work to something you'd rather keep to yourself depending on the audience. Sort of like if you work for Raytheon.
I highly recommend the book "Incorruptible" by Eric Ries. It is about how to build companies that don't become enshittified (financially, or product-wise) and it comes to a similar conclusion: the mission needs to be about "human flourishing". A great book.
This article has useful insights. It points out that capitalism is an optimization engine for something. Shareholder value and CEO compensation, mostly. It reminds us of something that many people would prefer be forgotten - capitalism operates under rules set by governments, and those shape the market. Tax law matters - what's deductible warps spending patterns. Bankruptcy law matters - who gets left holding the bag? Corporation law matters - how immune are executives from the consequences of their decisions? Labor law matters - do people go to jail for wage and hour law violations?
All of those are changeable by legislation. Most have become much more favorable to companies since 1970. That's not necessarily permanent. We had a New Deal in the 1930s, following the Gilded Age of the 1920s and the crash of 1929.
The thing I find most convincing about these arguments is that I often hear many "free market boosters" say that any intervention in markets is a foolhardy idea, and that individuals are best equipped to make their own decisions. But IMO this is a BS argument because it pretends that we don't already have tons of rules built into the system that prioritize specific values and outcomes (e.g. how corporate formation often protects individuals from some disastrous consequences of their decisions).
I think that's the most important idea in these discussions: we do not live in some "default" state of fair, open competition. The system we built is a consequence of the rules we laid down to bring it into creation, and we can change those rules if we decide the current outcome is not what we desire.
The 1920s boom was mostly due to electrification.
When you say "capitalism", what do you mean? How would you determine whether any given phenomenon was or wasn't "capitalism"?
Asking people to put a bunch of work in that you're just going to dismiss with a no true scotsman is rude. I'm not saying that you're doing that, but it sure looks like you're doing that.
Had we explained to them what modern capitalism is doing, they would be probably quite impressed by the productivity gains, but utterly perplexed at the offshoring and outsourcing to other countries.
Yes, it makes things cheaper, but they also knew about manufacturing bases required for defence, and innovation.
But the market requires oversight to ensure that you have meaningful choices. Otherwise, monopolies and corruption will work to remove your preferred options, forcing you to pay rents.
This is also true of a wide variety of small businesses.
More often than we think, monopolies arise out of well-meaning legislation that ends up restricting competition, resulting in oligopoly/monopoly
"If a tavern-keeper does not accept grain according to gross weight in payment of drink, but takes money, and the price of the drink is less than that of the grain, she shall be convicted and thrown into the water."
Any of those you would call counterproductive?
A market is a human invention and construction, unless you are talking about a kindergarten versio of it that is so simplified it borders to parody.
it's that the cost of money was cheap so the opportunity cost of investing in a startup was cheaper
when you can get a decent rate for free, you're less interested in chasing unicorns
No, at its core enshittification is just capitalism in the absence of consumer bargaining power. Every year, companies alter the deal and it gets a bit worse for consumers and a bit better for shareholders. Consumer protection is asleep at the wheel, antitrust is shot dead at the wheel (thanks Ronald Reagan and Robert Bork), and the hands-off approach enters its Nth year of doing what you expect it to do if you are smart and the opposite of what you expect if you are the opposite of smart.
The world for those us who started in the 2010s is much more competitive than it was for the anachronistic minority of hippie types you'd see in SV when growing up (PS - Dario Amodei's parents were similar).
Since the early 2010s, getting into a target or semi-target CS/ECE/EECS/EE/CE program required having a 1500/2200+ SAT; 12-14 APs; honors societies like NHS & CSF; multiple semi-competitive ECs like Speech&Debate, Sports, Orchestra; and being competitive in Olympiads. Forget CS@Stanford - you won't get into CS@SJSU or UCR without a similar profile now.
To succeed this requires a deeply competitive, vicious, and tactical personality and mindset.
Those of us who came out of this background are now leadership (it only takes 10-20 years out of college to reach that position). Either you should be able to compete or you will go the way of the dinosaurs or the Aztec.
The engineer in me needs proof for this.
Heck, your undergrad alma mater's CS and ECE program has been a target program for VC analysts and PMs for over a decade now. And take a look at the trajectory of most of your school's alma maters now in tech leadership, founder, and VC positions.
1) When Unix was made the hardware gap between Bell Labs and a uni comp-sci dept was a single minicomputer. Bell Labs was also legally barred from commercializing it due to the 1956 AT&T Antitrust Consent Decree. Today the frontier AI requires gigawatt-scale data centers costing a couple of tens of billions of dollars. No open software protocols or MCP servers can decentralize a market when the underlying compute substrate in itself is an industrial monopoly.
2) The author argues that governance fails unless we reprogram the corporate incentive machine, yet he expects AI labs to voluntarily build open ecosystems inside this capitalist market he just proved is inescapable with the shareholder arguments. Even if every lab agrees that displacing consumers ruins the macroeconomy, any lab that sacrifices margins for "human flourishing" will just be out-competed.
---
Like, he is against government regulations, but at the same time fails to see that Bell Labs would have probably commercialized UNIX if it was not forced to not do so by the antitrust law. So the previous solution is ignored, and the new solution 'reprogram the machine' has no plan or suggestion on how to do it.
Tim implies this is great, but...
Advertising on the commercial web led to adtech, which "perfected" surveillance capitalism, which is one of the things he is complaining about.
Perhaps he felt it was acceptable until "don't be evil" (rather inevitably) started limiting Google's growth (and was therefore abandoned.)
P/E ratio of 600 seems crazy unless the assets balance it out somehow.
We are now about 15 years past when most Americans could be expected to have a smartphone, and nearly every kind of long form content is having a little trouble breathing like dinosaurs after a giant meteor. The entire movie industry appears broken and is basically taking its content from youtubers (a refreshing update to a long suffering dinosaur), only a very few (ie: none) musicians can make a living from selling records, and with the exception of perhaps you folks on hacker news, few people read books.
- the average tech bro
Hell, it can be both within a single company. A commercial airline is both a famously terribly unprofitable venture and one that employs insane rent-seeking tactics (premium tiers, seat shrinkage) routinely. Hospitals and other healthcare industry participants are the same way: massive cost collapse leads to businesses closing and PE buyouts, at the same time as parasitic rent-seeking ("Office Visit" ICD codes, anyone?).
The recipe is usually: extremely inelastic demand/captive audiences for a service means that competition-based pricing assumptions pretty much immediately lead to rent-seeking and/or monopolistic behavior, and at the same time the margin pressure on the business might be intense.
Or, even more briefly: slim or negative margins don't always indicate a pareto optimum or the presence of competition. Sometimes, that just means that a competition/market-based pricing model isn't a good fit for a given industry.
I think you can make a base thermodynamic argument. Value creation means climbing hard up against entropy. It means searching large solution spaces. It means risk with unknown payoff.
A rough analogy would be drilling a gas or oil well versus replicating what plants do with solar panels or harnessing atomic energy directly. Those are way harder than burning what was already deposited at high concentration. An imperfect but not totally wrong comparison I think.
Value extraction just requires that you identify value already there and make a plan to get and store or monetize it.
A lot of the debate over the merit of capitalism would be clearer if we made a hard distinction between value creating capitalism and value extracting capitalism. The latter is closer to imperial conquest or raiding the next village but with many more steps and in the modern world often without direct force.
The most common way to sack the populace today is trickery and exploitation of the human dopamine system. Financial con games rank highly too.
Value creating capitalism gave us modernity and a higher standard of living than Pharaohs. Raiding and conquest was what humans were jacking off doing for all the preceding 100k+ years when standard of living barely moved for most people.
"Today" being the last 35 years at least...