- What if Google didn't have more money than god, and couldn't afford to bankroll Waymo ~10b?
- Would TSMC have gotten as far if Apple couldn't reliably buy out months of capacity on the next node, bankrolling TSMC's tech development?
- Would we even have >10B-param LLMs at all?
There is something to be said for the concentration of resources, such that they can be deployed on projects with payoffs years or decades later. The same could be said for all the tech that came out of Bell Labs or PARC. Advocating for smaller businesses is advocating for shorttermism to some degree; even startups today are funded based on the premise that they could potentially capture an entire market in a few years.
Large projects need resources, but who decides how those resources are deployed, to what end, and who benefits from them is the important part.
All of your examples are profit-driven, and not necessarily (even if we do benefit) done for the greater good of all or advancement of society.
We can still accomplish big innovations without those innovations coming from, or being controlled, by nation-state private companies. We've ran high-risk R&D projects successfully before as public projects - Manhattan project, the space race & moon landing, ARPANET, etc.
Waymo could still happen, only it'd be a publicly funded project and the societal benefit would be broader and not tied to a single company's market dominance.
> but who decides how those resources are deployed
The current system selects people that have allocated resources effectively in the past by providing them more resources to allocate.
> We've ran high-risk R&D projects successfully before as public projects
And what is stopping countries from doing this today? This isn't an either or thing, public projects can still exist, there is no law of nature saying that massive companies are the sole source of innovations but for some reason people treat it like they are mutually exclusive. You bring up projects from decades ago but are there any modern examples?
Since when has throwing money at systems that haven't shown success worked? And you are suggesting that we take money from others to throw it at a system that doesn't work.
"Allocated resources effectively" is doing a lot of heavy lifting in your description. The current system also rewards large-scale cons and rent extraction with the political power to do more of the same.
It is classic circular reasoning. Why should they have all money? Well because they are the best ones at allocating the resources. How do we know they are best at allocating resources? Because they have all the money.
"To him that hath, more shall be given
and to him that hath not, more shall be taken away"
I suppose it was more popular in seemingly simpler times when the playing field was more even and the players more evenly matched and distributed. But here we are in the future, and that game seems to have been concluded.
I myself rather preferred the view from the shoulders of giants than the undersides of their feet.
There is a agent simulation paradigm where a large amount ofagents engage in one-on-one transactions and wealth is transferred. I haven't followed but decades ago there was a simulation where with rather simpler rules, in the end all the wealth concentrated to one agent.
In this model, it's the rules of the game, not the morality of the players that causes the effects.
It seems you're implying that since being a billionaire is not 100% heritable, it must be a result of individual merit (leap 1) which is purely synonymous with skillfully allocating resources (leap 2) more than any other type of person (leap 3) rather than any other explanation for how they got there, say, sociopathy, right place right time, etc.
> It seems you're implying that since being a billionaire is not 100% heritable, it must be a result of individual merit
Nope, another strawman, I never claimed it was 100% merit. You claimed that wealth is purely heritable which is easily proven false by observing reality.
> sociopathy
So your claim is that sociopathy is more important for getting rich than skill?
> right place right time
Choosing the right business to build given the state of the world is a skill. Business requires luck the same way poker requires luck, and let me tell you, if you play poker vs a skilled player you will lose a lot of money.
Yes exactly. Also, we should stop pretending that the money supply is fixed and that everyone exists on the same monetary playing field.
Arguments about efficient allocation are laughable when you consider that someone who is socially 6 steps removed from an institutional 'money printer' lives in a monetary environment where money is 10 times more scarce than it is at the source (due to taxation between each hop). Few people are so far removed in practice but the effects are still very powerful even with less distance. Taxation brings all economic activities closer to the government and banking sector.
In competitive industries were profits are paper thin, monetary asymmetry can fully determine business outcomes. The company receiving government contracts on the side has a massive upper hand over its competitors during a monetary contraction. Same can be said about companies which operate in environments where their customers have access to large amounts of credit by virtue of their highly valued collateral. Their success has little to do with optimal allocation and a lot to do with socio-economic positioning and monetary system design.
> The current system also rewards large-scale cons and rent extraction
I never claimed it was a perfect system and I am more than willing to admit rent extraction, scams, and monopoly power are massive issues with capitalism. It still hasn't been shown that replacing that with government is better.
> You bring up projects from decades ago but are there any modern examples?
There are no modern examples in the USA precisely because there is no political will to fund them. That political will is undermined because private companies are large and strong enough that they can influence politicians and prevent projects they they would have to compete with.
Your argument has the implication the wrong way around.
Meanwhile, look at China. The vast majority of China's cometlike economic success can be directly attributed to state funding, and many of its successful projects are directly state-run. That's because China still plays industrial politics that are concerned with economic growth and public welfare instead of rent-seeking.
> The current system selects people that have allocated resources effectively in the past by providing them more resources to allocate.
The current system selected people that have maximized shareholder value and financially engineered it into other financial assets that provide power under the capitalist system. This includes private equity services that have simply squeezed money out of consumers for no increase in quality of life, or companies that managed to avoid the consequences of the externalities of their economic activity.
> And what is stopping countries from doing this today? This isn't an either or thing, public projects can still exist, there is no law of nature saying that massive companies are the sole source of innovations but for some reason people treat it like they are mutually exclusive. You bring up projects from decades ago but are there any modern examples?
Regulatory capture and lobbying that attempts to force a profit motive behind every large government initiative when the profit motive substracts value away from society at large.
To be concrete, I guess in a different time and society someone like Sam Altman would have been a successful politician or perhaps like Hyman Rickover or Marcel Boiteux working within the government to cause colossal steps in progress.
> We can still accomplish big innovations without those innovations coming from, or being controlled, by nation-state private companies. We've ran high-risk R&D projects successfully before as public projects - Manhattan project, the space race & moon landing, ARPANET, etc.
So you prefer nation-state nations over nation-state private companies. :-)
Is that a bad thing? Democracies, even if flawed, are accountable to their citizens at least to an extent. Their power structures aren't dictatorial, like in a company, and can be steered and course-corrected. Private companies are accountable to no one, their only motivator and reason for existence is profit. If they're allowed to run our society, the outcome that we're seemingly inching towards, there will be nothing to stop their inevitable abuse for the sake of value extraction.
It seems to me none of this is as clear cut as it seems. Government entities may hold shares in private companies, companies may act on voter's demands by accepting government grants. For some companies, government contracts are actually a main revenue stream - shareholders can jump up and down, if their supporters are voted out they will falter.
Corporations are still subject to law and ultimately under the control of the government. The current set of rules just gives them a fair amount of freedom to operate.
This only makes sense in a spherical-cow-in-vacuum world where government and business are somehow barred from communicating with one another. In reality, the "current set of rules" in many countries is a result of companies relentlessly trying to and succeeding in finding ways of influencing government. Political advertisements, campaign funding, lobbying, corruption, underhanded favoritism, countless other methods that are an amazing RoI for any business large enough to engage in it. Large enough corporations are resembling governments more and more in terms of value and power, and they use all of that power to endlessly try to bend the rest of society into serving their profit motive.
1. Not all companies are publicly traded, and they don't have to be
2. The 'voting' and 'steering' in a corporation is also completely dependent on money. The value of your 'vote' is proportional to how much money you have. This isn't a democracy or some sort of equal system that will converge on serving people, it will converge on serving money. I'm genuinely baffled at how "you get one vote per person" and "your value and voting power is directly tied to your net worth" are in any way comparable. You and I have zero effective power over them, and always will.
> The 'voting' and 'steering' in a corporation is also completely dependent on money.
It is not, eg. Zuck didn't control Facebook because he was a priori rich, he became rich because he controlled Facebook in a successful way. He gained those shares and that control with his skill and labor (and maybe one symbolic dollar or something).
The unit of power in stock corporations is dollars, while in democracies it is personhood. In the former, one person can acquire multiple units, while in the latter they cannot. There is an obvious difference.
No, a corp is only accountable to their board and the LARGEST shareholder. A single person can control an otherwise publicly traded company. Zuckerberg, for example. And not everyone can afford to spend their earnings owning companies. So what you get with a democracy is that power is spread out by default rather than concentrated in a single element. Default enfranchisement rather than the polar opposite. One at least nods politely at the idea of upward social mobility in passing while the other eschews all pretense as to the status of its party invitation.
Corporations only have any accountability in so far as the law of the nation-state they're incorporated in grants it.
Corporations only exist as a legal construct of other entities. Absent government, they wouldn't be corporations since there'd be no law to create them.
The problem is, when the majority is held by pension funds, ETFs and Blackrock... there isn't much governance in practice, particularly from the low-fee purely passive ETFs. And since government run pension schemes are on their way to the gutter in favor of stonk market private pensions, the share of such dumb passive capital will only grow.
If there is one thing that we learned from the 20th century it is that societies that gave more private control over how resources were used did better than those that had more state control over how resources were used. Perhaps in the 21st century that has changed, but for me that falls under the "extraordinary claims" category with the corresponding evidence requirements.
The structure of giant corporations today is like those centralized societies that were so inefficient in your example. The mandates to put AI in everything are one example of out of touch leadership throwing money and effort blindly towards things of dubious value. The sycophantic managers, afraid that they will be eliminated for insufficient fervor for the board’s latest fascination, will seize upon anything to prove themselves loyal and useful to those above them.
By moving the locus of control, whether it be considered the ceo or shareholders, so far from the actual business and implementing mandates based on whatever the current fancy is and meaningless targets of growth on such a giant scale you get the same sort of excesses.
The current system is marked by irrationality and uninformed and ill considered decision making. With smaller organizations and actual business competition they would be held to account by their competitors or just by running out of money before something catastrophic for the greater economy happened.
This is an excellent point. Thank you for posting it.
Large monopolistic mega-corporations do tend to have the same issues that one would see in the old 20th century planned economies like the Soviet Union.
Not sure why this obvious fact is being down-voted. The comments above don't mention that the killer feature of private orgs is the ease of exit, and therefore, the enormous risk of failure. This remains the dominant feature of private orgs, even if we can argue about certain orgs on the margin. For every example of "users are locked into either the Apple or the Android phone platform", I can think of several crappy Google and Apple products which failed and were withdrawn from the market (e.g. Google Wave).
It is much easier to exit from or steer a private org. For example, it is very possible to run a company which caters to 10 percent of a consumer base by providing niche products which may be slightly more expensive. Those 10 percent will simply consume less of some other good. It is very difficult to do an analogous thing at the state level, because we generally don't get individual "ticket books" which we can "spend" on more of one state service vs. another. The democratic model is that you first get 50+ percent support and then your coalition decides how resources are allocated for almost everyone.
> If there is one thing that we learned from the 20th century it is that societies that gave more private control over how resources were used did better than those that had more state control over how resources were used.
In some metrics (such as GDP), yes. And in other metrics (such as wealth inequality and health care), the answer is less clear-cut.
The West encompasses a wide gradient of private vs. state control over resources, and there are states which aren't typically considered Soviet or Western (e.g. Nordic states.)
I do, when citizens have the ability to steer their nation state (i.e. the ability to vote.)
On this note, I've lived in a couple states with ballot initiative processes and while they are not perfect, I think they are absolutely necessary for citizens to truly be able to hold their elected representatives accountable (i.e. override them) and I wish we had them at a federal level.
You also have a "vote" as a consumer. The market could be much more responsive than a "democratic" system.
For instance, say you think pesticides are a bad thing. You can get 49% of the population to vote to the ban them and what do you accomplish? Nothing
No wonder people look at politics with despair.
If you can get 5% of the population to eat organic food on the other hand, you've reduced pesticide use by 5%. You create trade associations, the idea of organic food spreads more widely and maybe someday you get enough support that you can change the law.
>
You have to buy your way into voting as a shareholder. In a democracy, it's just your given right as a citizen.
In a democratic country, only the people who have citizenship are allowed to vote. In a shareholders meeting, only the shareholders are allowed to vote.
You sometimes cam buy your way into citizenship. As a shareholder, it is your given right to vote in a shareholders meeting.
Even if you think there's no qualitative difference between the 2 (which I think is a deeply immature idea, but whatever), there's an obvious quantitative difference: In practice democratic voting power is much more socioeconomically spread and shared than shareholder voting power.
> Shareholders receive power proportionate to their buying power. Citizens get a single vote.
Historically, there did exist experiments that not each person has the same voting power (for example the Prussian "Dreiklassenwahlrecht" [three-class franchise]):
Depending on the amount of taxes you paid, you were assigned to one of three classes. The sizes of each of these classes were chosen so that each class paid 1/3 of the whole tax volume. The votes in each class elected representants for this class.
The idea is obvious: those who pay a lot more taxes should have more influence.
Thus: each citizen has the same voting power is just the "currently fashionable" implementation of democracy.
I think California's system is a mess. It's been overrun with private interests bankrolling ballot initiatives and steamrolling them through. Add to that the Government itself sponsoring ballet initiatives that sale bonds to finance things that people don't understand really loans and it turns into a mess. I do like the ability to remove politicians from office via ballot though.
One idea that I think is reasonable is to use some kind of actual meetings.
Dividing people into groups of 50 or 100. Initiatives are voted on in these groups, if they are passed they go to the next level, 1000 people.
Sort of like that idea in the Yes, Minister episode about 'genuinely democratic local government'. The idea here is the tree structure is to prevent people to push initiatives other than as individuals.
I think that's unavoidable, but also not a bad thing. If we were to undertake any similar large scale public projects today, it would also have significant private sector participation. But, that drives job creation and positive effects on the economy (i.e., new deal).
The main difference is, ideally, the project was voted on by the public, and is being steered as such. A public-private collaboration, with the public driving it rather than it being entirely the domain of a single private entity for their own profit.
Might I note that we're building down nation states at a rapid rate at least in the west trough migration. The vast majority of my capital is not comprised of belgian ethnicities for example and the same will start to hold trough for a number of the other biggest cities.
"All of your examples are profit-driven, and not necessarily (even if we do benefit) done for the greater good of all or advancement of society."
Without profit as a motive, innovation would be decades behind (if not longer than this). Governments can barely afford crumbling infrastructure maintenance as it is. I seriously doubt they are going to invest in projects for the 'greater good'.
"We've ran high-risk R&D projects successfully before as public projects - Manhattan project, the space race & moon landing, ARPANET, etc."
Yes, for defensive or offensive military purposes. Not much beyond this.
Even big pharma supplies the world. The rest of the world with socialized medicine create knock-offs at a fraction of the cost, because they didn't have to spend decades going through testing and billions of dollars developing it.
> Without profit as a motive, innovation would be decades behind (if not longer than this). Governments can barely afford crumbling infrastructure maintenance as it is. I seriously doubt they are going to invest in projects for the 'greater good'.
The reason that governments have such a restrictive budget in the first place is people are individually profit-motivated. Governments do invest in projects for the greater good - you yourself note "big pharma" research, and in fact historically the US gov provided more than half the funding of all basic research nationally.
> Yes, for defensive or offensive military purposes. Not much beyond this.
Shinkansen.
Anyways, governments across the world are driven by incentives that do recognize long-term economic/strategic interests. You can see it with AI, with climate change, even with the broad desire to create a "homegrown" Silicon Valley.
> The reason that governments have such a restrictive budget in the first place is people are individually profit-motivated
You've got the cart before the horse; the government would not have a budget at all if people were not individually motivated to generate taxable events.
Profit is the practice of accumulating more resources than you immediately need in the anticipation of their future use and enjoyment. Without a government, a profit makes the bearer a target for anyone who can overpower you. So the essential purpose of a government is the preservation of profit by opposing the forces that would destroy or carry it off: criminals, scammers, foreign militaries.
Governments did not command the invention of penicillin, powered flight, electric light, transistors, the blue LED, or the majority of software products that are essential to society today. But it protected individuals to invent with the knowledge that their work could be rewarded on some timeframe rather than being immediately destroyed by an interloper.
Maybe not government, but much public resource was not driven by profit.
The entire software industry relies on a foundation of free and open source software. The profit model wouldn't even exist without the work of people who do it purely for passion.
Penicillin was university research and given away freely.
The Wright brothers weren't trying to start an airline, that just wanted to fly.
ironically, tho not recently government, the majority of your examples were not discovered/created in the pursuit of profits.
Adam Smith points out that it doesn't matter if things are done for the greater good or not. Them existing will improve lives for everybody either way.
The problem with government funded large project is that they often monopolize. In the Soviet Union to much investment was flowing into the wrong stuff. And small scale innovation didn't get the resources to grow.
While mistakes on a high level lead to total stagnation. For example NASA building the Shuttle crowed out almost everything else, and because the Shuttle was the wrong way to go the US has spent 50 years doubling down on that.
If a large company makes a really, really big bet on something, they can pay a very large cost if they are wrong. And this has been historically the case, large projects that don't get anywhere are canceled. Governments can double down almost forever.
So I think we do need everything, innovation from maverick individuals, innovation from smaller companies, large innovation form big companies, innovation from government project, innovation from private/government funded universities.
There is no perfect 'this is the way to get innovation' that we have yet discovered.
Greater Good of All is a bit nebulous, and quite often translates into rather concentrated good of a few well-connected players.
When you mention the space race, you should also add that once the Moon landing was over, the government-supported part of space activity got mostly bogged down in cost-plus boondoggles (see: Space Launch System, also called Senate Launch System), and without a vibrant private sector with deep pockets, the US would be launching maybe some twenty rockets a year now, more likely twelve, each at an extreme cost and without much technological progress. And American capability of supporting human spaceflight would be tenuous at best, or possibly nonexistent.
(NASA is not at fault here. The politicians which command it, though... they seem to love giving Boeing et al. expensive projects.)
Major weapons development (or dual use) programs — especially for weapons of mass destruction — probably have to be government run due to national security concerns. But the notion that governments can effectively manage technology R&D projects is ludicrous. Look at what happened when Japan's MITI tried to run a Fifth Generation Computer Systems (FGCS) project: total failure and waste of tax money.
In general, economic central planning is a dead end. People keep trying to claim that it would be more efficient or benefit society but it just doesn't work. Bureaucrats and politicians can't be trusted with resource allocation decisions.
> for the greater good of all or advancement of society
I'm not sure we agree enough on the definitions of these things to justify a democratic redistribution of resources towards them. Tyranny of the majority is still tyranny after all. The nice part about private enterprise is that it's hard to argue they didn't earn their money. Google, Apple, et al provided some value to some folks who volunteered to pay for it in a free exchange. Their claim to use their earned wealth as they see fit is much easier to substantiate than a government intervention which is neither voluntary nor obviously providing value to the people who pay for it.
> or being controlled, by nation-state private companies
You don't need to publicly fund something to "control" it. You can pass laws about it, you're the government!
I always find this kind of rhetoric from Western socialists strange, it's like they've forgotten they aren't libertarians.
The other issue being that they've forgotten there are downsides to owning something, because you don't always want to own all the risk. eg if you're a tech employee and paid in your employers' shares, you own some of your means of production. But you /shouldn't/, it's way too risky! You should sell it as soon as you get it and put it in an index fund.
> Waymo could still happen, only it'd be a publicly funded project
You really think so? All of the examples you gave are military technology during wartime, which the government does tend to be able to do since the existential risk motivates the organization to root out graft and free riding.
I could see some kind of alternate reality future government funded Waymo being spun out of drone tank tech from WWIII but we wouldn't have it today.
Wait, could you remind me what war was going on when NASA took us to the moon?
Could you remind me what war was going on when the CDC eradicated malaria from the United States?
Could you remind me what war was going on when FDR build our basic social safety nets?
Broadly speaking, people have 3 ways to organize large groups: business, government, and (organized) religion. Each has strengths and weakness. To say that only one can produce social good is a bit of a stretch.
The whole thing is worth a read too, it explains all the other military use tech that will arise from the self driving car ecosystem, further justifying the investment.
You cannot be serious, the whole thing was called the space race for a reason. Space tech has always always been primarily a military venture, and it remains so to this day.
> Malaria
Glad you asked, chloroquine was developed during WW2 for soldiers, and chloroquine resistance of soldiers in Vietnam drove the creation of mefloquine and artemisinin.
> Social safety nets
Not a science breakthrough
> To say that only one can produce social good is a bit of a stretch
I 100% agree. It's not "everything ever created was because of war". It is rather that "a lot of difficult amazingly unimaginable things i.e 'root node science' would have never been created had it not been for war, and this is what unlocked an exponential number of amazing things we have today". We would certainly have scientific advancement even without war, just exponentially less.
Also, we need to count derivative works of these works as primarily existing because of war reasons too.
This is not an American specific or 20th century specific phenomenon either. Science and war have always been friends, and to my point, with reciprocal benefits, not just war benefiting from science. For example, Fourier was part of napoleons egypt expedition. Euler worked for the Russian Navy, and even has a direct book "Neue Grundsätze der Artillerie" (“New Principles of Artillery”) (1745). Lagrange similar: a lot of his projectile analyses arose out of problems posed by the Turin artillery school.
Most crucially, Euler and Lagrange and many other household names were entirely funded by the military complex. Ecole polytechnique which employed Lagrange was a military engineering school[1], and St. petersburg academy which employed euler[2] was heavily supported by the navy and army.
That said, there are also examples of people creating science for purely fun -- most of gauss' work, galileo's work and a lot of 1300-1600 era indian mathematics arose purely out of astronomy studies, and, I suppose, rolling random crap down a slope for the funsies(galileo) and visions from a goddess (ramanujan). I'm sure there are a gajillion other examples too, of "root node" science being created for non-war reasons. But it's also true that a massively larger number of insanely cool things we have today only ever existed because of war.
[1] and it remains under the French defense ministry [whatever it's called] to this day!
[2] fun story, he was employed by both Frederick the great in berlin and by Catherine I in St. petersburg at different points in his life. He was even accused of espionage.
Multiple edits: looked through my notes and edited some inaccuracies.
To add to your examples, Neil Tyson wrote a book entitled “Accessory to War: The Unspoken Alliance Between Astrophysics and the Military” (https://en.wikipedia.org/wiki/Accessory_to_War)
> The book chronicles war and the use of space as a weapon going as far back as before the Ancient Greeks. [It] includes examples such as Christopher Columbus' use of his knowledge of a lunar eclipse, and the use of satellite intelligence by the United States during the Gulf War.
Much more science than people tend to realize is military-funded.
automating logistics lines does have military potential -- a waymo doesn't have to be holding bob and sara on the way to mcdonalds, it could also be long-hauling thousands of pounds of troop equipment and logistics needs.
the lack of public funding towards automated cars isn't due to a lack of potential, it's due to a lack of focus and lower-hanging-fruit.
After the DARPA Urban Challenge of 2007 I naively thought that commercial self driving urban vehicles were about 5 years away. It actually took until 2020 for Waymo to offer services to the public, and just in one city to start:
That's a long timeline from "tech demo" to usable technology. I don't know how to maintain government funding for that long in a democratic system. No president, senator, or representative goes that long without fighting for re-election. Any technology that still isn't working after 12 years is likely to be considered a dead end and canceled. The big impressive government projects of the 20th century delivered results faster; there were only 7 years between Kennedy's "We choose to go to the moon" speech and NASA actually landing on the moon.
Companies with large resources can behave more like "planned economies" that aren't subject to short term whims of the electorate. Of course they can also exhibit even more short-term orientation -- the notorious "next quarter's earnings report" planning horizon.
* TSMC: Wouldn't we all be better off if the entire world weren't so dependent on a single company, located in a such a geopolitically sensitive territory?
* 10B-param LLMs: Wouldn't it have happened regardless, once everyone realized that increasing the scale of early models like GPT-2 and GPT-3 was key to improving performance? I'd add that the model that launched the deep learning craze (AlexNet) and the model that launched the LLM craze and (the Transformer) were developed by tiny teams on the cheap.
> Waymo: EVs were repeatedly killed by corporations
Not to detract from your point, but Waymo's happen to use electric iPace Jaguar cars in several cities that they serve, but their original self driving taxi service used gasoline minivans at beginning in Phoenix, AZ. EV vs ICE is orthogonal to Waymo's self-driving car technology. Waymo was a pure R&D self driving project for 15 or so years that Google/Alphabet dumped insane amounts of money into before a car ride was ever sold to the public. The are a few competitors to Waymo, at various stages, so market forces likely still would have resulted in self driving car technology eventually arriving, but as its competitors are also well funded, so it seems like it still takes a large org to turn university prototypes into a real live product.
"Wouldn't it have happened regardless, once everyone realized that increasing the scale of early models like GPT-2 and GPT-3 was key to improving performance?"
Notice how it was massive private spending that uncovered the power of scale in the first place. Would've been hard to say, get a federal grant for that. Would've probably happened gradually, with some gains from moderate scale justifying slightly larger grants for successively larger scale.
As for TSMC, the counterfactual assumes such technology would've happened regardless. Just because technology seems to happen inevitably, doesn't make it so. We have evidence of one approach (private) giving incredible results. And also some examples of public (in wartime) giving incredible results. I don't know the evidence for peacetime public incredible results. Maybe warpspeed?
The cost of developing GPT-2 and GPT-3 was on the order of millions of dollars, well within the budget of most tech organizations. See https://arxiv.org/abs/2005.14165 for the total compute invested in them. OpenAI had raised only a few tens of millions in donations at the time, as a non-profit organization.
The increase in performance of computation has been happening for so many decades now that it's been given names like "Moore's Law." People like Hans Moravec predicted way back in the 1980's that the cost of compute would continue to decline and become cheap enough for AGI by the 2020's or 2030's. That's half a century ago!
Fair point, the first few gens weren't that expensive. And like I said I'm certain the scaling would've been discovered soon with some time lag. But the transformer paper was 2017, right? Just from a benefit -to- society perspective, assuming LLMs are a net positive in terms of productivity, perhaps reducing time to drug approval or improving government efficiency (1).. Isn't getting there just one year earlier worth it, if the gains really are that big? We could be talking lives saved via faster approvals or more efficient spending. My point is that a private company made it happen faster, as evidenced by them doing it first. A good thing sooner is valuable.
(1) I'm convinced at the very least LLMs can feasibly speed up paperwork.
I believe that LLMs are detrimental to government efficiency, because they distract from solving the underlying problems (such as porkishly complex laws or a lack of unique identifiers).
The reason the world is so dependent on a single company is because it costs country-breaking amounts of money to keep-up with the semiconductor manufacturing technology. You can only have cheap semiconductors if there are very few entites building them.
Before WWII, middle-class married women were strongly discouraged from working for pay outside the home. If their husbands could provide, "respectable" women were expected to stay home as homemakers.
One could argue the opposite: that the mass entry of women into the paid workforce expanded the labor supply, contributing to wage stagnation and, eventually, the erosion of the middle class. But that wasn’t the only cause. Globalization, declining unions, automation, and regressive taxes were also factors.
> As for the middle class, most of the reason for the decline is people moving into the upper class.
Thats not what I get from the source you provided.
It shows that middle (and lower class) are massively losing income share:
Ine 1971, you have 88% of population in middle class or below, with 72% of total income.
81% of population remain in that bracket, but now they only get 51% of total income. That is massive, and also bad for the economy as a trend because rich people spend less of their income.
The conclusion I draw from this is that middle-class (and below) is in decline because the rich "upper-class" is soaking up much of their income.
Also worth mentioning that in that time period the rest of the world was recovering from devastation. Either the devastation of two world wars or the devastation of imperialism.
Following your argument we should just outright reject progress because, for the most part, humanity has been really really shitty. Also, how much thought did you put into it before writing that this type of society isn't sustainable? Can't the things that happened since (mostly, a massive wealth consolidation) be undone? Why?
>
And equal rights for minorities, sexual or not, were achieved in a handful of countries for the past 40 years.
> Surely you're not suggesting...
Indeed I see the evidence on the side that these ideas were some temporary fads that might get out of fashion in the foreseeable future. This is clearly not a suggestion, I just see the signs on the horizon that this is indeed plausibly to happen.
And it wasn't as good as it's often mythologised to be. Back in the 60s (in Australia) people weren't going on holidays overseas, they lived in houses under half the size of modern builds, they had worse healthcare and lived 13 years less, they had relatively monotonous diets (growing up, my mother's staple food was bread with dripping), they weren't buying new clothes, furniture, knickknacks all the time.
And my grandfather, as a farmer, was up early in the morning and worked all day, never got weekends off. My grandmother was also working all the time - cooking, cleaning, sewing things, gardening. She wasn't employed but that didn't mean she was idle. The kids had to work when they were old enough too.
That was also a pretty decent income for time as well, there were a lot of urban poor living in tiny, crowded little houses.
It's not to say that it's never going to be possible for the mythical postwar boomer lifestyle of leisure (with modern standards of living) to actually available to the bulk of the population but it's going to need a lot more automation and productivity increases (like AI and self-driving cars) to get there, there's no "just tax the billionaires" one-simple-trick or policy that will immediately bring it in.
While rose-tinted glasses are a huge factor, I feel like a big part of what people dream of when looking back on the last century is not just leisure, but stability and dignity.
Stability in that you had jobs that lasted a lifetime and paid a pension once you retired, not layoffs every couple years. Dignity in that anyone could get a real, important, meaningful (and very rough, once you take off said rose-tinted glasses) job as a factory worker, farmer, coal miner, whatever, instead of what, working at Walmart or 7-11?
I do agree with you though, especially your last paragraph.
Coal used to power entire countries, and it still runs our steel mills. American industrial might and American quality of life was only possible because of our coal miners. A Walmart stocker puts cheap Chinese stuff on shelves.
The typical romanticized coal miner is a masculine figure, a breadwinner, the representative of an industry that might have been core to the family and town for generations. A Walmart stocker is a guy in a T-shirt. Walmart itself is famous for... pricing out local businesses whenever it comes to a small town.
I'm not at all denying that it's a culture and glorification thing. I just think this is a factor people sometimes miss when looking at how a lot of the country is nostalgic for the 20th century economy, and why people keep wanting (mostly via vibes) to reindustrialize America.
That's the wrong question. The competition is a family with one earner vs the counterfactual /same/ family with two earners.
The second one will have more money in the modern context so they're better off. At some times in the past, they wouldn't have been better off because their expenses would increase more than their income. Basically it's about cost of childcare.
They say that, but when you point out that they could have that if they accept a lower standard of living they lose interest (and if possiple downvote or otherwise try to shout you down)
They want a trad farming lifestyle without technology but they get mad when you tell them that they have to work 4am-10pm in the summer and one child dies per winter
A secure lifestyle and a good lifestyle are not mutually exclusive. We have the tech to enable something which at least somewhat approximates, and even if we didn’t, it’s easy to imagine a world in which the trillions of dollars spent on wasteful garbage like surveillance, ads, engagement-farming, etc were instead redirected towards research and development of technology which enables a secure _and_ good lifestyle.
What you really need to live, and the luxury you want can be very different. I've lived in a one bathroom house, I'm willing to pay for more. I can eat "beans and rice", but I want more (not just meat, there are vegetables that are more expensive). Most people are not willing to live without a lot of luxury and honestly would choose both parents working a full time job to get more luxury.
This (socalled "luxury space communism") is impossible insofar as a good lifestyle includes positional goods and social status. Demand is infinite, even your own demand, and you have to be able to outrun it.
The best technology I know for this is Ozempic. If there was a way to ban yourself from getting loans that would also help, but you wouldn't like it.
I think the problem is that you are proposing a false dichotomy: that if they do not want one consequence of the current system, they should eschew the entire system.
But in actuality, I like some parts of how society is organized, and dislike some other parts. I don't want to leave society - I want society to be better.
>They want a job market where one single breadwinner can support their house, spouse and kids
If society also wants women to be able to have the same income earning opportunities as men and hence have financial freedom.
Animals compete and compare themselves to others, and so everywhere, dual earning households will outcompete single earning households, and so most market participants will be incentivized to be dual earning households.
No? An easy comparison would be a world where the both partners work 20hrs/wk each, for a total of 40hrs, with the rest devoted towards, eg, childcare.
That addresses the reason for working (eg, pursuit of interests outside family-raising), while also eschewing the need for full time childcare.
You're basically talking about the shift system. A works for 20hrs a week, B works for 20hrs a week. A spends more time with spouse(A), who does the same at their workplace, and B spends more time with spouse(B), who does the same at their workplace. Sounds great.
But, it falls apart to the same logic GP proposed, that the reason you have dual income households is that they are richer than single income ones. Households where people both work 40hrs = 80hrs will be ahead of those that work only 40hrs total. So everyone will descend to working 80hrs too.
Of course, taking mine and GPs logic to it's conclusion is silly - people will have a point where they stop comparing with others and tradeoff less money for less hours. But looking at reality, it seems like that limit is very high! And it only happens at an already very high salary. A 40hrs/week SWE might not go to a high finance 70hrs/week job, because they're already comfortably paid. However these two are top 1% jobs in the world, and the quality of life is probably not too different. But if you go down to the lower rungs, people are more inclined to compare themselves with peers and tradeoff double hours for the next rung, which entails a much better quality of life (as a % increase)
> But looking at reality, it seems like that limit is very high!
Is it? 40hrs is quite low by historical standards. 100 hours per week was the norm in the pre-industrial era, and 60+ hours per week was still typical during the Industrial Revolution.
Labour advocacy groups were promoting 40hrs, much like the four day workweek is today, for a long time, but 40hrs didn't actually became the norm until the Great Depression, where capping hours was a tool used to try and spread the work out amongst more workers to try and resolve the high unemployment problem.
> But if you go down to the lower rungs, people are more inclined to compare themselves with peers and tradeoff double hours for the next rung
While that certainly happens, it seems most people in the lower rungs are quite content to work 40 hours per week, even though working more would put them in a much better position. I dare say you even alluded to that when you chose 40 hours in your example.
It is not like 40hrs is the perfect tradeoff or something. As mentioned before, labour advocacy groups have already decided that 32hrs is even better. I expect many people end up working 40 hours just because "that's what you do" and never give it another thought.
> the reason you have dual income households is that they are richer than single income ones.
If we assume both participants work 40 hours per week then it is true that the same household would have less income if one party stopped accruing an income and all else remained equal. But that doesn't necessarily hold true once you start playing with other variables. A higher income party, for example, may enable the household to have a higher income if they work 60 hours per week while the other party takes care of other life responsibilities to enable those longer hours.
A dual income household isn't necessarily the most fruitful option. In fact, marriage — which, while declining, is still the case in most non-single households — assumes that a single income is the ideal option. It seems that "that's what you do" without any further thought is still the primary driving force.
Lower rungs are definitely not content working 40 hours a week. They work crazy amounts (multiple jobs even!) just to get to the upper rungs of society.
I support the labour laws limiting an employer to 40hrs a week of a man's labour. This is important for people who really just want some employment and don't want to die. But the vast majority of people work two such jobs and try to get into the higher rungs of the financial ladder. Heck, even SDE3s in software companies work off-hours to become IC's and such, and I'm sure it's similar once you go down the executive route.
> "That's what you do"
That is definitely true, a lot of social fabric erodes when providing labour is turned into a psychotic thing. I'm not entirely convinced the labour laws we have today are enough to prevent this. My opinion is that we need to also have policies on the other side of the coin - i.e encourage family/extended family/communal/what have you living. Not "one child policy" level forced policies, but instead in the form of a good complement to strong labour laws.
> They work crazy amounts (multiple jobs even!) just to get to the upper rungs of society.
It does happen, as recognized before, but what suggests this is any kind of norm?
1. The median worker in the USA doesn't even make it to 40 hours of work in a week, only 34. What you say certainly doesn't hold true when dividing the latter in half.
2. Only 21% of the workforce normally puts in more than 40 hours per week. That could represent the lowest rungs, I suppose, but...
3. The data also suggests that those working long hours are more likely to be highly educated, high-wage, salaried, and older men. Does that really fit the profile of someone in the lower runs? Stereotypically, that is who most of us imagine is in the highest rung.
4. The upward mobility of which you speak is not typical. Most people will either stay on the same rung or find themselves heading lower.
First of all, I don't know where this specific example is coming from or how it relates to what I said exactly.
Secondly, when you look at the distribution of wealth in the US, and realize that the top 50% of Americans own 97.5% of the wealth, or that the top 1% owns over 30% of the country's wealth, or read a headline about Elon Musk's $1T pay package, conversations about "dual-earning families" versus "single-earning families" look kind of inconsequential.
The whole thread is about people who want to have a single wage earner lifestyle. That is where this all comes from, and how it relates. You too can live a single wage earner lifestyle in the US, but it will mean significant compromises to your standard of living.
Thanks, I missed the second part of this sentence:
> Quite a lot of people don't want TSMC, Waymo and LLMs. They want a job market where one single breadwinner can support their house, spouse and kids.
I stand by what I wrote above. I agree with you that it is possible today at a reduced QoL and I also would like to see society distribute wealth more equitably, which might also achieve the goal at a higher QoL.
It’s naive either in the way you put it or at the very least in your eyes. There is a lot to be said about the narcissism of innovators radically rethinking anything and everything traditional just because we think we can do better by our current metrics.
If things continue to be advanced haphazardly just because these companies have budget capacity what’s to say that in a hundred years the bulk of humanity will have lost capacity for independent critical thought? Is that really the world you want to create?
It’s not just a “ChatGPT will replace you”. Our humanity is potentially at stake if we don’t deliberately evolve this tech.
Our humanity is at stake no matter how we evolve this tech, because the tech evolves our humanity. It's not a one-way street. Culture, not genetics, is the dominant human evolutionary force today:
No I haven’t and you raise a good point. I mean I don’t know what else to do. The alternatives are to build it and not think about anything beyond the tech itself or to not build it out of an abundance of caution.
And I’m not really as concerned about the super intelligence. I’m more concerned about the impact on our culture as humans.
You ask your questions rhetorically as though the answer would be bad if they didn't. None of those really have had a profound good impact on the population at large and it's not clear that they're all that truly impactful in a good way to the population at large. I think we wouldn't notice or care by and large, and it's just because some nerd somewhere is excited for next product is not a good reason for centraliZation.
Self-driving EVs are not a solution to most of the externalities of personal automobiles. No means of transport is that requires twenty square meters of space and 2000 kg of matter to move 1.4 persons at a time on average.
It’s not even clear as of now whether they have a net positive effect on the society, even if they ever become viable in climates that aren’t always sunny.
> Self-driving EVs are not a solution to most of the externalities of personal automobiles.
Why do you think self-driving cars are meant to solve automobile externalities? I don't view them in that way. I view them as solving personal problems with driving - attention, safety, comfort, utilization rate, cost, etc..
> It’s not even clear as of now whether they have a net positive effect on the society
Why not? Personal automobiles have a utilization rate of around 5%. If Waymo can get that to even 40% that means you can reduce the number of vehicles by 8x (let's be conservative and say 4x to account for potentially higher use of them a la Jevon's Paradox). They're also infinitely safer and it's likely that they'll get to the point where there are no automobile related deaths. They'll replace the likes of Uber and taxis which have been known to be sources of assaults and other crimes. They'll unlock other opportunities too but regardless of how conservative you want to be with their effects, if they roll out to the point where one can forgo buying a personal automobile, it's very very hard to argue it won't be a huge net positive to society.
I bet your reason against it is that you believe something like trains and metros and bikes are better. And to an extent they are, I live in a city and prefer to ride the metro when I can. But it's pure hopium thinking they can ever fully obviate cars.
Reading your first three bullet points I thought you were dreaming about how much better the world could have been.
But your main paragraph following them reads to me like you want Waymo, a powerful TSMC, and huge LLMs.
If there is one thing concentrating power and wealth does it is preferring shorttermism. Growth in the next quarter trumps anything else. Humanity’s ecological niche is suffering long term. Civilization suffers as wealth inequality increases (which concentration of power makes happen).
I would gladly accept a bit of a slow down on progress if it meant my contributions to society were more meaningful. Additionally, I strongly believe this continuous dwindling of small organizations has resulted in an overall loss of community and a sense of belonging. In my opinion, this is what's causing the overall decline in health that we're seeing in developed nations.
For many, life seems aimless. Your future is to simply contribute what you're told to some faceless multinational for which after 20 years your only recognition will be a small piece of canvas with a mass produced screen printed design.
I think the key problem with this idea is taught in basic Microeconomics. A competitive market should have zero profit.
The choice of the government to allow a non competitive market is a choice to transfer consumer surplus to producers, which is effectively a tax-by-regulatory choice. So the counter argument to “what about Bell Labs” would be that the democratically elected government (in theory) can more efficiently gather that tax and pay for research.
Recognizing counter arguments about effective allocation of resources to useful research. But also recognizing that much R&D goes to future profits for the company rather than just societal benefit.
That's sort of the problem, right? Large organizations have become dominant precisely because of reasons like this, and there are indeed huge benefits. But if the hypothesis is correct that the crowding out of smaller organizations is fraying the fabric of society, that's a pretty significant drawback.
All those things could still happen if private enterprises pooled their resources and did R&D together. This is already routinely happening in the car industry, where companies band together to develop new engines, to make the R&D expense hurt less.
You're talking about large-scale hyperexpensive infrastructure research projects with uncertain payouts, which is something that the state already excels at. Public grants are responsible for bankrolling an enormous amount of the research that gets done, and infrastructure projects are money-sinks with a very long timescale. I don't really see how this is an argument functions in favour of big corporations and against big government.
Not a single one of those projects is preferable to a more equitable society. Maybe self driving, yes, but the impact of AI is highly dubious and uncertain at this point in time.
Self Driving cars are very much unproven, and without google people would still be working on them. And so far its not clear that this investment was worth it.
TSMC had many costumers willing to buy new chips.
But I agree, sometimes, larger companies very much do make sense.
I would compare "larger businesses" with "socialistic planning system" while "smaller businesses" with "free market economy". There are examples when centralized planning got good results - NASA's Apollo project is an example. There are also examples when market economy - eventually, in the long term, not short - prevailed: the Cold War is an example here.
It's also quite possible the analogy is flawed though.
It is also very hard to separate “socialist planning system” from “Russian Empire”. It isn’t like things were going swimmingly in Tsarist Russia and then they reentered a golden age in the 90s.
If anything this socialist system brought them to the height of their power and influence.
Much of that was due to foreign aid and lend-lease from the US.
Of course, we also destroyed Russia and made it an oligarchy by giving them bad advice ("shock therapy") after it collapsed. Like basically everything else (including the existence of Facebook) this is Larry Summers' fault.
"Socialist planning system" can also be - approximately, of course - compared to modern, 2020-s Russia, and modern China (PRC, not Taiwan). The renaissance of economy by the end of 1990-s was significantly assisted by moving towards "free market economy".
I would argue that "socialist planning system" had its successes, but in the long run didn't survive the competition, at least in the case of the USSR.
> - Would TSMC have gotten as far if Apple couldn't reliably buy out months of capacity on the next node, bankrolling TSMC's tech development?
TSMC would have had other customers. And if not them, then e.g. Samsung or Intel would have something to offer.
Sure, without this investment the pace of development wouldn't be as fast, but are chips from e.g. a decade ago already utterly useless? Of course not. Life would be largely the same, only somewhat slower.
Perhaps without hardware updates we would finally start thinking long and hard about performance optimisations?
All of the things you mentioned serve the primary purpose of making the rich and powerful more rich and powerful, not improving the lives of the majority.
We could live without self-driving cars, and most of the world still does; faster chips are nice but not revolutionary when we’re using them to waste away watching six hours of ten seconds disinformation videos a day; LLMs are literally telling people to eat glue, convincing people to kill themselves, making cocky ignorant assholes more sure of themselves, and increasing the spread of lies and misinformation.
Humanity would probably benefit from moving slower, not faster.
I very much is not. That logic could be used to justify just about anything, from slavery to littering.
You don’t live in a world populated by just yourself, there are people around you affected by your actions. Be respectful and mindful that what you do affects others. And yourself as well, even if you don’t immediately see it.
- Would TSMC have gotten as far if Apple couldn't reliably buy out months of capacity on the next node, bankrolling TSMC's tech development?
- Would we even have >10B-param LLMs at all?
There is something to be said for the concentration of resources, such that they can be deployed on projects with payoffs years or decades later. The same could be said for all the tech that came out of Bell Labs or PARC. Advocating for smaller businesses is advocating for shorttermism to some degree; even startups today are funded based on the premise that they could potentially capture an entire market in a few years.