This was a well-written and thought provoking article. The USA bashers out there should contemplate a world where the Chinese PLAN has assumed the role of the US Navy. Hint: it wouldn’t be pretty.
Where the essay goes off the tracks is in its conspiratorial view that the US no longer wants to perform the role of securing maritime trade. Instead, the US is intent on expanding its Navy, and the Trump administration clearly recognizes that trade is crucial for future American and worldwide prosperity.
"America first, not only" was Obama's policy [0], not Trump's. The US wouldn't be starting trade wars with its oldest, closest allies if the current executive believed in multipolarity.
One issue with private, for-profit organizations is their major focus is “creating shareholder value”, which is apparently the modern way of saying “making money”. That, along with competition (if applicable, see below), drives them towards minimizing costs, increasing efficiency, and innovating.
Government has the major focus of taking money, and largely spending it arbitrarily. It’s inarguable that the US Federal Government is bloated, corrupt, and vastly inefficient. Despite taking in $trillions in revenue every year, the US has $40 trillion in debt. That’s utterly inexcusable, and heads should roll.
Getting back to the discussion about private enterprise, would you characterize Amazon as bloated, corrupt, and inefficient, despite being an effective monopoly? I would not, though I’ll admit I have no data on the corruption aspect…
Often "minimizing costs" for a for-profit organization means providing worse services to their customers, though.
> would you characterize Amazon as bloated, corrupt, and inefficient, despite being an effective monopoly?
It's funny you should bring this up in the context of healthcare. Amazon acquired One Medical a few years ago. Within no more a year, One Medical got so much worse. Harder to book appointments, wait times increasing, and care team not responding to in-app messages as quickly. (And sometimes being significantly less helpful than they used to be; e.g. "you don't need a recommendation or a referral for a general surgeon since you're on a PPO; just look through UCSF Health's directory and pick one". And this, after I specifically asked for a recommendation, twice. Not to mention most of UCSF's departments will turn you away without a referral, regardless of what kind of insurance you have.)
So I don't know if Amazon is bloated, corrupt, or inefficient, but Amazon acquiring what used to be a genuinely useful medical group made it much, much worse, and I assume that was all in the name of "minimizing costs".
Amazon also owns whole foods and it seems to be doing just fine. Maybe not the exact same as when it launched, but in many places it is one of the best places around to buy groceries.
One Medical, I had before/after Amazon and was never that impressed. They had nice offices and a good booking system, but the GPs were a couple notches above webmd IMO.
> That, along with competition (if applicable, see below), drives them towards minimizing costs, increasing efficiency, and innovating.
Highly debatable. Not making a profit on paper (or moving it to low tax-environments) is a great way to dodge paying taxes. Innovation is a side effect of competition, not of profit-seeking. Plenty of examples of oligopolies and monopolies where innovation ground to a halt in favor of siloization, buying up competitors, dark patterns, regulatory capture, etc.
The right way too look at this is that the organisation of private industry creates a selection pressure on "making money" (not profit!), with the minimum being "able to cover costs" and thus stimulates evolution of strategies to do so. Like in biological evolution, there are many, many pathological situations that can occur.
> Government has the major focus of taking money, and largely spending it arbitrarily.
This is false. The US is a bad example because it has an incredibly corrupt political system, especially since Citizens United. In general in most modern developed democracies spending by the government is supported democratically to a very large extent. That is not to say that 80% of citizens support each bit of spending, but for 95% of the things where money is spent by the government, a majority agrees that money needs to be spent there. The word "arbitrarily" is definitely not warranted.
> would you characterize Amazon as bloated, corrupt, and inefficient, despite being an effective monopoly?
Don't know, yes, and don't know.
For corruption, I think it is necessary to determine how it would even look like for Amazon. Corruption in government is often "favoring entities friendly to the people making the decisions over objectively better options". I'd argue that that is wholly accepted in private enterprises and even encouraged. A "quid pro quo" in private enterprise is just "a business deal" or "customer relations". "Nepotism" is just "helping your friends out". Etc.
I would assume that things like that are very, very common within Amazon.
With regard to bloated/inefficient: This is incredibly hard to make definitive statements on without having worked there and having a lot of intimate knowledge, so I'll refrain from doing so. I think it is, but I don't know.
Also, whether an organization (private or public) is bloated and inefficient is often only a reference to what you think the goal of the organization is.
If the goal of the USPS was to turn a profit, then serving rural addresses is inefficient and the jobs and infrastructure to do so is bloat. Luckily that’s not the goal of the USPS.
I hesitate to mention Mamdani in this forum but it seems like he has done a lot to make NYC’s government “more efficient” in many ways. But also many people disagree on who that government should primarily serve.
I feel that a lot of government spending on roads is inefficient and would be better spent on transit systems. Many people disagree with my values tho.
Stop voting for politicians who keep starting pointless unwinnable wars and you can cut the military budget in half. Boom, 10% of the expenses gone, and $500B less of a budget deficit. Tax some billionaires and it'll be net-positive in no time.
No kidding. Eliminating the federal deficit isn't particularly difficult in mathematical terms; the political will just isn't there, and both Democrats and Republicans can blame the other party when they're in power for out of control spending.
(At least the Democrats seem to be more open to actually taxing to make up for shortfalls, unlike Republicans.)
I think you’re failing to factor in the rapid improvement in robotics.
I was at the hospital a few days ago, and was pondering the numerous nurses who were wheeling around portable nursing stations with monitors and various types of equipment. They’d stop at a room, take a patient’s blood pressure (among various other things), type some notes, and move on. I can easily imagine most nursing tasks being subsumed by robotics.
The same goes for brewers, soldiers, cops, and teachers. Soldier, cop, and teacher robots are already being tested, or actually deployed…and it’s early yet. If a safe, highly dense source of power is developed, that will vastly accelerate the process. At best, there will be far fewer humans involved, monitoring the robots, much as software engineers are now monitoring the agentic AIs.
I’m not sure how human society will adapt. The optimistic view (championed by Elon Musk) is that we’re headed for a “Star Trek economy” of essentially endless abundance for everyone. I hope things play out somewhere along those lines…
I’ve personally been impacted very heavily by LLM technology. Two of my major skills, coding and writing, have been greatly devalued by it. I’m relying on one of the truly core human traits to compensate - adaptability. That’s an area where LLMs currently lag far behind.
> I can easily imagine most nursing tasks being subsumed by robotics
I think you are greatly underestimating the difficulty of automating nursing tasks.
About teachers: some amount teaching profession can be automated, but the parts that cannot (in any reasonable way in the foreseeable future) are perhaps the most core aspects of teaching.
> About teachers: some amount teaching profession can be automated, but the parts that cannot (in any reasonable way in the foreseeable future) are perhaps the most core aspects of teaching.
Yeah, but "we" (note the scare-quotes) can always just decide skimp on those "core aspects of teaching," which may even be a tempting option when "we" decide "we" have less need for educated workers.
That’s not how technology works. We went from first flight in 1903 to a man on the moon in 1969. Then we basically did almost nothing in aerospace for the next 50 years. We never got past the supersonic barrier for feasible commercial air travel. Apart from somewhat worse fuel efficiency, a 747 from 1969 can achieve basically the same trips in the same amount of time as a 787 or A350.
What looks like exponential curves turn into S-shaped curves everywhere. In the last century, coal power plants have gotten less than twice as efficient. A professor at law school was giving a lecture on technological singularity back in 2011. And I pointed out that Moore’s law was already dead. And in fact in the decade since then improvements came far slower than they did from 1990-2000.
It allows people with half the available money to do what others could do before. That's more than twice as many people.
It allows you to move freight really fast at half the cost. That's a lot of things that can move around the world 100X faster than they could on the next-fastest available mode: ships and trains.
It allows you to move freight really fast at half the cost
That's leaving out fuel prices. The planes may use half the fuel, but aviation fuel has increased 7x in price since 1990. I couldn't find prices for 1969, but if you take the price of crude since then you're looking at a CPI-adjusted increase close to 100x.
Great point. Even more value to halving the fuel cost, then, but fewer products able to take advantage of it.
Of course, inflation should be canceled out before evaluating the overall benefit. If you can self the goods for 7X then the fuel savings is back to 2X.
When things are spread out over time people are able to adjust to the point that society, as a whole, might not even notice there was a problem, even if for some segment of that society it might feel like the end of the world - or at least an incomparable inflection point. The only real problem you have is when things come on hard and fast, like a tsunami.
We shall see about how much of nursing can be automated. I do think that the human presence of a nurse (particularly if of the opposite sex to the patient) is helpful. That could still be provided on a less frequent basis.
As to the rest of their function, I think probably the biggest obstacle will be the ability to safely manipulate the human body, as in turning or moving the patient. That requires delicacy, advanced tactile sensors, and physical power. For fine detail work, such as finding a vein with a needle, the technology already exists. Elon Musk also predicts that robotic surgery will rapidly outpace human surgeons, and I agree there.
On the subject of human teachers, my comment is similar to that about human nurses. A human connection helps with motivation. That said, I think video serves much the same function, and with video one teacher can reach an unlimited number of classrooms. That’s independent of how humanistic the teaching robots will become, and I expect that will also improve rapidly. We are nearing singularity like situations with engineering problems, in other words recursive improvement without much if any human intervention.
It’s gonna be a long time before we have robots that can replace 100% of a nurse at a cost lower than a nurse. Right now we have nothing close to someone being able to replace those tasks. The industries that have had most robotics and automation such as warehousing still have human workers. Heck I would not be suprised if more people work in warehouses in USA today than 25 years ago.
It is sufficient to make existing workers more effective to be able to fire some of them. If turning a patient to change a diaper is a two person job and somebody buys a robot that assists to make it a one person job, suddenly you need fewer people during a shift.
There are a lot fewer workers per warehouse today than 25 years ago. There is fortunately much less flexible demand for changing nappies in elderly care homes.
They’re a threat to those fearing deportation or arrest for criminal activity. If you’re not in those categories, you have exactly nothing to worry about.
I’m a lot more worried about the giant NSA data centers like the one in Utah, funded under prior administrations. Those are definitely warehousing wide-ranging data about all Americans (and others), harvested right off the Internet backbone…
Oh yes, we definitely don't have news or straight up video of people who are regular citizens getting killed for no reason by ICE or even non-citizens getting effectively executed without trial.
> we definitely don't have news or straight up video of people who are regular citizens getting killed for no reason by ICE or even non-citizens getting effectively executed without trial.
You may have made-up “news” from the usual suspects, but you don’t have video. The videos of Good, Pretti, and the alien recently shot for using his vehicle as a weapon sure don’t show innocent victims dying at the hands of ICE. Any other law enforcement agency involved in the same incidents would have acted much the same.
In every case, the deceased put themselves in situations that lead to their deaths.
In two cases (Good and the alien), they attempted to use their vehicles as weapons.
In Pretti’s case he brought a pistol to a protest (illegal in his location), inserted himself into a detainment attempt, and struggled with officers. He may not have fired the shot that startled the officers surrounding him, causing them to fire (they all knew he was carrying), but he would sure be alive today if he hadn’t done the other things.
They say stupidity should be painful. In these cases, it was fatal.
I'm going to start with a link to the Pretti video analysis. If you don't trust the Times, feel free to ignore their assessments of motives and you can just look at the analysis of what events happened and when which are factual details: https://www.nytimes.com/video/us/100000010668660/new-video-a...
> In Pretti’s case he brought a pistol to a protest (illegal in his location)
I think you should know that the claim that it was illegal for him to carry that pistol is false and is a lie being told to you to deceive you into letting them sweep his death under the rug. Minnesota's concealed carry law (Pretti had a permit) does not restrict carrying in public or at protests. You can easily look that up if you care to. They don't think you will.
> inserted himself into a detainment attempt
This is also clearly false from all of the videos. There was no detainment attempt being made at that time. What is shown in all the videos is that an agent shoved a woman backward hard (to the ground) trying to get her to leave, not that they were trying to detain her. The woman was not part of any detainment attempt at that time. At that point she was merely a common victim of assault.
> and struggled with officers
Is that before or after he was being beaten while his hands were pinned down?
> He may not have fired the shot that startled the officers surrounding him, causing them to fire
The shot you're talking about was fired by one of those same border agents. It wasn't a magic bullet that came out of nowhere.
> They say stupidity should be painful
It is indeed true that some people enjoy pretending that cruelty is wisdom, but it isn't a good way to live.
Well, my informed (I guess? it's first hand) opinion says exactly what the PC said. And no the plan has been underfoot for so long that it really pretty much has nothing to do with the current regime even though I am sure like any regime they'd say they did it from the scratch.
I'd say we are getting really great at getting broadband to everyone than giving enough bread and education and healthcare to everyone :D (ignore the smiley, this sucks)
It allows technical folk to live wherever they’d like as long as they’re working remotely.
The mobile applications, particularly in the case of airline aircraft, have also been compelling and worth a lot of money to SpaceX.
Starlink has also brought broadband Internet to a vast number of people that would not have had it otherwise. This will boost the worldwide economy by an enormous amount.
I don't think rurp was saying there is no market, just that there was no obvious realistic TAM worth 1.6 trillion (going by the amount given by the S-1). How many people living remotely in an area with no fibre do you really expect there to be?
Starlink only has 10 Million customers, too expensive for most countries already.
Starlink brought internet to a lot of people who had it before already but made it easier for them.
Its still quite a interesting technology, given, but for the fact that he destroys potentially our atmosphere, has control over war critical tech, can do survailance and wants to send out A Lot MORE into our space, its a net negative for at least 7-8 BILLION people while 10 Millionen people benefit from it.
And they even increased the price just a few weeks back...
We should all keep in mind that foreign interests, primarily China, are astroturfing to drive resentment in the USA against AI and datacenters. National governments are largely viewing AI as a strategic capability.
That’s not to say there aren’t downsides, but they vary quite a bit depending on location and grid capacity.
I actually expect there will be major progress made on the energy efficiency and performance of LLM models using novel hardware. So, the buildout may overshoot by quite a bit… If so I hope we can find a use for all those racks…
We should also keep in mind that the problem isn't datacenters, but how they are built.
Datacenters do not have to be noisy. Datacenters do not have to cheap out on cooling solutions. Datacenters do not need to be powered by mobile gas turbines left on trailers to pretend they're not permanently installed to avoid having to get permits.
Those corners being cut is not what make AI datacenters possible or competitive. That race is purely chip supply.
For reference, I work in an industrial neighborhood where there are quite a few new datacenters from big providers. The buildings are ugly for sure, but unless you're staring at it you'd have no idea it was there. I could try to pay more attention to see if I can hear it if I focus on it, but I suspect the sound of nearby rustling leaves will be too deafening to make out anything.
The noise is often from backup power solutions like diesel generators that need to be turned on regularly. Maybe that’s what the turbines are? I can’t imagine a data center getting provisioned with a turbine as its primary power. But yes a data center can be silent most of the month.
Very open to being corrected on any of this, it’s just my own understanding but I’m not in the industry.
To my understanding (which can be wrong) these gas turbines are not backup generators, but used as primary or supplementary power sources during normal operation where the grid cannot supply (or when grid cost is unfavorable). In one clip (which is a bit sensationalist so I'll refrain from posting it as it may derail) shows a park of some ~24 large gas turbines - each of which sized like a small building - with about half in operation at that time.
Indeed, there is some noise from backup generators when tested (or in use), but with a bit of preparation, good palcement and sequential testing you can make that not too disturbing.
> Datacenters do not need to be powered by mobile gas turbines left on trailers
The others I agree with, but Im not sure this is true. The US government has proven itself completely incapable of expanding electricity production and grid infrastructure. How exactly are you supposed to power your datacenter when you cant access any electricity?
1. Build your datacenter near supply. If there were consequences to ignoring the rules or being a bad neighbor with municipality, state or federal government baring their teeth, more optimal locations with regards to supply and noise would also end up being the cheapest and safest location for them.
2. Build supply near your datacenter. Solar and wind are both very cheap right now, but requires buying more land, same for a proper gas power plant in its own building with appropriate noise and pollution treatment.
3. Make investment into infrastructure a prerequisite for the project instead of just complaining about it and making into an excuse for cutting corners.
Even if you need to have local supply, "mobile gas turbines left on trailers to pretend they're not permanently installed to avoid having to get permits" is never a necessity.
The US govt does not build out infrastructure. They at best subsidize it over long periods of time.
To answer your question, you build your datacenters where there is capacity or you plan with the utility in the area you want to build. This process takes years.
They are not doing that. They are building them as fast and as cheaply as possible.
As a result they are cutting corners and are leading to all of the complaints.
They are as a result stealing from the future to make a profit today.
The grid does not have gigawatts of extra base load capacity available as that has always been looked upon as wasteful.
We should all keep in mind that foreign interests, primarily China, are astroturfing to drive resentment in the USA against AI and datacenters.
AI companies are doing a fine job of driving the resentment all by themselves, they don't need China's help. Whether you're right or wrong is irrelevant, these companies are doing nothing, absolutely nothing, to help the public's view of them. They seem to think that they don't need public support.
Oh, they're trying alright. Google has been spamming Twitch with ads of teens and 20-somethings talking about how Gemini helps them study or "research" their paper, and the casting call clearly was for urban progressive LGBTQ types
> We should all keep in mind that foreign interests, primarily China, are astroturfing to drive resentment in the USA against AI and datacenters
There hasn't been any credible reporting on this so far. It's far more likely people are just mad that they have to pay for the AI boom both literally via increased electricity rates and via increased noise, water shortage and construction related pollution caused by the data centers.
If it were true that data centers are building "American dynamism" and China wants to suppress that, surely they would be building data centers the way they build trains.
There really isn’t a need for foreign interests to artificially boost the negative sentiments towards data centers, hyperscalers are doing a pretty terrible job at getting the public onboard
Why aren't we (in the US) building all these data centers in the Southwest where huge amounts of vacant land and the best wind and solar power options are?
With fiber connectivity it doesn't matter if they are in remote locations.
Probably the lack of water. I don't know why they need to continually pump it instead of doing a closed loop and cooling it through a few miles of pipe underground which is practically free, after you install the plumbing. Sure evaporative cooling is good, but certainly there are workable alternatives.
That or latent heat cooling, the other end of the phase-state spectrum: Instead of evaporating water during the day, make ice all night when ambient air temps drop and then use it to cool a closed-loop system during the day.
They are winning recall votes and getting a lot of pushback at in person city council meetings. People keep saying it's astroturfed (and I do expect a lot of communication online to not be in good faith) but you can also see the real grassroots pushback. I mean the article is talking about recalling mayors, you don't get that without pretty extreme local opposition.
Bullshit. Nobody needs to "drive" resentment. "AI" is touching everyone's lives and in not in a good way. And nobody gives a shit what privileged, wealthy, financially-benefiting political leaders want.
AI was used by the Israelis to decide what homes in Palestine to bomb.
AI was used by the Trump admin to decide what buildings to target, including a school of ~200 kids.
xAI was deploying numerous gas-turbine trailers at their datacenters in urban areas where they couldn't get enough power. Turbines that pump out huge amounts of NOx and noise...not to mention these DCs are generating so much heat they're raising the temperature of the areas they're in.
In every social group I'm in both IRL and online people hate AI shit.
They hate the monumental waste of energy, water, and land.
They hate the noise and pollution created by gas turbine generators being deployed at DCs because the DC can't get enough power
They hate that DCs consume water treated for DRINKING and pay a fraction of what individual ratepayers pay
They hate that politicians are handing out freebies to these DCs, whether it is tax breaks of infrastructure or discounted utilities or all of the above.
They hate AI phone system 'agents' that routinely get things wrong.
They hate AI "news stories."
They hate AI "agents" they have to deal with to get support for things which somehow manage to be even worse than someone who barely speaks English and has a decision tree in front of them, despite the company clearly being able to train it on much more data and refine it.
They hate it when media outlets use AI for "artwork" and graphics that look idiotic / have basic mistakes and clearly took work away from a photographer, editor, or graphic designer. Or all three.
They hate it when businesses, acquaintances, coworkers, friends reply to them with obviously-written-by-AI responses.
There are endless stories here about how AI is swamping open source projects with garbage bug reports and push requests.
In gaming circles they hate how prices on SSDs and RAM have tripled or more.
The programmers I know see endless slop and people in their field being fired because management thinks AI can do their jobs.
The list goes on. The only people who "like" AI are AI DudeBros and investors.
Since you have such strong (and numerous) opinions…
AI is a tool, and one that governments of all stripes will use for many things, including military applications. It’s overly simplistic to say “AI was used to bomb things”. There’s always a human making the decision to hit a target or not in Israel and America. Of course, that won’t necessarily be the case in totalitarian countries…
As to hitting the school, bear in mind that it was literally part of a military base.
On energy use, there’s actually no reason to limit it unless worldwide usage grows by orders of magnitude. What does need to happen is a transition to clean sources. Natural gas isn’t ideal, but still much better than coal. However, there’s a lot of interest in modular nuclear reactors for this application, which would be fine.
As to how well current AI works, the canard “Past performance does not guarantee future results” applies. We’re just at the infancy of the tech.
Your anecdotal screeds of hate towards AI should be balanced with the numerous stories here on HN and elsewhere of people using it very successfully and happily. As to the general populace, the daily usage figures for ChatGPT and other chatbots belie the hate you’re claiming to be universal.
Finally, yes the price increases on RAM and SSDs aren’t great. This too shall pass as more capacity comes online due to the wonders of capitalism. Don’t panic! ;-)
A much better, safer, and cheaper idea is dropping suitably packaged waste on the subducting side of the Mid Pacific Subduction Zone. It’d be inert long before seeing daylight again…
The only downside is the “waste” is quite valuable. MSRs can also use it, and their waste is only dangerous for ~300 years.
That's a bad idea. Subduction zones move very slowly, and material would come back up in "mud volcanoes", especially with the added heat input of the waste.
Nonsense. The sea floor in those areas is soft muck, 20,000 ft or more deep.
The waste, likely packaged in torpedo shaped structures made of corrosion proof, high density, high strength materials, would arrive at somewhere around 150 MPH. Being extremely dense and heavy, they would penetrate deeply into the sea bottom, estimated at 35-55 meters deep. There’s no chance they’d somehow be forced back to the surface. Even if they were, being sealed they would be harmless.
Even if by some miracle one broke open at the sea floor, the nuclear material is dense and wouldn’t disperse. There’s very little life to affect down there regardless.
My impression was that Gossamer has good Rust FFI rather than basic C FFI, sharing struct layout etc. I may play around with Gossamer and see how it really does performance wise…
The syntax similarities help make for a more painless two language approach.
Where the essay goes off the tracks is in its conspiratorial view that the US no longer wants to perform the role of securing maritime trade. Instead, the US is intent on expanding its Navy, and the Trump administration clearly recognizes that trade is crucial for future American and worldwide prosperity.
It’s America First, not America Only. ;-)
reply