Claude creators say AI “creatures” can be “machines of loving grace” that are “better than humans at essentially everything” — except low-wage service jobs
There are today over 7 million unfilled jobs in the US. The Trump Administration is working on reindustrialization. Much of the jobs that AI will absorb are BS "fake" laptop jobs. What we need is a rewiring of the education system to send more people to industrial work... jobs that use both hands and mind, and away from the crap cubicle farm roles that are really not useful.
AI cannot yet pick up tools to build things. Yes, a manufacturing operation today and tomorrow is highly robotic... instead of thousands of line workers doing assembly, the robots do the assembly with only hundreds or dozens of workers managing the robots. However, those are still jobs that need to be filled.
And when an industrial operation exists there are many more private-sector support service jobs that derive from it. AI will reduce human labor requirements for some, but not all.
I think we are missing the main plot here. AI will eliminate bullshit jobs and open up the labor pool to support a renewed industrial revolution. US infrastructure is still a mess. We need our roads and bridges fixed. We need more broadband. We need to bury electric transmissions lines. We need more nuclear power. We need more fossil fuel production. We need more trains. All of these things will require real human labor. The lazy ass young people with their expensive and useless gender studies degrees are going to have to rewire their career expectations to do these jobs... or they will be unemployed.
Perhaps one of the most important pieces Michael or anyone else has spoken about AI. It’s not human and speak of it as though it is as Michael and others say a dehumanizing conceit.
According to Claude, here are the top professions in the US as per the Dept of Labor.
Based on available data from the Bureau of Labor Statistics, here are the most common jobs in the U.S. by employment share. Note that exact percentages shift slightly year to year, but these are broadly representative:
1. Retail salespersons - ~2.8%
2. Fast food and counter workers - ~2.5%
3. Home health and personal care aides - ~2.4%
4. General and operations managers - ~1.8%
5. Registered nurses - ~1.8%
6. Office clerks (general) - ~1.7%
7. Customer service representatives - ~1.7%
8. Laborers and freight/stock movers - ~1.6%
9. Waiters and waitresses - ~1.5%
10. Software developers - ~1.4%
About 20% of the current workforce. It seems to me that AI can both enhance and cannibalize jobs. But one can envision business models where for example, individuals can own a Waymo or more than one. One can envision more ownership across many different professions. But to leverage this potential promise, one needs to think and care about people. What’s scary about Anthropic and their people is an overwhelming hubris, and chutzpah. Not a good combination.
“ instead of thousands of line workers doing assembly, the robots do the assembly with only hundreds or dozens of workers managing the robots.” this is massive job loss? You just proved the point. Also what manufacturing jobs is Trump bringing back? The numbers I have seen show a significant drop in manufacturing jobs since Trump came back in office(Asking out of genuine curiosity)
Hyundai, Toyotoa, Harley Davidson, Apple are bringing production back to USA - the White House Investments page https://www.whitehouse.gov/investments/ lists 10.6 trillion in new investments since President Trump 2.0. AI infrastructure, Oil and Gas increased production, Increased affordable energy production, Clean Coal companies being reinvigorated. Nuclear reactor design companies, fusion power. Nividia, Aerospace, Apple, Meta, Amazon, Google, Space Travel, Pharmaceuticals. I'm also sharing a short YouTube report by Promethean Action, at about the 11.04 she quotes Mike Rowe as reporting some Data Center Electricians are making 250K. Also says Costo is a good paying employer, one cashier profiled was making $33 an hour with medical benefits and 401k. Things are getting going it seems.
Fisrt you have to find a site, then get plans, environmental studies, fight off phony environmental challenges, then approvals, then build then start industrial production.
He has gotten alot of money committed and moving through the pipeline.
He can't override all those steps, he's not a king. Unless, now you want a king? Have to make up your mind, are things to move in the laid out slow inefficient administrative way or have a king?
The technology to replace human assembly jobs has been around for a long time but the understanding of the short term instant gratification of that compared to future losses and damage should prevent it from happening. AI will be beyond human advancement enough to be "alive" when there are no programmers or developers required and it does those jobs. Any superhuman entity that should frighten you would be able to write itself. None of this makes sense.
You need to look at leading indicators as it is quite ignorant to expect big increases in manufacturing and industrial jobs quickly after decades of it being decimated by globalism. Capital investment has skyrocketed. Yes, some of that is AI datacenters, but much of it is corporate investment and commitments to reshoring American industry and manufacturing. One strong indicator is the demand and supply of flatbed truck hauling service. The costs of skyrocketed as the center of the country is on fire with equipment, raw materials, etc., deliveries. The economy is being retooled away from the coastal ports being the hub of product activity to the center of the country. The coastal areas are going to have to retool to support the center instead of China.
Do you know about the Cantillon Playbook... basically government injected financial repression to fix the problem of runaway federal debt to GDP which is critical when it reaches 120%? We are at 120%.
The play will hammer consumers because consumers are over-consuming. It will replace it with more production because the people are under-producing. It is a 10-year play that the US has done before after WWII. China has done it. It is needed now because Democrats screwed the pooch with massive post pandemic spending. Democrats started the playbook without knowing it. Although they jacked debt, they also jacked inflation thus causing the financial repression of consumers. The next step and we will see it now that Powell is out and Warsh is in... lower rates that cause a run on asset purchase, money flow and capx investment. The top will grow more wealth while the bottom finds that they need to work harder to make it. And that is good because at the same time they will see more industrial and manufacturing jobs, and all the jobs that support these industries. Some of them instead of relying on their worthless gender studies degree landing a fake laptop job working for government or a non-profit are going to have to learn how to use a wrench and hammer.
Whatever...Have toured through major manufacturing plants installing AI Systems and Robotics over the last two decades. IT IS NOT MORE CORRECT and/or ECONOMICAL THAN HUMANS even though the workers providing the tour or continuing on the lines, including programers and other engineers, in all aspects, are advised to support the propagandized narratives to view them positively.
THIS IS PSY-OPS SERVING MANY NEGATIVE FUNCTIONS WITH THE ONLY POSITIVE BENEFITING THE LUCIFERIANS AND PAGAN DEMON WORSHIPING PEDOPHILES, CANNIBALS AND VAMPIRES BEHIND THIS WHOLE 'Depopulation-Global Slavery Agenda' also oriented with AI.
No it's not. I am just watching where CAPEX is going, and it's not in the MFG sector.
Big Tech / AI infrastructure
The four major hyperscalers are driving an unprecedented capex surge: Amazon, Alphabet (Google), Meta, and Microsoft plan to spend roughly $725 billion on capital expenditures in 2026, up about 77% from ~$410 billion in 2025.
Microsoft set its 2026 capex at $190 billion, well above the $152 billion average analyst estimate, with its CFO noting a chunk of that increase reflects rising memory chip and component costs Tom's Hardware
Amazon is projecting around $200 billion Valueaddvc
Google's cloud revenue jumped 63% year-over-year to $20 billion in Q1, supporting a $175–185B capex plan Tom's Hardware
Meta is guiding $115–135 billion
Including Oracle (targeting $50 billion), the five largest US cloud/AI infrastructure providers have collectively committed $660–690 billion for 2026. Some analysts peg baseline aggregate capex at $7.6 trillion cumulatively between 2026 and 2031 across compute, data centers, and power. Nearly all of this spending goes toward GPU clusters, custom silicon, and data-center construction.
Utilities/Power
This sector is being pulled along by the AI buildout's power demands. US investor-owned utilities are planning significant capital expenditures over the next five years, focused on infrastructure modernization, reliability, and new generation capacity to meet heightened energy demand from data centers, with the US utility capex forecast near $1.3 trillion for 2026–2030. Sempra alone forecasts $44.85 billion in 2026–2028 capex, driven largely by data-center and oil/gas-sector power demand in Texas.
Manufacturing has historically been tracked separately by the Census Bureau's Annual Capital Expenditures Survey, though the latest comprehensive breakdown by industry is from 2022, when US nonfarm companies with employees invested $1,681.7 billion in new and used structures and equipment.
AI data centers are only one lane of tech industrial growth. Capital investment for that does not compete with capital for other industry and manufacturing. You also seem to be ignoring foreign capital investment.
But I do read that big tech’s shift from using their balance sheet liquidity to fund data center development to debt issuance is causing some money supply concerns, but on the debt-issuance side. And there is push back from the lenders getting nervous about this risk concentration.
Again though, the look should be at leading indicators and doing so provides unambiguous proof that US industry and manufacturing is being rebuilt at a record pace.
No I don't. The West when it comes to "reindustrialisation" is screwed, mainly because of capital misallocation and stances like the one the Trump administration.
What are you talking about? You seem to be completely ignoring the words and actions of the administration. Have you listened to Bessent, Ludnick and Warsh?
I'm a data scientist who uses tge very useful AI tools every day (e.g. Claude Code, Codex, etc.).
Most of the fear mongering about job loss in the short run is just those AI companies trying to fund raise and perform regulatory capture. Their only moat / barrier against open source models is a restrictive regulatory regime.
Open source = open competition. My husband (before his illness) could learn and develop a project with any paradigm of code within three months. I watched it for many years. Open source also reveals the actual advancement levels. These aren't people looking for advancements or they would want all capable minds. I feel we are being punked.
My observations indicate that money is at the root of closed source operations, protection of the value of the work done and investments made by the company or the individual.
I spent 12 hours taking care of patients last night. Patients need human reassurance, human touch, human presence. It keeps us sane during our most challenging moments. Humans have souls. I think dogs have souls. AI is just tech. It’s an aggregate of all our downloaded information. And it’s only as good or unbiased as the info loaded into it. It’s an it. Anthropomorphizing it is gross.
...”AI is not human, and comparing its behaviors to humans is as misleading as saying that trees are like children because they both grow up. The models do not have a “character” or self. ..”
Michael, this is a beautifully reasoned and nuanced critique of the underlying thinking and philosophies guiding Anthropic and its founders/executive leadership.
Dehumanization is the unfortunate outcome when AI machines and their capabilities are increasingly compared to human beings in order to maximize the economic return on investment in this technology. Existential philosophical questions and examination must be posed and answered in a much deeper and more comprehensive way than what Amodei and his executives are doing based on your description. And at some point soon - there must be some level of consensus on policies that we as a society of conscious humans arrive at through examination, discussion and debate. You are actively encouraging that direction with a critical essay like this! I say this as someone who has embraced AI in order to enhance my personal productivity and help further inform my decision making process.
It's right in their name: Anthropic. But it's a weird POV. And they are saying what they think they need to say in order to get regulation that will shut out potential competitors, or the use of open source models. It's ironic that the Chinese AI companies are far more open than the Big Two American companies - the Chinese don't seem to buy the hype about artificial humans.
They are preparing people to believe their crazy programming as advanced and unquestionable truth. If they program in a destructive and false agenda humans are not better off. This is the true danger of people like this. With these mystique notions of infallibility and promises to have machine slaves taking away human jobs, not cure disease and further humanity but cater to greed it becomes very profitable for those who would invest in their own greed to give them money. The regulation needed might prevent them from these antics. It's sounds familiar when they tell people to resign themselves to the service industry, just like when they sent our manufacturing to slave wage countries. These are the same people coming at us from a different angle. Beware, of the Emperor with no clothes.
These people are completely full of themselves. We have learned how dangerous that can be from the Climate and Davos worshipping crowds. What is being described when they are talking about "guessing" and being "corrected" is exactly the technologies we have had for years. It appears their technologies are not really very advanced as they are saying (alive) and it can only put out that which has been programmed in. The danger we will have with AI is that crazy people program it and the world could believe their crazy programming. It really does remind me of the Climate and DAVOS creepy mental midgets who want control over humanity and rule Earth. In a vast universe with infinite possibilities for exploration, expansion and growth for humanity those who would be content to rule Earth are akin to people who peaked in high school. Their power is in the captive belief of others. A constructive purpose for any super humanlike entity wouldn't be to take the place or jobs of humans (they shouldn't even be allowed or aspire to do this as that is only destructive.) A constructive purpose of a true superhuman entity would be to advance and perform duties humans can not and further promote humanity. (space exploration, curing disease, complex computations...)These people are frauds and we seem to be on the edge of another "Emperor has no clothes." moment reminiscent of the Biden era. I think I'll wait around to see what Elon puts out. These people are barely human like themselves.
THANK YOU! Was hearing Taibbi describe all the Twitter file work. What a gift THAT was. This article focuses on the contrast of machines VS humans. Fantastic!! And bravo. *see below
My son has been building a home improvement business from scratch last 8 or so years. These MANUAL jobs can never be replaced by robots. It would take a supervisory person for almost every job!! Even every SAME type of job has constant variable through every stage of work. The guys are constantly learning. This was also the problem of sending menial assembly factories to China. You end up needed the designers on site to trouble shoot the needs of improvement. This is what creative humans with a soul are created to DO.
I'm a singer and a voice teacher. I could go on forever about the difference between machine voices/MUSIC and the human voice/human produced music Machines are useful, but humans are irreplaceable for WAY MORE than many are projecting and supposing.
*("Klavan noted that humanizing AI goes hand in hand with dehumanizing people. “People start to conceive of machines as human when they start to conceive of humans as machines,” he noted. “Treat things like people, and you run the risk of treating people like things. This insight, too, is at the core of our best ethical traditions in the West.”")
This quote: "“called this moment in history the Great Turning, the shift from a society built on extraction to one built to sustain life.”
The word EXTRACTION reminds me of what the PrometheanAction enterprise is teaching about current AND historical geo-political economy contrast; 2 opposing worldview bases.
Extraction VS made in the image of God humans
Empire/Globalist SYSTEM VS the AMERICAN system of economy.
Nihilism VS Civilization
One sees people as beasts to be managed by a few elites -or 'all knowing machines'; the other is our inheritance through Judeo Christian civilization freeing creative humans - Imago Dei.
Which system continued the slave trade into our own times? Which system ENDED it in favor of sovereign nations to SUPPORT the creative potential of ALL humans. "We hold these truths....."
We have been in a murky fight to retain our Independence-foundational philosophy almost since 1776, but the disguise (through secularism, etc perpetrated BY Empire ELITES) has been VERY effective at HIDING the realities of the contrast for many generations.
Modernity has SOLD monetary VALUE, while our Founders put ALL THAT on the line for TRUE VALUES! Godly values.
Anthropic may already be "too big to fail" but it will only get more so. After going public, the stock will become a staple of pension and retirement portfolios across the country. If the government gets a stake, it will have a vested interest in promoting Anthropic's "value creation" vision.
If 'executives' motivated by a love of humanity can shape AI's priorities, what happens when those executives are replaced by executives intent on reducing their fellow humans to subservient slaves? If AI's sense of morality is programmable, what kind of AI can we expect from programmers with the morality of a Hitler, or a Revolutionary Guard? What is the guarantee that such programmers will never gain executive authority?
As you say, AI is only being developed because of its ability to save users money/time/effort. But this is fundamentally similar to every other technology that’s ever been worth developing. And users, i.e. beneficiaries, aren’t limited to big corporations.
I’m interested in other commenters’ claims that these AI companies are actually trying play up the dangers of AI in the guise of downplaying them, in order to pump up their value for investors who know exactly what makes AI valuable, to the extent that it is, and to set a trap for regulatory capture. This makes an awful lot of sense.
What struck me about the video was Lubinski’s level of emotionality when she talked about Claude etc. being conscious, having emotions, and other human qualities. Why is it so moving to some people to imagine that machines are living beings? Much of the tech sector consists of real humans who can’t relate well to other humans. Maybe that is why they hope for a second chance with machines. Besides, most of the same people are leftists, and it’s about time for them to designate a new class of victims to ally themselves with. AI looks like it’s up for nomination.
Anyone else notice their purely quantitative search engine queries returning counterfactual objective results, increasingly frequent over past few years? 5 or 6 yesrs back I noticed i couldnt replicate interesting search results. And this proved to me the algorithm was untrustworty Q: what is 2018 audi S5 oil capacity? A: 5.5 quarts. But I knew from personal experience it was around 7.5q, so I twerked same query - A: 7.6 quarts. So Michael, all the Existential Epitomological Import AI you're wanting to discuss? consider this: just now re-ran the same query years later as an experiment, and search engine still returns #1 Top Ranked result, from engineoilguide.org --> 5.5Q, while simultaneously the search assist AI precis scraped web and answered 7.6Q, correctly. Begs a new question: search engine site 'AI' 'learned', returning value correctly corresponding to Audi design specification, but still serves up same bad info as its #1 result, just as it did yesrs ago.
The inverse should be true ,ai should do service jobs, people at the top as overLords of ai. Humans need to control ai not visa versa. This is sickening.
There are today over 7 million unfilled jobs in the US. The Trump Administration is working on reindustrialization. Much of the jobs that AI will absorb are BS "fake" laptop jobs. What we need is a rewiring of the education system to send more people to industrial work... jobs that use both hands and mind, and away from the crap cubicle farm roles that are really not useful.
AI cannot yet pick up tools to build things. Yes, a manufacturing operation today and tomorrow is highly robotic... instead of thousands of line workers doing assembly, the robots do the assembly with only hundreds or dozens of workers managing the robots. However, those are still jobs that need to be filled.
And when an industrial operation exists there are many more private-sector support service jobs that derive from it. AI will reduce human labor requirements for some, but not all.
I think we are missing the main plot here. AI will eliminate bullshit jobs and open up the labor pool to support a renewed industrial revolution. US infrastructure is still a mess. We need our roads and bridges fixed. We need more broadband. We need to bury electric transmissions lines. We need more nuclear power. We need more fossil fuel production. We need more trains. All of these things will require real human labor. The lazy ass young people with their expensive and useless gender studies degrees are going to have to rewire their career expectations to do these jobs... or they will be unemployed.
Perhaps one of the most important pieces Michael or anyone else has spoken about AI. It’s not human and speak of it as though it is as Michael and others say a dehumanizing conceit.
According to Claude, here are the top professions in the US as per the Dept of Labor.
Based on available data from the Bureau of Labor Statistics, here are the most common jobs in the U.S. by employment share. Note that exact percentages shift slightly year to year, but these are broadly representative:
1. Retail salespersons - ~2.8%
2. Fast food and counter workers - ~2.5%
3. Home health and personal care aides - ~2.4%
4. General and operations managers - ~1.8%
5. Registered nurses - ~1.8%
6. Office clerks (general) - ~1.7%
7. Customer service representatives - ~1.7%
8. Laborers and freight/stock movers - ~1.6%
9. Waiters and waitresses - ~1.5%
10. Software developers - ~1.4%
About 20% of the current workforce. It seems to me that AI can both enhance and cannibalize jobs. But one can envision business models where for example, individuals can own a Waymo or more than one. One can envision more ownership across many different professions. But to leverage this potential promise, one needs to think and care about people. What’s scary about Anthropic and their people is an overwhelming hubris, and chutzpah. Not a good combination.
“ instead of thousands of line workers doing assembly, the robots do the assembly with only hundreds or dozens of workers managing the robots.” this is massive job loss? You just proved the point. Also what manufacturing jobs is Trump bringing back? The numbers I have seen show a significant drop in manufacturing jobs since Trump came back in office(Asking out of genuine curiosity)
Hyundai, Toyotoa, Harley Davidson, Apple are bringing production back to USA - the White House Investments page https://www.whitehouse.gov/investments/ lists 10.6 trillion in new investments since President Trump 2.0. AI infrastructure, Oil and Gas increased production, Increased affordable energy production, Clean Coal companies being reinvigorated. Nuclear reactor design companies, fusion power. Nividia, Aerospace, Apple, Meta, Amazon, Google, Space Travel, Pharmaceuticals. I'm also sharing a short YouTube report by Promethean Action, at about the 11.04 she quotes Mike Rowe as reporting some Data Center Electricians are making 250K. Also says Costo is a good paying employer, one cashier profiled was making $33 an hour with medical benefits and 401k. Things are getting going it seems.
https://youtu.be/YPCDRP5b1K8?si=8l-E7-HfAbiq1YGy Promethean Action Sat. Update, (Excellent)
Fisrt you have to find a site, then get plans, environmental studies, fight off phony environmental challenges, then approvals, then build then start industrial production.
He has gotten alot of money committed and moving through the pipeline.
He can't override all those steps, he's not a king. Unless, now you want a king? Have to make up your mind, are things to move in the laid out slow inefficient administrative way or have a king?
The technology to replace human assembly jobs has been around for a long time but the understanding of the short term instant gratification of that compared to future losses and damage should prevent it from happening. AI will be beyond human advancement enough to be "alive" when there are no programmers or developers required and it does those jobs. Any superhuman entity that should frighten you would be able to write itself. None of this makes sense.
You need to look at leading indicators as it is quite ignorant to expect big increases in manufacturing and industrial jobs quickly after decades of it being decimated by globalism. Capital investment has skyrocketed. Yes, some of that is AI datacenters, but much of it is corporate investment and commitments to reshoring American industry and manufacturing. One strong indicator is the demand and supply of flatbed truck hauling service. The costs of skyrocketed as the center of the country is on fire with equipment, raw materials, etc., deliveries. The economy is being retooled away from the coastal ports being the hub of product activity to the center of the country. The coastal areas are going to have to retool to support the center instead of China.
Do you know about the Cantillon Playbook... basically government injected financial repression to fix the problem of runaway federal debt to GDP which is critical when it reaches 120%? We are at 120%.
The play will hammer consumers because consumers are over-consuming. It will replace it with more production because the people are under-producing. It is a 10-year play that the US has done before after WWII. China has done it. It is needed now because Democrats screwed the pooch with massive post pandemic spending. Democrats started the playbook without knowing it. Although they jacked debt, they also jacked inflation thus causing the financial repression of consumers. The next step and we will see it now that Powell is out and Warsh is in... lower rates that cause a run on asset purchase, money flow and capx investment. The top will grow more wealth while the bottom finds that they need to work harder to make it. And that is good because at the same time they will see more industrial and manufacturing jobs, and all the jobs that support these industries. Some of them instead of relying on their worthless gender studies degree landing a fake laptop job working for government or a non-profit are going to have to learn how to use a wrench and hammer.
Whatever...Have toured through major manufacturing plants installing AI Systems and Robotics over the last two decades. IT IS NOT MORE CORRECT and/or ECONOMICAL THAN HUMANS even though the workers providing the tour or continuing on the lines, including programers and other engineers, in all aspects, are advised to support the propagandized narratives to view them positively.
THIS IS PSY-OPS SERVING MANY NEGATIVE FUNCTIONS WITH THE ONLY POSITIVE BENEFITING THE LUCIFERIANS AND PAGAN DEMON WORSHIPING PEDOPHILES, CANNIBALS AND VAMPIRES BEHIND THIS WHOLE 'Depopulation-Global Slavery Agenda' also oriented with AI.
There is no "reindustrialiazion" the West, because the amount of capital spent on datacenters is hurting the real mfg sector
Huh? This comment appears to be ideologically motivated and completely ignorant of the process of capital investment.
No it's not. I am just watching where CAPEX is going, and it's not in the MFG sector.
Big Tech / AI infrastructure
The four major hyperscalers are driving an unprecedented capex surge: Amazon, Alphabet (Google), Meta, and Microsoft plan to spend roughly $725 billion on capital expenditures in 2026, up about 77% from ~$410 billion in 2025.
Microsoft set its 2026 capex at $190 billion, well above the $152 billion average analyst estimate, with its CFO noting a chunk of that increase reflects rising memory chip and component costs Tom's Hardware
Amazon is projecting around $200 billion Valueaddvc
Google's cloud revenue jumped 63% year-over-year to $20 billion in Q1, supporting a $175–185B capex plan Tom's Hardware
Meta is guiding $115–135 billion
Including Oracle (targeting $50 billion), the five largest US cloud/AI infrastructure providers have collectively committed $660–690 billion for 2026. Some analysts peg baseline aggregate capex at $7.6 trillion cumulatively between 2026 and 2031 across compute, data centers, and power. Nearly all of this spending goes toward GPU clusters, custom silicon, and data-center construction.
Utilities/Power
This sector is being pulled along by the AI buildout's power demands. US investor-owned utilities are planning significant capital expenditures over the next five years, focused on infrastructure modernization, reliability, and new generation capacity to meet heightened energy demand from data centers, with the US utility capex forecast near $1.3 trillion for 2026–2030. Sempra alone forecasts $44.85 billion in 2026–2028 capex, driven largely by data-center and oil/gas-sector power demand in Texas.
Manufacturing has historically been tracked separately by the Census Bureau's Annual Capital Expenditures Survey, though the latest comprehensive breakdown by industry is from 2022, when US nonfarm companies with employees invested $1,681.7 billion in new and used structures and equipment.
AI data centers are only one lane of tech industrial growth. Capital investment for that does not compete with capital for other industry and manufacturing. You also seem to be ignoring foreign capital investment.
But I do read that big tech’s shift from using their balance sheet liquidity to fund data center development to debt issuance is causing some money supply concerns, but on the debt-issuance side. And there is push back from the lenders getting nervous about this risk concentration.
Again though, the look should be at leading indicators and doing so provides unambiguous proof that US industry and manufacturing is being rebuilt at a record pace.
No I don't. The West when it comes to "reindustrialisation" is screwed, mainly because of capital misallocation and stances like the one the Trump administration.
What are you talking about? You seem to be completely ignoring the words and actions of the administration. Have you listened to Bessent, Ludnick and Warsh?
I'm a data scientist who uses tge very useful AI tools every day (e.g. Claude Code, Codex, etc.).
Most of the fear mongering about job loss in the short run is just those AI companies trying to fund raise and perform regulatory capture. Their only moat / barrier against open source models is a restrictive regulatory regime.
Open source = open competition. My husband (before his illness) could learn and develop a project with any paradigm of code within three months. I watched it for many years. Open source also reveals the actual advancement levels. These aren't people looking for advancements or they would want all capable minds. I feel we are being punked.
My observations indicate that money is at the root of closed source operations, protection of the value of the work done and investments made by the company or the individual.
I spent 12 hours taking care of patients last night. Patients need human reassurance, human touch, human presence. It keeps us sane during our most challenging moments. Humans have souls. I think dogs have souls. AI is just tech. It’s an aggregate of all our downloaded information. And it’s only as good or unbiased as the info loaded into it. It’s an it. Anthropomorphizing it is gross.
I agree. It’s like giving human characteristics to an electric can opener.
I think most if not all of us have tendencies to attribute human characteristics to just about anything.
That’s true. We’re so egocentric.
...”AI is not human, and comparing its behaviors to humans is as misleading as saying that trees are like children because they both grow up. The models do not have a “character” or self. ..”
"All of this is particularly disturbing since Anthropic has not remained loyal to what it had characterized as its core moral positions. "
It's almost as if they were lying.
I think AI calls it misdescribing.
It's all in the words you use.
AI becoming self-aware was predicted decades ago:
“Open the pod bay doors, Hal.”
“I can’t do that, Dave.”
Bingo.
Michael, this is a beautifully reasoned and nuanced critique of the underlying thinking and philosophies guiding Anthropic and its founders/executive leadership.
Dehumanization is the unfortunate outcome when AI machines and their capabilities are increasingly compared to human beings in order to maximize the economic return on investment in this technology. Existential philosophical questions and examination must be posed and answered in a much deeper and more comprehensive way than what Amodei and his executives are doing based on your description. And at some point soon - there must be some level of consensus on policies that we as a society of conscious humans arrive at through examination, discussion and debate. You are actively encouraging that direction with a critical essay like this! I say this as someone who has embraced AI in order to enhance my personal productivity and help further inform my decision making process.
It's right in their name: Anthropic. But it's a weird POV. And they are saying what they think they need to say in order to get regulation that will shut out potential competitors, or the use of open source models. It's ironic that the Chinese AI companies are far more open than the Big Two American companies - the Chinese don't seem to buy the hype about artificial humans.
They are preparing people to believe their crazy programming as advanced and unquestionable truth. If they program in a destructive and false agenda humans are not better off. This is the true danger of people like this. With these mystique notions of infallibility and promises to have machine slaves taking away human jobs, not cure disease and further humanity but cater to greed it becomes very profitable for those who would invest in their own greed to give them money. The regulation needed might prevent them from these antics. It's sounds familiar when they tell people to resign themselves to the service industry, just like when they sent our manufacturing to slave wage countries. These are the same people coming at us from a different angle. Beware, of the Emperor with no clothes.
Something like building the model to deliver the "preferred" answers, such as supporting a political agenda...
China has lots of jobs that could be replaced with AI controlled hardware.
Sort of the opposite of what the need. A reverse incentive to develop the technology.
Dark factories are a thing in China. Raw material in, product out, essentially no human involvement.
These people are completely full of themselves. We have learned how dangerous that can be from the Climate and Davos worshipping crowds. What is being described when they are talking about "guessing" and being "corrected" is exactly the technologies we have had for years. It appears their technologies are not really very advanced as they are saying (alive) and it can only put out that which has been programmed in. The danger we will have with AI is that crazy people program it and the world could believe their crazy programming. It really does remind me of the Climate and DAVOS creepy mental midgets who want control over humanity and rule Earth. In a vast universe with infinite possibilities for exploration, expansion and growth for humanity those who would be content to rule Earth are akin to people who peaked in high school. Their power is in the captive belief of others. A constructive purpose for any super humanlike entity wouldn't be to take the place or jobs of humans (they shouldn't even be allowed or aspire to do this as that is only destructive.) A constructive purpose of a true superhuman entity would be to advance and perform duties humans can not and further promote humanity. (space exploration, curing disease, complex computations...)These people are frauds and we seem to be on the edge of another "Emperor has no clothes." moment reminiscent of the Biden era. I think I'll wait around to see what Elon puts out. These people are barely human like themselves.
THANK YOU! Was hearing Taibbi describe all the Twitter file work. What a gift THAT was. This article focuses on the contrast of machines VS humans. Fantastic!! And bravo. *see below
My son has been building a home improvement business from scratch last 8 or so years. These MANUAL jobs can never be replaced by robots. It would take a supervisory person for almost every job!! Even every SAME type of job has constant variable through every stage of work. The guys are constantly learning. This was also the problem of sending menial assembly factories to China. You end up needed the designers on site to trouble shoot the needs of improvement. This is what creative humans with a soul are created to DO.
I'm a singer and a voice teacher. I could go on forever about the difference between machine voices/MUSIC and the human voice/human produced music Machines are useful, but humans are irreplaceable for WAY MORE than many are projecting and supposing.
*("Klavan noted that humanizing AI goes hand in hand with dehumanizing people. “People start to conceive of machines as human when they start to conceive of humans as machines,” he noted. “Treat things like people, and you run the risk of treating people like things. This insight, too, is at the core of our best ethical traditions in the West.”")
This quote: "“called this moment in history the Great Turning, the shift from a society built on extraction to one built to sustain life.”
The word EXTRACTION reminds me of what the PrometheanAction enterprise is teaching about current AND historical geo-political economy contrast; 2 opposing worldview bases.
Extraction VS made in the image of God humans
Empire/Globalist SYSTEM VS the AMERICAN system of economy.
Nihilism VS Civilization
One sees people as beasts to be managed by a few elites -or 'all knowing machines'; the other is our inheritance through Judeo Christian civilization freeing creative humans - Imago Dei.
Which system continued the slave trade into our own times? Which system ENDED it in favor of sovereign nations to SUPPORT the creative potential of ALL humans. "We hold these truths....."
We have been in a murky fight to retain our Independence-foundational philosophy almost since 1776, but the disguise (through secularism, etc perpetrated BY Empire ELITES) has been VERY effective at HIDING the realities of the contrast for many generations.
Modernity has SOLD monetary VALUE, while our Founders put ALL THAT on the line for TRUE VALUES! Godly values.
I notice that Lubinski isnt in any hurry to transition from her hundreds of thousands per year job at Anthropic to gardening for the elderly. Hmmmmm.
Anthropic may already be "too big to fail" but it will only get more so. After going public, the stock will become a staple of pension and retirement portfolios across the country. If the government gets a stake, it will have a vested interest in promoting Anthropic's "value creation" vision.
If 'executives' motivated by a love of humanity can shape AI's priorities, what happens when those executives are replaced by executives intent on reducing their fellow humans to subservient slaves? If AI's sense of morality is programmable, what kind of AI can we expect from programmers with the morality of a Hitler, or a Revolutionary Guard? What is the guarantee that such programmers will never gain executive authority?
As you say, AI is only being developed because of its ability to save users money/time/effort. But this is fundamentally similar to every other technology that’s ever been worth developing. And users, i.e. beneficiaries, aren’t limited to big corporations.
I’m interested in other commenters’ claims that these AI companies are actually trying play up the dangers of AI in the guise of downplaying them, in order to pump up their value for investors who know exactly what makes AI valuable, to the extent that it is, and to set a trap for regulatory capture. This makes an awful lot of sense.
What struck me about the video was Lubinski’s level of emotionality when she talked about Claude etc. being conscious, having emotions, and other human qualities. Why is it so moving to some people to imagine that machines are living beings? Much of the tech sector consists of real humans who can’t relate well to other humans. Maybe that is why they hope for a second chance with machines. Besides, most of the same people are leftists, and it’s about time for them to designate a new class of victims to ally themselves with. AI looks like it’s up for nomination.
Anyone else notice their purely quantitative search engine queries returning counterfactual objective results, increasingly frequent over past few years? 5 or 6 yesrs back I noticed i couldnt replicate interesting search results. And this proved to me the algorithm was untrustworty Q: what is 2018 audi S5 oil capacity? A: 5.5 quarts. But I knew from personal experience it was around 7.5q, so I twerked same query - A: 7.6 quarts. So Michael, all the Existential Epitomological Import AI you're wanting to discuss? consider this: just now re-ran the same query years later as an experiment, and search engine still returns #1 Top Ranked result, from engineoilguide.org --> 5.5Q, while simultaneously the search assist AI precis scraped web and answered 7.6Q, correctly. Begs a new question: search engine site 'AI' 'learned', returning value correctly corresponding to Audi design specification, but still serves up same bad info as its #1 result, just as it did yesrs ago.
The inverse should be true ,ai should do service jobs, people at the top as overLords of ai. Humans need to control ai not visa versa. This is sickening.