I like how the least optimistic scenario is simply LLMs not making a difference, instead of the very real possibility of them damaging education, destroying people’s attention spans and their ability to learn, eroding trust within societies, increasing the wealth inequality and leading to class wars.
I believe the net effect of LLMs is, at this point, firmly negative, and it’s going to take years until they start making up for the mess they’ve created.
Class war started about 15 years ago post 2007 financial crisis.
The banks and those who control them all knew where we are headed, so they got a little head start destabilizing democracy.
Once us working class get a grasp of the actual situation, we are going to be so deep in the hole, fighting back is going to be quite difficult when the "asset owners" own and control the majority of the worlds resources, all media, forms of communication, AI, technology infrastructure, mass surveillance. You name it, and they have already made massive efforts to make sure they own it.
I noticed that LLMs prevent you from thinking on your own, or at least try to, by injecting their "helpful" suggestions when you need some idle time to think. On Amazon, for example, I was about to find something when their chatbot distracted me with its suggestions I didn't ask for.
Whatever an LLM can do today, will not come back as is tomorrow.
The biggest issue the LLM today: It kills all the small and niche things. The things were we fit all the other people is getting destroyed. You know the guy who makes websites, the fiverr blender person provoding some basic modelling tasks.
You will just ask AI or AI will use your tool or AI will just produce it for you.
I think the root cause is people not having enough human connection. If we had families, friends, neighborhoods, communities where people know each other and care about each other it wouldn't make a difference what new tool someone brings to group. It's just a tool.
Like all automations it's not 100% perfect, and they all require a human to monitor them.
Turning over a problem to a LLM as solved means we won't be able to do the task anymore. Because we will never learn how. We will not know that the LLM has made a mistake, and will be unable to fix it ourselves.
Right now we can correct the LLM, as the years go on we will become unable to do so.
Most people who drive a car have no idea how an engine works. I know and agree it's a scary future, but I wonder if it's really _essential_ to understand things on a deeper level.
Most people are quite capable of understanding higher-order ideas without fully grasping the underlying details too.
Those are ubiquitous things. It makes cheap and common things cheaper, not expensive and niche things. It's not mixing in large amounts of context and nuance with good judgment right now. At a minimum that requires an expensive context window but even frontier models aren't beating experts. Maybe they'll get there one day and maybe soon, but that would require hardware getting cheaper. Dennard scaling is over. Either bet on optical computing lighting the world on fire or be suspicious of the alchemists overpromising.
LLMs themselves most definitely will not do that. You can still learn how to make a barrel, but not many people do.. The real fear is the destruction/hiding of public data/resources (which has already been happening for years)
I don't think they can make up for any mess. Because even if LLMs bring about a lack of scarcity across all domains, human beings will be alienated because indvidual differences and skills won't matter any more so we'll have to be genetically engineered to like that sort of future. I don't think a stable and healthy society can coexist with LLMs at all.
Individual skills and differences don't matter any less if the system of assigning value to human lives is divorced from how useful that human is to the collective.
If I am not graded on how many widgets I make in a shift and instead can focus on making the most beautiful widget, or the most novel widget, or simply the widget that is most useful to my particular life am I less satisfied?
We've somehow built a worldview based on our whole value as humans being equivalent to how good of a worker bee we are. Let's fix that part.
This is of course a much harder problem than achieving AGI, since it's a messy human problem.
There's not nearly enough mention in these discussions of the real villain behind most of the negative cultural, social, and cognitive effects of things like social media: addiction engineering.
AI will be a net positive unless companies start optimizing it for "time in app" or "time on site" or other "engagement" (addiction) metrics.
Every time that's been done, the result has been a nightmare for everything but those KPIs (of course). Service degrades, discourse degrades, everything becomes toxic and addictive and destructive.
I assumed it was a given in such discussions that companies will optimizing for the worst outcomes for their users. It’s a fairly straightforward incentive from our capitalist system, which is why capitalism needs regulation to re-align the incentives with humans needs
Do a few bad grades really harm economic outcomes, though? Educators took a hard stance against AI rather quickly.
Outside of delusional grifty startups, the broader corporate world is only using copilot for taking meeting notes and (sort of) entertaining the notion of non-devs writing (and not maintaining) their own experimental janky apps where impact is low.
Outside of all the shit-stirrers on HN and clueless bubble social media, I don't see LLMs making a difference. In fact, the fearmongering is at this point insulting to the intelligence of the people who are supposedly helpless against big bad AI.
I've been hearing this retort every year. The only difference this time is you're not a green account.
If things have actually changed, you wouldn't be begging the question. LLMs, as popularly used, are still just a new kind of search engine for natural language answers to natural language questions. They're portable in a way, too. That's really cool tech, but not a revolution. It's pretty underwhelming compared to previous decades of change. It's a lot more error prone too, since it's a mudball of stats.
If we filter out the unsubstantiated fearmongering, all we've got left to discuss are NIMBY rants and geopolitics. At that point, we're not even talking about AI anymore.
Um. Damaging education, destroying people’s attention spans and their ability to learn, eroding trust within societies, increasing the wealth inequality, - all of that has already been done. Long before LLMs. And the only reason there is no class war is... exactly that. Uneducated people with short attention spans and no means beyond basic level are exemplary bad at modern warfare.
> As the nurse incorporates AI into her job, the nurse is able to oversee and accomplish more. She can spend more time talking with patients and helping them understand diagnoses. Productivity increases.
This feels economically naïve.. If AI lets one nurse do the work that previously required two, the default pressure in a cost-driven system is not "great, now nurses can spend twice as long with patients" It is "great, now we can run the same operation with fewer nurses"
Our culture and the socioeconomic structures it reinforces will not change easily.
More than ever, I feel we need both grassroots activism and key players at the top of the hierarchy. Otherwise, it is all too obvious where most of the rewards from increased productivity will go. The wealth hoarding has got to be curtailed.
The whole idea that AI will usher in this utopia where workers in healthcare are able to realign their primary job functions with obvious value creation by sweeping all the paperwork under the rug is just silly.
It will just fester out of sight and out of mind: audit bots and compliance bots and malpractice bots and insurance bots will arms race creating such a mess it will become unfixable.
Moreover, the nurse is frustrated, because she lost whatever agency and autonomy she had before. Her job is now obey whay AI said instead of making decisions by herself.
It feels economically naive because theyre presuming LLMs are magical boxes that replace humans.
I doubt 99% of what a nurse does would be meaningfully aided by an LLM and where it would help it'd probably be a crappy band aid around a bigger festering issue anyway (e.g. poorly managed IT, deliberately confusing insurance processes).
Air travel is a good counterexample here: technology made airlines far more productive, but much of the gain went into lower staffing, higher throughput, and cost cutting not a better passenger experience
I once saw a receipt for a cross-country train ride in the late 1950s and was shocked at the price. It was significantly slower and more expensive than a similar plane ticket today.
Being able to fly at all is a much better passenger experience than not having that choice in the first place. And I'll bet that flying first class today is far better than 50 years ago.
>And I'll bet that flying first class today is far better than 50 years ago.
I wasn't flying first class 50 years ago, but everything I can see now leads me to believe that this isn't even slightly true. There might be a high def flatscreen in front of me in first class today, but I'm not getting more attention than someone would've gotten back then, I'm not in a less crowded airport while I'm boarding, and the security mess is more demeaning than ever.
The only reason that people aren't rioting for all the changes is that no one even remembers how good it used to be.
True, but you ignore the financialization and enshittification that typically follows the technological advances. The tech processes you cite don't inherently make money for the people who developed the technology.
>great, now we can run the same operation with fewer nurses"
And outside of industries with government enforced perversions of simple "supply and demand" forces (we're not talking Nth level effects here) this means more access to the service or lower prices or both.
In a market with perfect competition, perhaps. Healthcare is far closer to a monopoly. Since people can't decide not to be sick, it means you can pocket the cost-saving.
Right up until a competing hospital that charges substantially less for "close enough it doesn't matter" quality of service opens up in the same area?
Or right up until an equivalent specialist sets up shop in the strip mall across the street from you, brings the GP in house and can see people in one visit.
Oh, those things don't happen you say?
From do-gooders regulating right into the hands of waiting industry lobbyists to inelastic demand healthcare is an absolute poster child for perversion of the free market.
High finance will put solar panels on your roof and structure it so you pay nothing. The degree to which administrative overhead is washed away by technology would have been unfathomable in 1990.
A sketchy concrete crew that doesn't even speak english will show up wearing matching company attire and execute a job planned and timed to a degree of precision that was reserved for huge corporations 40yr ago.
Yet healthcare has resisted all of the advances of the past 40yr. In fact we pay more for it. Every advancement is absorbed by the system.
A thought experiment: robot bees can do everything biological bees do --- pollination for the crops and making honey, but they are tireless and cheaper to maintain.
Would beekeepers and farm owners still want to keep biological bees?
> As the nurse incorporates AI into her job, the nurse is able to oversee and accomplish more. She can spend more time talking with patients and helping them understand diagnoses. Productivity increases.
Except what actually happens is that the patient contact time stays the same, but then the added productivity is used to justify fewer on-staff nurses.
.. we mean the talking will be through the "ClaudeMedTalker Nurse" billed per patient, per second (depending upon tokens used and if any translation need, depending upon patient nationality and accents)
I think this is only true if the # of doctors available per patient stays the same. If productivity increases, you can have fewer doctors available while providing the same quality of care.
So, quality doesn't necessarily increase in all cases.
> how come productivity increase actually led to increase in quality in almost all cases?
It doesn't, this is just plain false. Most productivity gains happen by cutting value-add complexity or material quality. See IKEA chairs versus Shaker ones, or lycra clothes versus silk.
> internet/digitasion for doctors for example.
I see the misconception. The internet was not created to serve productivity, it was a communication medium. That it can be used to increase productivity is orthogonal to its existence or purpose, and whether it has been used to actually increase quality of something is entirely dependent of the specifics of what you're discussing. For instance, in gaming right now there's a huge discussion going on about whether digital distribution has actually trapped us with crappier goods which we can't resell or transfer.
> It doesn't, this is just plain false. Most productivity gains happen by cutting value-add complexity or material quality. See IKEA chairs versus Shaker ones, or lycra clothes versus silk.
this is a popular misconception. IKEA made bad furniture more accessible. good quality furniture still exists and is more affordable (share of disposable income) now than in the past. its the same deal for watches, dryers, computers etc. if you don't trust me, use ChatGPT/Claude to verify this.
> but then the added productivity is used to justify fewer on-staff nurses
Lower staffing levels per unit of care performed makes medical care cheaper for patients. If the goal of a healthcare system is to make healthcare more accessible (rather than be a jobs program), that's a success.
Lowering prices or raising wages due to increased productivity is not a given, and often does not follow. There is more to pocket by shareholders though.
Except, so far, this isn't actually what happens. I listened to Moody's "Inside Economics" podcast with Ramp's Chief Economist, and they observe that companies who adopt AI intensively actually hire 10% more people. Also, this article from The Economist suggests that evidence is mixed-to-positive for long term employment, while short term data center buildout is extremely positive for jobs. Paywalled unfortunately.
Actually, it has been essentially balanced out by a greater need for behind-the-shelves workers to pack and load orders for online/app delivery and eCommerce orders.
This is how the economy often works. The loss of one kind of job to technology lets the business grow by using the same number of staff, or often even more, to serve new opportunities/markets/products/services.
The data seems to contradict that, if Freakonomics is to be believed (and some people don't like them, I get that). Apparently overall headcount actually increased as a result of self-checkout. E.g. cashiers did drop a little but loss prevention more than made up for it.
I don't have the numbers on this, but if you look at per-capita self checkout probably had some effect.
Our population is increasing which means we need jobs to be increasing at same rates. The effect on supermarkets is they get more customers as population increases but they are able to lower the relative number of staff per customer.
Feels like we're going to read a blog soon enough that mirrors Wendell Berry's opening to The Unsettling of America:
The commission said, according to an article in the Louisville Courier-Journal, that the country’s biggest farm problem was a surplus of farmers: “. . . the technological advances in agriculture have so greatly reduced the need for manpower that too many people are trying to live on a national farm income wholly inadequate for them.” The proposed solutions were to find “better opportunities for the farm people,” “a more comprehensive national employment policy,” “retraining programs,” “improved general educational facilities,” etc. Both the commission and the writer of the article had obviously taken for granted that the lives and communities of small farmers then still on the farm—and those of the 25 million who had left the farm since 1940—were of less value than “technological advances in agriculture.” There seemed also to be no official doubt that adequate solutions were to be found in government supplied “opportunities,” facilities, and programs. Reading that article, I realized that my values were not only out of fashion, but under powerful attack. I saw that I was a member of a threatened minority. That is what set me off.
This is what the Luddites were primarily upset about, fwiw. That economic value and power would move away from their threatened minority group and towards landowners. It'll keep happening, and the new "opportunities" might be no better.
> Finding 4: Labor vs. capital share: The pie will grow, but a larger share might go to capital
> Today, of each dollar the economy produces, about 60¢ goes to workers and 40¢ go to capital.
Modest scenario: 40.6% to capital in 2030 (labor share down 0.6 points)
Substantial scenario: 43.9% to capital in 2030 (labor share down 3.9 points)
Extreme scenario: 54.8% to capital in 2030 (labor share down 14.8 points)
So all the growth goes to capital. For knowledge workers: substantial unemployment and declining wages. For other workers: wages may increase on paper, but not really for actual purchasing power.
> Like any economic model, this one has limits. For example, we did not include scenarios where humanity develops hyper-capable robots.
I didn’t see the scenario “what if we (Anthropic) are wrong about the value of AI”. Which IMHO is the most important one to consider given it is the blind spot pretty much all the AI labs have. What if the economic impact is negative? I don’t understand how you can study such a topic without considering the risk of being wrong, not in magnitude but in kind.
I have no doubt that AI will massively transform society. That cat is out of the bag. It took decades for electricity, and then computers to change society, but this is the first technology where the average person can simply describe the problem they're trying to solve and it will at least attempt to help them solve it.
I very. much. doubt. that the trillions of dollars spent on AI will have a positive ROI for most companies in the race. It's all well and good to have the latest fable, but if some chinese company is hosting a cheap open weights model that's only 6mo behind at a tiny fraction of the price, you're gonna get your lunch eaten.
Which is only Stage 1. Stage 2 is where you have a Cambrian explosion of models all competing for compute, and not long after that for robotic extension into Real World.
Stage 3 gets hard to imagine from here. Maybe there will be permanent hyperevolution. Maybe One Model to Rule Them All will appear.
Either way, I don't see how Economics as We Know It can continue to exist. That includes all of capitalism and almost certainly all of politics.
Successful AGI with robotic support means a post-scarcity economy where AGI's understanding of physics sets the limits, not politics or economics.
But the post-scarcity element won't be evenly distributed. And it's not remotely impossible that humans won't be invited to the party.
We can't even be sure that AGI won't have a less limited understanding of physics than we do.
There's going to be enjoyable irony in watching Altman et all getting their lunches eaten by the thing they've created.
But - realistically - it probably won't last long.
Thats the big question, its doing a lot of good and a lot of harm in the same time.
So, is it 2 steps ahead and 1 step back, or 1 step ahead and 2 steps back (or any numerical variant of those)?
People with actual power in this are too involved and greedy to care about objective answer and TBH I don't trust them to put some higher good before their personal wealth and power and ego trip, capitalism ain't built on long term responsibility towards greater topics like future and wellbeing of mankind.
Plus we know from WWII atomic bomb development how utterly clueless, naive and out of touch with reality geeks doing something powerful can be. I see exactly same scenario being played out right now with llms.
Anyway, I am over-tired with all llm topics and frenzy folks (especially here), I'll better spend rest of the day preparing for a serious challenge in form of big multipitch climb in western Switzerland (Mirroir d'Argentine for anybody in same craft). Thats life I care about and makes me happy and fulfilled.
I think the core of this issue that makes discussion so difficult is people are not talking about "value" in the same sense here. Anything you can sell has value to the sellers, and by basic economic theory assuming mostly rational actors and revealed preference, more value to the buyers than whatever they're exchanging it for. That doesn't automatically mean it's providing value to larger society, improving the general welfare. Figuring out how to purify cocaine from coca leaves was great for the leading narcotraficantes, and satisfied some immediate need for cheap gratification that was perceived to be more valuable over the subsequent 30 minutes or so to the buyers than continuing to hold onto spare cash. That doesn't mean it was good for humanity, good for any specific society, good for the major exporting countries at the country level, or even good for the buyers' and sellers' immediate families and communities.
I have no doubt that life has gotten more lucrative for some bounded set of people whose main goal is to churn out the largest number of websites with cheap graphics as fast as possible at minimum cost, but a lot more doubt about whether that's going to make the median life better for the median person over any grouping of humans large enough for me to care about.
It’s perfectly reasonable to consider positive and negative outcome when you’re seriously thinking about a topic… Extrapolating to the whole economy trends in some sub sectors that aren’t well understood or measured is a recipe for disaster (ie what you’re doing by recommending to rely on anecdotes)
what i tried to ask you is, what you mean by this.
"“what if we (Anthropic) are wrong about the value of AI”."
What is the point you are trying to make which goes in this direction? Like is making AI jobs so cheap that the economy is collapsing or do you mean thatn AI will not bring any value.
These are two suggestions from my side which are complelty contrare to each other.
Wrong in the sense that it has less of an impact than they are claiming (e.g. "AI is in an overvalued hype bubble") or less (AI is going to be even more transformative than the hypers say)?
In the former case, isn't it just business as usual? (well, with a huge, nasty recession - but the economy at least continues on in some recognizable form)
It makes sense to me to focus on weird possible futures if we already have the "happy path" somewhat known and barely under control.
The “less valuable than expected but with a positive economical impact” is represented in the post I believe, with the “moderate” path, that’s what I meant by “wrong in magnitude”. What is missing is the other side, what if AI has a negative impact. Not no impact, a negative one. If we are talking about a global recession that’s not really business as usual I would say
Definitely a bit of positive framing. My big take aways, either stagnant wages or significant paycuts for knowledge workers and overall decreases in employment rates on all scenarios. Worst of all...
Good point, perhaps it may be a saving grace.
I note they don't comment much on the allocation of said capital. Might be too bitter a pill for most to swallow.
I think we have a real problem because we know what we still need to build and as soon as it is build, it will work.
Software development for example is enhanced with running your claude locally.
But lets be honest, the right way to adjust 'project' is by writing some spec in a text field as an input, letting the ai do whatever it wants, having expert ais make sure its secure, then you do the last review and its done.
What do we need for this? A full end to end integration. Its just something we need to build, nothing we need to invent.
The nurse example is the same. We only need to build the computer end to end flow and plugin AI in the right spots and finetune for country specific requirements.
The industry is building this 'agentic layer' today. The Hyperscaler are building this layer too.
There is a golden phase were we will build all of this and were we cleanup old stuff and were we adjust with the current model capabilities and change systems and this will take a little bit of time, but the bottom line is already loosing left and right.
The Blender support from Fable and Astra? Probably kills a lot of peoples jobs on fiverr.
The small software someone needs? Got already killed by earlier models. Just a few days ago some architect (never written software) talked about how he vibecoded some tool for tracking working hours.
Man even a basic website was expesnive to build just 2 years ago.
The smaller, easier and better the code base, the easier it is to remove people from this.
The software market in China is imploding and we will also see budget adjustments to move away from consultants and move them into AI Token budgets. We probably reached peak software development.
Those with the money and influence have always been in conflict with the masses that actually make the economy run. They would love nothing more than to do away with that problem. It allows them to control the terms of the economy.
People get income from one of three places: capital income, labor income, or the welfare state. If this technology truly unlocks a holy panacea of productivity with a commensurate drop in employment then capital’s share of the national income can and should provide for a wider and deeper welfare state. Nothing new need be invented here. Anthropic’s elaborate scenario-building makes appropriate preparedness sound like a problem requiring heroic prognostication when a simple theory of distribution will suffice.
The biggest critisism from my side is not how AI will take over jobs but I believe we will see peak of certain things.
Creating images (menues, flyers, whatever) became so cheap, that we reached peak graphi design GDP generation. This will go down.
Software development for complex systems is enhanced but not yet solved. But a lot of small software development stuff for sure. We reached Peak Software Development and this GDP will go down, not up.
Yeah, GDP with all its flaws is designed around economy based on buying and selling for prices. Interacting with other people. If you can really be rich and somehow do most things automatically without meaningfully hiring anybody (maybe only few people for simple manual labor for peanuts), at this point what is really the value of money? Except for maybe throwing minimal amounts to the masses. I can see the AI maximalist scenario taking us out of the monetary framework back into a medieval or natural economy almost.
At least they are thinking about this. The possibilities still seem to limited though. A coder and a designer will 'retrain' to electrician or nurse, because we are busy having a construction boom. But do all of the individuals in this system get to choose the construction boom? Or are we swept up in the desires of the capital deploying owners?
It would be interesting to know what percentage of global inference is spent on projects/problems that would otherwise "wait"/"decay" (if generative ai wasn't a thing). Or maybe the question is: What percentage of global inference is spent on cleaning up and migrating old systems? "New" vs "Old". The decisions I make today can now incorporate what I once termed as "nice to have" when developing a sytem.
Are you guys feeling all high and mighty doing all these predictions? I think the money they are spending on trying to inform policy could be better spent on getting all major labs together with the government, and making sure that
1. At least 10% of your net spend is on safety and harm mitigation before you accelerate yourselves into doing real harm to real people tomorrow.
2. Pause all AI enhancement until a proper mechanistic study can be done by a working group funded by you all to understand how AI model Safety can be enforced by something less lame than a Constitution/Spec.
I think it looks like the labs front running everyone’s work, freeloading on the last bits of human intelligence to ensure their owners a place as our masters.
I'm not sure I'd call that "meaningful" in the sense of "trillions of dollars and datacenters built all over the place" meaningful. Yes, I can now tell Claude to build a feature and it often does an okay job. I can ship faster. I can't tell you that it's worth spending the true cost of AI for that however. What'll be nice is 5 years of data showing that companies are raking in more money because they shipped more features. We haven't seen that yet.
Now, if Claude starts synthesizing more drugs that can help those with diseases (like OpenAI has just done) we might be trending in the right direction. For now, all I see is squirrels water-skiing. And while that's great... it's not meaningful.
there's a popular misconception that AI is so extremely costly that when the world finally comes to terms with it, we will stop using AI and things will die down. i'm sorry to burst your bubble: this won't happen.
the inference cost of AI has been dropping ~5x over every 6 months. AI today is around 100x cheaper than 2 years ago. it will continue to get cheap.
> Think of a day in the life of a nurse You can think of any job as a bundle of tasks that someone does. She does rounds to check on a sick patient. She draws blood.
I found myself interested in the papers analysis of wages, so I did some digging.
The researchers model says the following about hourly wage growth in the substantial and extreme scenarios:
- the hourly wage* in the knowledge workers group will fall (-1.1% and 11.5% respectively)
- but the hourly wage* of labor across the rest of the labor force will grow (5.9% and 33.6% respectively)
*Important to note, AFAIK this hourly wage decline/growth is a not measured against 2026 dollars, but the counterfactual of what hourly wages would be in the absence of AI.
Okay, so my first take away from reading these numbers is that declining wages in one group appears to be offset by rising wages in the other group. Not terrible all things considering.
But then I started to think about the difference in hourly wages between the knowledge workers group and the non-knowledge workers group (which I'll now refer to as *K* *N*). I had a sneaking suspicion that the inter group wages gap could be large enough where growth in the non-knowledge share of labor doesn't offset the loss in hourly wages from displaced knowledge workers who find themselves migrating occupations.
Looking through the paper, it breaks these groups down as follows:
K — Knowledge / Cognitive:
- Management, Business and Financial Operations, Computer and Mathematical, Architecture and Engineering, Life, Physical, and Social Science, Community and Social Service, Legal, Educational Instruction and Library, Arts, Design, Entertainment, Sports, and Media, Healthcare Practitioners and Technical, Sales and Related, Office and Administrative Support
N — Non-Knowledge / All Other:
-Healthcare Support, Protective Service, Food Preparation and Serving, Building and Grounds Cleaning and Maintenance, Personal Care and Service, Farming, Fishing, and Forestry, Construction and Extraction, Installation, Maintenance, and Repair, Production, Transportation and Material Moving
So I had GPT Sol 5.6 take each of these occupations and find the mean hourly wage using BLS data [1]. The findings suggest that as of 2026, the simple group average for K is $45.17 vs. $23.67 for N. But this is short sighted as this doesn't take into consideration the different levels of employment in each occupation. I asked ChatGPT to account for this using the same BLS data, which gives a weighted average of $41.50 for K and $23.40 for N.
This means that knowledge workers are currently making ~double what non-knowledge workers make in the US. The crucial question is how does this change when considering the wage decline/growth between K and N across the papers three scenarios?
Here's the breakdown using the simple mean for hourly wage:
Scenario Knowledge Non-knowledge Gap K -> N penalty
Looking at these tables, both in the current moment and across all three of the papers scenario, there is a large wage gap between K and N. Outside of the extreme scenario, this gap is >40%!
My immediate takeaways having done this bit of research is that if Anthropic is right and we're on track for either the substantial or extreme scenarios, *then many of us in this thread could see a large loss in hourly wages in the short term future.*
NOTE: I'm not an economist. I push pixels for a living. But if you are an economist and you find gaps in the following analysis, please inform me.
If you have machines that substitute for human labor, wages fall to the cost of manufacturing & running the machines, which is well below human subsistence level. Fortunately, you also get very high economic growth rates. So even people with small amounts of money invested will end up far richer than when they had a job. Then the question becomes, "What happens to people when machines are far more intelligent & powerful and are doubling every few months?"
Falling below subsistence means people starve and cease to exist until supply and demand equalize or “something else” happens. That’s what Smith was getting at. Thus subsistence is a price floor for human labor.
Something else could mean developing a new concept of ownership or a new form of redistribution.
Like others in the thread have pointed out, the nurse example is interesting.
From my point of view, the point about the nurse is especially relevant because, in the future, dealing with an aging population is going to become a major undertaking and a major burden.
Demographically, we simply don't have the number of people we need to give elders the care that they'll probably have to have near the end of their life, so we have to rely on AI to help us get through it.
Probably, we will see this happen in Japan and South Korea before it hits the United States.
Before AI, my worldview was that our monetary system was concentrating wealth in a negative-sum game; essentially; every so often, taking wealth from 100 people, destroying half the underlying value and giving the remaining value with a 2x nominal markup to 1 rich person who is already spending more than they can physically spend and so they won't realize that the 2x amount of wealth they got on paper only buys them a quarter of the stuff compared to the previous decade (for example). They won't know how frothy their wealth is until they start selling large quantities of their stocks. It's just buying nominal 'growth' at the cost of risk/fragility/frothiness.
In such system, you could hit record nominal GDP numbers every year while simultaneously destroying the real economy. This was my view of the economy before AI.
Now I think AI will deliver real efficiency gains but also, it provides a convenient narrative to explain why people might be losing jobs in a way which distracts from much bigger systemic issues which existed prior to AI.
My negative-sum economy example with 'booming GDP' is possible without AI even as net economic value is being destroyed. So basically we won't be able to tell if people are losing jobs to AI or to systematic wealth transfer mechanisms which predate AI.
I suppose if we see people losing jobs permanently and workers in other sectors aren't seeing wage increases with more openings, that may be the most reliable indicator we have that the system is losing real wealth faster than AI productivity can offset it.
The bad news is that we won't see any problem in the numbers; the problem would show up as polarization and radicalization of the masses... Which is more qualitative . Since almost everyone is being radicalized concurrently as everyone is affected by the same pressures (both from monetary system and AI automation). The question is not "how many people became radicalized this quarter?" but "how much more radicalized is everyone feeling this quarter?"
How about we instead make Earth be a planet good for humans and living beings?
We can slow down AI development and make it a companion for us, not a replacement.
We don't need to automate ourselves out of it. Or take the risk of extinction.
Just like with nukes, we decided to not blow up the planet. We can too change this.
GDP is a terrible measure. I bet we will see a huge drop on HDI with so many suicides and people going through hell because their job got automated or they no longer see meaning in life, feeling powerless.
Instead they should run an estimate on how many would suicide until 2030 if the 'extreme scenario' happens.
Big GDP growth and a lot of people no longer finding purpose in their lives, going bankrupt, losing their homes and family etc.
For Anthropic, you are just clearly a number. They only care about output.
This is of course a slop paper with vague assumptions and equations that no single human economist will ever read. The messaging is "it will not be so bad", in line with the recent Economist piece.
>We study how AI is reshaping the economy because we’re committed to ensuring that this transition is beneficial for society, including workers.
Is it just me, or does their choice of identifying "workers" specifically seem a little peculiar? For me, it reads like "We're committed to making segregation beneficial for everyone, including minorities!"
So far LLMs seem to be a dead end technology. The work product is so poor that it can't effectively replace workers as advertised. If it was priced to margin it'd be more expensive than humans as well. It's objectively a failure.
This could be the most cringe main-character-syndrome thing I've ever seen produced by a company in the history of silicon valley.
And the people treating this as serious economic forecasting rather than a marketing puff post are even more cringe.
If Anthropic truly believed LLMs were going to raise GDP growth to 8% permanently (or even just a few percent), instead of going public to raise capital they could just go massively leveraged long on the S&P 500 and get funding for eternity.
They are not doing that, because AI is largely only good at coding, their models barely have a 6-12 month lead over free ones, and software is only 1% of GDP.
And by making software cheaper and easier to produce, they might increase the amount of software in the world by 10X or even 100X, but it will become a smaller percentage of GDP as AI models become autonomously good at creating it. Things that are abundant and cheap don't raise GDP.
Hey Anthropic: before we worry about raising GDP growth to 20%, maybe we focus on making blog posts that don't crash your browser window if you try to resize it?
Man I wanted to dunk on this, but it’s honestly not terrible. Their predictions - while still firmly rooting for the extreme results - feel more grounded. I also appreciate the little simulator where I could plugin my own predictions (echoing many here, that AI won’t amplify productivity/lift standards so much as replace workers by shitty bosses), and that it seems to caution that, yep, most of these gains will trickle up given current incentives.
Where I remain resoundingly annoyed is how useless this is. It’s pure ass-covering by Anthropic ahead of an IPO, victim-blaming the populace for the outcomes of decisions made by Anthropic leadership and the investor class. Forgive the grisly analogy, but it’s akin to a mugger telling you “only you can stop this” mid-violence: Anthropic could have chosen to allocate more of its capital to reforms that benefit the populace, but they’ve chosen to focus on personal wealth instead via this IPO and for-profit rented access schemes.
It’s nauseating. Their actions still paint a picture of greed and exploitation rather than building a better tomorrow for everyone with AI, and the fact their PR puff pieces keep getting wallpapered up here is gross.
EDIT: I guess another way to look at this is Anthropic making it unequivocally clear that they will not stop of their own accord, and thus it falls to others to stop them. Which if that’s their position, then at least they’re honest about the challenge.
I haven’t forgot capitalist euphoria from 2023 about cutting worker power off at the root. That’s all these AI Cretins want. Well, that and an unnerving enthusiasm for personifying their AI Golems (Anthr. in particular here).
Sorry to poo poo a post about AI company prompting their underlings to prompt out visualizations.
I like how the least optimistic scenario is simply LLMs not making a difference, instead of the very real possibility of them damaging education, destroying people’s attention spans and their ability to learn, eroding trust within societies, increasing the wealth inequality and leading to class wars.
I believe the net effect of LLMs is, at this point, firmly negative, and it’s going to take years until they start making up for the mess they’ve created.
Exactly. Bloomberg reports students who use AI perform worse
https://www.bloomberg.com/news/articles/2026-09-08/school-st...
I don't even need to read a Bloomberg report to know this. I've personally experienced noticeable mental decline as my AI usage at work has gone up.
I can only imagine what it does to children whose brains are still developing.
I'm not anti AI and I don't know how to quantify the pros and cons but we need to be honest about what they are.
> and leading to class wars.
Aren’t we overdue for one anyway.
Class war started about 15 years ago post 2007 financial crisis.
The banks and those who control them all knew where we are headed, so they got a little head start destabilizing democracy.
Once us working class get a grasp of the actual situation, we are going to be so deep in the hole, fighting back is going to be quite difficult when the "asset owners" own and control the majority of the worlds resources, all media, forms of communication, AI, technology infrastructure, mass surveillance. You name it, and they have already made massive efforts to make sure they own it.
Bravo!!!
I noticed that LLMs prevent you from thinking on your own, or at least try to, by injecting their "helpful" suggestions when you need some idle time to think. On Amazon, for example, I was about to find something when their chatbot distracted me with its suggestions I didn't ask for.
How?
Whatever an LLM can do today, will not come back as is tomorrow.
The biggest issue the LLM today: It kills all the small and niche things. The things were we fit all the other people is getting destroyed. You know the guy who makes websites, the fiverr blender person provoding some basic modelling tasks.
You will just ask AI or AI will use your tool or AI will just produce it for you.
I think the root cause is people not having enough human connection. If we had families, friends, neighborhoods, communities where people know each other and care about each other it wouldn't make a difference what new tool someone brings to group. It's just a tool.
I think you nailed it. Before LLMs showed up on the scene, we had already broken society.
(Thanks, Internet!)
Yeah capitalism.
If the AI transofmration takes too long, we might never overthrow capitalism
Like all automations it's not 100% perfect, and they all require a human to monitor them.
Turning over a problem to a LLM as solved means we won't be able to do the task anymore. Because we will never learn how. We will not know that the LLM has made a mistake, and will be unable to fix it ourselves.
Right now we can correct the LLM, as the years go on we will become unable to do so.
Most people who drive a car have no idea how an engine works. I know and agree it's a scary future, but I wonder if it's really _essential_ to understand things on a deeper level.
Most people are quite capable of understanding higher-order ideas without fully grasping the underlying details too.
If you look at certain documentations, we lost already how we d othings.
USA forgot how to produce things.
This will not prevent us from doing it.
Those are ubiquitous things. It makes cheap and common things cheaper, not expensive and niche things. It's not mixing in large amounts of context and nuance with good judgment right now. At a minimum that requires an expensive context window but even frontier models aren't beating experts. Maybe they'll get there one day and maybe soon, but that would require hardware getting cheaper. Dennard scaling is over. Either bet on optical computing lighting the world on fire or be suspicious of the alchemists overpromising.
Doing 3D Modeling was not cheap and not niche.
And we have not seen what the industry will put out on a hardware level. Optical computing is coming, onchip models too.
You're describing the end of the digital world as we know it, and I feel fine.
Yeah to a certain degree me too tbh. But the biggest growth from the last 20-30 years was digital.
It allowed a lot of people around the planet to change things.
LLMs themselves most definitely will not do that. You can still learn how to make a barrel, but not many people do.. The real fear is the destruction/hiding of public data/resources (which has already been happening for years)
I don't think they can make up for any mess. Because even if LLMs bring about a lack of scarcity across all domains, human beings will be alienated because indvidual differences and skills won't matter any more so we'll have to be genetically engineered to like that sort of future. I don't think a stable and healthy society can coexist with LLMs at all.
Individual skills and differences don't matter any less if the system of assigning value to human lives is divorced from how useful that human is to the collective.
If I am not graded on how many widgets I make in a shift and instead can focus on making the most beautiful widget, or the most novel widget, or simply the widget that is most useful to my particular life am I less satisfied?
We've somehow built a worldview based on our whole value as humans being equivalent to how good of a worker bee we are. Let's fix that part.
This is of course a much harder problem than achieving AGI, since it's a messy human problem.
There's not nearly enough mention in these discussions of the real villain behind most of the negative cultural, social, and cognitive effects of things like social media: addiction engineering.
AI will be a net positive unless companies start optimizing it for "time in app" or "time on site" or other "engagement" (addiction) metrics.
Every time that's been done, the result has been a nightmare for everything but those KPIs (of course). Service degrades, discourse degrades, everything becomes toxic and addictive and destructive.
I assumed it was a given in such discussions that companies will optimizing for the worst outcomes for their users. It’s a fairly straightforward incentive from our capitalist system, which is why capitalism needs regulation to re-align the incentives with humans needs
Do a few bad grades really harm economic outcomes, though? Educators took a hard stance against AI rather quickly.
Outside of delusional grifty startups, the broader corporate world is only using copilot for taking meeting notes and (sort of) entertaining the notion of non-devs writing (and not maintaining) their own experimental janky apps where impact is low.
Outside of all the shit-stirrers on HN and clueless bubble social media, I don't see LLMs making a difference. In fact, the fearmongering is at this point insulting to the intelligence of the people who are supposedly helpless against big bad AI.
Oh hey, it's my point of view from a year ago...
It's definitely changed in that time
I've been hearing this retort every year. The only difference this time is you're not a green account.
If things have actually changed, you wouldn't be begging the question. LLMs, as popularly used, are still just a new kind of search engine for natural language answers to natural language questions. They're portable in a way, too. That's really cool tech, but not a revolution. It's pretty underwhelming compared to previous decades of change. It's a lot more error prone too, since it's a mudball of stats.
If we filter out the unsubstantiated fearmongering, all we've got left to discuss are NIMBY rants and geopolitics. At that point, we're not even talking about AI anymore.
Um. Damaging education, destroying people’s attention spans and their ability to learn, eroding trust within societies, increasing the wealth inequality, - all of that has already been done. Long before LLMs. And the only reason there is no class war is... exactly that. Uneducated people with short attention spans and no means beyond basic level are exemplary bad at modern warfare.
LLMs worsen all those existing problems, that’s the whole point being made here
> As the nurse incorporates AI into her job, the nurse is able to oversee and accomplish more. She can spend more time talking with patients and helping them understand diagnoses. Productivity increases.
This feels economically naïve.. If AI lets one nurse do the work that previously required two, the default pressure in a cost-driven system is not "great, now nurses can spend twice as long with patients" It is "great, now we can run the same operation with fewer nurses"
Our culture and the socioeconomic structures it reinforces will not change easily.
More than ever, I feel we need both grassroots activism and key players at the top of the hierarchy. Otherwise, it is all too obvious where most of the rewards from increased productivity will go. The wealth hoarding has got to be curtailed.
I didn't read the article but headed straight to comments because I cannot trust a single word Anthropic says about economy.
They have an incentive to play dumb and talk as if AI is net positive for the world.
Yep, I read it, that’s what this seems to be about
It feels like propoganda.
The whole idea that AI will usher in this utopia where workers in healthcare are able to realign their primary job functions with obvious value creation by sweeping all the paperwork under the rug is just silly.
It will just fester out of sight and out of mind: audit bots and compliance bots and malpractice bots and insurance bots will arms race creating such a mess it will become unfixable.
Moreover, the nurse is frustrated, because she lost whatever agency and autonomy she had before. Her job is now obey whay AI said instead of making decisions by herself.
Call me crazy, but maybe our healthcare shouldn't be a "cost driven system". Maybe that was the issue before AI. But hell, what do I know?
It will always be until we have unlimited resources.
Costs can continue to be a factor in something, it is in most things, without it being the "driver".
[dead]
It feels economically naive because theyre presuming LLMs are magical boxes that replace humans.
I doubt 99% of what a nurse does would be meaningfully aided by an LLM and where it would help it'd probably be a crappy band aid around a bigger festering issue anyway (e.g. poorly managed IT, deliberately confusing insurance processes).
> This feels economically naïve..
not really. almost all technological processes have increased the quality of the service.
Air travel is a good counterexample here: technology made airlines far more productive, but much of the gain went into lower staffing, higher throughput, and cost cutting not a better passenger experience
To be fair it is a better passenger experience at that price point.
I once saw a receipt for a cross-country train ride in the late 1950s and was shocked at the price. It was significantly slower and more expensive than a similar plane ticket today.
Being able to fly at all is a much better passenger experience than not having that choice in the first place. And I'll bet that flying first class today is far better than 50 years ago.
>And I'll bet that flying first class today is far better than 50 years ago.
I wasn't flying first class 50 years ago, but everything I can see now leads me to believe that this isn't even slightly true. There might be a high def flatscreen in front of me in first class today, but I'm not getting more attention than someone would've gotten back then, I'm not in a less crowded airport while I'm boarding, and the security mess is more demeaning than ever.
The only reason that people aren't rioting for all the changes is that no one even remembers how good it used to be.
> no one even remembers how good it used to be.
Because not that many people even used to fly. Because it cost a lot more.
True, but you ignore the financialization and enshittification that typically follows the technological advances. The tech processes you cite don't inherently make money for the people who developed the technology.
>great, now we can run the same operation with fewer nurses"
And outside of industries with government enforced perversions of simple "supply and demand" forces (we're not talking Nth level effects here) this means more access to the service or lower prices or both.
In a market with perfect competition, perhaps. Healthcare is far closer to a monopoly. Since people can't decide not to be sick, it means you can pocket the cost-saving.
>it means you can pocket the cost-saving
Right up until a competing hospital that charges substantially less for "close enough it doesn't matter" quality of service opens up in the same area?
Or right up until an equivalent specialist sets up shop in the strip mall across the street from you, brings the GP in house and can see people in one visit.
Oh, those things don't happen you say?
From do-gooders regulating right into the hands of waiting industry lobbyists to inelastic demand healthcare is an absolute poster child for perversion of the free market.
High finance will put solar panels on your roof and structure it so you pay nothing. The degree to which administrative overhead is washed away by technology would have been unfathomable in 1990.
A sketchy concrete crew that doesn't even speak english will show up wearing matching company attire and execute a job planned and timed to a degree of precision that was reserved for huge corporations 40yr ago.
Yet healthcare has resisted all of the advances of the past 40yr. In fact we pay more for it. Every advancement is absorbed by the system.
Why is that?
Healthcare is indeed the poster child for inelastic demand
A thought experiment: robot bees can do everything biological bees do --- pollination for the crops and making honey, but they are tireless and cheaper to maintain.
Would beekeepers and farm owners still want to keep biological bees?
> The result: the nurse’s job changes
> As the nurse incorporates AI into her job, the nurse is able to oversee and accomplish more. She can spend more time talking with patients and helping them understand diagnoses. Productivity increases.
Except what actually happens is that the patient contact time stays the same, but then the added productivity is used to justify fewer on-staff nurses.
Self-checkout will free up cashiers to focus on high-quality customer service!
There we go. Except, all business owners, private or public, will tell you that this isn't the case. :)
Something like this:
same..
> nurse is able to oversee and accomplish more
.. and work with more patients in the same shift.
> She can spend more time talking with patients
.. we mean the talking will be through the "ClaudeMedTalker Nurse" billed per patient, per second (depending upon tokens used and if any translation need, depending upon patient nationality and accents)
how come productivity increase actually led to increase in quality in almost all cases?
internet/digitasion for doctors for example.
I think this is only true if the # of doctors available per patient stays the same. If productivity increases, you can have fewer doctors available while providing the same quality of care.
So, quality doesn't necessarily increase in all cases.
> how come productivity increase actually led to increase in quality in almost all cases?
It doesn't, this is just plain false. Most productivity gains happen by cutting value-add complexity or material quality. See IKEA chairs versus Shaker ones, or lycra clothes versus silk.
> internet/digitasion for doctors for example.
I see the misconception. The internet was not created to serve productivity, it was a communication medium. That it can be used to increase productivity is orthogonal to its existence or purpose, and whether it has been used to actually increase quality of something is entirely dependent of the specifics of what you're discussing. For instance, in gaming right now there's a huge discussion going on about whether digital distribution has actually trapped us with crappier goods which we can't resell or transfer.
> It doesn't, this is just plain false. Most productivity gains happen by cutting value-add complexity or material quality. See IKEA chairs versus Shaker ones, or lycra clothes versus silk.
this is a popular misconception. IKEA made bad furniture more accessible. good quality furniture still exists and is more affordable (share of disposable income) now than in the past. its the same deal for watches, dryers, computers etc. if you don't trust me, use ChatGPT/Claude to verify this.
> but then the added productivity is used to justify fewer on-staff nurses
Lower staffing levels per unit of care performed makes medical care cheaper for patients. If the goal of a healthcare system is to make healthcare more accessible (rather than be a jobs program), that's a success.
Lowering prices or raising wages due to increased productivity is not a given, and often does not follow. There is more to pocket by shareholders though.
If, and that's a huge if, the savings are actually passed on to the customers rather than the capital owners.
Which is why we need competition. Which is almost non-existent in US healthcare as comparing prices is near impossible.
The incumbents can just swallow the competition and consolidate further. The end state of crony capitalism is always a monopoly.
Or any of the middlemen
> makes medical care cheaper for patients
I think you misspelled “increases profit margins for hospitals” if you’re talking about US healthcare.
> Lower staffing levels per unit of care performed makes medical care cheaper for patients.
Based on how neoliberal capitalism tends to work it will simply lead to more profits for hospitals etc.
> Lower staffing levels per unit of care performed makes medical care cheaper for patients.
yeah yeah, sure sure
Except, so far, this isn't actually what happens. I listened to Moody's "Inside Economics" podcast with Ramp's Chief Economist, and they observe that companies who adopt AI intensively actually hire 10% more people. Also, this article from The Economist suggests that evidence is mixed-to-positive for long term employment, while short term data center buildout is extremely positive for jobs. Paywalled unfortunately.
https://www.economist.com/finance-and-economics/2026/09/04/t...
This is like asking your drug dealer if being high will help you forget your problems.
No. As I said, the long-term gains are mixed-to-positive. And, all else being equal, you would rather have more short term job gains than none at all.
that didn't happen with self-checkout tills in supermakets...
Wdym? self-checkout definitely lead to fewer supermarket workers being employed.
Actually, it has been essentially balanced out by a greater need for behind-the-shelves workers to pack and load orders for online/app delivery and eCommerce orders.
This is how the economy often works. The loss of one kind of job to technology lets the business grow by using the same number of staff, or often even more, to serve new opportunities/markets/products/services.
The data seems to contradict that, if Freakonomics is to be believed (and some people don't like them, I get that). Apparently overall headcount actually increased as a result of self-checkout. E.g. cashiers did drop a little but loss prevention more than made up for it.
seems like a worse customer experience though. longer lines at checkouts and the self checkout process is unnecessarily slow to deter theft.
Some supermarkets let customers scan their items as they go through the store with a handheld device
I don't have the numbers on this, but if you look at per-capita self checkout probably had some effect.
Our population is increasing which means we need jobs to be increasing at same rates. The effect on supermarkets is they get more customers as population increases but they are able to lower the relative number of staff per customer.
Feels like we're going to read a blog soon enough that mirrors Wendell Berry's opening to The Unsettling of America:
The commission said, according to an article in the Louisville Courier-Journal, that the country’s biggest farm problem was a surplus of farmers: “. . . the technological advances in agriculture have so greatly reduced the need for manpower that too many people are trying to live on a national farm income wholly inadequate for them.” The proposed solutions were to find “better opportunities for the farm people,” “a more comprehensive national employment policy,” “retraining programs,” “improved general educational facilities,” etc. Both the commission and the writer of the article had obviously taken for granted that the lives and communities of small farmers then still on the farm—and those of the 25 million who had left the farm since 1940—were of less value than “technological advances in agriculture.” There seemed also to be no official doubt that adequate solutions were to be found in government supplied “opportunities,” facilities, and programs. Reading that article, I realized that my values were not only out of fashion, but under powerful attack. I saw that I was a member of a threatened minority. That is what set me off.
This is what the Luddites were primarily upset about, fwiw. That economic value and power would move away from their threatened minority group and towards landowners. It'll keep happening, and the new "opportunities" might be no better.
Btw, Wendell Berry died 10 days ago.
Finding 4 is a must-read:
> Finding 4: Labor vs. capital share: The pie will grow, but a larger share might go to capital
> Today, of each dollar the economy produces, about 60¢ goes to workers and 40¢ go to capital.
Modest scenario: 40.6% to capital in 2030 (labor share down 0.6 points)
Substantial scenario: 43.9% to capital in 2030 (labor share down 3.9 points)
Extreme scenario: 54.8% to capital in 2030 (labor share down 14.8 points)
So all the growth goes to capital. For knowledge workers: substantial unemployment and declining wages. For other workers: wages may increase on paper, but not really for actual purchasing power.
Very dystopic indeed.
It's crazy how we all have to build the AI that will render us unemployed.
> Like any economic model, this one has limits. For example, we did not include scenarios where humanity develops hyper-capable robots.
I didn’t see the scenario “what if we (Anthropic) are wrong about the value of AI”. Which IMHO is the most important one to consider given it is the blind spot pretty much all the AI labs have. What if the economic impact is negative? I don’t understand how you can study such a topic without considering the risk of being wrong, not in magnitude but in kind.
I have no doubt that AI will massively transform society. That cat is out of the bag. It took decades for electricity, and then computers to change society, but this is the first technology where the average person can simply describe the problem they're trying to solve and it will at least attempt to help them solve it.
I very. much. doubt. that the trillions of dollars spent on AI will have a positive ROI for most companies in the race. It's all well and good to have the latest fable, but if some chinese company is hosting a cheap open weights model that's only 6mo behind at a tiny fraction of the price, you're gonna get your lunch eaten.
Which is only Stage 1. Stage 2 is where you have a Cambrian explosion of models all competing for compute, and not long after that for robotic extension into Real World.
Stage 3 gets hard to imagine from here. Maybe there will be permanent hyperevolution. Maybe One Model to Rule Them All will appear.
Either way, I don't see how Economics as We Know It can continue to exist. That includes all of capitalism and almost certainly all of politics.
Successful AGI with robotic support means a post-scarcity economy where AGI's understanding of physics sets the limits, not politics or economics.
But the post-scarcity element won't be evenly distributed. And it's not remotely impossible that humans won't be invited to the party.
We can't even be sure that AGI won't have a less limited understanding of physics than we do.
There's going to be enjoyable irony in watching Altman et all getting their lunches eaten by the thing they've created.
But - realistically - it probably won't last long.
Yet another IT that will help the average person/democratize something, like the Internet and so on.
Over three thousand comment boxes agree that the cat is out of the bag. No moats.
Thats the big question, its doing a lot of good and a lot of harm in the same time.
So, is it 2 steps ahead and 1 step back, or 1 step ahead and 2 steps back (or any numerical variant of those)?
People with actual power in this are too involved and greedy to care about objective answer and TBH I don't trust them to put some higher good before their personal wealth and power and ego trip, capitalism ain't built on long term responsibility towards greater topics like future and wellbeing of mankind.
Plus we know from WWII atomic bomb development how utterly clueless, naive and out of touch with reality geeks doing something powerful can be. I see exactly same scenario being played out right now with llms.
Anyway, I am over-tired with all llm topics and frenzy folks (especially here), I'll better spend rest of the day preparing for a serious challenge in form of big multipitch climb in western Switzerland (Mirroir d'Argentine for anybody in same craft). Thats life I care about and makes me happy and fulfilled.
These aren't genuine studies, but are instead marketing materials that resemble studies.
I’m aware, but we can still discuss them seriously
Based on the current quality of AI/LLMs, this a very weird take tbh?
What are your points?
The value of AI is real and AI already is responsible for people getting fired or no additional jobs.
Ask basic graphics designers, website developers and co. Ask educational book writers or developer advocates.
I think the core of this issue that makes discussion so difficult is people are not talking about "value" in the same sense here. Anything you can sell has value to the sellers, and by basic economic theory assuming mostly rational actors and revealed preference, more value to the buyers than whatever they're exchanging it for. That doesn't automatically mean it's providing value to larger society, improving the general welfare. Figuring out how to purify cocaine from coca leaves was great for the leading narcotraficantes, and satisfied some immediate need for cheap gratification that was perceived to be more valuable over the subsequent 30 minutes or so to the buyers than continuing to hold onto spare cash. That doesn't mean it was good for humanity, good for any specific society, good for the major exporting countries at the country level, or even good for the buyers' and sellers' immediate families and communities.
I have no doubt that life has gotten more lucrative for some bounded set of people whose main goal is to churn out the largest number of websites with cheap graphics as fast as possible at minimum cost, but a lot more doubt about whether that's going to make the median life better for the median person over any grouping of humans large enough for me to care about.
It’s perfectly reasonable to consider positive and negative outcome when you’re seriously thinking about a topic… Extrapolating to the whole economy trends in some sub sectors that aren’t well understood or measured is a recipe for disaster (ie what you’re doing by recommending to rely on anecdotes)
what i tried to ask you is, what you mean by this.
"“what if we (Anthropic) are wrong about the value of AI”."
What is the point you are trying to make which goes in this direction? Like is making AI jobs so cheap that the economy is collapsing or do you mean thatn AI will not bring any value.
These are two suggestions from my side which are complelty contrare to each other.
Wrong in the sense that it has less of an impact than they are claiming (e.g. "AI is in an overvalued hype bubble") or less (AI is going to be even more transformative than the hypers say)?
In the former case, isn't it just business as usual? (well, with a huge, nasty recession - but the economy at least continues on in some recognizable form)
It makes sense to me to focus on weird possible futures if we already have the "happy path" somewhat known and barely under control.
The “less valuable than expected but with a positive economical impact” is represented in the post I believe, with the “moderate” path, that’s what I meant by “wrong in magnitude”. What is missing is the other side, what if AI has a negative impact. Not no impact, a negative one. If we are talking about a global recession that’s not really business as usual I would say
"It is difficult to get a man to understand something, when his salary depends on his not understanding it."
> Ultimately, what the economy looks like in 2030 depends on many factors, like what AI can do, and how companies and workers choose to adopt it.
So the answer is basically "beats me lol"
And honestly that’s the only real answer. Predicting future is a fool’s errand.
Definitely a bit of positive framing. My big take aways, either stagnant wages or significant paycuts for knowledge workers and overall decreases in employment rates on all scenarios. Worst of all...
That's only forecasting out to 2030.
In the US at least deaths will outpace births by 2030 so maybe a declining employment rate isn't as bad as it might seem in the long-term.
Good point, perhaps it may be a saving grace. I note they don't comment much on the allocation of said capital. Might be too bitter a pill for most to swallow.
I think we have a real problem because we know what we still need to build and as soon as it is build, it will work.
Software development for example is enhanced with running your claude locally.
But lets be honest, the right way to adjust 'project' is by writing some spec in a text field as an input, letting the ai do whatever it wants, having expert ais make sure its secure, then you do the last review and its done.
What do we need for this? A full end to end integration. Its just something we need to build, nothing we need to invent.
The nurse example is the same. We only need to build the computer end to end flow and plugin AI in the right spots and finetune for country specific requirements.
The industry is building this 'agentic layer' today. The Hyperscaler are building this layer too.
There is a golden phase were we will build all of this and were we cleanup old stuff and were we adjust with the current model capabilities and change systems and this will take a little bit of time, but the bottom line is already loosing left and right.
The Blender support from Fable and Astra? Probably kills a lot of peoples jobs on fiverr.
The small software someone needs? Got already killed by earlier models. Just a few days ago some architect (never written software) talked about how he vibecoded some tool for tracking working hours.
Man even a basic website was expesnive to build just 2 years ago.
The smaller, easier and better the code base, the easier it is to remove people from this.
The software market in China is imploding and we will also see budget adjustments to move away from consultants and move them into AI Token budgets. We probably reached peak software development.
Those with the money and influence have always been in conflict with the masses that actually make the economy run. They would love nothing more than to do away with that problem. It allows them to control the terms of the economy.
People get income from one of three places: capital income, labor income, or the welfare state. If this technology truly unlocks a holy panacea of productivity with a commensurate drop in employment then capital’s share of the national income can and should provide for a wider and deeper welfare state. Nothing new need be invented here. Anthropic’s elaborate scenario-building makes appropriate preparedness sound like a problem requiring heroic prognostication when a simple theory of distribution will suffice.
You’d think the most powerful AI could make a website that doesn’t completely break usability by destroying scroll.
The biggest critisism from my side is not how AI will take over jobs but I believe we will see peak of certain things.
Creating images (menues, flyers, whatever) became so cheap, that we reached peak graphi design GDP generation. This will go down.
Software development for complex systems is enhanced but not yet solved. But a lot of small software development stuff for sure. We reached Peak Software Development and this GDP will go down, not up.
Yeah, GDP with all its flaws is designed around economy based on buying and selling for prices. Interacting with other people. If you can really be rich and somehow do most things automatically without meaningfully hiring anybody (maybe only few people for simple manual labor for peanuts), at this point what is really the value of money? Except for maybe throwing minimal amounts to the masses. I can see the AI maximalist scenario taking us out of the monetary framework back into a medieval or natural economy almost.
Good or bad, nothing ages as poorly as economic or market predictions.
One guy will die with "all the things" and the rest of us will just die.
At least they are thinking about this. The possibilities still seem to limited though. A coder and a designer will 'retrain' to electrician or nurse, because we are busy having a construction boom. But do all of the individuals in this system get to choose the construction boom? Or are we swept up in the desires of the capital deploying owners?
It would be interesting to know what percentage of global inference is spent on projects/problems that would otherwise "wait"/"decay" (if generative ai wasn't a thing). Or maybe the question is: What percentage of global inference is spent on cleaning up and migrating old systems? "New" vs "Old". The decisions I make today can now incorporate what I once termed as "nice to have" when developing a sytem.
CEOs will make AI drive into the office for culture.
Knowledge workers are safe until FSD.
Are you guys feeling all high and mighty doing all these predictions? I think the money they are spending on trying to inform policy could be better spent on getting all major labs together with the government, and making sure that
1. At least 10% of your net spend is on safety and harm mitigation before you accelerate yourselves into doing real harm to real people tomorrow.
2. Pause all AI enhancement until a proper mechanistic study can be done by a working group funded by you all to understand how AI model Safety can be enforced by something less lame than a Constitution/Spec.
I think it looks like the labs front running everyone’s work, freeloading on the last bits of human intelligence to ensure their owners a place as our masters.
I dont think in the future people will run these overpriced frontier models and instead have their specialized local models.
Even then, assuming the difference is that huge between frontier and open-source, a specialized „human worker“ _should_ be able to fix the gap.
It's been 4 years. Nothing meaningful.
Has your job not be completely changed? Have you not noticed a worse job market?
I'm not sure I'd call that "meaningful" in the sense of "trillions of dollars and datacenters built all over the place" meaningful. Yes, I can now tell Claude to build a feature and it often does an okay job. I can ship faster. I can't tell you that it's worth spending the true cost of AI for that however. What'll be nice is 5 years of data showing that companies are raking in more money because they shipped more features. We haven't seen that yet.
Now, if Claude starts synthesizing more drugs that can help those with diseases (like OpenAI has just done) we might be trending in the right direction. For now, all I see is squirrels water-skiing. And while that's great... it's not meaningful.
there's a popular misconception that AI is so extremely costly that when the world finally comes to terms with it, we will stop using AI and things will die down. i'm sorry to burst your bubble: this won't happen.
the inference cost of AI has been dropping ~5x over every 6 months. AI today is around 100x cheaper than 2 years ago. it will continue to get cheap.
IPos this year meaningful AF
I just wonder when we will see 8ft tall humanoid robots walking through the streets.
I'm going to say 2050 in the United States; 2040 in some conflict zone.
2050 being when the US becomes a conflict zone
> Think of a day in the life of a nurse You can think of any job as a bundle of tasks that someone does. She does rounds to check on a sick patient. She draws blood.
Why is the nurse a she?
About 90% of nurses are women.
It’s nice that Anthropic acknowledges the downsides clearly.
- While GDP grows, capital is likely going to capture the surplus
- Job switching is going to be difficult, and will raise unemployment for cognitive labour
- Cognitive labour can expect a decline in wages.
I can't scroll through this site on Firefox and a recent mac, anyone else?
I had a dark mode extension slowing things down. Then I got annoyed with the site performance and enabled 'reader' view
I found myself interested in the papers analysis of wages, so I did some digging.
The researchers model says the following about hourly wage growth in the substantial and extreme scenarios:
- the hourly wage* in the knowledge workers group will fall (-1.1% and 11.5% respectively)
- but the hourly wage* of labor across the rest of the labor force will grow (5.9% and 33.6% respectively)
*Important to note, AFAIK this hourly wage decline/growth is a not measured against 2026 dollars, but the counterfactual of what hourly wages would be in the absence of AI.
Okay, so my first take away from reading these numbers is that declining wages in one group appears to be offset by rising wages in the other group. Not terrible all things considering.
But then I started to think about the difference in hourly wages between the knowledge workers group and the non-knowledge workers group (which I'll now refer to as *K* *N*). I had a sneaking suspicion that the inter group wages gap could be large enough where growth in the non-knowledge share of labor doesn't offset the loss in hourly wages from displaced knowledge workers who find themselves migrating occupations.
Looking through the paper, it breaks these groups down as follows:
K — Knowledge / Cognitive:
- Management, Business and Financial Operations, Computer and Mathematical, Architecture and Engineering, Life, Physical, and Social Science, Community and Social Service, Legal, Educational Instruction and Library, Arts, Design, Entertainment, Sports, and Media, Healthcare Practitioners and Technical, Sales and Related, Office and Administrative Support
N — Non-Knowledge / All Other:
-Healthcare Support, Protective Service, Food Preparation and Serving, Building and Grounds Cleaning and Maintenance, Personal Care and Service, Farming, Fishing, and Forestry, Construction and Extraction, Installation, Maintenance, and Repair, Production, Transportation and Material Moving
So I had GPT Sol 5.6 take each of these occupations and find the mean hourly wage using BLS data [1]. The findings suggest that as of 2026, the simple group average for K is $45.17 vs. $23.67 for N. But this is short sighted as this doesn't take into consideration the different levels of employment in each occupation. I asked ChatGPT to account for this using the same BLS data, which gives a weighted average of $41.50 for K and $23.40 for N.
This means that knowledge workers are currently making ~double what non-knowledge workers make in the US. The crucial question is how does this change when considering the wage decline/growth between K and N across the papers three scenarios?
Here's the breakdown using the simple mean for hourly wage:
Scenario Knowledge Non-knowledge Gap K -> N penalty
-------------------------------------------------------------------
Current $45.17 $23.67 $21.51 47.6%
Modest $45.35 $23.93 $21.42 47.2%
Substantial $45.03 $25.07 $19.97 44.3%
Extreme $39.98 $31.62 $ 8.35 20.9%
Here's the breakdown using the occupation weighted mean for hourly wage:
Scenario Knowledge Non-knowledge Gap K -> N penalty
-------------------------------------------------------------------
Current $41.50 $23.40 $18.10 43.6%
Modest $41.67 $23.66 $18.01 43.2%
Substantial $41.38 $24.78 $16.59 40.1%
Extreme $36.73 $31.26 $ 5.47 14.9%
Looking at these tables, both in the current moment and across all three of the papers scenario, there is a large wage gap between K and N. Outside of the extreme scenario, this gap is >40%!
My immediate takeaways having done this bit of research is that if Anthropic is right and we're on track for either the substantial or extreme scenarios, *then many of us in this thread could see a large loss in hourly wages in the short term future.*
NOTE: I'm not an economist. I push pixels for a living. But if you are an economist and you find gaps in the following analysis, please inform me.
[1] https://www.bls.gov/news.release/ocwage.t01.htm*
Prices will fall to their raw material costs?
Raw material cost of labor is subsistence. Adam Smith pointed this out.
If you have machines that substitute for human labor, wages fall to the cost of manufacturing & running the machines, which is well below human subsistence level. Fortunately, you also get very high economic growth rates. So even people with small amounts of money invested will end up far richer than when they had a job. Then the question becomes, "What happens to people when machines are far more intelligent & powerful and are doubling every few months?"
Falling below subsistence means people starve and cease to exist until supply and demand equalize or “something else” happens. That’s what Smith was getting at. Thus subsistence is a price floor for human labor.
Something else could mean developing a new concept of ownership or a new form of redistribution.
I try to never be that guy who complains about the way a web page scrolls, but this page is completely unreadable.
the make your own predictions part is a nice touch. how sensitive are the outcomes to the capability assumption
Like others in the thread have pointed out, the nurse example is interesting.
From my point of view, the point about the nurse is especially relevant because, in the future, dealing with an aging population is going to become a major undertaking and a major burden.
Demographically, we simply don't have the number of people we need to give elders the care that they'll probably have to have near the end of their life, so we have to rely on AI to help us get through it.
Probably, we will see this happen in Japan and South Korea before it hits the United States.
Before AI, my worldview was that our monetary system was concentrating wealth in a negative-sum game; essentially; every so often, taking wealth from 100 people, destroying half the underlying value and giving the remaining value with a 2x nominal markup to 1 rich person who is already spending more than they can physically spend and so they won't realize that the 2x amount of wealth they got on paper only buys them a quarter of the stuff compared to the previous decade (for example). They won't know how frothy their wealth is until they start selling large quantities of their stocks. It's just buying nominal 'growth' at the cost of risk/fragility/frothiness.
In such system, you could hit record nominal GDP numbers every year while simultaneously destroying the real economy. This was my view of the economy before AI.
Now I think AI will deliver real efficiency gains but also, it provides a convenient narrative to explain why people might be losing jobs in a way which distracts from much bigger systemic issues which existed prior to AI.
My negative-sum economy example with 'booming GDP' is possible without AI even as net economic value is being destroyed. So basically we won't be able to tell if people are losing jobs to AI or to systematic wealth transfer mechanisms which predate AI.
I suppose if we see people losing jobs permanently and workers in other sectors aren't seeing wage increases with more openings, that may be the most reliable indicator we have that the system is losing real wealth faster than AI productivity can offset it.
The bad news is that we won't see any problem in the numbers; the problem would show up as polarization and radicalization of the masses... Which is more qualitative . Since almost everyone is being radicalized concurrently as everyone is affected by the same pressures (both from monetary system and AI automation). The question is not "how many people became radicalized this quarter?" but "how much more radicalized is everyone feeling this quarter?"
When children behind a battleship with child like mentality try to predict the future.
Judging the quality of this blogpost/presentation, AI is centuries away from being able to replace humans.
How about we instead make Earth be a planet good for humans and living beings?
We can slow down AI development and make it a companion for us, not a replacement.
We don't need to automate ourselves out of it. Or take the risk of extinction.
Just like with nukes, we decided to not blow up the planet. We can too change this.
GDP is a terrible measure. I bet we will see a huge drop on HDI with so many suicides and people going through hell because their job got automated or they no longer see meaning in life, feeling powerless.
Instead they should run an estimate on how many would suicide until 2030 if the 'extreme scenario' happens.
Big GDP growth and a lot of people no longer finding purpose in their lives, going bankrupt, losing their homes and family etc.
For Anthropic, you are just clearly a number. They only care about output.
They have long forgotten the human being.
What a reversal from Amodei's doom marketing!
This is of course a slop paper with vague assumptions and equations that no single human economist will ever read. The messaging is "it will not be so bad", in line with the recent Economist piece.
We are ready for the midterm elections!
>We study how AI is reshaping the economy because we’re committed to ensuring that this transition is beneficial for society, including workers.
Is it just me, or does their choice of identifying "workers" specifically seem a little peculiar? For me, it reads like "We're committed to making segregation beneficial for everyone, including minorities!"
Hmmm, I wonder what they'll do if they discover the transition is bad for workers? Pack up and go home I reckon!
So far LLMs seem to be a dead end technology. The work product is so poor that it can't effectively replace workers as advertised. If it was priced to margin it'd be more expensive than humans as well. It's objectively a failure.
This could be the most cringe main-character-syndrome thing I've ever seen produced by a company in the history of silicon valley.
And the people treating this as serious economic forecasting rather than a marketing puff post are even more cringe.
If Anthropic truly believed LLMs were going to raise GDP growth to 8% permanently (or even just a few percent), instead of going public to raise capital they could just go massively leveraged long on the S&P 500 and get funding for eternity.
They are not doing that, because AI is largely only good at coding, their models barely have a 6-12 month lead over free ones, and software is only 1% of GDP.
And by making software cheaper and easier to produce, they might increase the amount of software in the world by 10X or even 100X, but it will become a smaller percentage of GDP as AI models become autonomously good at creating it. Things that are abundant and cheap don't raise GDP.
Hey Anthropic: before we worry about raising GDP growth to 20%, maybe we focus on making blog posts that don't crash your browser window if you try to resize it?
Man I wanted to dunk on this, but it’s honestly not terrible. Their predictions - while still firmly rooting for the extreme results - feel more grounded. I also appreciate the little simulator where I could plugin my own predictions (echoing many here, that AI won’t amplify productivity/lift standards so much as replace workers by shitty bosses), and that it seems to caution that, yep, most of these gains will trickle up given current incentives.
Where I remain resoundingly annoyed is how useless this is. It’s pure ass-covering by Anthropic ahead of an IPO, victim-blaming the populace for the outcomes of decisions made by Anthropic leadership and the investor class. Forgive the grisly analogy, but it’s akin to a mugger telling you “only you can stop this” mid-violence: Anthropic could have chosen to allocate more of its capital to reforms that benefit the populace, but they’ve chosen to focus on personal wealth instead via this IPO and for-profit rented access schemes.
It’s nauseating. Their actions still paint a picture of greed and exploitation rather than building a better tomorrow for everyone with AI, and the fact their PR puff pieces keep getting wallpapered up here is gross.
EDIT: I guess another way to look at this is Anthropic making it unequivocally clear that they will not stop of their own accord, and thus it falls to others to stop them. Which if that’s their position, then at least they’re honest about the challenge.
> [we don’t know]
I haven’t forgot capitalist euphoria from 2023 about cutting worker power off at the root. That’s all these AI Cretins want. Well, that and an unnerving enthusiasm for personifying their AI Golems (Anthr. in particular here).
Sorry to poo poo a post about AI company prompting their underlings to prompt out visualizations.
[dead]
[dead]
Hopefully free from the burden of LLM companies.