Anthropic's Index Shows Job Evolution Over Replacement 19
Anthropic's fourth installment of its Economic Index, drawing on an anonymized sample of two million Claude conversations from November 2025, finds that AI is changing how people work rather than whether they work at all. The study tracked usage across the company's consumer-facing Claude.ai platform and its API, categorizing interactions as either automation (where AI completes tasks entirely) or augmentation (where humans and AI collaborate). The split came out to 52% augmentation and 45% automation on Claude.ai, a slight shift from January 2025 when augmentation led 55% to 41%.
The share of jobs using AI for at least a quarter of their tasks has risen from 36% in January to 49% across pooled data from multiple reports. Anthropic's researchers also found that AI delivers its largest productivity gains on complex work requiring college-level education, speeding up those tasks by a factor of 12 compared to 9 for high-school-level work.
Claude completes college-degree tasks successfully 66% of the time versus 70% for simpler work. Computer and mathematical tasks continue to dominate usage, accounting for roughly a third of Claude.ai conversations and nearly half of API traffic.
The share of jobs using AI for at least a quarter of their tasks has risen from 36% in January to 49% across pooled data from multiple reports. Anthropic's researchers also found that AI delivers its largest productivity gains on complex work requiring college-level education, speeding up those tasks by a factor of 12 compared to 9 for high-school-level work.
Claude completes college-degree tasks successfully 66% of the time versus 70% for simpler work. Computer and mathematical tasks continue to dominate usage, accounting for roughly a third of Claude.ai conversations and nearly half of API traffic.
It's not generosity (Score:2)
Re: (Score:2)
If people get more efficient using AI, then people will be fired.
Greed is blinded by greed. Doesn’t matter how smart they are or were when that shit takes over.
The pathetic Catch-22 being presented to graduates today, is you better be able to prove you can augment your job with AI. Otherwise, you’re not “future-proof” enough for Greed to pretend it will hire AI augmentees just long enough for AGI to come along.
Then, every motherfucker who ain’t a cyborg is getting fired. Because of annoying meatsack needs like sleep, food, and medical ins
Re: (Score:2)
The pathetic Catch-22 being presented to graduates today, is you better be able to prove you can augment your job with AI. Otherwise, you’re not “future-proof” enough for Greed to pretend it will hire AI augmentees just long enough for AGI to come along.
Then, every motherfucker who ain’t a cyborg is getting fired. ...
This assumes that AGI will come along, and soon. IMO, it's more likely that LLM's continue to augment the way work is being done, rather than replacing real jobs (which seems to be what the article is saying as well).
Case in point - TFS: "Claude completes college-degree tasks successfully 66% of the time versus 70% for simpler work."
Barring true AGI, in a future where LLM's are much better at predicting the right completion, they'll still need hand holding. This is certainly a significant change, but so was
Re: (Score:2)
Say you don't get "true AGI" (which by my definition humans don't have), but only twice the efficiency and scope of current AIs. (I'm NOT limiting this to LLMs, which are a subset of AIs.)
Then they will probably be able to to 75% more of the work, so, after job restructuring, you'll need 75% fewer people. This *will* make you more efficient (see Jevon's paradox) so more jobs will become available, but I really doubt that 3 times as many jobs as currently exist will be created. (Yeah, that's a lousy way to
Re: (Score:2)
Want a good example? We went through one already - industrialization.
I can't compare your numbers, cause they don't make sense ("able to (do) 75% more of the work"... is a comparison to the current amount of the work they can do, which may sum pretty closely to zero). But let's say it IS able to do 10% of "the work", and let's assume it will be able to do 75% more of the work, it would then be doing 17.5% of "the work". 17.5% of the work does not account for 75% of the people.
So say half as many people will be employed...or people will be employed half as much of the time. 20 hours/week sounds like an ideal solution, but not one we're likely to get to without a lot of social unrest. And different jobs will be automated/restructured at different times, so a legislated work week isn't a plausible answer.
Regardless, if it is actually a
Re: (Score:2)
Where was all the whining when autocomplete and predictive text rolled out?
Autocomplete and predictive text didn't ax entire departments in multiple industries over the course of 2-5 years, jackass.
The only people who sidestepped that were the ones closest to the development, knew when to jump ship, and knew what jobs would take awhile to be replaced. Those guys got their degrees and jumped into the labor market for their new jobs just as the worst hit.
Everyone else got caught with their pants down, and are currently scrambling to find something to do that LLMs won't suddenly
Re: (Score:2)
Where was all the whining when autocomplete and predictive text rolled out?
Autocomplete and predictive text didn't ax entire departments in multiple industries over the course of 2-5 years, jackass.
Industrialization did though. Also, I'm not convinced LLM's have actually been responsible for that net loss - there is SO MUCH upheaval from so many directions these days.
I feel for college grads. Chasing the next job market has always been such a shit show. If a prediction is right and followed, too many people go into it and flood that market. If the prediction fails but people followed it, loads of people end up with "useless" degrees. I went in for Fine Art, so I knew I had zero market from the get go,
Re: (Score:2)
If people get more efficient using AI, then people will be fired.
If people get more efficient at doing work, then people will be expected to do more work. For the same pay. This has always been the way of work.
If management doesn't have a backlog of projects for you to work on, they are incompetent. That is a sign that the company is going under. Time to flee the sinking ship!
Good E. Nuff. (Score:3)
When the threat of AI/AGI first came on, most arrogant meatsacks assumed it would take Perfect AI to replace them at their job.
Most fallible humans don’t realize it will only take Good E. Nuff AI to replace them. Because that is exactly how they landed their human job in the first place. You were good enough to hire. Nowhere near perfect.
The need for UBI will happen far sooner than Greed N. Corruption realizes. That doesn’t put Greed in charge. That puts Greed on the fucking menu. With Mass Chaos in charge.
Re: (Score:3)
It did not hurt to have a new circus in charge of destabilizing the economy to the point most companies would cut back at the same time, to greater lengths than before.
Good E Nuff was also Real E Cheap too.
Re: Good E. Nuff. (Score:2)
You'll have to forgive him: He's been afraid of robots ever since he first saw Chuck E Cheese.
We investigated ourselves and found.. (Score:2)
Re: (Score:2)
I don't think that's quite accurate...but it's sure in the right ballpark. IIUC Anthropic tries to be "the good guys" and tries to be honest. But they *do* have a highly biased viewpoint.
i don't understand (Score:1)
Re:i don't understand (Score:4, Insightful)
Re: (Score:2)
Note that their study was based on those using their chatbot. They've got a highly biased data sample.
that's what I expected and what I observe (Score:1)
I see people, who are not programmers at all, creating automation solutions for their businesses:
https://slashdot.org/comments.... [slashdot.org]
People are using AI to help them create content that otherwise wouldn't have been created ever https://youtube.com/shorts/un8... [youtube.com] just because this is in someone's fantasy does not mean they have resources to make it real until AI stepped in.
We generate and put edit videos for advertising that otherwise wouldn't even be possible, because it would have required a movie crew and I
Re: (Score:2)
We generate and put edit videos for advertising that otherwise wouldn't even be possible, because it would have required a movie crew and I cannot afford that but I do hire people to make these new types of video ads, so some jobs are created due to AI.
At far lower pay and employing far fewer people.
Fun fact: What's important is whether or not the majority of your population can support themselves and have their basic needs met. If that's not the case, it doesn't matter how "efficient" your businesses are, because they are putting themselves in front of the firing squad.
Re: (Score:2)
Which part of "advertising that otherwise wouldn't be possible, because it would have required a movie crew" is unclear? I never hired a movie crew and I would never hire a movie crew, so there wouldn't be any advertising shot with a movie crew, because I wouldn't be able to pay for it.
I can, however, pay for a couple of guys making videos based on my requirements by using tools that were previously unavailable and thus costing one thousand of what it would cost with movie crews and traditional filming tec