I found this interesting: "Still, despite all the hype about how AI coding tools will replace software engineers, software engineering is still one of the most secure jobs you can have today, relative to most other white-collar jobs."
Completely absent is quantity of jobs. If e.g. ml engineer job postings go +40% from 200 to 280 and writer job postings go -50% (over two years) from 20000 to 10000, then we have a better idea of the impact.
Without those data this report isn’t really quantifying impact on “180M jobs.”
> First, let’s establish our benchmark: job postings dropped 8% in 2025 compared to the same period in 2024. Indeed reported a 7.3% year-over-year decline for US jobs recently, so this was a good sanity check, and told me that my data most likely was comprehensive.
Doesn't this also suggest that the job market is in such an unusual state, that any further analysis makes little or no sense?
I've been getting lots of value out of AI coding tools; especially in the last few months. My assessment is that this will lead to more work to do, not less. It's not a zero sum game. More of what I do is baby sitting AIs. But together we do more than me with my team before. It will reflect in my hiring decisions as well. I want people that can do this responsibly (are able to tell good from bad code, are able to think for themselves, are able to get stuff done).
My observation so far is that micromanaging AIs still sucks
I wonder what the numbers look like if you compare pre-pandemic to now.
To me it feels like during the pandemic companies found out they can fire a lot of people without a lot of backlash as long as they have a weak excuse.
I feel like AI taking over jobs is such an excuse, especially considering creative jobs and the layers that are supposedly being replaced by AI in the article they call it "Creative roles that “execute” ".
Comparing US job openings from 2024 to 2025 seems to be selective. Is the job in sudden decline? Or a reset to a norm? Or is the decline something that has been happening over even a longer timeframe?
Taking two years and drawing conclusions from those two years seems to miss larger pictures - especially when the mess of Covid and the pandemic years and that job market is mixed in with this. Yea, that's half a decade back but companies are still trying to figure out how much staff they need where.
Photography? Yes, that's likely been impacted. Compare https://web.archive.org/web/20230124085038/https://www.bls.g... and https://www.bls.gov/ooh/media-and-communication/photographer... and the rate of growth has changed. However I'll also draw attention to the estimates that there's only a few thousand open positions with that classification each year. This includes self employed stock photographers and artists - corporate photographers have been a hard thing to get for a much longer timeframe (I looked into it a little bit back in '09). Additionally, artistic photography is impacted by the amount of money that regular people are willing to spend at art festivals and the like - that's gone down irrespective of AI.
I agree with this. I'm a product designer, and apparently there's "only" been a decrease in job postings of 2.6%. That doesn't seem bad, however in late 2022 through all of 2023 people were getting canned left and right and the market was incredibly tough. Can't really have big decreases when you're near the bottom of the hole.
I perform a different analysis, focused more on the type of blue collar jobs that most people believe are the prime candidates for AI/robotic job replacement:
You switch from global automation to US car manufacturing in the first paragraph without a even a line break, basically equating them.
The first graph you show is missing the corresponding table from the source which shows the exact opposite of what you have stated: the workforce had been steadily increasing until ~2019 and is in decline since then. Not massively but still it contradicts what you stated. Next you switch to Korea and the wage increase without aligning a corresponding price increase or adjust for inflation. Then you switch to the history on the industrial revolution. At that point I gave up because I know the point you are trying to make and I can see, you adjusted your samples to match that point.
There is no causality analysis here, and really no justification for why AI is the reason for these job losses. An alternative explanation is the tariffs.
Same is true for sustainability related jobs. Given the backlash first from Texas, Florida, and other red states and now from the Trump administration, it is not at all surprising that these jobs are in decline. AI has nothing to do with it.
The decline in frontend engineering jobs matches my personal impression from the past year. Smaller companies can do a lot with vibe coding alone, while larger companies can multiply their FE team productivity without making any new hires.
Frontend code can be very repetitive / labour intensive, I bet that has made this more attainable than for other layers in the stack. Most mistakes in UI code are also easily corrected and have a very tiny impact radius.
AI is taking jobs but for a different reason. Big Tech is in an AI race so they will shrink their staff or move jobs to India as much as possible to free capital for GPU datacenters. To hell with the consequences.
Meta, Alphabet, and Oracle are even issuing almost $60 billions in bonds. Unheard of.
> While I acknowledge not all job postings result in a hire, and some are ‘ghost jobs’, since I was comparing the relative growth in job titles, this didn’t seem like a big issue to me.
Uhhh... That's a very big issue. In addition, there is also a glaringly obvious analysis error here related to layoffs/attrition.
No offence but is this analysis vibe analyzed or something?
All this is actually measuring is how the number of job listings in a specific industry have changed globally. That's not the same thing as "what jobs AI is replacing" at all!
Like nurses decreasing 11%. There is 0% chance that AI is disrupting nurses. If anything, nursing is probably in more demand due to continual aging. It might just be that fewer people are quitting because of covid being over. The total number of nurses might be going up still!
Are we just upvoting anything with AI in the title now? HN used to be pretty respectful of scientific methods. The methodology section here just reads like personal resume filler for showing that you can use AI. I mean I get it and I admire the hustle, but the quality here is pretty lacking overall. https://bloomberry.com/blog/i-analyzed-180m-jobs-to-see-what...
Would be cool to see cybersecurity covered here as well. I find it interesting how the machine learning engineer is exploding, but every engineer I talk to about it suggests these positions are mostly doing non-productive "bullshit" work and not really contributing much (arguably this applies to many tech jobs). Obviously that's anecdotal and probably not the real story, but it does seem like "machine learning engineer" is a broad term and probably doesn't tell the full story in itself.
The US is in a recession when the AI bubble is taken out. So obviously jobs are in decline. Machine learning hires are up because there are billions of dollars in circular financing.
While consumers are generally receptive to assistive machine learning like speech recognition, translation, limited auto-complete I think there is a huge backlash towards it being used to generate content and as far as I know all the major AI companies are losing huge amounts of money with every transaction. I am sure AI is being widely used as a reason for staff reductions but the reality is more likely sales are down, costs are up and head count is too high for current business requirements.
This is part of the picture, but in terms of impact one also needs to look at this in terms of absolute numbers not just percentages. Two job positions can both drop by 10% and in one case it represents 100,000 jobs and in the other case 500,000 jobs.
29 comments
[ 8.7 ms ] story [ 187 ms ] threadWithout those data this report isn’t really quantifying impact on “180M jobs.”
Maybe Apps are less of a priority now?
I'm not familiar in this industry so someone help me out here
Doesn't this also suggest that the job market is in such an unusual state, that any further analysis makes little or no sense?
My observation so far is that micromanaging AIs still sucks
I doubt they have this data. I highly doubt they have this data.
Taking two years and drawing conclusions from those two years seems to miss larger pictures - especially when the mess of Covid and the pandemic years and that job market is mixed in with this. Yea, that's half a decade back but companies are still trying to figure out how much staff they need where.
Photography? Yes, that's likely been impacted. Compare https://web.archive.org/web/20230124085038/https://www.bls.g... and https://www.bls.gov/ooh/media-and-communication/photographer... and the rate of growth has changed. However I'll also draw attention to the estimates that there's only a few thousand open positions with that classification each year. This includes self employed stock photographers and artists - corporate photographers have been a hard thing to get for a much longer timeframe (I looked into it a little bit back in '09). Additionally, artistic photography is impacted by the amount of money that regular people are willing to spend at art festivals and the like - that's gone down irrespective of AI.
This subset is not perfect, but good enough.
https://emusings.substack.com/p/is-automation-going-to-eat-y...
The first graph you show is missing the corresponding table from the source which shows the exact opposite of what you have stated: the workforce had been steadily increasing until ~2019 and is in decline since then. Not massively but still it contradicts what you stated. Next you switch to Korea and the wage increase without aligning a corresponding price increase or adjust for inflation. Then you switch to the history on the industrial revolution. At that point I gave up because I know the point you are trying to make and I can see, you adjusted your samples to match that point.
If you have additional data, I would be grateful for it and I will amend my article.
Frontend code can be very repetitive / labour intensive, I bet that has made this more attainable than for other layers in the stack. Most mistakes in UI code are also easily corrected and have a very tiny impact radius.
Meta, Alphabet, and Oracle are even issuing almost $60 billions in bonds. Unheard of.
Uhhh... That's a very big issue. In addition, there is also a glaringly obvious analysis error here related to layoffs/attrition.
No offence but is this analysis vibe analyzed or something?
All this is actually measuring is how the number of job listings in a specific industry have changed globally. That's not the same thing as "what jobs AI is replacing" at all!
Like nurses decreasing 11%. There is 0% chance that AI is disrupting nurses. If anything, nursing is probably in more demand due to continual aging. It might just be that fewer people are quitting because of covid being over. The total number of nurses might be going up still!
Are we just upvoting anything with AI in the title now? HN used to be pretty respectful of scientific methods. The methodology section here just reads like personal resume filler for showing that you can use AI. I mean I get it and I admire the hustle, but the quality here is pretty lacking overall. https://bloomberry.com/blog/i-analyzed-180m-jobs-to-see-what...
While consumers are generally receptive to assistive machine learning like speech recognition, translation, limited auto-complete I think there is a huge backlash towards it being used to generate content and as far as I know all the major AI companies are losing huge amounts of money with every transaction. I am sure AI is being widely used as a reason for staff reductions but the reality is more likely sales are down, costs are up and head count is too high for current business requirements.
O’NET taxonomy is better for this task, IMHO.