I know this doesn't make things better, but it bears repeating: much as the media would like it to be, this is not caused by AI. Two major things happened in 2022 that everyone knew would cause this if they stuck around:
* The Section 174 changes went into effect, requiring all software to be expensed as R&D, which means it can't be deducted from profits directly and instead has to be amortized over 5 years [0].
* Interest rates went back up starting in Feb 2022 and continuing through Aug 2023. This chart [1] is a near-perfect mirror of the job trends.
I know this doesn't help anyone who's struggling right now, but it does serve to emphasize that this didn't just happen, and it wasn't the result of technological advances making us obsolete. Federal policy was set up to encourage an enormous amount of investment into software development, and it's now... not. Turns out politics matters.
Yes but there is also another force that I see playing out now. With two decades of continuous training companies have grown an army of people overseas who can do any software or IT jobs at a level good enough to really be competitive to domestic workers. Offshore locations today are not the same it was twenty years ago both from quality and quantity perspective. No point in staying blind to this phenomenon.
I think the most competition to domestic workers is offshore workers working in their same timezone.
Hiring within the country of the company is a bit of a prestige move a more profitable / affluent company can do to afford political power. Similar to German car manufacturing
> * The Section 174 changes went into effect, requiring all software to be expensed as R&D, which means it can't be deducted from profits directly and instead has to be amortized over 5 years [0].
I get the short term initial issue. I don't see why it would be an issue to stick current level of staffing until it normalizes 5 years later. I'm not sure this was ever a major issue. It just seems like a convenient excuse to cover up something else.
Interest rates are obviously definitely an issue.
> this is not caused by AI.
It may not be caused by AI, but it is going to affect hiring numbers from now on.
I feel like the 174 issues might be more painful for smaller teams? Like if almost all your costs are in salaries, and you bring in a new person, suddenly you're booking profits despite not having cash profits?
That was my understanding, at least. That R&D salaries make this all weird.
The Section 174 changes are terrible for startups, but yes, they're only inconvenient for established companies. Many startups won't even last long enough to amortize all of their dev expenses, especially if they owe taxes on "profits" earned during those first few years. The whole startup ecosystem (which is what this set of job listings is largely measuring) relied on being unprofitable and therefore untaxable.
> It may not be caused by AI, but it is going to affect hiring numbers from now on.
That remains to be seen. I can say with high confidence that we did not have tech that was capable of reducing the need for engineers from 2022 to 2024. It's hard to predict the future.
How many developer jobs are due to startups? I can't imagine enough to really affect the job market all that much. There's a lot of people out there looking.
>> It may not be caused by AI, but it is going to affect hiring numbers from now on.
> That remains to be seen. I can say with high confidence that we did not have tech that was capable of reducing the need for engineers from 2022 to 2024. It's hard to predict the future.
If I can run some small models locally which helps me significantly right now, I don't really think it is hard to make the jump to needing fewer engineers overall. I've seen people paying for the big models do even more.
We will see by how much. I seriously doubt it improves productivity by a huge margin; if I had to guess, it would be in the ballpark of 10-20%. In the end, it's not the main driver for hiring or lack thereof. Companies that will need to get stuff done and that will be able to make enough money will hire as much as they need to get it done.
Thank you for posting this! I can’t believe section 174 is not more well known. I run a small software company and hiring an engineer is literally 5X more expensive as hiring a marketing or support person in a post section 174 world.
> If anything, AI hype is actually helping keep many of us employed.
AI is also going to keep many of us unemployed. I'm running a small version of Deepseek R1 locally and it does an extremely good job of predicting of where I'm going in a number of situations. I went to pull out some code into another function. As soon as I finished typing the name of the new function, it immediately suggested the exact code I was planning to move into the new function. All I had to do was hit tab. I didn't have to do the copy and paste myself. This was thirty lines of code and it chose the correct starting line and ending line.
It's been really quiet. I was laid off twice in 2024. First in April after 10 years at the same place. There wasn't a lot. Got a new great opportunity, but restructurings happened, I was the last to join and... again laid off.
But now it feels even more quiet than April-May-June. I've been debating cutting down my cover letter into a couple of lines just telling the company why I find the role interesting, and removing the rest.
The goal at this stage isn’t to get a job offer, it’s to start a conversation. If you’re being asked why you want to work somewhere, I’d consider that a winning cover letter/resume.
That said, I’ve job hoped my entire career and have never once written a cover letter since a college workshop where I had to in order to earn course points.
Web 2.0 came out of the dotcom bust. As a HN oldtimer, I really wish I was seeing more posts on HN about people partnering up to build cool stuff right now given there’s ton of collective idle bandwidth out there. Maybe it’s happening somewhere else and HN is not the place for that these days?
Well, I tried to do something cool but I had rent to pay and kids to feed so I needed to get a day job (I was very fortunate to find one, on an HN who's hiring thread no less).
However, I don't have the savings a lot of people here seem to, perhaps because I'm in Europe.
I think it's pertinent. Ppst 2008 was essentially disinflation economy whereas inflation has been high running into this and people are struggling a bit. To launch something you gotta be hungry but not too hungry. Maybe people are either retiring on thier 401ks or uber driving 24 hours a day to make rent. That or the demo here got old and acquired responsibilities.
There's still some large enough percentage of laid off folks that perhaps are looking but living on reserves instead of 24hr gig work. This is the pool of idle bandwidth I'm speaking of. G'parent that found another job; not idle bandwidth. Someone that had no reserves and is flipping burgers; not idle bandwidth. Refreshing Indeed, LinkedIn Jobs, etc. daily after about a week or two while living on savings, moving back with parents, living off spouse's income only, and an infinite number of other possibilities; idle bandwidth.
I find if funny you mention economics of 2008 but don't also mention how tech salaries changed over that same time. At a high level, this is the exact population of people that should have some dry powder in the keg and be able to weather this storm. I'm not saying that to be dismissive of how dire prolonged unemployment is to one's personal finances, even with savings. But I am being dismissive of the idea that most of these people are doing gig work so much they can't hack on something.
I do also think the demo here has changed a lot and you're onto something with that comment.
Fair point, I think the idea was more about people who had the ability to work on their own project unpaid for a decent amount of time perhaps being scarce. But if you're on unemployment now seems like a great time to bootstrap something.
Ah, I didn't know this was such a trend on twitter as I've never used it. Probably makes most sense to do it that way. I suppose I used to see it manifest as a "Show HN" and hope perhaps there's a wave of that coming
I should caveat that "Show HN" has turned more into "look at this github repo" and I'm talking about actual businesses as HN used to be a bit more entrepreneur oriented and has become a bit more maker oriented in this regard
Neither did they? If you have time and skills, you just build and share to try to gain traction and/or partner up with people that can cover your skill gaps.
I feel like your comment is the root difference in what I see here. You must not be entrepreneurial at all if you think money is the barrier to building something people want. Sorry if that's harshly worded but I've started a business with a $5/month DO droplet and $15/year domain as have many others.
A lot of talented people are struggling right now, but could take the time and energy they're spending on fruitless job applications and devote it to some modest-scale side project with commercial potential.
Worst-case scenario, you get an interesting portfolio project and something to talk about at a future interview. Possibly, you get something that pays the bills and avoids having to deal with everything that's gone wrong with tech hiring.
If you've been laid off from a technical role, you probably came away with enough knowledge to build a competing micro-service.
Few people, even in decently earning roles like in tech, will have the runway to try and make something successful on their own all of a sudden after job loss.
Starting a freelancing practice is more likely to bear fruit, but it's a very different ballgame of overheads than "just" the core job itself, if you want to get the full rewards of being a freelancer.
"Making money is easy, just go out there and make money" is effectively what you're saying.
It takes time to find and execute these ideas. Yes the tech can be cheap, maybe even building it can be cheap, but the time to grow your customer & client base from scratch can be highly varied.
If it were really this easy, you'd have every person on IndieHackers having ditched their jobs already because their ideas have taken off. Yet very few have.
Go out and execute, yes, but it can take many iterations to get anywhere.
Even Pieter Levels has about a 10% hit rate on his projects being successful.
I think you're being pretty dismissive here. That person is talking about someone who currently isn't working and is struggling to find work. "idle bandwidth". Nobody is saying it's simple or easy, but if you're in this field you have the skills to at least try. I've been laid off for 9 months so I understand the toll it takes mentally, but having a pessimistic attitude won't help you in the short or long term.
Looks like the downtrend has slowed, and almost stopped. This aligns with my observations on local job boards to me, and some programming language specific Slack channels I'm on that have #jobs boards, in that there's a slight uptick in jobs right now. This is good news, it seems.
Somewhat correlates with the Layoffs.fyi trends [1][2].
Extreme lack of layoffs Q3 2020 to Q1 2022. Big rise through 2024, peaking at Q1 2023 and then slowly declining. Mostly tapered off to 2021 and 2022 levels near the start 2025.
Also, complimentary chart from Stackoverflow about the general trends in questions related to these categories [3]
What a great reply this graph can make the next time we see VCs on X complaining how, "Everyone I talk to has so, so many engineering positions they just can't fill". By these measurements, the job market looks worse now than it did at the troughs of lockdown.
Personally, I'm getting close to just giving up. I've completed hundreds of applications, contacted hundreds of recruiters, and spent hundred of hours studying. All with nothing to show for it. Every single job I see has hundreds of applications. I have no idea what to do.
As lolinder stated, one of the main causes of this is Section 174 where software is now amortized as R&D over 5 years which means companies can't treat is as a regular expense. This leads to companies hiring less.
This is effectively a subsidy for the tech industry, why is this incentive necessary here? It's not specifically targeting startups from what I can tell and potentially just a kickback to large employers and investors.
All software jobs, including those in startups, are affected by the amortization rule. It makes it harder for startups to afford more software engineers because they have to pay taxes on cash they don't actually have.
The only companies the new amortization rule helps is big established firms that can afford to amortize because they have excess cashflow.
2021 was such a bizarre year in so many ways. It feels like a post COVID fever dream in retrospect, and seeing that massive jump in job listings that year is a reminder of how weird it was.
My armchair macroeconomic take is Covid should have been an enormous financial depression but they flattened the curve and caused bubbles in some cases by printing so much money we’ll be paying the price for decades. It was certainly a huge reallocation/consolidation of wealth. I even think the somewhat untold story is just how much less well off a lot of Americans are. Like in my world, what I’ve seen amongst my access to HR data at a handful of large companies, the executive class (maybe VP and up) did great, the low/minimum wage class did great (labor shortage helped wages net of inflation), but there’s a huge mass of people in the middle that perhaps are only making 10% more than their 2019 wage right now after all the inflation. House prices ~doubled so unless they were already a homeowner acquiring one is basically impossible now, especially with rates were they are. They didn’t own much stock so largely did not participate in that either. It’s a different world and I think we can’t say for sure what will happen next.
Yeah, when people talk about the emptying of the middle class, I never really thought about it much till I saw it happen with COVID. You either made it out one end, or are sliding down to the other.
The job market and the capital markets are still insane. Unfortunately, I don’t think we reached the bottom yet given the astronomical real estate prices relative to incomes and high market valuations relative to earnings. Some signs of reality are starting like Real estate prices starting to fall in the Sunbelt of the US. The obvious commercial real estate bubble has not completely burst yet. So many markets are hanging like guillotine’s ready to drop. Markets have been buoyed by excess capital and rampant speculation by the wealthy for so long now; its shocking how long the charade has lasted.
The chart correlates well, with similar references from the US and similar related markets (see [1], [2], [3] and [4] ).
At a more macro level, for the United States, I looked at U.S. unemployment data, for the last 80 years. Starting from 1945 until today Jan 2025. All jobs, not only Technology. What jumped out immediately is a clear cyclical pattern: Unemployment peaks occur, on average, every 7–8 years, with some variability.
- Shorter Gaps (3–5 years) showed up in the 1950s and early
1960s, often tied to small recessions.
- Longer Gaps (10–11 years) happened before the early 1990s
recession, again around the early 2000s, and more recently
between 2009–2010 and 2020.
- Outliers include the early 1980s (double-dip recession
pushing unemployment above 10%) and, the 2020’s pandemic
spike, which briefly soared past 14%.
Overall, the business cycle hasn’t radically changed over 80 years: Expansions that push unemployment down to historic lows, followed by recessions that create sharp spikes in unemployment.
Assuming 7–8 years is a decent rule of thumb, and based on US history of the last 80 years, we are in for a next continued peak into growing unemployment, until at least 2027–2028.
At a smaller scale, local trends in geopolitics, tariffs, AI adoption, and political dysfunction reflect (or confirm) the broader macro trend.
Interesting take on deepseek. I think deepseek may be a breath of fresh air slightly leveling the playing field for small startups and individual developers, Assuming the claimed reduced training cost are real.
* The Section 174 changes went into effect, requiring all software to be expensed as R&D, which means it can't be deducted from profits directly and instead has to be amortized over 5 years [0].
* Interest rates went back up starting in Feb 2022 and continuing through Aug 2023. This chart [1] is a near-perfect mirror of the job trends.
I know this doesn't help anyone who's struggling right now, but it does serve to emphasize that this didn't just happen, and it wasn't the result of technological advances making us obsolete. Federal policy was set up to encourage an enormous amount of investment into software development, and it's now... not. Turns out politics matters.
[0] https://qqrl.tk/item?id=34627712
[1] https://fred.stlouisfed.org/graph/?g=1DfOX