34
votes
Software engineer lost his $150K-a-year job to AI—he’s been rejected from 800 jobs and forced to DoorDash and live in a trailer to make ends meet
Link information
This data is scraped automatically and may be incorrect.
- Published
- May 14 2025
- Word count
- 771 words
I just don't really believe this as someone that uses AI coding assistants every day. They only provide a marginal improvement. I have to believe he's got some other issues finding a job but is blaming AI instead.
The headline is absolutely false on the face of it. I can definitely believe that (a) someone had a cushy, high-paying software job, and then (b) the AI bubble got real big, and then (c) now they don't have that job anymore. The causative connection between (b) and (c) is absolutely not there, though. (Insofar as there is a connection, it's that slop has utterly saturated the input side of hiring pipelines. But, like, this guy has zero connections he could use as a reference to separate himself from the ocean of garbage?)
More than the kind of farcical article itself, I'm interested in why Yahoo published it. And there's one super obvious reason: your boss wants you to be afraid for your job, so you stop telling him when he's wrong and making ridiculous demands like being treated with basic human decency. Software engineers have retained a very disproportionate level of labor power in our post-Reagan job market, and employers hate that.
The answer to this is, of course, the same as it's always been: unionize.
FYI, Yahoo didn't actually publish this, per se. Yahoo News is just a third-party news distribution platform, similar to MSN News, that other publishers often distribute their articles through. And in this case it was Fortune magazine that authored and published the original article. See: https://fortune.com/2025/05/14/software-engineer-replaced-by-ai-lost-six-figure-salary-800-job-applications-doordash-living-in-rv-trailer/
Your overall point about why you think this was published still applies, but just to Fortune rather than Yahoo.
I doubt that Yahoo published it because of that reason. Regardless of whether or not they report on it, the latent fear of being replaced by AI is everyone's minds, and online news has been a race to the bottom in churning on content that preys on that latent fear.
Yeah, this doesn’t pass the sniff test. The job market for devs might not be as hot as in the past, but 800 denials for an experienced dev? That seems like there’s a bigger issue at play here.
800n denials doesn't sound too off to me given that in 2021 I applied to over 1K jobs and only got one offer as someone highly experienced in my field (albeit, my field is struggling), and Mr. Tired went through the entirety of 2024 without a job and applied to a similar number of jobs and only got two offers after mostly being ghosted.
What doesn't pass the sniff test to me is that he's blaming AI for taking his job rather than blaming AI for rejecting him, since ATS are basically an application of AI that have been around and have recently become even more difficult to get past.
Well no wonder he can’t get hired. The real game is in employers inferring how good your LLM prompting is from your job applications.
My thoughts exactly.
As LLMs exist right now, if they’re replacing anybody it’s interns and maybe fresh grads, not mid-level devs and certainly not experienced veteran engineers.
I believe that the state of the economy is a much stronger factor for the overall tech sector job market. This person in particular either has undisclosed problems marring his resume and/or is making demands that are unlikely to be met (article comments call out full remote work as a likely candidate).
Man refuses to move for work. Fails to find work.
To be clear, I currently work remotely and would prefer to continue to do so, but if it came down to it I would return to office if I had to. Especially in the current economic climate, the salary is worth it and at the end of the day remote work is more of a nice to have than a must.
It’s one of the reasons I’m now living in a mid-CoL area with some tech industry presence instead of somewhere cheaper. Much better chances of finding a local non-remote position should that become necessary.
I’m out of touch, but the software development job market does seem harder to get into than it used to be, particularly if you’re limited to working remotely.
The connection between “finding a job is hard” and blaming it on AI is mostly a news hook, but there do seem to be companies that aren’t hiring as much and say it’s due to AI.
There's still piles and piles of independent/freelance/temp work that you'd fall back on before you're forced to door dash. Yes demanding remote however is not going to help your chances. I do think remote is something that more companies can and should do, but there's a lot of people who think that's all anyone needs and it's usually just not true, and they aren't willing to adapt or tack a sacrifice (lower salary).
Where are the best places to look for work like that? I know someone who might need it.
I get random emails fairly frequently through the various hiring/consulting firms out there. They're all smaller/midsized firms so I can't just recommend something as I think it's region and work dependent.
That said there's companies like Robert Half for hiring/placing and there's places like freelancer for more direct freelance work. Reaching out to consulting companies to get on their lists is also likely an option.
I'm not going to say that any of these are guaranteed to be stellar jobs, but they're very likely to pay the bills a hell of a lot better than door dash. I have known some people who swear by freelance/independent contractor work. It's a bit of a pain because taxes are annoying and you really need to factor in things you aren't used to like health insurance, but at the same time with the nature of "hurry up and wait" in developing they often juggle a few gigs at a time, with most of them in holding patterns of "i'll get this done in 3 weeks" and they finish in 2 but spend a week waiting for feedback and the like.
Again the main point here being, all of this is going to be available to you before you jump into door dash.
For me independent contracting work only became available after I moved to SF and built a professional network here. There's so many startups building things out and I can help out having done that a few times before. I really like the flexibility.
Using an LLM is like having a really, really fast intern. With the caveat that the intern doesn't really "learn" and improve over time like a real one. An LLM can't engineer anything, certainly not without an actual engineer on hand to fix up all the slop it spits out.
Agreed. Especially when you consider the longer term support and maintenance of code LLMs are still not nearly enough competent for that. It reduces my labour hours, but not really my actually labour. While it did take me a long time to find a job, that was as a starter. Going from a 150k job to nothing...? If you're that good I'd also think you would have connections to potentially find something else, no offence towards his suffering. It's difficult to square the circle in my mind.
In case anyone was interested, the guy in question wrote a blog post about this that was posted on HackerNews yesterday.
HN Post: https://news.ycombinator.com/item?id=43963434
His blog: https://shawnfromportland.substack.com/p/the-great-displacement-is-already
The author comments about the post in the comments. Let's just say that I'm not terribly sympathetic to his plight. It's clear he has other issues going on. For example, his resume has listed as his #1 skill "Vibe coding". It is an interesting read, though I have to admit mostly for armchair psychoanalysis.
As someone only a few years beyond my time as one of the 20s interviewers he has that condescending attitude toward, I'm guessing is his problems are a combination of:
Of note is these are all shallow opinions derived from his blog post and resume.
I'd argue that knowing how to use AI well—knowing how to maximize the benefits and minimize the pitfalls—is a skill. An AI is a messy, smudgy map of all human knowledge but not (quite yet) its territory: skillful users can use it to rapidly traverse vast landscapes of knowledge.
That being said, "vibecoding ecosystem" is a rather embarrassing presentation of such a skill.
Having a legal last name of "K" and being a "professional vibe coder" have a startlingly similar energy. Someone needs to write a book on this man.
Yeah, the name that he himself had chosen... I didn't mention it elsewhere because it is so highly subjective, but it would be a huge red flag for me. It might be a coincidence, but the most toxic (and I am talking actual personality disorder level of toxic) people I met in my life had unusual attention-grabbing names or nicknames that were not given to them; they themselves 'renamed' themselves.
I will defend him a bit here. He says that he picked K because he didn't want his father's name following him around because that man was never part of his life. He also went by K for a long time so it seems like a natural change to make.
Fair enough, I think someone with your username ought to be an authority on this matter.
Well... that's what's red-flaggy about it, isn't it? Does that explanation not sound weird to you? Maybe it's just me. As I said, it's subjective. I might be totally wrong.
I can see where you are coming from, but I would say that it's not a red flag to me. I think taking charge of what people call you is reasonable. I actually find going with just K better than picking a random last name that starts with K. Because he might not immediately respond to the random name!
If you actually go to the pdf of his resume itself, I really don't believe for a moment he should struggle to this degree. That much JavaScript experience, along with SQL, API and a few other things should net you interviews, easily. Weirdly enough, its when it gets to the things like 'vibecoding ecosystem' and some AI jargon that I raise my eyebrows?
In a sense he may actually be sabotaging himself without realising it. As you put it though, that's mostly armchair psychoanalysis. Still, if his resume is correct he does have valuable, hard skillsets that LLMs won't be able to perform for the next few years at least.
The doubt here is really depressing.
I haven't counted, but I have applied for a lot of jobs in the last year and only got maybe 5 interviews. It wasn't as many as this person, possibly because I spend a lot of time tailoring my resume and application to highlight experience in the job listing. One of the interviewers told me that ai applicants are a huge problem. Apparently they're having to fly candidates out to make sure they're humans, despite being a remote first company.
Yes, I have contacts in the industry. I've applied with all of them that I knew were hiring and they all passed, so on top of everything else I'm now wondering if the job i loved was full of people who thought I sucked. Maybe I did, but I would have worked at getting better if I'd known what to work on.
There are so many confounding factors. How much experience do I list so that I look experienced, but they won't think I'm old? Is everyone only hiring from prestigious schools now? Is it a mistake to put my obviously female first name on my resume with the anti DE&I environment that we seem to be in? I don't have a choice with some applications that ask for a legal name, but not all do. Do I come off as inhuman? I've tried my own words and I've tried using help from ChatGPT, it doesn't seem to make a difference.
How many of those applications I spent real effort on were for jobs that didn't even exist?
Perhaps it's time to switch careers.
I'm on team "this sounds suspicious" but I had a passing wonder if the computer science/programming/Tech/and broader STEM evangelists will pause and give a nod to fields like psychology, social work, education or even nursing which rarely seems to be included in STEM, especially by the feds, where there's a plethora of needed and valuable (if not valued) work that LLMs cannot do.
If LLMs (autocorrect wants this to be llamas) have zero haters I'm dead so I'm not overselling what they can actually do in programming. But let's say there is a shift and AI can and does do more - we did work retraining for people whose ceramic plant jobs left the country or who were coal miners (my partner became a chef this way) - there are a lot of options other than doordash out there. They don't pay as well, but they don't make you live in the most expensive parts of the country either.
Idk it was a passing thought.
If I lose my job I'm headed towards mental health and therapy. -..-
I mean, same. Except I'm already in therapy and won't be able to afford it or my meds anymore. But it was just a passing thought of if this field no longer has the prestige/pay/job opportunities will other subjects, where humans are less replaceable become options, or would (in this hypothetical) it be more like coal mining where the goal seems to always be to bring those specific jobs back.
In reality, I don't think those jobs are actually replaceable by AI as a whole, and most of not all employees feel pretty disposable/replaceable if only because they're treated that way by corporations. But I've disliked the STEM or nothing sort of mentality for a while (and obviously this is only one field, if not a subset of a field).
I think I'm just piling on, but this is a completely nonsensical statement.
On one hand, tech certs are typically not highly valued (especially for folks who already have significant professional experience). But even the most sought after certs tend to cost no more than $100-300 to sit for and have plenty of self-study materials freely available online.
On the other hand, I'm not an expert on trucking but 60 seconds of googling tells me that training + exam for CDL can cost anywhere from around $1-8k depending on if you go to a community college or a private program. There also appear to be numerous companies that will pay for your training.
Plenty of people also predicted it would be an enormous waste of money!
This reminds me of those really indignant stories about, say, getting banned from Youtube. I have bad news for you, Shawn, it has been going on for years. You just didn't pay attention until it was your turn...
My friend who worked at Google went through this.
Anthropic CEO Dario Amodei has predicted that AI will be doing all coding tasks by next year—but an existential crisis is already hitting some software engineers. One man who lost his job last year has had to turn to living in an RV trailer, DoorDashing and selling his household items on eBay to make ends meet, as his once $150k salary has turned to dust.
And musk said we'd be on mars and have full self driving cars by now. CEO timelines, especially when they stand to profit from the hype they generate, are almost always bullshit.
Musk has never made good on any of his grand promises. Anyone who believes in him is a fool.
AI doing all coding requires AGI. AGI is what cold fusion was a few decades ago.
May I ask some advice from all the coders here as a sidetrack to this piece? I've recently went back to school and will graduate with a software engineering degree next May but I find it hard to practice coding based on what the college has taught me. What would be some good ways to practice? I guess I'm having trouble connecting the syntax and basics we've learned and turning it into something "bigger"
The best answer is to find something that you think you can do, something within your capabilities that excites you to think about, and build that. Could be a game, could be a website, could be a video chat app for backgammon enthusiasts, could be a to-do list. The process of building something that you want to exist is more educational than any video or book.
Slack/discord chatbot, text editor plugin, whatever it is, make it something small that affects something or someone you care about. Momentum builds from there.
I would also add to not let, “but that already exists” to get in the way. If one limits themselves to things that don’t exist in some form or another, they’re ruling out innumerable experience-building projects.
When I first started dabbling in programming ages ago, my reasoning for selecting the projects I did was that I wanted to create a version of whatever pre-existing thing tailored exactly to my specifications, and even now some 20 years since I started tinkering and 10 years in the industry, many of the side projects I start are still of that nature.
I think most of us who’ve used computers extensively have opinions about the software we use and have ideas about how we’d change it if given the opportunity. Dig into that and use it as a source of inspiration.
That's good information. Thanks, i'll try to put it into practice .
I hate to give this answer, but you should probably do leetcode if your primary goal is passing interviews and getting a job. If you deeply understand the problems and solutions, you'll be much better off than most grads. A good book to start with is Cracking the Coding Interview, or there are a few Grokking the Coding Interview courses which cover things thoroughly.
I deeply hate leetcode, but it does give you some skills on how to practically approach algorithmic challenges. The best engineers identify the patterns and think about how they can apply tools like hashmaps to massively speed up their code.
You could also read a book on design patterns like Head First Design Patterns to really understand what things like observers, factories, and strategies actually are. Alternatively or additionally, a deeper architectural book like Designing Data Intensive Applications will give you a massive leg up as you're starting out.
My last recommendation would be to read some books like the Mythical Man Month or Clean Code by Martin to better understand the field and art of software engineering. People and coding styles are the fuzziest things in the field, and they're frequently the source of problems you'll encounter at work. The absolute best engineers know how to communicate, whether that's talking with management or using readable variable names!
Thank you for the recommendations, I will add them to my reading list.
I agree with the other responses here. Building something to show is always a good idea. I would add that whatever you choose to build, should be ambitious or challenging enough that it gives you a better idea of the challenges faced by developers in that area.
This can help you feel one step ahead of the interviewer, and better guess what they want to hear. As an example, no interviewer actually wants to hear about the basic backend API you threw together. They want to hear about how you load balanced huge numbers of requests to those endpoints and got the database to handle concurrent requests (disclaimer, I’m not a backend person).
I do embedded and operating systems work. When I’m interviewing a candidate, I like to hear how they accidentally deadlocked 2 threads and went about fixing their synchronization primitives. I don’t really care that they wrote X-many lines of C and they blinked some LEDs.
At first, it may be difficult to get a sense for what the interesting challenges are. For that, you really just have to expose yourself to the subspecialty in more depth. Whether that’s talking to experienced folks, reading, etc. That bit is up to your preferences and situation.
Leetcode is good for interview prep. If you want something a little less leet and a little calmer, I enjoy completing old Advent of Code problems as a way to practice a new language. Beyond that, building something small with the language is probably the best choice, though choosing a project that targets the right skills to practice can sometimes be non-trivial.
In addition to the other replies, consider contributing to existing open source projects if you have the bandwidth. Working in someone else's opinionated code base is a skill you'll need. These are also places that you'll get direct feedback. If someone doesn't like your approach to solving a problem they'll usually tell you why they don't like it and what they think you should do instead. Just be careful to not get too attached to your ideas because they will get adjusted or shot down entirely and you'll need to not take those instances as personal attacks.
This is more of job prep than interview prep. Interviewers don't have the time to get into problems that can't be solved in under 20 lines of code. You may be able to drag some of those contributions into your interview through other methods though. For example, you could be asked a question about how you interact with others and use it to drag in some cool feature you implemented in some project by way of using it's review process to discuss interpersonal interactions.
This works even for "little" things, like updating documentation. Critically this kind of stuff forces you to read a lot of real-world code, which should help with seeing how things fit together.
I'm gonna go against the grain here and suggest you follow a tutorial that builds a simple product from ground up. Something like "build a Twitter clone in nodejs + React" if you're interested in working on web apps. The reason being that if you have zero experience, it's really hard contributing meaningfully to existing projects.
Understanding someone else's code, the framework they're using, and architecture decisions is difficult even once you've had some experience. Working on something from ground up with guidance will show you the different parts of a typical web app: the frontend client, the backend server, the database, maybe a caching layer as well. Once you've seen that, you can start to identify similar patterns in other codebases. The best part of this is that a tutorial will have an end state and you can see how far the end of the tunnel is.
Also, learn git and +1 to practicing leetcode problems on the side.
Edit: given your timing, it sounds like you won't have an internship before you graduate? In that case, it's imperative you have some projects outside of school. Do a tutorial and build something, then build something without a tutorial. You really need to do something that wasn't a school project in order to stand out.