In “Ask HN style”, let’s say that software engineering does go extinct in the next X years, what would you do?
I’ve thought about psychology. I know LLMs can work as pseudo-therapists but I feel like that’s a field where the human connection / human element will remain important.
On one hand some jobs with human element are safe, at first. Think of artists being made obsolete by the camera. Portrait artists became mostly obsolete, but we still pay for art. It's the story behind the art that became important. Or, I still go to cafes with nice atmosphere and friendly staff. There are restaurants with robot staff here in Japan, much cheaper. After the meal you pay at the table without ever talking to a person. But it does not feel nice to sit in there, so I gladly pay a premium for the nice coffee.
On the other hand, it is not only software jobs in danger, but all office jobs. So a lot of people may suddenly be out of money. Let's say you open a cafe, but no one has money to come and pay. Society has to change a lot from the current model to be able to handle this.
I actually think AI has an unfair advantage with software that is making it seem far more capable that it is. Software is entirely text based, and producers have been putting their outputs and problem solving online for free for decades.
I think applying AI to other white collar roles that also require problem solving but do not have as much training, will prove much more difficult. Even coding on proprietary dominated domains is a much, much worse experience than people have with more accessible code. Using it for electronics has been hit or miss, embedded software is a bit shakey, game development is also challenging to use it for etc.
I get what you’re saying, but it’s not as though there are trillions of books and blog posts and stackexchange questions about excel and the handful of other things that most office workers do, too.
I honestly figured that’s why everyone is coming out with MS Office plugins for all the models, and MS itself is putting it in the tools.
So if most any company only needs one person to solve limited IT issues, prompt code production and deployment, generate the usual truckloads of excel spreadsheets, and do most of the finance and accounting… it starts to look pretty scary.
Then, what about the people making and maintaining all the facilities for these people we don’t need anymore? The world flipped its lid about commercial real estate when wfh became a thing. That was relatively small and temporary.
> Then, what about the people making and maintaining all the facilities for these people we don’t need anymore?
And all the small businesses like local restaurants and coffee shops that they frequent, etc, etc.
There are so many 2nd order contagion impacts if the knowledge work economy implodes that very few people won't be negatively impacted to some significant degree.
And some people seem to think that outcome means the government will step in and engineer some sort of soft landing. And outside of the US this may very well be true, but here in the US? Seems unlikely.
While I see your point, I think AI may be banned in those universes for the same reason that time-travelling devices tend to be written out of sci-fi and fantasy: Stories are better without them. (If one counts the ever unfolding history of the world as a story, my counter-point actually validates your original point. Touché.)
Those broads limitations also make writing the stories easier, since the author/worldbuilder doesn't have to come up with 100 different reasons why it can't be done for each particular case. He only have to do it once.
I know a lot of people working across portfolio management and tax accounting. Nobody I know of is using LLMs much and frankly their management has started to back-off pushing it more in the workplace.
LLMs suit some jobs more than others. Its quite possible SWE's are the only profession massively affected - whether that means a evolution of the role or decline/death is another question.
> I know a lot of people working across portfolio management and tax accounting. Nobody I know of is using LLMs much and frankly their management has started to back-off pushing it more in the workplace.
I could say the same thing for software engineers I know as recently as the middle of last year, things can change very quickly.
Up until about December 2025 the fact that LLMs would replace us all (SWEs) was the punchline to a joke for most working developers I know. But most of the ones I know aren't laughing anymore, unless its a nervous laugh.
LLMs may (likely will) disrupt software developers first, but I don't think we are particularly unique and I don't see any reason why the same risks won't spread to virtually all knowledge work, especially if executives in those fields see a significant amount of SWEs being replaced by LLMs as an initial test case.
There are still a few quantum leaps needed. I have had great results with Opus 4.6, in particular in green field. But it behaved real messy in some professional real life projects.
It seems you also need to tell it very specifics things some times but for that you need to do a software developer in the first place.
LLMs certainly aren't ready to replace all software developers yet.
They may never reach that point.
But even if they never get good enough to replace all software developers, they can still cause massive job losses by allowing companies to do the same work with far fewer developers.
Sorry man. This feels like you are hoping SWE alone isn’t affected.
Unfortunately the workflow of a software engineer has been to do things like asking questions on stack overflow to do their job - to use digital resources scattered across the web - to show examples of code freely across the web.
The workflow of an accountant, portfolio manager etc has nothing to do with accessing and using the web in the same manner. If you did their jobs you’d know this, but you don’t. Right?
Is it really a surprise? Nope. Thankfully writing code isn't enough. So your job is still somewhat safe for now.
Tax and accounting is rule based reporting. With formal authorities and openly available rulesets on right and wrong. There’s judgement in it, but that’s even less true than in development. Maybe someone makes the case that there’s art and ergonomics in it too, but not more than swe.
Professional accreditation and responsibility is its only real moat. And those are “yeah but!” issues we hand-wave in discussions around swe too.
Accounting. In my region tax law can suddenly change outside of the usual annual update cycle and, like everywhere, is riddled with edge cases and unclear interpretations.
Most importantly there's often a period of general uncertainty and adoption, during which the new law is already in force, but LLMs will rely on whatever there was previously.
Most people find this job stressful and boring, but the same can be said about software engineering. Regular people pay money to have it dealt with.
Overall I think there will always be demand for handling the messiness of the real world and humans have the upperhand here because they learn as they go, not via release cycles costing a sizeable sum and taking months.
If LLMs can handle software engineering well enough to no longer need engineers, accounting will be solved by the same model version.
Seriously if the future manifests, all of these standard effort based jobs would become redundant...
The issue with outdated information is way overstated, it'd just add the current rules to the context when evaluating and be done with it. We're already at 1 million context size... That's enough for a lot of rules - and the number will likely go higher as time progresses
You would die rather than move somewhere cheaper? What an odd take. I live in the midwest and pay $700/mo for a perfectly fine apartment in a clean and safe suburb.
Housing is expensive in Canada, but it's absolutely not uniform. $2500/mo starting is crazy, which city are you sourcing these claims for? I live in a major city (but not Vancouver or Toronto, obviously) and if you're just trying to survive, you can live with roommates for $700-900, possibly less depending on your luck. Apartments, studios and other types of housing for one are about $1500 and up. Then you can go to Quebec and enjoy slighter cheaper housing still, even in the big cities. There's some middle ground between downtown Toronto and some mining town in northern Manitoba.
In this theoretical scenario where AI displaces everyone, the only thing with value will be housing and physical necessities, so I think housing prices will go up.
I think its somewhat comparable to cutting grass in the cities.
It was manual labour first. Then there were teactors. Now robots join in - does that mean that personel cutting grass is obsolete? No , you need all of them. That means that city becomes nicer.
With software and AI I somehow feel the same will happen. How many features have you skipped just because it would help some niche set of users and PM or Management would not approve the spending. It is low priority. Or bugs that were annoying but financially not bringing much value.
I hope switching some work to AI , some companies will capture opportunity to make software better while others will make the same software cheaper
I have some ideas for small business, but am also keeping an eye on jobs that could be prominent and enjoyable. I would much prefer a small business (have done it before) but financial realities may dictate a regular job at first.
I think many of us who have been in software for a while will fantasize about low-tech jobs, I imagine there will be a bunch of hobby farms...
I am also quite interested in psychology, but at least in my realm it would take something like 5–7 years of studies etc to become a licensed psychologist, if you want to work directly with humans. That is quite the investment and I am not sure that I have a long enough personal runway for that.
Now is the time to start thinking about being more than just IC, and thinking in terms of an entrepreneur. Call it a "lifestyle business" or whatever, but what can we work toward today that enables us to call the shots? Just don't fall down the trap of making it developer-related.
I've decided to look into becoming either a landscape designer or an electrician. Worst case scenario, there is always a nursing shortage and it's really not very hard to get that degree.
I suspect the number of startups will skyrocket the nexr few years. Fired engineers will start to compete against the establishment that fired them. Competition may get a lot more fierce for a while.
I’ve thought about psychology. I know LLMs can work as pseudo-therapists but I feel like that’s a field where the human connection / human element will remain important.