I too am not a SWE. This confuses a lot of SWEs, recruiters, interviewers, etc.
My domain used to have a name, but that name got changed to something that has the word AI in it. And I don't really identify with that either, although I use parts of that.
Hacker (in the HN sense, obviously) actually captures fairly perfectly how I approach using computers/tech.
Good piece for those with such an identity crisis to re-find the word for it.
He’s a perfectionist idealistic artsy software engineer. Surely some 10x engineers can find some roles where they can totally avoid agentic AI coding and leave that to their coworkers to manage for him.
We need people like him.
We also need pragmatic efficient generalist software engineers who have a get shit done / FAFO mindset, accept the probabilistic nature of this enterprise and focus on what customers really need
(Hint: your customers don’t care about your beautiful TDD OOP SOLID DRY code)
I don't know where he lives, but I am Canadian so I find it laughable that someone with a few-week bootcamp and a few years of on-the-job coding experience can call himself an engineer in the USA. In Canada "engineer" is a title with serious requirements and responsibilities. I don't see the Bay Area coder crowd rushing to take responsibilities like "if the bridge I design falls, I can be sued and/or lose my license."
I love it! I had this convo with many people before, I consider myself a hacker too, not the Hollywood cliche, but the concept of hacking things and repurposing them to do another goal or a new one. I get bored quickly doing the same thing over and over so sitting and writing code is a soul crushing task, rather, I have a goal to do something and in the process I learn all the new exciting stuff while building it, of course, you have to be a quick learn and adaptable to do that. It’s perfect except when it comes to get a job, most companies still follow the traditional concept of you being a cog in a machine, and whatever you are doing will be used in our work, they don’t look at the attributes you have but rather how many X years in writing Z language or FPGA design.
I’m not sure what his point is? This just looks like grief-posting. Keep up with the times and don’t be incurious.
If you don’t like AI then you should find a niche in software development that may still require rawdogging. I think kernel development and making hand crafted libraries are good options.
One of their primary gripes seems to be that coding via AI is not deterministic - what you get is somewhat up to chance.
But how is this any different from delegating to someone else? They're not going to produce precisely what you told them to.
In either case, you can quite obviously reduce the variability by writing better prompts/instruction, using specs, tests etc... And, likewise, you are going to need to iterate anyway.
The problem, in either case, is when you don't give good instructions, don't carefully review, etc.
So, one might say that a defining characteristic of a proper "engineer" - be it software or otherwise - is conscientiousness. If you combine that with creativity, communication skills, determination, etc then you can produce good results.
This is perfect!
I too am not a SWE. This confuses a lot of SWEs, recruiters, interviewers, etc.
My domain used to have a name, but that name got changed to something that has the word AI in it. And I don't really identify with that either, although I use parts of that.
Hacker (in the HN sense, obviously) actually captures fairly perfectly how I approach using computers/tech.
Good piece for those with such an identity crisis to re-find the word for it.
He’s a perfectionist idealistic artsy software engineer. Surely some 10x engineers can find some roles where they can totally avoid agentic AI coding and leave that to their coworkers to manage for him.
We need people like him.
We also need pragmatic efficient generalist software engineers who have a get shit done / FAFO mindset, accept the probabilistic nature of this enterprise and focus on what customers really need
(Hint: your customers don’t care about your beautiful TDD OOP SOLID DRY code)
I don't know where he lives, but I am Canadian so I find it laughable that someone with a few-week bootcamp and a few years of on-the-job coding experience can call himself an engineer in the USA. In Canada "engineer" is a title with serious requirements and responsibilities. I don't see the Bay Area coder crowd rushing to take responsibilities like "if the bridge I design falls, I can be sued and/or lose my license."
I find the American version far more attractive to my ideals. No fake credentials whatsover. And who ends up making better software? Americans.
Are you American? What does that mean to you?
I’m not American but I don’t think that’s relevant here.
I love it! I had this convo with many people before, I consider myself a hacker too, not the Hollywood cliche, but the concept of hacking things and repurposing them to do another goal or a new one. I get bored quickly doing the same thing over and over so sitting and writing code is a soul crushing task, rather, I have a goal to do something and in the process I learn all the new exciting stuff while building it, of course, you have to be a quick learn and adaptable to do that. It’s perfect except when it comes to get a job, most companies still follow the traditional concept of you being a cog in a machine, and whatever you are doing will be used in our work, they don’t look at the attributes you have but rather how many X years in writing Z language or FPGA design.
> How AI software is made possible through mass exploitation.
Throwing around vacuous accusations like this dilutes the effect of flagging real exploitation.
I’m not sure what his point is? This just looks like grief-posting. Keep up with the times and don’t be incurious.
If you don’t like AI then you should find a niche in software development that may still require rawdogging. I think kernel development and making hand crafted libraries are good options.
Product engineers will be more exposed to AI.
One of their primary gripes seems to be that coding via AI is not deterministic - what you get is somewhat up to chance.
But how is this any different from delegating to someone else? They're not going to produce precisely what you told them to.
In either case, you can quite obviously reduce the variability by writing better prompts/instruction, using specs, tests etc... And, likewise, you are going to need to iterate anyway.
The problem, in either case, is when you don't give good instructions, don't carefully review, etc.
So, one might say that a defining characteristic of a proper "engineer" - be it software or otherwise - is conscientiousness. If you combine that with creativity, communication skills, determination, etc then you can produce good results.