This is the kind of stuff AI-skeptics have warned about daily on this very site for the last year or two. We are met with "you'll be left behind" and "you're just coping" and "the future is coming".
This shit is blatantly obvious. And if people keep pushing and pushing for LLMs to generate things that are actually used directly, the problem will grow so large that most people will barely remember a time when their job was something that could be understood.
In the lawyer's case one might argue that it's just a time quantization inaccuracy. There's no smaller unit of time than an hour for them.
One can argue that there should be, but billing in 10 second increments might also be a bit much to ask.
One would hope that they would at least aggregate multiple smaller-than-resolution time blocks into one before billing though. (Five documents over the course of a week ends up costing the same)
i think they meant the hardware becomes more fungible. This is what made IBM lose their competitive advantage when it came to PCs, because others were able to make IBM-compatible ones that are cheaper and still able to use the same old software.
Right. But viewed differently, that made everyone have to compete at actually making good PCs, instead of relying on customers unable to escape to another vendor due to lock-in.
We want this. Software and other modern tech desperately need this! How many companies are stuck with an absolutely terrible suite of MS products because you can't easily rip out and replace just the IM or videoconferencing part, for example?
Its good for the buyers, and good for broader society and the economy.
Its bad for vendors.
Why would vendors do something that is bad for them? HN is full for people talking about how to create a "moat" for their products. Why do you expect hardware vendors not to want to do the same?
Except if the moat is something like "have managed to attract and retain stellar workers who make fantastic products with excellent durability and support at a reasonable price, resulting in such a happy customer base that any upstart will struggle to gain a foothold". The horror!
I've been saying this for years – it's a huge societal problem that most people, including powerful people, view technology in general and software in particular as something slightly magical, something that cannot be truly understood. We're seemingly forever doomed to accept that technical things are a bit shitty and that we'll have to do a rain dance every now and then to appease them. Instead of, you know, throwing out the shitty products and making or buying good ones.
Lovely post, and lovely phrase (normalizing of inexplicable failures).
Agree. It's like people who dream of becoming middle managers in order not to have to do things themselves anymore.
I mean, you do you, I don't mean to shame people for whom management is the dream – I just find it weird and sad that they're trying to pretend that this is somehow how people with a hacker/geek spirit think or should think. To me it's antithetical to the hacker spirit.
I'm becoming more and more certain that there are two kinds of people in the world: those for whom the end result is the main point, and those for whom the journey there is. Generative AI has revealed how stark this divide is.
I'm not saying either type is better than the other, but the recent technological changes certainly have put a lot of wind in the "end resulters'" sails. They must not be allowed to hijack the definition of programmer, hacker, geek.
It also makes me think there's a divide on how many steps people choose or are able to see ahead.
They tell people who liked the journey that they're way is inefficient and you need only prompt to build, not seeing that in two years they're boss will be saying the same thing to them because they don't warrant the salary to pay them to do the same thing anyone with an idea can do.
> Who likes typing code, writing boilerplate, reading bad documentation for nights looking for that small thing, asking around in forums, reading dependency source code, writing trivial unit/ui tests.
Me! I do!
> And who likes planning, architecting, directing a team, steering and giving advice, reviewing code, designing interfaces and APIs. Coding became more mentally exciting.
I'd rather eat rusty barbed wire. I'd avoid your middle manager hellscape even if if doubled my salary, I swear!
> I don't understand what's difficult about enjoying programming now that LLMs are everywhere. It's akin to musicians who still play their instruments despite having the ability to stream any music they'd like. Nothing's stopping them from playing their instrument. Why is programming any different?
Because a great majority of our colleagues and compatriots – people we believed viewed the art somewhat similarly to how we do – have shown themselves willing to replace their cellos with CDs, their tubas with tapes and their vocal chords with voice-overs!
We find ourselves in an orchestra in which it seems the loud majority will happily proclaim that they're proud to just press play! It's heartbreaking.
I can understand pressing play to put food on the table. That wouldn't make me sad and angry. Seeing so many so proud and happy about it, and doing it also when the stakes are zero, does.
Good points yet how others view the art doesn't affect my enjoyment nor practice of it. Like music, others can enjoy listening to a CD or play their own instrument; I don't care. Why allow others' style and practice influence how we enjoy programming?
Absolutely not unthinkable! We're paid for exactly that in research, for example.
It's terrifying to hear that 5 hours of focused thinking is considered crazy. I can understand a budget battle if it's 5 months, but 5 hours?! Everyone, in every profession (or none), would be better off mentally if they "wasted" 5 hours thinking about something every now and then!
If 5 hours of thinking time to solve a problem is unthinkable to some people does that mean they have some form of Attention Deficit Disorder, be it a natural thing or introduced by an over-dependence on LLMs? Like stopping for an hour to think through a problem is unfathomable and instead they would rather hit small dopamine spikes every few minutes from talking to a chatbot?
Sometimes you gotta do things that others don’t value highly. A long career needs you to act with good judgment. Even - sometimes - when the optics are bad.
If you were in need of legal services in a high-stakes case, or if you needed brain surgery, would you want "fast visible progress" or "I need to think this through"?
I'd say both relevant professions are highly regarded.
A very good friend, who is not a brain surgeon, but designs and plans brain surgeries for clinical trials with the very brain surgeons doing the surgeries. So yes they do think it through and even do entire dry runs through every step.
This shit is blatantly obvious. And if people keep pushing and pushing for LLMs to generate things that are actually used directly, the problem will grow so large that most people will barely remember a time when their job was something that could be understood.
This is why generative AI sucks.
reply