An Anthropic study released last month found AI is already theoretically capable of completing the majority of tasks associated with computer science, law, business, and finance, and other major white-collar fields
There’s a huge difference between “capable of completing the majority of tasks” and “capable of completing the majority of tasks WELL”.
Sure, you can have an AI code your web app or mobile app, due example. But it will be riddled with bugs, and bloated with inefficient code.
And from what I’ve seen, it’s not getting noticeable better at that.
But the AI companies won’t acknowledge that, of course. They will continue selling the snake oil that cures everything that ails you.
Honestly, the only reason I use that LLM shit now is because the job market in tech is getting a bit like the hunger games and my employer strongly encourages its use - and even then, the vast majority of my usage is “fancy search engine”. Even the boilerplate it gives me sometimes is like… really weirdly styled and quite often has to be corrected to be fit for purpose. I cannot understand people who just slap that shit in without even bothering to check it.
I cannot understand people who just slap that shit in without even bothering to check it.
Really? Because that’s exactly what I do. I’m metered by my spend, so I produce just absolutely wholesale volumes of slop, don’t review it or even look at it and just am like fine here you go, this is what you want. No one else reviews it, it gets piled on top of other slop, shits going to really start piling up and breaking. Execs won’t care, they’ll start firing people anyways, it’ll continue to pile up, problems will multiply…
I mean the alternative is do nothing, get fired. Or contribute to a downfall, survive a little longer maybe, still get fired. Or fix it, produce really good work at antagonizing levels of fixing everything, and still get fired. So what’s not to understand here?
Well, I do my best to fight the good fight and make systems that aren’t impossible to maintain, because I work at an oncology biotech and despite all its flaws it’s a very compelling and worthwhile mission to me overall. So I do my best to not treat things too transactionally.
From the article:
There’s a huge difference between “capable of completing the majority of tasks” and “capable of completing the majority of tasks WELL”.
Sure, you can have an AI code your web app or mobile app, due example. But it will be riddled with bugs, and bloated with inefficient code.
And from what I’ve seen, it’s not getting noticeable better at that.
But the AI companies won’t acknowledge that, of course. They will continue selling the snake oil that cures everything that ails you.
🙄
Honestly, the only reason I use that LLM shit now is because the job market in tech is getting a bit like the hunger games and my employer strongly encourages its use - and even then, the vast majority of my usage is “fancy search engine”. Even the boilerplate it gives me sometimes is like… really weirdly styled and quite often has to be corrected to be fit for purpose. I cannot understand people who just slap that shit in without even bothering to check it.
Really? Because that’s exactly what I do. I’m metered by my spend, so I produce just absolutely wholesale volumes of slop, don’t review it or even look at it and just am like fine here you go, this is what you want. No one else reviews it, it gets piled on top of other slop, shits going to really start piling up and breaking. Execs won’t care, they’ll start firing people anyways, it’ll continue to pile up, problems will multiply…
I mean the alternative is do nothing, get fired. Or contribute to a downfall, survive a little longer maybe, still get fired. Or fix it, produce really good work at antagonizing levels of fixing everything, and still get fired. So what’s not to understand here?
Well, I do my best to fight the good fight and make systems that aren’t impossible to maintain, because I work at an oncology biotech and despite all its flaws it’s a very compelling and worthwhile mission to me overall. So I do my best to not treat things too transactionally.