Ask HN: GPT4 Broke Me

It's super useful for finding out how things work like asking it how to use an SDK to do something it works really well for it even provides stuff I searched the docs repeatedly for.

But for complex logic that I want to be too lazy to do, it literally just wastes my time. I tried for an hour to get it to do something and tried to modify the code to get it to work. Ended up just doing my own version completely.

Stop reading tech for a few weeks and you will get another perspective and you won't feel so neurotic.

A professional translator I’ve known for more than twenty-five years has, within the last few months, seen all of his sources of work dry up. He worked in a field that required specialized knowledge and experience and for which there was steady demand, but there was little human interaction: clients would e-mail him texts, and he would e-mail back the translations. He is angry and upset because he believes—probably correctly—that he can produce more accurate translations than GPT-4. But the price difference is so great between machine and human translation, and the average quality difference now is so small, that his clients have dropped all of their human translators. They have offered him some work checking machine translations, he says, but the pay is much lower. He is now, in his sixties, trying to start a new career.

I myself was a professional translator for twenty years, until 2005, when I took a university job. Most of the translation work I did—for which I earned a good income—can now be done much faster and more cheaply by LLMs. In the mid-1990s, I started also working as a lexicographer—writing and editing entries for bilingual dictionaries. That was difficult work, but I enjoyed it and felt proud to be able to do it. Before I retired from my academic position earlier this year, I had been thinking of returning to dictionary work part-time in my retirement. There’s no future in that now, though, because LLMs have pretty much eliminated the need for human-edited language dictionaries.

Both my daughter and the daughter of a friend of mine are freelance illustrators. While both of them are continuing to get work now, they are following developments in Midjourney etc. with growing worry.

In contrast, a couple of weeks ago I got together with an old friend about my age whom I hadn’t seen in a while. He worked for many years as a glazier, and since he retired on a good union pension he has kept busy repairing regulators for scuba-diving equipment. I tried telling him how excited and worried I was about AI, but he was dismissive. It took me a while to understand why: The work he has done—installing windows and mirrors, fixing precision devices on which people’s lives depend—is not likely to be done by machines any time soon. It’s those of us who have done most of our work in front of computers who need to worry.