On the usefulness of AI…

Today I gave AI a chance. It didn’t go well.

It’s not the first time I give it a shot. In previous attempts I have found that AI, in that case CoPilot, is OK at coding, as long as you personally figure out the exact algorithm you want to turn into a python script and tell the bot exactly what you want it to do step by step. So I’m not necessarily opposed to the technology altogether.

However today I tried to get the Claude bot to translate a document that was maybe half a page long. I sat there and after a while thought “well this is taking a while”. But I was patient. And after several minutes, voila! I got a message that I was “out of messages”.

And even worse, there was seemingly no way for me to save the prerequisites and strategy for the translation so that the next time I have to deal with the same situation, Claude wouldn’t have to do all this extra processing.

The source and target languages are very common. The file format was very common. The amount of text was frankly trivial. And I have nothing to show for all this. At least if a human was working on this I would have had at least a partial translation.

Words fail me in expressing how much of a disappointment this whole experiment turned out to be. This is why companies are spending hundreds of billions of dollars? This is the “future of work”? This is why it’s practically impossible to find a job in the software industry these days?

What a sh*t show.