If you don't expect the unexpected, you'll have nothing to write about.
The reason I haven't written much about Devin lately is not because I don't use it anymore, but because I use it on a regular, daily basis and there are no more events that I particularly want to write about anymore from Try Devin.ai 4/1~.. I thought the factor "As quality improves, it becomes invisible." was "When you start doing a lot of AI work, you can't talk about AI," but it's not just that, it's that when you think, "If I do X, it will be Y," and you do it, and it actually becomes Y, he as a person, it's so obvious that there's no motivation to write about it. As the user becomes more proficient in the technique, the estimate of what it can do becomes more accurate.
For example, when they first ride a bicycle, they will say, "I can ride a bicycle! And for a while, they will say, "Now I can ride a bike, I can go to places I couldn't go before! But eventually, the fact that I can ride a bike and that I can go there by riding a bike becomes natural, and I don't feel like talking about it anymore. The 4o illustrations were also more than I initially expected.
I use ChatGPT every day, and Devin almost every day, but I don't bother to write about that.
Even when the AI does something unexpectedly stupid, unpredictability gives you an opportunity to write about it. Unexpectedness decreases as behavioral predictability increases.
After this "signaling period of beginning use", the opportunity to know decreases
---
This page is auto-translated from /nishio/意外性がないと書くことがなくなる using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.