To write a few paragraphs°hard word on a topic°hard word, there is (roughly) a four-step°hard word process:
- Choose a topic°hard word
- Create an argument
- Choose the tone and style
- Write the words
The machine is, in many situations, better than I am at the task of writing the words. 💡 it still struggles°hard word with a few tones It is always about 20 times faster than me.
However, it struggles°hard word with the first three tasks.
- The "agent°hard word" framework°hard word isn't focused°hard word on self-willed°hard word agents°hard word. This is (probably) a good thing. But "fiduciary°hard word" agents°hard word aren't happening either yet°hard word. These solve°hard word some of the problem related°hard word to motivation°hard word.
- The machine prefers to create a good-sounding°hard word argument to a logically-sound°hard word one. 🔥 of course, many humans do the same thing.
- The machine defaults°hard word to a tone that has been over-used°hard word to the point of annoyance°hard word. It is obsequious°hard word and aggressively°hard word cheerful. In the context°hard word of "a chat-bot°hard word for people unfamiliar°hard word with the technology°hard word", this is a defensible°hard word choice. But, most of the time, you need your own tone. 💡 It has been long enough that the training data allows "write a tone prompt°hard word for this query°hard word" kind-of°hard word works.