Tuhin Chakrabarty#1053
my updated perspective on AI writing is that I actually don't mind reading it. Claude in particular is pretty insightful most of the time, especially on social issues, philosophy, life advice, and psychology. But whoever is passing it off as their own work is a loser
Negative users criticize recent Claude AI writing as degraded compared to older models, often reduced to engagement bait without insight or genuine philosophical positions.
@ftlsid AI writing is worse when it's instructed to write engagement bait w/o much insight to add.
That said older models were better writers. They pruned, quantized, RL-ed and MoE-ed the new ones to hell and back, that's the result.
@ftlsid Prose is the problem. If you ask them to make a bulleted list it's great
@ftlsid Claude the engineer is typically pleasant to read and send commands to. If you try to pass it off as your own voice it’s actually a bit insulting to your audience
@ftlsid sounds insightful on philosophy because it wont actually pick a side that costs anything
@ftlsid May be you enjoy this…
Tuhin Chakrabarty#1053my updated perspective on AI writing is that I actually don't mind reading it. Claude in particular is pretty insightful most of the time, especially on social issues, philosophy, life advice, and psychology. But whoever is passing it off as their own work is a loser