Raw LLM Responses

10:00 it sounds like he's trying to say that prompt manipulation is similar to clear communication, more precise wording, or even just being aware of your own emphasis. 13:30 YES. It's just "knowing that other people can't read your mind" and "I am fallible, therefore if I need to change something if I dislike the outcome".  14:45 omfg the thigh lighting is so wrong. I'm not great at colors but jfc. 28:00 this feels like it encapsulates capitalism as a parasite on human wellbeing. 33:00 omg yes. There is no way to be polite about Shad's arrogant misappraisal of his own artistic skills. It's okay to be imperfect.  I have to keep skipping his talking because the egotism is nauseating tbh. Therapy is a public service I s2g. 49:00 yet again "AI prints money" is its own argument for UBI and UBS. If you remove the need for subsistence as a perverse incentive, then people act better. Universal Basic Income and Universal Basic Services are the natural next step in societal sustenance. He also ignores that progress, like entropy, tends to increase as you work on the art (until you reach burnout or boreout). If you spend your time on the menial tedium of fixing mistakes you didn't even make, then what joy is there anywhere? They don't accept transition as an end-goal, or experiential intent as being able to have meaning. They think "good" means something is highly detailed, takes a very short or very long time, is completely exceptional, or otherwise "obviously" good. If all you expect to see is hyperrealism and everything else is a bad imitation, then yes, of course you're going to be miserable and hate every step of any artpiece. (Same thinking as "your wording is imprecise or ignores a niche edge-case, therefore your entire argument is wrong". Yes, valid and necessary for programmers working with literal binary and hard rules. Yes, annoying af you have to translate manager-speak and allistic self-importance into discrete parameters. Still wrong to throw out the starting framework if the assumptions and how the person reached them could be useful information themselves.) Also, can you share the public domain-only trained AI please?