It's very dishonest to me to liken LLMs to something like the railroad, steam engine, electricity, etc. as though a software autocomplete machine does anything to elevate the standard of living, despite the fact that the proliferation of software has coincided with a general decline in the quality of life, at least in much of the west. It's obviously not causal but it's certainly true that material energy/industrial/biological problems won't be solved by producing more software faster, since that hasn't been the bottleneck in decades. When it comes to the average person, the roadblocks to a quality software experience have mostly been with discernment and organization rather than generating code per se anyway.
I think the negative responses people have to AI are often really inappropriate (or at the very least disproportionate) and are the result of anxiety with the fact that very serious issues are being dismissed in favor of what is basically a toy outside of the hands of already capable people in industry and academia which make up a relatively small share of the "total addressable market" that the big labs purport to have as the basis for commandeering the entire U.S. economy. The fact that it's a) basically the only thing we have going for us and b) promoted by intensely unlikable people, is bound to ruffle the feathers of people who don't need whatever capabilities it may have and who have to deal with all the friction an increasingly slop-filled internet entails. Considering how much of life is online now for the majority of people, that last part probably isn't as much of a triviality as it sounds to me writing it out. My perception is that there's definitely also an element of resentment that arises from people who have so many tools at their disposal and yet don't really even try to do anything with them.
TF2 and especially TFTV is such a small and dying community that I don't think the costs of having to dismiss or ignore TF2 statistics & News & Pugsite #1723678913 or retarded slop like TF2-in-browser warrant treating interesting projects with hostility. I don't have a use for it but the Demo Salvage doesn't deserve outright hostility, and it makes sense that the author might be prickly in this thread given that was the response to his relatively benign project. On the other hand, AI boosters underestimate the degree to which the value of something (especially in niche communities like this) comes from the person behind it rather than the thing itself. If someone were to manually rewrite TF2 for the browser, it would demonstrate an intense love of the game, dedication large amounts of one's finite personal time, the ability to reconcile complex systems, etc. even if it is completely and utterly useless in and of itself. Telling an LLM to do it means that you can't ascribe that meaning to the author and leaves you with nothing but a completely and utterly useless product. There's often a huge disconnect between the person who asked the LLM to make something broadly useless (to whom it's not useless, it's a great piece of entertainment) and everyone else, who finds it both pointless and meaningless. Projects that aren't useless can still be disheartening because the basis of our socialization is interactions with one another rather than the material outcomes of the interactions themselves, focusing on which would be considered antisocial or sociopathic.