There's no agent no agency. It's just auto-complete. It's fundamentally not intelligent, no problems are being solved.
Chat-gpt is garbage for search because it doesn't evaluate or compare.
]]>But that's about it. For about 10k US you can merge some of these features, and for 200k US maybe all of them.
The LLM'S are not going to wake up, they're not AI, they're also not good search engines. LLM'S aren't man-splaining at scale they are auto complete at scale. Really good for regularizing some light scripts, pumping out the most predictable scenarios (i.e. porn), or repeating what a lonely person needs to hear (but what do you think about that?) But pretty shit at everything else in the same way the viral "auto complete this sentence" trends of social media are hilarious and absurd.
The de-noisers are great tools for creating art! For experienced digital artist they are a godsend. For amateurs they enable the creation of plausibly nice to look at pieces, but they are trying to do the same thing as the LLMs, take a prompt remove all of the noise from static oerba training regime. The SD's aren't creative, and certainly don't understand subject material all they can do is create a probabilistic image out of random image substrate, which is why they create so much nightmare fuel, for real this is what 10,000 fingers look like in superposition!
There are absolutely twchno-evangalists, prosperity theology futurists and hucksters, buuuut in a more limited way these tools are accessible, and powerful. Ignore them at your own peril.
]]>