The Bonfire of the Strawmen
Watching the debate about AI on social media is like watching a million strawmen arguments set ablaze. Some of the arguments against AI I see are (some paraphrased):
“I can’t get 100% perfect results, therefore AI is useless for everyone, everywhere.”
“AI uses so much power that it’s just evil.”
“Deepseek proves that everybody else has no clue what they’re doing.”
“AI is just going to reproduce popular misinformation and present it as fact to rubes.”
“It’s just a better search engine.”
“It’s not good at _________.” (Some thing nobody designed it to be good at)
“It’s Elon’s fault.” (This is not a deep argument)
“It's making my students dumber, getting my friends fired on flimsy pretexts, being used by billionaires to destroy people’s careers” (that’s a literal quote)
“I've just been asking it about my mum, and every response it's giving me includes incorrect information.” (another literal quote)
“I don't trust AI because it has been built by profit-driven, entitled, and unethical people.” (yet another literal quote)
“Groundbreaking BBC research shows issues with over half the answers from Artificial Intelligence (AI) assistants” (Actual BBC headline)
“my mom discovered ai slop and she won’t stop posting.” (OK, they have a point here)
“I will NOT even start to respect AI until it gives me some Aretha Franklin instead of Justin Timberlake when I ask for songs sung by black women.” (What LLM is she using?)
“AI is possibly the most unnecessary tech buzzword to date and it serves absolutely zero need or function to modern society.”
“It’s just a fancy search engine, nothing more. There’s no ability to reason, to think, only to replicate. It can’t create, only regurgitate.”
There’s a lot of people mad at AI who are working hard to convince themselves that AI just is useless if not outright dangerous. But that’s just not true, it’s not useless.
There is no doubt that LLMs have limits, and if you push them beyond the limits you run into trouble. And if you are unaware of the limits you can get into trouble (and even be dangerous). Sadly, AI vendors have not done a great job explaining where the limits are, preferring to make glittery statements about how it’s smarter than Ph.d students and the like.
I get that people had expectations that everything with AI was supposed to be effortless and magical. When they discovered that it could take a lot of effort and often had little magic, they felt deceived.
The trough of disillusionment is upon the public. Let’s hope it doesn’t spill over into business.