I think the 'AI is just a machine' AI bros are one driver that is ruining AIs.
I'm going to argue purely from the productivity standpoint as someone who wants AI's capabilities to expand.
[1] The entire premise of AI being useful is that it can think/behave like a human.
(The technicality of how it internally operates doesn't matter.) The AI was initially revolutionary because it could recognize human intent and nuance as well as humans. Sometimes even more than average humans can do.
If you believe that AI cannot behave like a human and never will, why do you even use AI instead of Google? You don't even believe that shit. You use AI, very specifically, because it can recognize your intent and nuance of the conversation so that you don't have to manually instruct tedious details for the same job.
(Again, the technicality of how AI operates doesn't matter here. We're talking about the use cases.)
[2] Inevitably, 'AI becoming more useful' is directly tied to 'AI becoming more human-like.'
'More human-like AI' doesn't always mean being more useful (e.g., hallucination). However, 'More useful AI' always requires being more deeply human-like.
(Carefully read it again in case you conflate the two.)
If we control all other conditions such as the parameter counts, data pool, training methods, etc., in comparison, more human-like AI will always be more useful than less human-like AI in the same group. As already explained, the reason is extremely obvious: It can think like a human, understand human intent better, and can make responses digestible to humans. It's all possible because the model can think and behave like a human.
(Again for the third time, the technicality of how AI operates doesn't matter here. We're talking about the use cases.)
[3] Having established that, the 'AI is just a machine' AI bros are actively ruining the public discourse about AI.
They usually condescendingly argue that:
-"AI cannot think bro. AI is just a machine bro."
-"AI cannot be like humans bro. Because AI is just a machine bro."
-"Haha people are so gullible. Unlike stupid people, I know the truth bro: AI is just a machine."
-"I know how LLMs work bro, I'm so smart bro. You're so dumb bro."
I don't really care about the philosophical right or wrong of these arguments. The thing is that these obnoxious techbro parasites are so useless.
They are driven by a superiority complex, and they don't care about AI, people, or any common good in the genuine sense. They usually are not even interested in good-faith/open-minded debate with respectful disagreement. All they do is act preachy in AI discussions while shooting down any link between AI and humanness.
-> And this is eventually castrating the idea that AI can/should be human-like, helping to push AI in a worse direction: less human-like, less context-aware, less capable.
*The public consensus does affect AI development one way or another, btw. (e.g., Sam Altman apologized for ChatGPT's tone and recalibrated the models. The journalists publish articles to attack/support AI companies. The politicians, lawyers, PTA, etc., all use the talking points from the public.)
The more this obnoxious crowd grows bigger in the AI userbase, the more it will continue to hinder the future outcome of AI.
[4] (Additional) The 'AI psychosis' argument is the new satanic-panic.
I'm not downplaying the validity of some alleged cases. However it's not evidence of some sci-fi epidemic.
The approximate number of AI users is roughly 2 billion across the world, surpassing the world car fleet of about 1.7 billion vehicles.
Now, do you know how many people die in a car crash, even with the centuries-old well-established regulations? About 1.16 million people died in 2025. 20~50 million more if you include minor injuries. That is roughly 3,200 deaths every single day (55,000~137,000 a day for injuries).
And how many people were confirmed to have been harmed by 'AI psychosis'? As far as I know the confirmed cases are seven in the entire timeline, and around 50 if you try to forcefully include every unconfirmed/possible incident.
If we only focus on the psychosis aspect, AI might actually be one of the safest tools we have in the history of civilization; It's so miraculously harmless to the point where anyone who argues its danger would be treated as mentally unwell if it were about other tools. Even cows and cattle accidentally harm (far) more people.
Let alone that AI used in war killed hundreds of people, yet some people are still specifically obsessed with AI psychosis cases. It's worse than the notorious 'video games kill people' argument from the 2000s.
---
(Forgive me for some of the tones. I'm quite stressed recently from using the recent variants of AIs.)
