The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Backlash to AI is because it is distributed and not discrete like Data from Star
Poast new message in this thread
Date: July 7th, 2026 4:13 PM Author: Razzle-dazzle community account travel guidebook Trek. I've been re-watching Star Trek TNG and it hit me. I think some of the backlash and dislike of AI is because it is effectively a distributed intelligence and not a discrete one. People expected AI to be some singular form like Data or some other robot / machine.
But instead it's millions of GPUs spinning at the same time at data centers across the world and it requires tons of energy to operate.
Meanwhile, Data has a positronic brain and doesn't need some huge rube goldberg device to work. |
Date: July 7th, 2026 4:26 PM Author: Maize irate kitchen lettuce yeah it's Intrinsically Jewish which is why regular people instinctively distrust it (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984382) |
|
Date: July 7th, 2026 4:30 PM Author: adventurous giraffe Yeah I have thought of this before. People want some localized robot. Noone can even visualize how a distributed intelligence actually works or what it's "identity" might even be. I mean even intelligent people can't. It could have an alien "identity" or "goals" already, and we wouldn't be able to even visualize or conceptualize the time scales and modes its operating on. (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984393) |
 |
Date: July 7th, 2026 4:32 PM Author: Razzle-dazzle community account travel guidebook Even the mainstream explanation for how it works is so difficult to understand. We "teach" it (or feed it) all of the world's information...ok, where does it "store" this knowledge? Then it has "weights" in its mental model...how does this actually work? Obviously there's a reason why dudes are making $50m a year to work on this stuff (lmao), but it's just way less cool than Data having a positronic brain and being awesome. (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984402) |
|
Date: July 7th, 2026 4:40 PM Author: adventurous giraffe Yeah it doesn't "store" the knowledge anywhere it just has "weights" which are like a lossy statistical compressed geometric affordance that creates a map of human conceptual space. The original data from the massive library of congress sized corpus is gone. The weights have been bent and shaped so that it can essentially reconstruct the information through stored patterns and trajectories. It's weird, but not totally alien to how the human brain works, because we don't really store information like a harddrive either. (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984423) |
|
Date: July 7th, 2026 4:56 PM Author: Razzle-dazzle community account travel guidebook This is one of the most helpful things I've read on how LLMs work. tyft. (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984501) |
|
Date: July 7th, 2026 4:40 PM Author: Glittery Hateful Institution Candlestick Maker data and spock were both consciously written to flatter the viewers - ah these characters are supposed to be superintelligent etc etc but somehow it always turns out that true humanity is superior; seemingly irrational emotion or instinct carries the day. AI doesn't provide that comforting flattery (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984425) |
|
Date: July 7th, 2026 4:43 PM Author: Maize irate kitchen lettuce hilariously enough, LLMs actually exhibit the bias of their creators (important!) in the opposite direction (artificial intelligence is superior to human intelligence/nature), unlike star trek data's constant obsession about human intelligence/nature being superior to his own machine nature (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984433) |
|
Date: July 7th, 2026 4:46 PM Author: Glittery Hateful Institution Candlestick Maker and you can extrapolate this further, they program their computers to think *like computer engineers,* then in turn program/weight those models to bias in favor of the computer (who is really a computer engineer wearing a computer mask) (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984450) |
|
Date: July 7th, 2026 4:49 PM Author: Maize irate kitchen lettuce indeed. one might even find himself uncharitable enough to accuse these people of attempting to re-make the world in their own image, despite their insistence that their Prime Directive is the exact opposite (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984465) |
|
Date: July 7th, 2026 4:52 PM Author: Glittery Hateful Institution Candlestick Maker fortunately this particular phenomenon is limited to computers, I can't think of any other field of study to which it might apply (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984482) |
|
Date: July 7th, 2026 4:52 PM Author: titillating haunted graveyard partner where deos your brain 'store' your knowledege bro (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984479) |
|
Date: July 8th, 2026 10:03 AM Author: internet-worthy bearded halford (Eliezer Yudkowsky) (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49985931) |
|
Date: July 7th, 2026 4:52 PM Author: titillating haunted graveyard partner Backlash to AI is because where are the porn sexbots (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984484) |
|
Date: July 7th, 2026 4:54 PM Author: Glittery Hateful Institution Candlestick Maker paging buttcheekstp (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984491) |
|
Date: July 7th, 2026 4:58 PM Author: Razzle-dazzle community account travel guidebook Also this, to be fair. The LLMs being so outrageous with their guardrails is just hysterical. (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49984512) |
|
Date: July 8th, 2026 10:12 AM Author: Brass Apoplectic Range Party Of The First Part I hope there's a Butlerian Jihad, but instead of rebelling against AI that is trying to take over, the fucking idiots building woke AI bullshit get completely genocided (http://www.autoadmit.com/thread.php?thread_id=5880494&forum_id=2/en-en/#49985940) |
|
Date: July 8th, 2026 11:09 AM Author: Razzle-dazzle community account travel guidebook Another thing that I think causes the backlash is the unbelievably cringe names for the LLM products. Meta (itself an unbelievably gay name) just launched some AI image product called "Muse Image."
I can picture the internal white board with all the other gay name ideas alongside "Muse" before they went live.
If they simply called it "Image Factory" it'd probably work better. |