ThaBigChirp (99dd2c) in Ai1 mo ago

New Anthropic Ad Implies AI Could Kill Us All

Open link at futurism.com

New Anthropic Ad Implies AI Could Kill Us All

A new commercial from Anthropic about asking the "hard questions" about AI implies that the tech could kill us all.

100
💬 4

Comments

  • concerns over AI ... "erode or ability to think"

    this is absolutely what I see happening everywhere I see Ai in widespread use. People will never turn down something that allows them to be lazier and this is laziness on hyper max strength steroids. If you're proficient in using Ai, you can run it in circles so that you don't even have to be involved like 90 plus percent of the time.

    Where I see hope is when you find something you really really love working on, and you use Ai to help you - as long as you stay curious and inquisitive so that you can truly be the creative force behind what Ai is focused on - it can be powerful and freeing.

    The track record of people taking the easy way out is where you gotta scratch your head and wonder if people are going to outsource all of their thinking at some point - but even in that I see opportunity for the small percentage of people who fight against it and truly find inspiration and purpose through their use of Ai to accelerate their creative juices.

    Coincidentally, I see @The_Weird_Canadian posting their music all the time and you can really tell that Ai has unlocked something in him.

    +3
    • Post creator action: highlighted

      AI in its current form (LLMs) is unreliable - it can give wrong answers and make mistakes, but not by design, it's simply a result of imperfect training data and inference algorithms. The people who control LLMs are the dangerous ones. They have their own agendas, and we cannot trust them unconditionally because, after all, they are just human - imperfect and fallible.

      I have nothing against using AI to finish a job faster or better, to unlock skills I didn't even know I had, or to make my ideas a reality. But that is fine only as long as I am in command. The moment I delegate my thought process, my creativity, and my agency to an AI to do things for me, problems arise. I slowly begin to lose myself. Instead of being the driver, I become just a passenger. That is neither good nor something humanity should strive toward.

      In the long run, as people delegate more and more tasks to AI, because, let’s be honest, humans are inherently lazy and prefer to achieve goals with minimal effort, we will create a culture of black boxes and cargo cults. We feed inputs into a black box, receive results, and mimic operations without understanding what happens behind the scenes. If something breaks, we simply hope the machines will fix it. One day, there will be no one left who is actually capable of fixing these systems, let alone creating new ones.


      This shift seems inevitable. We have already seen this with calculators (people rely on them for basic arithmetic, and younger generations often struggle with long multiplication or division by hand). Kids don't read books anymore because it is 'boring and slow,' preferring summarized digests instead. We see it in programming, too: 40+ years ago, I knew exactly how a program would run, even down to predicting the exact CPU instructions. Today, young coders often don't even know why a specific pattern is used, they simply copy those because 'that is how others do it.'


      The decay and loss of institutional knowledge is inevitable - not because we won't build new systems, but because we will no longer have the need to understand anything, nor will we be capable of comprehending the massive outputs of these machines. Cognitive decline is almost certain.


      Therefore, AI as a tool or an assistant is highly beneficial, but it must never be allowed to take the driver's seat.

      N.B. Not native English speaker, rarely have opportunity to use English to write/speak so don't judge me too harsh if I'm a bit incoherent.

      +5