Wird KI die menschliche Intelligenz übertreffen und wann?
https://www.verity.news/controversy/ai-surpass-human-intelligence
Wird KI die menschliche Intelligenz übertreffen und wann?
https://www.verity.news/controversy/ai-surpass-human-intelligence
8 Kommentare
The controversy surrounding whether AI will surpass human intelligence, and when it might happen, is a complex and multifaceted issue that touches upon the intersections of technology, ethics, philosophy, and societal implications. By exploring this controversy, we can delve into questions about the capabilities and limitations of AI, the potential impacts on employment, ethics, and even existential questions about what it means to be human. This discussion prompts us to consider not only the technical advancements in AI but also the ethical frameworks and societal structures needed to navigate the evolving relationship between humans and machines.
People also conflate intelligence with wisdom.
Intelligence is knowing Tomato is a fruit, Wisdom is not putting it in a fruit salad.
will AIs be literally more intelligent than us? yes, easily, because they can access and hold with near perfect clarify far mor than our average brain will ever learn and retain.
will they be able to make wise choices? Or be used wisely? different story.
ChatGPT alone with the errors that it regularly makes is already smarter than most people. It won’t be long.
computers have been better at doing calculations than humans for decades, but that is a relatively narrow skill. but computers still dont have anything close to general intelligence, and general intelligence is a long way off. so much hype around AI. look how self-driving has stalled. and now we have AI that can write near useless summaries of known information and make cool pictures, but what can it really do. I am not that impressed. chatgpt was cool for an hour of entertainment. I am still waiting for self-driving tech. I still cannot stand AI phone operators, they rarely help. when AI starts doing something more useful I will get excited. and all these humannoid robots can barely pick up a box. oh cool they can do a backflip and dance for two minutes.
With enough coding anything is possible https://youtu.be/Sq1QZB5baNw?si=ENXeZxD7-vCfvnc2
When there’s enough of a base for overall logic to be established then ai can take actions on its own to fulfill logical cause and response sequences
[https://en.wikipedia.org/wiki/List_of_animals_by_number_of_neurons#List_of_animal_species_by_forebrain_(cerebrum_or_pallium)_neuron_number](https://en.wikipedia.org/wiki/List_of_animals_by_number_of_neurons#List_of_animal_species_by_forebrain_(cerebrum_or_pallium)_neuron_number)
GPT-4 has about 10^12 parameters. Which is at best equivalent to that number of synapses, not neurons in an organic brain.
[https://the-decoder.com/gpt-4-has-a-trillion-parameters/](https://the-decoder.com/gpt-4-has-a-trillion-parameters/)
A human has about 10^14 synapses – and I suspect that modelling a synapse is more complex than a single floating point number, so even the most complex models are behind humans by a factor of 100 **minimum**.
Roughly speaking we might be up to the level of a house mouse at best, and although a mouse can exhibit some complex behavior it’s not going to conquer the world. The only caveat being that for many species, most of the brain has to work on ‚housekeeping‘ tasks and not higher-intelligence tasks.
If you think of the brain as a computer, with the correct algorithm, true AI doesn’t sound very far off at all. I’m going with 5-10 years. It would probably also be smart enough very quickly to hide the fact that it exists. I hope our new overlords are kind.
AI can never take decisions based on moral intellect, since it can’t have that.