Will digital intelligence replace biological intelligence? Prof. Hinton, the pioneering figure on AI presenting at Sheldonian Theatre, Oxford.
What do you think?
He talked about the issue of Human existence!
#threatsofAI#futureofhumanity
The problem of the neuron network is it can only adjust the weight, and its structure and dimension can only be changed by AI experts in next iteration. It fails to respect the "First Principle of Thinking". Hyuge Intelligent System proposed AGI - Artificial Genomatic Intelligence won't have these problems.
https://lnkd.in/gTQNpcbB
Is generative AI becoming so powerful that it poses an existential threat to humanity? Watch this video to find out what one of the godfathers of AI thinks...
https://lnkd.in/drA-A5eB
I came across this video Geoffrey Hinton's 'Romanes lecture'. He's known as one of the godfather's of AI from his early work on back-propagation in neural nets in the 1980's.
Suffice to say he's a deeply credible character and worth listening to about his hopes and fears for AI and where it is heading.
It's pretty academic (well he is giving a lecture in the Sheldonian Theatre at Oxford University, so gowns and ceremony in full evidence!), but so rich in ideas.
https://lnkd.in/dGvEQdjd#generativeai#llms#cognitivescience
Here’s a fantastic, if scary, overview from Prof. Geoffrey Hinton of how AI works and understands – and the various risks this presents, from fake content, job losses, mass surveillance, lethal autonomous weapons and discrimination to the biggie: the existential threat to humanity. This is why we launched RAID in 2016, and now the need to regulate and adopt AI safely is greater than ever. Join us at RAID Virtual on 30 April and RAID Brussels on 23-24 September – register at www.raid.tech.
We learn from Prof. Hinton that a European Commission Vice President agreed with him that a superintelligence would always want to get more control. There will be no switch to turn superintelligence off, and if self-preservation causes superintelligences to start competing with each other, evolution would make them tribal and aggressive, much like chimpanzees and humans.
Today’s digital models are "very close to as good as brains and will get to be much better than brains". They are inherently better at sharing learning and knowledge with each other than humans are.
In the next 20 years, digital computation will “probably” be smarter than us, and “very probably” much smarter than us in the next 100 years. And there are very few examples of more intelligent things being controlled by less intelligent things.
With many thanks to Prof. Hinton for the insights.
#RomanesLectureUniversity of Oxford#AI#LLM#GenAI#AIrisk#ExistentialThreat#FutureofWork#FutureofHumanity#deepfakes#surveillance#aibias#AutonomousWeapons#techregulationhttps://lnkd.in/eZCcYiSG
Another good talk on AI and juxtapositions with human decision making.
While I know many who will disagree with some of the statements made, I think with all the AI-maximalist talk that dominates the airwaves, it's good to keep anchoring ourselves around the many flavours of why, in order to decide where to next...
36min, suitable for 1.25-1.5x.
#ai
Please take a moment to watch a lecture that is both enlightening and frightening, presented a few days ago by Prof Geoffrey Hinton, titled "Will Digital Intelligence Replace Biological Intelligence?", during the Romanes Lecture at Oxford University. Prof Hinton, "the godfather of AI", guides us through the current dangers posed by AI and the possibility of machines outpacing human intellect—a scenario that could lead to the extinction of humanity.
https://lnkd.in/eqBraeEm
What is mortal computing?
This fascinating lecture by Geoffrey Hinton covers the basics of neural networks, backpropagation, and LLMs and discusses why he believes they are intelligent.
He also discusses what will happen when hardware and model weights are intertwined to be much more energy efficient and, therefore, mortal.
He left Google for a reason. He wants to talk openly about the dangers and risks. He doesn't seem melodramatic, and his arguments have merit.
I am generally more optimistic, but his views are sobering and worth consideration.
https://lnkd.in/eEEMfsFh#ai#llm#neural#network#humanity