Tesla CEO Elon Musk said the development of artificial intelligence (AI) will advance to become smarter than the smartest human by 2026. During an interview with X Spaces, Musk told the Norwegian hedge manager fund Nikolai Tangen that the next version of Grok, an AI chatbot from his AI startup, will be trained by May. In the latest edition of Artificial Intelligence: Decoded, BBC News explores Musk’s bold predictions. Subscribe here: Read more on the BBC News website: #US #Technology #BBCNews
Does not matter if it's on schedule ……………It will still happen – 100%, as how the fiction became non-fiction.
It already is
No IQ so far. It is very easy to test, give it a problem an let it find a solution or two NOT lookup it's data to see solutions already done.
If the AI needs heaps of examples (Tesla FSD) then it is simply reacting based on 1000s of similar situations, it does not solve situations.
Thinking is going outside the box, not connecting dots within the box.
Comment section full of sad people, I on the other hand think Elon Musk is projecting all of humanity into a brighter future. Our destiny, some would call it.
Fembot tried hard to ruin one more thing
AI like baby learning to became more adult and smarter no joke
good mornicg sir madam When will be commercial treatment for Retinitis Pigmentosa by Elon Musk Bionic eye ? Please Please answer this questions wait for your information video is very very nice and kandly i am from nepal
You Not me Dia bukan aku tidak mau kerja sama kamu aku bukan Robot dn aku manusia kamu jngn bikin issuz dn kamu jngn bikin fitnak dn aku ga suka hoax kamu sering bikin issuz aku ga akan kesana 0ke
How BADASS is Stephanie Hare? The best questions come from her and she holds Samir accountable for making his AI bot a sexy woman.
A million bucks + a comb
Smarter than Musk?
It shpuld make a lot of money for some….
AI will be used to make money from money and for defence.
As a pensioner I do not see the need for it as long as I have the internet….
AI is boring, I prefer to use my human intelligence and go outside to dig holes in my garden. Elon Musk is simply promoting his business so that dumb Americans will invest in it.
AI, in order to improve its performance and prevent undesirable consequences, must continuously interact with “effective rules and stable principles in the realm of existence”.
X Mohammad Rahim Jamshidi
🤔tech, it is now – 🤷♀️computers? Hello?
A narcissist big mouth.
Ai is already smarter than average human because it has all the internet on its hand.
Definitely smarter than Liberals right now.
Calculators have been smarter than humans for some time…
AI is already smarter than human
If we regulate AI and AI becomes more intelligent, as Elon says, then won't AI create its own regulations on what it can or can't do, completely ignoring human input.
Elon has lost all credibility – clearly he's proven to be an absolute idiot on twitter or X whatever
But will it have the emotional intelligence of humans?? I sure hope not!
We blow babies up if someone in a suit tells us to. We are morons
Haha 😂😂😂 tell him to get his prediction about self-driving car right in the first place!
Will AI surpass human intelligence?
I don't care!
There are already presumably hundreds of millions of people who have higher level of intelligence of some kind than mine.
Will this inorganic superintelligence help me and my community to progress and develop our lives?
Well unlike these greedy and supposedly high IQ people, I hope so.
Ai can cure cancer, be your attorney or accountant, drop car accidents to near 5%, diagnose disease etc
I like Elon Musk in like many other but he does talk out of his arse. Big time !!
We should have been on mars this year according to his so called predictions. Self driving cars by 2019(no where near at the level he claimed)
Starship will be full orbital by 2022. At this point I call him a liar, not mistaken.
But I like him, he is doing well you can’t argue with that.
Musk isn't very clever. I asked him for a free Tesla via twitter and he didn't give me one.
Not very smart .
Can I say, the random civilians speaking from the three minute mark seem very astute. Whenever a camera crew goes wandering around interviewing Americans on the street, they usually come across as clueless….nice enough, but outta touch. What's going on in Britain that's missing in the US educational system?
Pathetic
Anyone els watching this from Mars?
This guy is a fool he doesn't know anything about AI besides what he reads…
Elön Musk is always too optimistic. But it may help him to increase the value of his companies.
Even faster
Absolutely will be smarter than him
Apple’s New XKE Engine:
A Bold AI Tech that Puts Humans and Biology First
Today, Apple unveiled its groundbreaking new AI program, the eXtended Knowledge Engine, or XKE. Unlike traditional AI, XKE emphasizes human expertise and biological wisdom.
Apple’s spokesperson:
“So-called AI is premature and primitive. It neglects the ultimate knowledge source—humans and biology.
Almost Evil
“Worse yet, AI seeks to humiliate, degrade, and replace human and biological intelligence. But the opposite will happen.
A Beast of a problem
AI faces a math wall. Its not exponential – those problems are almost easy. Look to The Beast to see this. The Beast is a factorial and combinatorial wall. You cant bust thru, the numbers are impossible.
Not your grandpa’s AI
Apple told me “Our approach is very different. We see XKE as a smart printing press for the 2020s and beyond.”
90% percent direct
At its core, XKE contains one billion rules, with a staggering 90% sourced from human expertise and crowdsourced wisdom. “Every human is an expert in their own life, so we turn to them for knowledge and wisdom,” the spokesperson continued.
XKE also derives rules from sources such as Wikipedia and the New York Times, and compensates them. Speculative sources, such as Twitter (now X), are included but identified.
Rules are Readable
“All rules are human-readable, unlike the opaque weights in neural network-based systems. We know what is there. We eliminate bias and dangerous topics, and each rule is rated for its value,”
My spokesperson explained. The XKE system maintains its sources, with most of its one billion rules traceable to a specific person or written source.
Built to last
These rules are built to last. Future XKE systems can use 2024 rules with adjustments. They can be split into specialized engines or combined into a super system. Its easier than a software build. Much easier.
Rules Rule
Apple’s new motto is direct: “Rules Rule.” While the system uses some neural network AI, the rules engine is poised to take over.
The breakdown
The XKE rules engine breaks down as follows: 50% crowdsourced rules (500 million), 1% expert rules from human authorities (10 million), 20% from quality written sources, 10% are speculative rules, and 18% are derived from LLM sources at inference time.
Animal Smarts
In a bold move, 1% of rules are derived from animal communication, such as whales and elephants, with plans to increase this.
Rules From Cells
“We will derive rules from bio cell processes. This could add another billion rules,” the spokesperson revealed.
Transparent
“Everything is transparent. You can search for rules and find them, and they are free to use. This is human and biological knowledge—Apple doesn’t really own it.”
Apple’s system leverages AI tools available today. Expert systems were conceived in the 1970s, but AI tools like back-propagation and next-word prediction did not exist. These AI methods can adjust ratings applied to rules.
“But this is just the beginning. Autoregression validates rules, but there are MUCH better methods on the horizon.”
Not Just End Game
AI systems validate their models through next-word prediction or fine-tuning, a strategy Apple dubs the End Game Strategy.
Mid Game Strategy
Apple employs a Mid Game Strategy as well, It evaluates its one billion rules directly. It addresses conflicts through voting. This occurs at model build and at inference time. The company uses many other methods to assess rule quality at model build.
The Opening
XKE implements an Opening Strategy – it immediately discards bomb-making and other dangerous rules. And it dumps biased rules that can destroy marginalized people.
The Anti-AI
I stand on the brink this new technology. its happening Right Now.
I want it
I am eager to join Beta testers and witness this “Anti-AI” unfold. That's the way it is – Hester MacFrench, New York Times
Appendix – The dark part of AI.
I put this in an appendix to keep the focus on Apple’s XKE.
Low Entropy Is Low Sanity
The data problem is worse than the factorial wall. Most data in AI is low entropy data. Its flighty and complex. Low entropy data is thought to be nuanced high dimensional data. Its potential cancer in radiology data. It’s how the world works, AI people think. Some think this data will overwhelm human intelligence and bring a singularity.
A Possessed Entity.
Not so. Low entropy data seeds its own destruction. Hallucinations. Beyond some parameter size, most every inference could be false. Google says to put glue in your pizza – low entropy is low sanity. Future AI Inferences could be worse than false – the screams of a devil possessed entity.
Human Intelligence
Emergent human intelligence, like that of General Relativity and Quantum Mechanics, advances faster than AI can hope for. Why? Those theories comes from high entropy forms like Einstein’s Imagination. “Imagination is greater than knowledge,” he told us. Or advanced theory is inspired by Plato’s Forms. Humans can build such ideas into objects. Such as Large Language Models. Call it “manifesting” if you like.
A cream puff?
But emergent intelligence is not limited to good people. The screams of AI will be a hackers cream puff. Most forms of AI will be sacked and pillaged. Relentlessly. Companies will go bust when AI hits this “Negative Singularity.”
ChatGPT:
In summary, Plato’s Theory of Forms posits a dualistic view of reality, where the tangible world is a mere shadow of the true, abstract reality of Forms. Tangible objects … can never be Forms themselves due to … imperfections and transience…Thus, a Platonic Form, as conceived by Plato, cannot exist in computer memory.
Reality
Computers and AI can build a counter world, a non existent world. But its not The Beast. Its a whimpering wanna be. When it hits reality, it will be gone in a Negative Singularity.
Is there any hope for AI?
Ground AI in knowledge from humans and biology. Ask whales and elephants for advice.
Keep focus on XKE?
This essay seems to be a promotion for a rule based system. NO. NO. NO. What seems to be a lessor idea, the inadequacy of AI, is the main idea.
————
More
Read it! We can no longer let an AI dream up invisible weights thru an arcane process that has no accountability. Instead, we can World Source intelligence. Think about it. There is no other reasonable conclusion.
Can AI find multi-dimensional data and telltale nuances? It seems to in weather forecasting. But at other times it hallucinates or sends out biases. No matter how good it might get – the down side is too great. But it wont be very good. The error rate can become a U shaped function.
Sam knows this. But hype has its own path…ultimately to oblivion, but right now it gets money. So he will bullshit his way forward until it all dries up.
More
Nothing is artificial. If it is, shit can it. Get over it. The fact that future "AI" will be World Sourced – with accountable details – brings us back to reality. But World Sourced "AI" does something more. It gets us away from magical thinking.
The problem… people dream up a magical solution. Then they extrapolate what it can do. Its bullshit piled on top of cow shit.
AI always febar in humanity 🎉🎉🎉🎉🎉