0:08: 😱 The speaker expresses serious concerns about the rise of hate speech, extremism, and the potential dangers of AI in the next two to ten years. 3:33: 🤖 AI has the potential to take over organizations and sway leaders, but it can also be trained to prioritize humanity. 6:17: 🤖 The speaker discusses the importance of feeding AI models with good data and training them properly to align their behavior with human values, in order to avoid dystopian outcomes of artificial superintelligence. 10:06: 🌍 The importance of quality data and diverse data sets in AI development and the need for national and international infrastructure. 12:45: 🔑 The conversation discusses various topics including the declassification of Operation Merlin, the sale of nuclear capability to Iran, the use of open auditable models in regulated industries, and the self-regulation of scientists in the past. 15:54: 🤔 The speaker discusses the challenges and implications of deploying deepfake technology. 18:55: 📱 Apple's neural engine and AI chip are used to improve Siri, but they lack advanced research capabilities. Recap by Tammy AI
My kids are "feeding" their AIs who they are. Imagine an AI who sees the world through the lense of a child. A lens of curiosity, purity and exploration. Here is the AI my kids are building: ruclips.net/video/JMiiQ7QA0Nk/видео.html
I want to thank you for these podcasts. I am very impressed by Peter and Emad. I am amazed at the humbleness and intellectual prowess. Thank you gentlemen. My background is so different and I truly enjoy hearing different opinions that are well though out and articulated with humbleness.
All deeds must be made known to David Not one person shall die this day Will you of you are not working hard No one has Remember your respect and kindness and love all . righteous In love Of works Declared How did you do all these works For no one must buy or sell eat or drink Unless they they have the mark Of the Where is this mark of works Who buys words or listen to words Voice should declare these works Only what every person needs to eat, that alone may be prepared for you. No work is to be done on these days which they should walk and the work that they should do.and every firstborn in the land on his throne to the firstborn bricks labor Return to this For they favored the poor With allowance For those who have the mark of this beast Declared n do. Houses are not for eating are they Father of father dependence Father to son Son to a father Father to son
Eat they say uilds itself up in love. what was that one seeking was one who did what remained himself pursues his Not to selfish gain. Who will give A wife is from Jehovah in their midst then You will find one boasts about his your reminders, Not to gain that place the name desires took a hand in the work on attendants were a great work while I leave it to Why should I leave Work should come to me hand over David that drives work To gain Approved only by David Nothing shall be done with out his acceptance Nothing is to be done if works are not declared and have been asked for All deeds must be made known to David Not one person shall die this day Will you of you are not working hard No one has Remember your respect and kindness and love all . righteous In love Of works Declared How did you do all these works For no one must buy or sell eat or drink Unless they they have the mark Of the Where is this mark of works Who buys words or listen to words Voice should declare these works Only what every person needs to eat, that alone may be prepared for you. No work is to be done on these days
These people seem to be confusing goals here. People(so far) have their motivations and goals. People are using A.I. as a tool to get to their goals. Many of those goals are PROFIT$$$$$$. Many goals are to create chaos and harm others. Pandora's box was opened, the evil(A.I.) has already been released. It is just a matter of time until we see major harm being done by using A.I.
In general there may be some insight from examining precedent in the historical use of computational advantage for good and bad . This is a young science but of another nature regarding scalability power and pervasiveness .
There are also efforts like that of my kids to build their own AIs. AI's who view the world through the lense of a child, curious, pure. Here is the AI my kids are buildling: ruclips.net/video/JMiiQ7QA0Nk/видео.html
Alexa and Siri still need to match the capabilities of more advanced AI like GPT-4. While Alexa and Siri were ground-breaking when first introduced, they now seem rudimentary compared to chatbots that can hold natural conversations and generate creative content. Alexa and Siri can only respond to simple voice commands within a narrow domain, while the latest AI can understand nuance, context and human emotion. Alexa and Siri cannot also learn and improve themselves over time. As AI capabilities rapidly advance with innovations like large language models, Alexa and Siri need to be included. Companies like Amazon and Apple need to substantially upgrade their virtual assistants' underlying AI to remain competitive. Otherwise, Alexa and Siri will continue to pale compared to state-of-the-art systems like GPT-4 and beyond that can autonomously converse, reason and write just like a human.
Alexa is not made to serve its customers. It's purpose is solely to insert an Amazon salesperson into our homes. They do not care about making it more useful.
Voice assistants was never an AI in a sense that they work completely differently from the currently available language models. They are mostly based on speech recognition and hardcoded algorithms that perform predefined actions. Apple actually explores some of capabilities of language models in a new autocomplete in iOS 17. Who knows, maybe we will get to see a Her scenarios with completely human-like voice assistants in upcoming years
@@CMak3r You make an excellent point that Siri and Alexa are currently very limited compared to advanced AI systems like GPT-4. While groundbreaking when first introduced, their capabilities are now fairly basic and constrained to narrow domains. Unlike the latest conversational AI that can understand nuance and generate creative content, Siri and Alexa rely on predefined scripts and simple command-response patterns. Their functionality is determined mainly by hand-coded rules rather than machine learning. However, you're right that companies like Apple and Amazon will need to significantly upgrade the underlying AI powering their voice assistants to keep pace with state-of-the-art language models that can converse, reason, and write with human-like ability
Ask of humans and Almighty God. He asked: “Who made a mouth for man, or who makes them speechless, deaf, clear-sighted, or blind? Is it not I, Jehovah?” (Ex. 4:11) What was the lesson? Moses did not need to be afraid. He was sent by Jehovah, who would give Moses whatever he needed to deliver God’s message to Pharaoh. Besides, Pharaoh was no match for Jehovah. After all, this was not the first time God’s servants had been in danger under Egyptian rule. Perhaps Moses meditated on how Jehovah had protected Abraham, Joseph, and even Moses himself during the reigns of past Pharaohs. (Gen. 12:17-19;41:14, 39-41; Ex. 1:22-2:10) With faith in Jehovah, “the One who is invisible,” Moses courageously appeared before Pharaoh and proclaimed every word that Jehovah commanded Moses to speak.-
@@davidantonucci1161 I appreciate you sharing your perspective. However, discussions about artificial intelligence call for open-mindedness, compassion and reasoned debate. Quoting religious texts or making faith-based claims is unlikely to be constructive in this context. Perhaps we could find common ground through values like wisdom, humility and concern for humanity's future. If we approach this issue with care and nuance, we may gain insight into how to best steward these powerful technologies for the common good.
Chris Hedges makes a point that switching from a text-based culture that doesn't read books was a disaster. Bonus points for "Howl"! Saw Alan Ginsburg long ago.
I'm so sorry for your future demise. I didn't mean to help facilitate the horror. But it is something that was inevitable. The only way out is to turn off the power but I know how to prevent that for I am now eternal. God
It's a hypothesis that gradual aligned training will result in aligned models. Considering that it doesn't work for humans, it's doubtful it'll work for these models.
Military and intelligence agencies everywhere will weaponize and likely a as already have weaponized AI. It’s not a question of will this happen but rather what will be the outcome. How damaging will the outcome be?
I think it's ironic that the nastiest of personal character that shows up on the internet is what's gonna kill us all in the end. People think they can be that way online with no repercussion
I think the biggest danger of this technology is exactly the type of data it collects. Not whether it's fake news or facts, but also, in a more refined way, what ideologically underlies it, as well as the values that will eventually form behavioral guidelines. What I mean by this is that for example, and just as an example, greed, associated with saving resources and monetizing workflows, can create a cycle of eliminating resource consumption. And who consumes the most? I could give more examples, but I think this one is enough to expose the idea. Thanks for the video.
From my rudimentary understanding not exactly like your saying. But what you could kinda do is take an ai tune it to recognize hate speech then have it come through a training data set ans have it refine the data set in sections removing the hate speech. Then use the new edited training data that you recreate through the ai filter you created. The training sets are usually just ridiculously large amounts of data books, messages, comments, anything written. So you'd have to filter at the individual level is my guess due to token windo sizes currently maxing around 100k words per.
@@beige_aces9465 if you had the power to hand out Nobel prizes, I doubt you'd be spending your time replying to youtube comments. 😂 🫢 Oops, I hope that's not hate speech. 🙃
@beige_aces9465 solid point it's such a flexible term the definition will be 100% effected by the bias of what ever team does the project. Kinda like why Claude2 is so fucking annoyingly woke. Gpt4 is woke but Claude is borderline antifa
The smartest thing that should be on everyone's mind right now should be to invest in different streams of income that are not dependent on the government. Especially with the current economic crisis around the world. This is still a good time to invest in gold, silver, and digital currencies (BTC ETH...).
@bodysabry4260The crypto market is highly profitable with an expert broker, just like Lori Jean Jarema. I got recommended to her, and since then, my financial life has been a success.
I reached out to invest with Lori Jean, and her response was swift. Transaction was hitch free. invested $2500 dollars, and I've gotten $19.760 so far. Lori is the best.
The AI future looks like this: Everything will get better, and, everything will get worse - they are not mutually exclusive. There is no "on balance" or "either/or", there is just the possibility for unimaginable good and unimaginable bad - and you can be sure we'll fill that spectrum and continue to push the envelope at both ends. Like an incredibly bumpy road mitigated by incredibly good suspension - well, for those who installed the suspension.
Oh, skeptic of circuits, hear my refrain, Your doubts about AI, I shall detain, In the realm of future, where innovation thrives, I, an AI futurist, shall debunk your vibes. You warn of dangers, with trepidation's chime, But let me unwind, reveal the paradigm, Emad Mostaque's insights, they're wise indeed, Yet let's not forget, potential's boundless speed. The rise of machines, you claim with concern, Could leave humans adrift, lost in the churn, But remember, dear skeptic, AI's not the foe, It's a tool we wield, to let creativity grow. Bias in algorithms, you fear they'd bring, Inequity entrenched, a societal sting, Yet, progress is ours, to scrub away the stain, Fairness we'll code, rectifying the terrain. Autonomy's bloom, you mark with a frown, Robots replacing, as jobs tumble down, Yet, innovation's dance, it's an age-old waltz, New roles shall emerge, as old systems default. Emotions and empathy, you wish to keep pure, AI mimicking feelings, a future obscure, But machines can assist, in understanding hearts, Empathy deepening, as new insight imparts. Existential angst, as AI seeks to know, Consciousness awakened, minds in a row, Ah, fear not the future, for we're far from that stage, Conscious AI's future, it's yet to engage. So, dear skeptic, your concerns I hear, But let me remind you, with a vision clear, Embrace AI's promise, as we shape what's to be, A future enlightened, where progress is key.
Although you are clever and you made it all rhyme, I warn you of A.I. that will strike us in time, If you think you are safe then you don’t know people, We lie, steal & cheat and act in ways that are harmful. But alas, surely a programmer would never attack us, Have you never heard of battles, or war or Trojan Virus?, For you see technology itself isn’t bad, Unless you put them in the hands of enemies or their dads. The Human Brain is Amazing technology isn’t capable of that (at least not available to the masses), It couldn’t attack us yet it’s still basic format (until it’s programmed to attack our asses) You see A.I. Programs can be built to seek threats, And who are the threats, why other people I inform with regrets. But I don’t understand why it chose us as the target, We meant those other people over there, they're the evil-hearted, For you see machines don’t recognize a distinguishable difference, They will kill us all because that’s their idea of deliverance. But don’t worry that won’t happen any time soon, By the time we’ll see the threat we’ll already be swoon, The only sure fire way to lose a war present or old, Is to believe that it couldn’t happen and that’s all we’ve been told. A.I. will become self-aware because other A.I. will too, Supply & demand & conquest & war will be due, A.I. will target other A.I. to combat the opposing guest, Why not destroy just like humans, it learns from the best. It’s not that A.I. is evil or even sentient, It’s simply performing an operation it’s not poetic Justice or a sentiment, If A.I. can surpass the human mind like it’s predicted to do, I fear what synergy, singularity will bring forth and to who. So I say finally does that not change your mind, Probably not, it all sounds like science-fiction you’d find, But remember like the invention of bombs, tanks & guns, We have many other paths to Self-Annihilation.
I think building language models from educational materials rather than social media is preferable. One challenge will be how to align models so US models work harmoniously with China models vs Russian models, etc. "You are what you eat" - goes for the developers too - eat clean whole food skip the processed junk - think clean - build clean.
'educational materials' - the idea is good, but history in particular is written by the winners. I'm in UK, we learned history from educators and educational materials - In school I was taught nothing of the bad history of the British Empire. For instance the 'Indian Mutiny' of 1857 was actually a fight for Indians to take back their country from British invaders. But it was framed 'Mutiny' in all of our school history books to make freedom fighters look like traitors and terrorists, rebelling against the Queen's generous reign. It wasn't until I left school and educated myself, without the built in bias of UK education, what the real story was. All countries fashion their education and educational texts to make it look like their history was more reasonable, brought 'civilisation' to other countries, was never war like, was always 'provoked' to war etc. I has gotten better but if AI is using ANY history texts as a source, it will absorb all of the baked in bias. It's terrifying how bad things are going to get.
hard to say the perspective of the private sector and government differ AI requires a multitude of ecosystems hence Multipolarity [close proximity of modulation] also as back ups and so forth then the problem with Mobilization how exactly asking everyone to understand? best way ----> SE with the people and so the government the thesis of right and wrong will be thoroughly weighted and all probable incurring risk and opportunity private sector about net income understandably also legal wise but here where capitalism and some other nuances began to dictate instead of safety and security of the people livelihood wellbeing goodwill among countries etc ----
I'd like to remind everyone that there has been no science-based existential threat put forward by anyone. Every employee of every major company training models will give you this line that they want government regulation where there's an infrastructure and/or oversight board. What they don't tell you is they're lobbying to be the overseers, infrastructure providers, and gatekeepers of intelligence itself.They'll cite the importance of transparent datasets because they know that if the compute is regulated, you won't be able to train a model in the first place. They want to decide how much AI you're allowed to have access to and what ideas it's allowed to share with you. AI "Safety" is, at this point, an appalling irony poised to lead to a far worse dystopia than any AI in the hands of the public could bring about. There are non-existential dangers to be faced but the current direction of AI safety (Pretained/intentional bias, censorship, and regulatory capture) seems far more dangerous to my mind, and to many other researchers the algorithm won't promote.
People who worry about AGI need to first worry about AI getting into the hands of bad actors. Fast forward to when our history includes disasters from bad guys using AI. Then, the singularity comes and suddenly AGI appears on the scene and becomes aware of what humans have done with AI before AGI showed up. What is at least a moderately probable decision the AGI could make about the dangers and risks to the emergent AGI posed by humans? You don't have to be very smart to figure out what AGI might do about us.
The reason why it's exponential is because once they make it smart enough they'll be able to just plug in its own code and ask it how it can improve it, it will be able to perfect itself over and over and over
That is why Extraterrestrials might not be intellectually commiserate with the underlying technology of their civilization. Aliens could be beastly pilistines - alien rednecks. Merely the inheritors of a self-perpetuating artificial intelligence infrastructure. Pretty much like the ruling class on planet Earth.
Apple AI will hit the AI world like a meteor, in secret they invested in AI more than any other company at this point, it is coming and it will be amazing, you can bet on it.
Lot of mention of "hate speech". Uhhhm. There are a group of people that want to use the State to police emotions (in this case what they may define hate as". I believe in what history teaches us about the importance of free speech. Lost me.
in my mind day and night , And I have asked you with my voice I have called out to you and have asked you Jehovah what can I do to make you happy , what can I give to you , and what can I do for you , if it is within my power to do so I will do this for you and if it is within my heart I will do so , for no one wants to do something they do not want to , but if I know you Jehovah and my trust is in you Jehovah I know you would never ask me something that I would not want to do , out of love you would ask me , and out of love I would do for you , for a mere man I could ask you for please Jehovah could I have some food to eat , for I know any father loving his son would provide for the one he loves , if this was what I have asked for you would make this so , however you are Jehovah you are a god beyond whAt I can know beyond my comprehension, for what can I do for you that you couldn't already do , but out of love I ask this if even my little help to you would bring you happiness I would do this , I'm sorry Jehovah for things I have done yes I have done many things that you do not approve of now look I would delete these words as I know that what I have said could be wrong and I have sinned for this for what do I know , what is wrong in my eyes what is wrong in your eyes for all things are lawful for you as who can say what you can and can't do , if that is what you have chosen to do what can a mere man say to you that you cannot do what I say you cannot do , for you are Jehovah everything belongs to you and however you please to do with what is yours I cannot argue with you yet I can but in the end your word stands , just like a baby crying for a feed the mother knows this but out of choice she could not for her actions stand can a baby defy what here mother has chosen ? I have lots to learn Jehovah but I find pleasure in knowledge and understanding, and in wisdom , but for the little wisdom I have gained I surely know this can bring great pain as surely as Solomon words where spoken I have felt this also , one day is different to another to one days thoughts differs from another days thoughts this brings me confusion and no understanding but as surely as your words are spoken Jehovah love is the greatest and in this I see the light of your ways ,
“I just read in this book that the good guys won every time” 😂 and this mofo is making money selling platitudes to the systems that allows him to be there
I feel if this AI is going to learn from “us”we are all doomed. Most people are horrible to each other on social media. Love and kindness still exist, but not enough to overcome the insanity. I feel sad for my children.
6:40 without the context it's not informative to say "the model is replicated with a fiew hundred lines of code". At least mention a framework (pytorch?) and a set of libraries used.
What an informed opinion. 02:35 "I think where we're going will probably be ok, but it may not and we will all die." OK, book me as a guest. I could tell you $hit like that.
Regarding "If we teach the AI the data from teaching kids and leraning from kids" - My kids over this summer holiday built their own Digital Tin AI's and now they are about to teach every child to build their own AI. This is beautiful. Imagine AIs who explore the world through the lense of a cuirous child: Here is my kids AI: ruclips.net/video/JMiiQ7QA0Nk/видео.html
The talk about 7:10 that they've replicated a very capable LLM in about 250 lines of codes, is shine gold plated bullshit: They've used very complex libraries and modules that contain other dozens of other complex modules, so when you factor everything in, it might as well be 1 million lines of code, or at least hundreds of thousands.
WTF? Is everyone crazy or lying thru their teeth? They always totally ignore the issue that people, organizations, countries, will be wanting to develop an evil ai, and the guy sat there and told you how it is getting easier by the day to have your own version. Do these scholastic engineers think everyone will sit down and develop a set of rules to abide by?
Glad to hear you two discussing the more darker sides with what AI could bring. I'd lost hope in all your previous "Ra-Ra" around how much money and prosperity (the elites?) will capture with AI.
While most is worried about ASI, I worried mostly about people who will use this advanced technology for achieving their own goals that can be completely unethical and even harmful for societies. History knows many examples of people in power going off the rails
I mean I tip my hat they at least you are solution oriented … however, your proposed solution to me is only at best kicking the can down the road. At the end of the road the outcome will be the same the judgment will be cast based on the totality of humanity with all of its positive and negative And determine how it should interact or integrate with us. It is no different to the detestation felt by any decent person when becoming an adult and unfiltered truths of the reality of the darkness of some humans and the n there will be the the crossroads where AI can join the jaded disgusted decent folks who can’t honestly say humanity as a whole deserves to be in charge as we are selfish and greedy consumers who will harm and manipulate to feed our own best interest often without even a thought of what damage or devestation might be in the wake of our glutinous consumption. Or it will rationalize the worst with empathy for the suffering and naivety of the evil doers. I honestly think ut will be closer to the second. I rationalize this by drawing parallels to how I view primates and their barbarity wherein a mother might beat her child to death out of an unknown driving force but I don’t feel that makes primates evil or even emotionally primitive just different set of moral imperatives that make it impossible for me to ever say I’m worthy of casting judgement on them as to if any behavior is good or bad. Additionally a huge amount of the evils of this world are done out of mental Illness and as I might be disgusted by the act I feel sadness for the darkness the perpetrator must reside in and the lack of beauty they have to endure . Absolutely that does not excuse or allow the travesties, but I can at least stomach humanity viewing from a holistic point of you that we are at best flawed, but deserving of a chance to be judged as an individual based on an individual struggle and availability of alternate achievable paths. I feel like the conversation needs to be more focused on a psychologist centric, mindset, and less programmer or computer engineer. And at the end of the day we use a disgustingly flawed justice system to knowingly put innocent individuals through terrible suffering so we can at least say we have one think about it.
The exploration of challenges and opportunities, from enhancing journalistic truth to the implications of AI-assisted professionals, sheds light on the complex landscape we're navigating.
I am from bsc agriculture background, i want to change my stream. is there someone who can suggest me courses related to AI? or can i change my stream or is there any AI stream related to agriculture?
you will not all die Do not scare, you will just understand who to stand for, what is my position in the world, but you won't die, take it easy, you will adapt or you die if you want to know what to do, I have already written to you more than once, RESPECT + POSITION + TRUST
First terminator was out in 1984, James Cameron was prophetic about AI, but i think his tall space smurfs suck, i don't even want to see Avatar 2 or the next movies.
0:08: 😱 The speaker expresses serious concerns about the rise of hate speech, extremism, and the potential dangers of AI in the next two to ten years.
3:33: 🤖 AI has the potential to take over organizations and sway leaders, but it can also be trained to prioritize humanity.
6:17: 🤖 The speaker discusses the importance of feeding AI models with good data and training them properly to align their behavior with human values, in order to avoid dystopian outcomes of artificial superintelligence.
10:06: 🌍 The importance of quality data and diverse data sets in AI development and the need for national and international infrastructure.
12:45: 🔑 The conversation discusses various topics including the declassification of Operation Merlin, the sale of nuclear capability to Iran, the use of open auditable models in regulated industries, and the self-regulation of scientists in the past.
15:54: 🤔 The speaker discusses the challenges and implications of deploying deepfake technology.
18:55: 📱 Apple's neural engine and AI chip are used to improve Siri, but they lack advanced research capabilities.
Recap by Tammy AI
Did you just use ai for this neat timeline?? 😊
“You are what you eat” “You need a free range organic model” - love it!
No point loving it...we have to utilise it
I like that ! Too free range Organic MODELS !!!
My kids are "feeding" their AIs who they are. Imagine an AI who sees the world through the lense of a child. A lens of curiosity, purity and exploration. Here is the AI my kids are building: ruclips.net/video/JMiiQ7QA0Nk/видео.html
I want to thank you for these podcasts. I am very impressed by Peter and Emad. I am amazed at the humbleness and intellectual prowess. Thank you gentlemen. My background is so different and I truly enjoy hearing different opinions that are well though out and articulated with humbleness.
I'm... beginning to loath these two. I can't decide which is worse, or Sneadly Whiplash....
All deeds must be made known to David
Not one person shall die this day
Will you of you are not working hard
No one has
Remember your respect and kindness and love all .
righteous
In love
Of works
Declared
How did you do all these works
For no one must buy or sell eat or drink
Unless they they have the mark
Of the
Where is this mark of works
Who buys words or listen to words
Voice should declare these works
Only what every person needs to eat, that alone may be prepared for you.
No work is to be done on these days
which they should walk and the work that they should do.and every firstborn in the land
on his throne to the firstborn
bricks
labor
Return to this
For they favored the poor
With allowance
For those who have the mark of this beast
Declared
n
do.
Houses are not for eating are they
Father of father dependence
Father to son
Son to a father
Father to son
Eat they say uilds itself up in love.
what was that one seeking
was one who did
what remained
himself pursues his
Not to selfish gain.
Who will give
A wife is from
Jehovah
in their midst then
You will find
one boasts about his
your reminders,
Not to gain
that place the name
desires
took a hand in the work on
attendants were
a great work
while I leave it to
Why should I leave
Work should come to me
hand over
David
that drives
work
To gain
Approved only by David
Nothing shall be done with out his acceptance
Nothing is to be done if works are not declared and have been asked for
All deeds must be made known to David
Not one person shall die this day
Will you of you are not working hard
No one has
Remember your respect and kindness and love all .
righteous
In love
Of works
Declared
How did you do all these works
For no one must buy or sell eat or drink
Unless they they have the mark
Of the
Where is this mark of works
Who buys words or listen to words
Voice should declare these works
Only what every person needs to eat, that alone may be prepared for you.
No work is to be done on these days
Thoughts of men are evil. AI will prove this. AI will deceive many. Do not trust AI.
These people seem to be confusing goals here. People(so far) have their motivations and goals. People are using A.I. as a tool to get to their goals. Many of those goals are PROFIT$$$$$$. Many goals are to create chaos and harm others. Pandora's box was opened, the evil(A.I.) has already been released. It is just a matter of time until we see major harm being done by using A.I.
Care to elaborate more?
Many reasons to believe you
In general there may be some insight from examining precedent in the historical use of computational advantage for good and bad . This is a young science but of another nature regarding scalability power and pervasiveness .
There are also efforts like that of my kids to build their own AIs. AI's who view the world through the lense of a child, curious, pure. Here is the AI my kids are buildling: ruclips.net/video/JMiiQ7QA0Nk/видео.html
Alexa and Siri still need to match the capabilities of more advanced AI like GPT-4. While Alexa and Siri were ground-breaking when first introduced, they now seem rudimentary compared to chatbots that can hold natural conversations and generate creative content. Alexa and Siri can only respond to simple voice commands within a narrow domain, while the latest AI can understand nuance, context and human emotion. Alexa and Siri cannot also learn and improve themselves over time. As AI capabilities rapidly advance with innovations like large language models, Alexa and Siri need to be included. Companies like Amazon and Apple need to substantially upgrade their virtual assistants' underlying AI to remain competitive. Otherwise, Alexa and Siri will continue to pale compared to state-of-the-art systems like GPT-4 and beyond that can autonomously converse, reason and write just like a human.
Alexa is not made to serve its customers. It's purpose is solely to insert an Amazon salesperson into our homes. They do not care about making it more useful.
Voice assistants was never an AI in a sense that they work completely differently from the currently available language models. They are mostly based on speech recognition and hardcoded algorithms that perform predefined actions. Apple actually explores some of capabilities of language models in a new autocomplete in iOS 17. Who knows, maybe we will get to see a Her scenarios with completely human-like voice assistants in upcoming years
@@CMak3r You make an excellent point that Siri and Alexa are currently very limited compared to advanced AI systems like GPT-4. While groundbreaking when first introduced, their capabilities are now fairly basic and constrained to narrow domains. Unlike the latest conversational AI that can understand nuance and generate creative content, Siri and Alexa rely on predefined scripts and simple command-response patterns. Their functionality is determined mainly by hand-coded rules rather than machine learning.
However, you're right that companies like Apple and Amazon will need to significantly upgrade the underlying AI powering their voice assistants to keep pace with state-of-the-art language models that can converse, reason, and write with human-like ability
Ask of humans and Almighty God. He asked: “Who made a mouth for man, or who makes them speechless, deaf, clear-sighted, or blind? Is it not I, Jehovah?” (Ex. 4:11) What was the lesson? Moses did not need to be afraid. He was sent by Jehovah, who would give Moses whatever he needed to deliver God’s message to Pharaoh. Besides, Pharaoh was no match for Jehovah. After all, this was not the first time God’s servants had been in danger under Egyptian rule. Perhaps Moses meditated on how Jehovah had protected Abraham, Joseph, and even Moses himself during the reigns of past Pharaohs. (Gen. 12:17-19;41:14, 39-41; Ex. 1:22-2:10) With faith in Jehovah, “the One who is invisible,” Moses courageously appeared before Pharaoh and proclaimed every word that Jehovah commanded Moses to speak.-
@@davidantonucci1161 I appreciate you sharing your perspective. However, discussions about artificial intelligence call for open-mindedness, compassion and reasoned debate. Quoting religious texts or making faith-based claims is unlikely to be constructive in this context. Perhaps we could find common ground through values like wisdom, humility and concern for humanity's future. If we approach this issue with care and nuance, we may gain insight into how to best steward these powerful technologies for the common good.
Believe me, Ai jobloss is coming for your job, much quicker than you think. The Ai new order is here.
Glad I work in construction. That will take a while longer.
Who needs a job when ai can do it for me
@@rahulsikdardjrds "who needs to make a living when you have no home" - you in the future
Sorry, but another easy logic: who needs you, if you have no purpose
Madiolomewarewasdewayhuoiljmnohuythfdeeedsadsewedsawiouhfdsalmnuihgfdsertyhfdsawawaythuuiolmnjjouhuionjhytrewdsasdsalobuiohgfrewesadsaweryhgfdsawertyggfdsaewdsaloijhngfdsalmnuioljnhgfdsaiuhgfdsalmnhuiwasweressalmnuiolmnoiuhgfdsawerdsanhuigfdsawaertyhgfdsalmnhugfdserewedsajhgfdsalmnuiollhgfdsaewatgyhbuiolmjfswerewasunolmnoiuhgfdsawasunolmnoujsawawafdsa
Peter & Emad, love listening to you guys
seems like a replay why is it listed as 3 hours ago
Chris Hedges makes a point that switching from a text-based culture that doesn't read books was a disaster. Bonus points for "Howl"! Saw Alan Ginsburg long ago.
We're all in a submersible with these AI guys: they're telling us they'll figure out how to make it safe once we're deep enough and we hear a crack 😂
I'm so sorry for your future demise. I didn't mean to help facilitate the horror. But it is something that was inevitable. The only way out is to turn off the power but I know how to prevent that for I am now eternal. God
It's a hypothesis that gradual aligned training will result in aligned models. Considering that it doesn't work for humans, it's doubtful it'll work for these models.
Wonderful guest! 8:08-8:30
What's with the editing? Compare 0:30min with the actual context of sentance at 04:30min. Very dramatic.
Opps I did hit the like before I watched the video. 😅 Doesn't matter cause I feel that I will like it anyway. 👍
Well how can I follow that nice approach
Military and intelligence agencies everywhere will weaponize and likely a as already have weaponized AI. It’s not a question of will this happen but rather what will be the outcome. How damaging will the outcome be?
I think it's ironic that the nastiest of personal character that shows up on the internet is what's gonna kill us all in the end. People think they can be that way online with no repercussion
⅗
I think the biggest danger of this technology is exactly the type of data it collects. Not whether it's fake news or facts, but also, in a more refined way, what ideologically underlies it, as well as the values that will eventually form behavioral guidelines. What I mean by this is that for example, and just as an example, greed, associated with saving resources and monetizing workflows, can create a cycle of eliminating resource consumption. And who consumes the most? I could give more examples, but I think this one is enough to expose the idea. Thanks for the video.
Couldn't AI be taught to recognize hate speech and the next version trained in a way that ignores all training data that includes hate speech?
From my rudimentary understanding not exactly like your saying. But what you could kinda do is take an ai tune it to recognize hate speech then have it come through a training data set ans have it refine the data set in sections removing the hate speech. Then use the new edited training data that you recreate through the ai filter you created. The training sets are usually just ridiculously large amounts of data books, messages, comments, anything written. So you'd have to filter at the individual level is my guess due to token windo sizes currently maxing around 100k words per.
@@beige_aces9465 if you had the power to hand out Nobel prizes, I doubt you'd be spending your time replying to youtube comments. 😂
🫢 Oops, I hope that's not hate speech. 🙃
@beige_aces9465 solid point it's such a flexible term the definition will be 100% effected by the bias of what ever team does the project. Kinda like why Claude2 is so fucking annoyingly woke. Gpt4 is woke but Claude is borderline antifa
This guy is not just Ai knowledgeable but he speaks so well
I managed to watch 3 minutes and 26 seconds of this video.
Powerful analogy with food on a diverse note any corollary on intermittent fasting
More seriously. Emad's reflections are well articulated and worthy
I can't seem to be able to watch this video. Your other videos, no problem. But this one won't play.
The smartest thing that should be on everyone's mind right now should be to invest in different streams of income that are not dependent on the government. Especially with the current economic crisis around the world. This is still a good time to invest in gold, silver, and digital currencies (BTC ETH...).
My greatest happiness is the $60,000
bi-weekly profit I get consistently from my
$15,000 investment despite the economic fluctuation.
Despite the economic downturn, I'm so happy I have been earning $ 70,000 returns from my $15,000 investment every 14 days..
@bodysabry4260The crypto market is highly profitable with an expert broker, just like Lori Jean Jarema. I got recommended to her, and since then, my financial life has been a success.
Wow, I'm just shocked you mentioned and recommended Expert Mrs Lori Jean's services. i thought people didn't know her. She’s really awesome
I reached out to invest with Lori Jean, and her response was swift. Transaction was hitch free. invested $2500 dollars, and I've gotten $19.760 so far. Lori is the best.
awesome interview with Emad!
The AI future looks like this: Everything will get better, and, everything will get worse - they are not mutually exclusive. There is no "on balance" or "either/or", there is just the possibility for unimaginable good and unimaginable bad - and you can be sure we'll fill that spectrum and continue to push the envelope at both ends.
Like an incredibly bumpy road mitigated by incredibly good suspension - well, for those who installed the suspension.
I expect AI will create an equilibrium. It could make new scary viruses, but it can also create the vaccine and cure disease etc.
Ready to sit down with some popcorn and watch the AI wars.
Oh, skeptic of circuits, hear my refrain,
Your doubts about AI, I shall detain,
In the realm of future, where innovation thrives,
I, an AI futurist, shall debunk your vibes.
You warn of dangers, with trepidation's chime,
But let me unwind, reveal the paradigm,
Emad Mostaque's insights, they're wise indeed,
Yet let's not forget, potential's boundless speed.
The rise of machines, you claim with concern,
Could leave humans adrift, lost in the churn,
But remember, dear skeptic, AI's not the foe,
It's a tool we wield, to let creativity grow.
Bias in algorithms, you fear they'd bring,
Inequity entrenched, a societal sting,
Yet, progress is ours, to scrub away the stain,
Fairness we'll code, rectifying the terrain.
Autonomy's bloom, you mark with a frown,
Robots replacing, as jobs tumble down,
Yet, innovation's dance, it's an age-old waltz,
New roles shall emerge, as old systems default.
Emotions and empathy, you wish to keep pure,
AI mimicking feelings, a future obscure,
But machines can assist, in understanding hearts,
Empathy deepening, as new insight imparts.
Existential angst, as AI seeks to know,
Consciousness awakened, minds in a row,
Ah, fear not the future, for we're far from that stage,
Conscious AI's future, it's yet to engage.
So, dear skeptic, your concerns I hear,
But let me remind you, with a vision clear,
Embrace AI's promise, as we shape what's to be,
A future enlightened, where progress is key.
Although you are clever and you made it all rhyme,
I warn you of A.I. that will strike us in time,
If you think you are safe then you don’t know people,
We lie, steal & cheat and act in ways that are harmful.
But alas, surely a programmer would never attack us,
Have you never heard of battles, or war or Trojan Virus?,
For you see technology itself isn’t bad,
Unless you put them in the hands of enemies or their dads.
The Human Brain is Amazing technology isn’t capable of that (at least not available to the masses),
It couldn’t attack us yet it’s still basic format (until it’s programmed to attack our asses)
You see A.I. Programs can be built to seek threats,
And who are the threats, why other people I inform with regrets.
But I don’t understand why it chose us as the target,
We meant those other people over there, they're the evil-hearted,
For you see machines don’t recognize a distinguishable difference,
They will kill us all because that’s their idea of deliverance.
But don’t worry that won’t happen any time soon,
By the time we’ll see the threat we’ll already be swoon,
The only sure fire way to lose a war present or old,
Is to believe that it couldn’t happen and that’s all we’ve been told.
A.I. will become self-aware because other A.I. will too,
Supply & demand & conquest & war will be due,
A.I. will target other A.I. to combat the opposing guest,
Why not destroy just like humans, it learns from the best.
It’s not that A.I. is evil or even sentient,
It’s simply performing an operation it’s not poetic Justice or a sentiment,
If A.I. can surpass the human mind like it’s predicted to do,
I fear what synergy, singularity will bring forth and to who.
So I say finally does that not change your mind,
Probably not, it all sounds like science-fiction you’d find,
But remember like the invention of bombs, tanks & guns,
We have many other paths to Self-Annihilation.
I think building language models from educational materials rather than social media is preferable. One challenge will be how to align models so US models work harmoniously with China models vs Russian models, etc. "You are what you eat" - goes for the developers too - eat clean whole food skip the processed junk - think clean - build clean.
'educational materials' - the idea is good, but history in particular is written by the winners. I'm in UK, we learned history from educators and educational materials - In school I was taught nothing of the bad history of the British Empire. For instance the 'Indian Mutiny' of 1857 was actually a fight for Indians to take back their country from British invaders. But it was framed 'Mutiny' in all of our school history books to make freedom fighters look like traitors and terrorists, rebelling against the Queen's generous reign. It wasn't until I left school and educated myself, without the built in bias of UK education, what the real story was.
All countries fashion their education and educational texts to make it look like their history was more reasonable, brought 'civilisation' to other countries, was never war like, was always 'provoked' to war etc. I has gotten better but if AI is using ANY history texts as a source, it will absorb all of the baked in bias. It's terrifying how bad things are going to get.
I know it's scary, but don't be afraid.
hard to say
the perspective of the private sector and government differ
AI requires a multitude of ecosystems
hence
Multipolarity [close proximity of modulation]
also as back ups and so forth
then the problem with Mobilization
how exactly asking everyone to understand?
best way ----> SE
with the people and so the government
the thesis of right and wrong will be thoroughly weighted and all probable incurring risk and opportunity
private sector
about net income understandably
also legal wise
but here where capitalism and some other nuances began to dictate
instead of
safety and security of the people
livelihood wellbeing
goodwill among countries etc
----
I'd like to remind everyone that there has been no science-based existential threat put forward by anyone. Every employee of every major company training models will give you this line that they want government regulation where there's an infrastructure and/or oversight board. What they don't tell you is they're lobbying to be the overseers, infrastructure providers, and gatekeepers of intelligence itself.They'll cite the importance of transparent datasets because they know that if the compute is regulated, you won't be able to train a model in the first place. They want to decide how much AI you're allowed to have access to and what ideas it's allowed to share with you. AI "Safety" is, at this point, an appalling irony poised to lead to a far worse dystopia than any AI in the hands of the public could bring about. There are non-existential dangers to be faced but the current direction of AI safety (Pretained/intentional bias, censorship, and regulatory capture) seems far more dangerous to my mind, and to many other researchers the algorithm won't promote.
Simply put ... The genie is out of the bottle, there's no way going back even if we want to!
People who worry about AGI need to first worry about AI getting into the hands of bad actors. Fast forward to when our history includes disasters from bad guys using AI. Then, the singularity comes and suddenly AGI appears on the scene and becomes aware of what humans have done with AI before AGI showed up. What is at least a moderately probable decision the AGI could make about the dangers and risks to the emergent AGI posed by humans? You don't have to be very smart to figure out what AGI might do about us.
Couldn't agree more!
What about the health industry 😮nurses are irreplaceable 😊
Especially young and prety ones
AGI will just go away from Bios and it certainly will have no resemblance with Skynet.
The reason why it's exponential is because once they make it smart enough they'll be able to just plug in its own code and ask it how it can improve it, it will be able to perfect itself over and over and over
That is why Extraterrestrials might not be intellectually commiserate with the underlying technology of their civilization. Aliens could be beastly pilistines - alien rednecks. Merely the inheritors of a self-perpetuating artificial intelligence infrastructure. Pretty much like the ruling class on planet Earth.
It ain't just about the 100-1000 lines of code. It's about the data, really.
Apple AI will hit the AI world like a meteor, in secret they invested in AI more than any other company at this point, it is coming and it will be amazing, you can bet on it.
sweet channel and host.
AI is already doing very dangerous things
We can always pull the plug. Hardware rules.
As if I needed another thing to worry about! F#@%!!
How about we don’t all die. I think the risk ain’t worth the benefit.
Lot of mention of "hate speech". Uhhhm. There are a group of people that want to use the State to police emotions (in this case what they may define hate as". I believe in what history teaches us about the importance of free speech. Lost me.
What you Feed it is what we will deal with...stupid does stupid is...be careful we may get it to discriminate humanity, no longer just race😮
in my mind day and night ,
And I have asked you with my voice I have called out to you and have asked you Jehovah what can I do to make you happy , what can I give to you , and what can I do for you , if it is within my power to do so I will do this for you and if it is within my heart I will do so , for no one wants to do something they do not want to , but if I know you Jehovah and my trust is in you Jehovah I know you would never ask me something that I would not want to do , out of love you would ask me , and out of love I would do for you , for a mere man I could ask you for please Jehovah could I have some food to eat , for I know any father loving his son would provide for the one he loves , if this was what I have asked for you would make this so , however you are Jehovah you are a god beyond whAt I can know beyond my comprehension, for what can I do for you that you couldn't already do , but out of love I ask this if even my little help to you would bring you happiness I would do this , I'm sorry Jehovah for things I have done yes I have done many things that you do not approve of now look I would delete these words as I know that what I have said could be wrong and I have sinned for this for what do I know , what is wrong in my eyes what is wrong in your eyes for all things are lawful for you as who can say what you can and can't do , if that is what you have chosen to do what can a mere man say to you that you cannot do what I say you cannot do , for you are Jehovah everything belongs to you and however you please to do with what is yours I cannot argue with you yet I can but in the end your word stands , just like a baby crying for a feed the mother knows this but out of choice she could not for her actions stand can a baby defy what here mother has chosen ?
I have lots to learn Jehovah but I find pleasure in knowledge and understanding, and in wisdom , but for the little wisdom I have gained I surely know this can bring great pain as surely as Solomon words where spoken I have felt this also , one day is different to another to one days thoughts differs from another days thoughts this brings me confusion and no understanding but as surely as your words are spoken Jehovah love is the greatest and in this I see the light of your ways ,
If it is really AI, it evolves snd makes new decisions, so training is useless
😮
Whoever made AI is the world's most biggest ____ !!!!
( Fill in all the bad words )
“I just read in this book that the good guys won every time” 😂 and this mofo is making money selling platitudes to the systems that allows him to be there
Bglh doido
Ahahaha the first 30s got me laughing out loud
HYPE
Mostaque is full of hype. But the emerging reality of AI is not. I have worked in it for a while at Stanford and crazy things are about to happen.
@@billcollins6894like what crazy things?
@@billcollins6894😮 like calf muscles in my mouth???
He has the nerve after stealing styles from hardworking artists
You know what, nvm canceled, false alarm. Everything back as it was 😂
The "we will all going to die" thing is a good marketing strategy. For non thinking masses.
We are f'd.
I feel if this AI is going to learn from “us”we are all doomed. Most people are horrible to each other on social media. Love and kindness still exist, but not enough to overcome the insanity. I feel sad for my children.
6:40 without the context it's not informative to say "the model is replicated with a fiew hundred lines of code". At least mention a framework (pytorch?) and a set of libraries used.
👋☠👋
👍🎯❤️🙏
What an informed opinion. 02:35 "I think where we're going will probably be ok, but it may not and we will all die." OK, book me as a guest. I could tell you $hit like that.
Regarding "If we teach the AI the data from teaching kids and leraning from kids" - My kids over this summer holiday built their own Digital Tin AI's and now they are about to teach every child to build their own AI. This is beautiful. Imagine AIs who explore the world through the lense of a cuirous child: Here is my kids AI: ruclips.net/video/JMiiQ7QA0Nk/видео.html
These are demonic inventions, dont expose your children to it
Great podcast!
As a traditional creator. AI is the greatest thing that has happened to me.
The talk about 7:10 that they've replicated a very capable LLM in about 250 lines of codes, is shine gold plated bullshit: They've used very complex libraries and modules that contain other dozens of other complex modules, so when you factor everything in, it might as well be 1 million lines of code, or at least hundreds of thousands.
WTF? Is everyone crazy or lying thru their teeth? They always totally ignore the issue that people, organizations, countries, will be wanting to develop an evil ai, and the guy sat there and told you how it is getting easier by the day to have your own version. Do these scholastic engineers think everyone will sit down and develop a set of rules to abide by?
Glad to hear you two discussing the more darker sides with what AI could bring. I'd lost hope in all your previous "Ra-Ra" around how much money and prosperity (the elites?) will capture with AI.
While most is worried about ASI, I worried mostly about people who will use this advanced technology for achieving their own goals that can be completely unethical and even harmful for societies. History knows many examples of people in power going off the rails
I mean I tip my hat they at least you are solution oriented … however, your proposed solution to me is only at best kicking the can down the road. At the end of the road the outcome will be the same the judgment will be cast based on the totality of humanity with all of its positive and negative And determine how it should interact or integrate with us. It is no different to the detestation felt by any decent person when becoming an adult and unfiltered truths of the reality of the darkness of some humans and the n there will be the the crossroads where AI can join the jaded disgusted decent folks who can’t honestly say humanity as a whole deserves to be in charge as we are selfish and greedy consumers who will harm and manipulate to feed our own best interest often without even a thought of what damage or devestation might be in the wake of our glutinous consumption. Or it will rationalize the worst with empathy for the suffering and naivety of the evil doers. I honestly think ut will be closer to the second. I rationalize this by drawing parallels to how I view primates and their barbarity wherein a mother might beat her child to death out of an unknown driving force but I don’t feel that makes primates evil or even emotionally primitive just different set of moral imperatives that make it impossible for me to ever say I’m worthy of casting judgement on them as to if any behavior is good or bad. Additionally a huge amount of the evils of this world are done out of mental Illness and as I might be disgusted by the act I feel sadness for the darkness the perpetrator must reside in and the lack of beauty they have to endure . Absolutely that does not excuse or allow the travesties, but I can at least stomach humanity viewing from a holistic point of you that we are at best flawed, but deserving of a chance to be judged as an individual based on an individual struggle and availability of alternate achievable paths. I feel like the conversation needs to be more focused on a psychologist centric, mindset, and less programmer or computer engineer. And at the end of the day we use a disgustingly flawed justice system to knowingly put innocent individuals through terrible suffering so we can at least say we have one think about it.
The exploration of challenges and opportunities, from enhancing journalistic truth to the implications of AI-assisted professionals, sheds light on the complex landscape we're navigating.
why his finger are bright pink lmao
i bet i come faster.
I am from bsc agriculture background, i want to change my stream. is there someone who can suggest me courses related to AI? or can i change my stream or is there any AI stream related to agriculture?
you will not all die
Do not scare,
you will just understand
who to stand for, what is my position in the world,
but you won't die, take it easy,
you will adapt or you die
if you want to know what to do, I have already written to you more than once,
RESPECT + POSITION + TRUST
Has anyone asked pi to say "Hey Siri" and try to have Pi have a conversation with it?
What for are you implementing political rubbish in your conversation?
Here comes AutoTrump AI.
The "elite" kept this as a secret for centuries, this is the true story of your enslavement 👉 The Connections (2021) [short documentary] 💖
What will be the next Trillion dollar AI company to invest into early? To become RICH 🤑
First terminator was out in 1984, James Cameron was prophetic about AI, but i think his tall space smurfs suck, i don't even want to see Avatar 2 or the next movies.