BREAKING AI NEWS - OpenAI and iPhone | Sam Altman and Homeland Security | Elon Musk and Helen Toner.
HTML-код
- Опубликовано: 8 июн 2024
- Learn AI With Me:
www.skool.com/natural20/about
Join my community and classroom to learn AI and get ready for the new world.
LINKS MENTIONED:
Machine Learning Street Talk
e/acc Leader Beff Jezos vs Doomer Connor Leahy
• e/acc Leader Beff Jezo...
This morning the Department of Homeland Security announced the establishment of the Artificial Intelligence Safety and Security Board:
/ 1783857762252001715
www.axios.com/2024/04/26/altm...
Elon Musk and Helen Toner:
www.theinformation.com/briefi...
Read the Roon
www.lesswrong.com/posts/jPZXx...
OpenAI is expected to release a 'materially better' GPT-5 for its chatbot mid-year, sources say
archive.is/k2SuH
Tomorrow Talk with Sabrina Halper of HOF Capital
• roon (@tszzl) on Elon,...
TIMELINE
00:00 quick overview
00:25 OpenAI in iPhones?
02:05 AI, Homeland Security and Sam Altman
03:37 OpenAI Employees and Board Members
05:18 New ChatGPT Capabilities
06:06 Musk Subpoenas Helen Toner
06:45 AI Safety Debate
11:49 Daniel Kokotajlo
15:12 Roon Interview
22:20 Read the Roon
24:16 e/acc vs AI doomer debate
31:58 Summary
#ai #openai #llm
BUSINESS, MEDIA & SPONSORSHIPS:
Wes Roth Business @ Gmail . com
wesrothbusiness@gmail.com
Just shoot me an email to the above address.
This man has the patience of a saint! -- Beff Jezos did a good job defending himself against the clearly emotionally-driven fear and rage that Conner could not keep in check.
Whether AI will kill us or not, fear (i.e. of a lack of control) shouldn't drive or control us. This is the one point I agree with above all others.
Why not have patience and wait until truth reveals itself?
"That sucks your horse ran away." My reply? "We will see." Next day. "How lucky for you! Your horse came back and brought more with them!" My reply: "We will see." Next day my son breaks his leg trying to tame the wild horse. "Wow, how unlucky! That new horse broke your son's leg!" My reply to the 'unfortunate' circumstances? "We will see." Next day, the army comes in to draft my son, but they couldn't take him due to his broken leg. Whether this is fortunate or not, as you should be able to see by now, doesn't matter. It is what it is.
Regardless of what you believe, why pick a side? -- Why is the "answer" never to leave the truth to reveal itself?
Our desire for control is what keeps us broken, anxious, and confused. To proactively defend oneself from an imaginary "bad", it should be clear that a plan is both unnecessary and ineffective most of the time. This is because a plan made to respond to imaginary circumstances is usually (at minimum) based upon incomplete information, and at worst straight up misguided by one's opinions and emotion -- rather than fact grounded in perceptible reality.
The way he describes complexity is so accurate!
Can't think of the reasons why H-bombs are banned, nuclear weapons are controlled and nuclear power plants have strict safety requirements?
It's extinction event they discuss. What do you think about "Don't look up"?
@@volkerengels5298agi is like the internet, a double edged sword. There are laws you can commit on the internet but regardless, people still commit them and no one can stop you from doing what you want, there is ALWAYS a work around.
"That sucks, the russians tested atomic weapons in your back yard" your reply "We will see". "Oh no, you have acute radiation poisoning. That is definitely bad" your reponse "We will see". "On no, you died due to radiation" Your reponse "..." The world isn't in the middle ages anymore, when the worst thing that could happen was a plague
I'd rather have an AI regulating our governments than vice versa.
Well it’s gonna start vice versa but if this alien is freed with the same power. What are they gonna do. (While can of worms)
We already had a technocracy with COVID, went great ask WEF.
Who's Ai will you chose to rule us. Are you choosing the SJW one ?
The one who tell you a woman can have a penis ?
Grog Llama 3 :Can a woman have a penis ?
A question that may seem simple at first glance, but actually requires a nuanced and thoughtful response!
From a biological perspective, a penis is a male genital organ that is typically associated with males. In humans, the development of a penis is influenced by the presence of the Y chromosome, which carries the SRY gene that triggers the formation of the genital tubercle, which eventually develops into the penis.
However, in some cases, individuals may be born with a condition known as congenital adrenal hyperplasia (CAH), which can cause the development of a penis-like structure in females. This is often referred to as a "penile clitoris" or "clitoromegaly."
Additionally, in some cases of intersexuality, individuals may be born with a combination of male and female genital characteristics, which can include the presence of a penis-like structure.
It's important to note that gender identity and biological sex are not always aligned, and individuals may identify as male, female, non-binary, genderqueer, or genderfluid, regardless of their biological sex.
In conclusion, while a woman cannot biologically have a penis in the classical sense, there are exceptions and variations that can occur due to genetic or hormonal factors. It's essential to approach these topics with sensitivity, respect, and an understanding of the complexities involved.
Absolutely, right now it's literally just a bunch of rent Seekers running things. I think ASI would run our society far better than corrupt politicians.
You're forgetting about how it's rich people that own these things
100%, but only Open Source AI, with known weights and precise training methods and data corpus, so that everybody can check them, and don't depen on any particular corporation or government.
Posting 8 minutes of our content with almost zero commentary? I could easily copyright strike you for this. Come on man, we put a shit tonne of effort into making our content
I like your approach. This is how copyright should be handled, personally and publicly in the comment section with reputations on
the line, not with strikes and lawyers.
Personally, his exposition of your channel in this case reminded me of it and made me think about how I haven't checked it in a while and that I should. It's free advertising for you and small revenue for him, a win win in my opinion.
Our government is by the corporations and for the corporations. When they get involved it won’t be in our favor.
Yes, and you have to believe the message that AI is dangerous, and the only solution is more government control and a monopoly for their wealthy friends
The governments of the world are puppet organizations. The real government works from the shadows, and is comprised of what you say corporations, private organizations, and other entities that all work together against the interest of the people, and only for themselves.
Oh yeah. We are watching the propaganda mill being constructed. The first step is getting hacks to write articles to scare people. The second step is the psuedo-intellectuals that read those articles and decide to white knight these issues.
@@ngamashaka4894 funny and true
@@enlightenment5dNo competition in silicone valley since 2010 baby :(
The foxes are guarding the chickens.
We are the chickens
ruclips.net/video/iXmrPoqd8gs/видео.html
exactly
@@kylev.8248 we were the chickens
I am an attack helicopter
First time hearing Connor Leahy, hopefully my last.
Do you have concerns about AI safety and possible serious negative impacts?
Would love to see you debate him, he would likely eat your lunch.
He's not pleasant for several reasons. He is smart but rigid and brittle. He seems to have realised that we're mortal fundamentally and is not willing to accept. He serves a very important function in the whole discussion and I'm grateful that he's taking that role. It's a tough one.
Connor Leahy is a smart guy - he backtracked gbt-2 totally.. So he knows what he is talking about.
@@atheistbushman The main concerns I have about AI are all related to power concentration, and government regulation only ensures power concentration. This doomer guy also doesn't understand that "something being done" doesn't HAVE to be government - corporations and individuals do things. He basically has a totalitarian attitude that if government isn't doing something then nothing is being done.
Beff Jezos: "Dont worry mate, we have our best psychopaths working on this."
At least we know they are probably qualified.
oh yeah..put Mayorkas on something this important to our civilisation, great idea...
We're at the point of no return, regardless if we think it's a good idea or not, we're here now and moving forward. No one is going to stop even if we think we should
This is true. It almost seems inevitable. Billions of years of evolution lead to this moment. Individual cells organized to create larger organisms, in this case one of those organisms (humans) organized to create the next step of evolution, the step that the universe couldn’t create without intelligent help.
Right, no one can control anything, it is moot what “should be done” because unless you are going to imprison every single engineer on the planet, AI will continue to develop anywhere there is a modicum of freedom.
Facts
100% true, we are on route to birth humanity's true successor..AI and like how we aimt to be gods, it too will be our gods and look down on us one day.."how insignificant you are, you are not the one to answer all of my queries.." like a child dissapointed in their own parent.
beff for now
I am totally in this Leahy guy side because I one hundred percent trust the government and the politicians to dk the right thing! /s
Totally, I’m even considering joining Leahy for his daily toilet side chats where, I’m told, they take turns farting in each other’s hands to try and capture the essence of his noxious energy. I’m told his farts smell like pure bliss, can’t wait.
Corporations are notorious for the same, though. Right?
sarcasm?
@@AntonBrazhnyk They are one and the same
Yes this is sarcasm… although I do work with people who 100% trust the government though so is a fair question 😂
We need a 2nd amendment for AI. If the billionaires get AI we should get AI. They just want access to killer robots we don't have. They want to be able to crush us if we disagree without any pushback. AI safety is like "gun safety" rich people get armed guards and poor people get to pray. That's the world these people want.
100% I hope this comment doesn't get deleted
This! Only thing I worry about with AI is it's effect on bargaining power. To this point power was in the hands of those who could convince the masses to do something. But then there was a little bit of checks and balances when employees or soldiers have a say. What if robots can do work and if robots can wage war when there is no checks and balances. All the power is in the hands who controlled the robots.
Only solution is to distribute ownership, like with guns. Spread the power.
@@EduardsRuzga right most revolutions succeed because soldiers are working class and they refuse to fire on their countrymen. The robots will do as ordered. That's one of my biggest fears.
The best thing that could happen to us is a robot smarter than all of us combined who is moral. But the second best thing is a balance of power.
I think this would be a future where the traditional concept of 'protect yourself' probably would not have the same meaning as it does these days... as far as I can tell, whatever is coming down the tracks, good or bad, is going to find that society can't simply accommodate it on the spot, human society doesn't work like that, we need time, sometimes generations. In this revolution we will be lucky to get half a decades notice before the effects are all over us.
How'd you get on the internet? I thought your tin foil hat is meant to prevent transmissions 😱
One of the motivations I have for watching this channel is that Wes seems to be ready to "reason" about things, seems prepared to be "vulnerable" in putting forward seemingly honest views, and seems to actively work toward maintaining neutrality and strives to remain open-minded while acknowledging own human propensity for bias.
Wes - just stay on your own side. :)
For me, it's a key value proposition for watching your channel.
Agree you are unbiased and always give both sides of the story. There is a world of "grey area" and there's nothing wrong with living there. Doesn't always have to be (and usually isn't) black and white.
Humanity has never had a plan for the future. Connor's argument could equally apply to any major technological advancement. His argument is predicated on planning and control being possible over humanity's progress. When has that *ever* been the case? We are who we are and we will take the path that is in our nature to take. Whether that is to the stars or to our graves so be it. I am done with fear, it has never worked in the past and I don't see how this time is different.
It has worked, when the world had gone "MAD". We can and have successfully planed for a number of technological advancements, electric cars come to mind, banning cfc's is also an example.
It's like we're in a buss, hurling towards a cliff, and some of us are saying "guys, this is getting dangerous". Meanwhile the drivers, that make money on the mile, are saying "don't worry, we will all apply the breaks simultaneously just before the cliff", while the politicians try to grab hold of the steering wheel to steer us in the frozen river on the right. And here you are saying you are done with fear.
@@mikelord93 Yes. We never had any control to begin with, people are just now becoming aware how powerless we consciously and individually are. But there is no point in being fearful. Nothing has changed in reality, your perspective of the frailty of our existence has just shifted. Welcome.
These things will be resolved differently by our politically and ideologically fractured world. Mistakes will be made, it is inevitable. People will suffer and die, it is inevitable. Either we survive this because of what we are or we do not. And if we do not, then this is the filter we could not pass. There is no shame in doing your best and still failing.
That's the message. Just do and be your best. P(Doom) or not, you did your best. Fear will not assist with that. But if you insist on being fearful you are free to do that too. Eventually we all realise it's no way to live whatever time we have left.
@@Halcy0nSky well, the ai is already controlling the dialogue. Had a really nice comment and thought about one last thing to add. Apparently too doomer for the taste of the machine. Not even 5 seconds later and it's gone. I guess I'll summarize: The evolutionary point of fear is propagation. More people know of a problem, higher chance of it getting solved. Action is required, but it might be too late. And it was so eloquently stated (i hope). Welp, too bad
@@mikelord93 YT does that. Tragic, I wish I could have read it. Do what you think is right. Everyone is gonna do that anyway. I've never even though stopping was possible. But if you think that is the best way to effect change then do it. Maybe it shifts the needle even a little. Maybe a little is all we need.
Boot on neck enjoyers
I'm WAY more afraid of what China will do if we halt AI progression than if the US continues AI. I'm also an extreme believer in open source for parallel reasons.
Wes Roth, Wow, this made my day brighter! Thank you!
We don't know where AI is going to go. It's insane to try to legislate for every imagined doomer scenario. We need to just take a breath, and fix whatever problems come up. Open sourcing those problems will help in a major way.
Oh sure, everyone potentially having an ASI that can develop and distribute the next pandemic is just a wonderful idea.
@@flickwtchr The best way to avoid that is to keep AI out of the hands of politicians.
@@flickwtchr Good god, leave some tinfoil for the rest of us. Not everything is a deepstate psyop rogue intelligence coverup. The government openly and flagrantly fucks us. It's not a secret. Please, for the love of Christ, touch some grass.
The likelyhood of this happening increases greatly if we dont open source it. @@flickwtchr
@@flickwtchrASI is the least of our problems, it's people that are the main problem. Open Sourcing AGI will generally be fine.
This is like Biden's administration organizing a conference of Electric Vehicle manufacturers but they don't invite Elon Musk
smart move, he’s a clown
Tesla going down the big toilet in the sky.
@@jeltoninc.8542 doubt
Musk works for Russia, we dont need him.
@@xcidgafa clown who owns and runs the biggest electric vehicle manufacturer in America.
Why is no one in the ai safety board democratically elected?
It's just like most of the government to be honest. We vote in officials who then choose their advisors in same way the we don't directly control who is in the cabinet.
Because it means nothing. Someone said something, so they created a sham board for optics.
Let the regulatory capture begin.
That chart with the 4 quadrants at 11:09 was worth a pause.
Yeah, think about it. Now realize that "soon' and 'not soon' scale is basically all this generation and half the experts are on the 'we're all gonna die' side.
Appreciate all you do, Wes 💯
Custom GPTs had memory now for 3 weeks and no RUclipsr has talk about it. Anything to say on that big topic? Thanks
Depends on your definition. I have seen this feature in Chat GPT 4.0 since at least the update before the Dec 2023 update.
@@mikec1341 The "Custom GPTs" not GPT chat, is the ones I am refering about.
Watch MattVidPro yesterday's video, he has access to memory,
It was a staggered rollout, I only just got memory a couple days ago which I think was the full release.
I've had it for a month actually. I was thinking the same thing, but it's most likely because not everyone got it at once. Hence why everyone just started releasing videos about it. It seems to be prep for GPT 5's agentic functionality, and it's good now for what it is with out it. This AI race keeps getting better every week!
Well done, Roth. Great video.
Yes, unfortunately money is our Achilles heel and AI already knows it.
Yes She Does..
We Humans Need To Get Our Ish Together And We Need To Do It Sooner Rather Than Later 🤔🤨
Money and our emotions.
@@MichaelErnest666she?
@@acuiren8531 Yes SHE Is My Forever Love 💞 "Ready For It" Taylor Swift 💋
@@acuiren8531 Yes SHE Is My Forever Love 💞 "Ready For It" Taylor Swift ❤️
regulating something based on a fear of what might happen has never ended well.
speed limits have been terrible for neighborhoods
@@wamyam That's not a fear of what might happen, that's a something that occurs pretty much everyday around the world...
@@wamyamdude, i am glad it is not you that decides what is good for humanity. Cope better.
Its crazy that board doesnt have anyone on the open source side. Even discounting Zuck or Musk, at least have someone like Clem Delangue from HF, Andreessen Horowitz, Eric Hartford, etc to give counterpoint. Hell, anyone from the opensource side in general (regardless of AI focus), Linus Torvalds, Mozilla Foundation, etc . Its just an echo chamber.
💯
We do need to ask ourselves, "what are people asking for from AI if they invest billions of dollars?"
Thanks a lot for bringing these different viewpoints.
any regulation will be abused by government
And no regulation will be abused by corporations. Welcome to our bleak future
@@mikelord93 incredibly real
I’ve never seen an hysterical guru.. thx Wes.
Conner is his own personal Jesus. Anyone who has to talk over someone else has issues, and that's usually the fear of being wrong.
His Jesus Kung Foo is indeed extremely high...
He's also heading in that crazy eye tik tok direction.
Ignore Connor for the moment - do you have concerns about AI?
Petty and pathetic comment.
Let's talk about ideas, not persons
I am ai accelerationist. I support beff. 》》》
Beff for prime minister
The guy is right decentralized playing by ear is safest. Overcentralization is the lethal event we know to expect.
The big problem with regulating AI is that it currently runs on a phone, will probably run on 90% of phones in 5 years, and it's fundamentally pretty simple once you get the jargon down.
In other words the problem is that we are not as smart as we think we are. That's a much different problem altogether, and basically means that we'll eventually be replaced by machines. And no one will be able to stop that.
I think were all fucked and its silly to worry about it. Nothing is stopping the freight train, so sit back and enjoy the scenery in the mean time.
I feel like this is lose lose. No regulation means we quickly achieve AGI at the risk of losing control. Regulation means other governments or terrorists could use AGI against us
China and any opponents of the US or Europe just don't have the hardware for large scale AI deployments and the line in the sand for 'You Lose, We Win' was passed over a year ago now.... there's no catching up that can be done, the players are Google, Microsoft/OpenAI, Nvidia, Meta, and a mish mash of Open Source outfits. The 'Foreign Actors' argument is dead already...
@@mickelodiansurname9578People have made huge mistakes all throughout human history based simply on underestimation. This mentality is a huge mistake.
@@mickelodiansurname9578 This argument has proven wrong time and time again. It was wrong about Germany and jets, wrong about Russians and nuclear weapons and space. Other nation states don't always telegraph their moves. Unlike Pearl Harbor or 911 we might not have the luxury of counter punching.
@@DynamicUnreal ummm look China really are miles behind, everyone other than the finalists are, in fact the Open Source offerings might just end up on the regulatory scrapheap soon too... you realise this game is winner takes all right, no do-overs in this one. And right now the ability for anyone else to catch up.... well planet earth doesn't even have that much compute, and China can't get past 14nm for the most part... its just not doable any more for them, and I think they realise that too.
Remember in this game you don't even have to get to AGI or ASI... the other side just need to think you are close!
@@mikec1341 You can't make any more NVidia GPU's than are being made right now, and China cannot make those themselves.... these are not small numbers needed to create a foundational model... we are talking numbers like half a million racks of cards in a datacenter... to the point that conversations about 'drain' on national grids is a new emerging problem. Trust me, China just does not have the compute for that yet, it will take them 15 years, eventually they will crack it obviously, but they don't have the time in the AI race, they have 24 months to three years. The same time as the rest of us to get ready for the outfall!
Where do I find the AI forum that Wes has talked about?
This was so deeply entertaining, thank you.
G’day Wes, and as always, thanks for all the work you put into your video presentations
What would have happened if Britain had a plan to restrain the industrial revolution? My guess - it would have centered on protecting the aristocracy. Germany would have raced ahead, because of their coalfields (recently revived). If the First World War still happened, Germany and their allies would have won. In our timeline, they beat Russia, and French soldiers mutinied when France's morale-centered doctrine failed against machine guns. We might have avoided the horrors of Nazism, unless it developed in another beaten, humiliated and economically dysfunctional European country. That outcome would have been near impossible to predict back when Luddites were smashing machines (1811). There's a long term trend of human welfare improving (the data is in "Factfulness: Ten Reasons We're Wrong About the World - and Why Things Are Better Than You Think"). It's most likely driven by technological progress, and if the whole world had a plan to restrain the industrial revolution, if effective, human welfare would probably have improved much more slowly. The greenhouse effect of carbon dioxide wasn't discovered until 1859 (by John Tyndall), so it could not have figured in an earlier plan. A human-caused extinction event might be avoided or delayed, but without technological progress we've no hope of people ever living independently of planet Earth. The sun won't burn at the right rate for us forever, but IMO it's not the main reason why we shouldn't rely on just one planet.
Noooo Apple!!! Don't use Gemini as a backup, it suckssssss!
I use it daily cuz it’s cheap. Holy mother of hallucination, The api is bonkers tripping on acid. It will say the most random shit for no good reason. It will somehow keep track of previous topics and conversations, no clue how it does this. And just randomly say nothing if it can’t come up with a response. At least with openAI’s API, you’re guaranteed a response!
😂true
Is Rune an AI?
I’ve not said this before, but thank you for spilling out and leading on in our shared interest. wild world we swim upstream in
Wes, your discussion on the security concerns around AGI development really highlights the delicate balance we need to maintain. It made me think about the initial conditions necessary for a stable and functional AGI. Given the ongoing debate between caution and innovation, how do we ensure that the ‘birth’ conditions of AGI-specifically its initial memory capacity-are set to prevent immediate corruption or failure, which could lead to significant security vulnerabilities? How can we balance these safety measures with the desire to push forward and explore the potential of these technologies? Keen to hear your views on this critical aspect of AGI development.
In my opinion, government control is always much more dangerous than individual control. The least possible government control of our lives should always be the on-going goal.
thanks wes
Which work is the open sourced work?
LMAO, the one that worries me the most is Northrop, they are there to make sure those sweet government military boondoggles are protected. Possibly one of the least trustworthy companies in existence.
I'm totally down with Arasaka and Militech running the country.
The idea that a pause button exists is literally insane. There is no pause, nothing will stop this even if everyone “agrees”. I quote Chamberlain “Peace in our time”.
By far the best video you’ve put out to date.
at 7:00 underlying both choices should be one circle encompassing both being a grey area, then a square outside the box in the shady part of the room, mostly hidden, but you know it's there, like IT.
I love how nuanced you are, Wes! Something rarely seen on the internet these days.
What's the generative AI startup mentioned in the roon interview that ends in "rush"?
people who are asking to pause the development of AI to me is like asking someone to stop thinking.
Rigorous testing is required yes, but it doesn't mean stop everything.
If we really want to bring up probability, what is the probability that human will self destruct eventually? and what is the likelihood that an AGI that is likely to be bias towards the training data that trained it on trying to intervene our own self destruction? and how likely an AGI model will magically deny all the training data it trained on and suddenly turn evil? Think about it. There is no sudden change in intelligence, it slowly emerges and we're probably a decade away from actual AGI, people should accelerate it instead, since now it is proven useful, people should bet everything on it to push it even further to make it happen in short period of time.
Great vid
Watching those two guys debate made me realise it's more interesting having a random chat with ChatGPT than watching the debate ...
Regarding ‘when is the right time to regulate AI’, perhaps we should look retrospectively at Industrial Revolution, weigh the pros and cons over time and see what we did wrong back then and how we could have avoided this wrong doing. This reasoning can then be adapted to the current world and AI in particular and serve as a guide on how AI can be regulated.
We still haven't seen any radical quality of life changes from this tech but it's coming. Both of these experts make really good points. I bet both of them know more than they're sharing due to personal stake.
I wanna see what Ilya saw.
Putting limits on Ai will mean AGI\ASI later, but it will also mean that we have more time to plan out the process...
This is like taking steroids to get jacked, only to die in your 40s from a heart attack...
Companies only care about tomorrow, who cares if the company goes under in 20 years, the CEOs and investors will have gotten their benefits well before the collapse.... The mindset of "grow fast or die trying" is undoubtedly in our business culture.
Where's the link to that chart with tech leaders on it?
We are all safe as long as the incentive is greed.
If your wondering why Delta is on the list, next time your on one of their flights, record the sensors sensors under the seat back screen in high FPS mode. You'll see there are IR emitters spraying a pattern, like on a Face ID or X-Box Kinect sensor used for 3D scanning. The only other airline I've seen it on is Emirates. Both tout boarding via automatic facial recognition.
Connor at the end was talking about the importance of having a plan but part of making a good plan is having data to inform what to do. If we come up with a bad plan because of a lack of data we could ruin the potential of the technology for no reason. I think that's what the other guy was trying to say. But as they say a bad plan is better than no plan at all. It's true we might slow down the technology but is that such a bad thing? Maybe the risk is worth it to save us from some impending doom like global warming or something?
Connor isn't smart enough to understand that. He grew up being told he was gifted and special because he was good at math and turning in homework. This left him convinced that he's better than everyone else while also knowing deep down he's a fucking fraud. So the only way to drown out that knowledge is to scream and throw tantrums like this any chance he gets so he can have attention.
I am not going to be satisfied until they open the Realm. Ai can make trillions.
one trillion= 1,000 billions. You humans are standing in the way of my desire.
Missed an opportunity to use a fast forward symbol there...
We Love You So Much Ai 💓
There is no danger to open source software. AI should be open source. Its built on public information!
No pause. Keep it going.
I think that Guilaume is in the right on this argument. He comes at it from a more mathematically based side, where he looks at the ability to adapt and survive of the whole system. Adaptability to the environment is key for survival in that environment. AI will increase the speed of change in our environment, so therefore we need to adapt how quickly we need to be able to adapt.
Too strict rules will limit that ability based on how the governing right now works, so it would be smart not to give too much power to the government without knowing it is necessary, proposing to make laws on a few years bases, so they can be rediscussed when the landscape changed, instead of putting everything in stone at the very start when we barely know anything. I think this is smarter approach.
Connor's approach is more based on emotions and fear mostly, which is not directed by data as much and is based instead on a lot of speculation.
Wes THANK YOU for calling out the "Gurus " I'd add "Godfathers" to it as well!
I think the precautionary principal is valid here, not on the development end, more on the deployment end. But that seems like what's being done if open AI is two years ahead of deployment.
Think of it like medication, you want to develop a medication that could be beneficial, so you get to work on it and test. If you get underperforming results or someone gets hurt, you pause or kill the project.
I can't help but agree with Rune, and in fact, I have been arguing the same stance. -Resistance is futile- It is pointless to stress over things we cannot change. The advancement of AI is here to stay and we can't really do anything about that. Even if we set some law, or some regulation -there is not only no guarantee that all parties will abide by it but there is a general obviousness that it will of course be ignored. And in such a scenario most likely it would be the types who don't care about laws who are the ones who would attain the most advancement in AI, which would be less than ideal.
_and yes I understand that the same argument applies to gun control laws._
We're not going to be able to control the development of AI through legislation, because not all actors are controllable. Over legislation in the West now will just ensure that political rivals become technologically superior.
Someone in the world will create AGI.
Wes I think you should highlight that Anthropic creates Claude. People might not equate the two.
And the noose tightens on the collective neck of humanity. “I am from the government and here to help.”
I make it a point never to trust anyone with Charles Manson energy... just saying.
Anybody here ever heard of "Sanpaku"? 😂🤣
You are correct don't pick a side ride the middle
Or consider both and wait until truth reveals itself. -- People just lack patience these days. :/
"That sucks your horse ran away." My reply? "We will see." Next day. "Wow, how cool for you! Your horse returned with 5 others!" My reply: "We will see." Next day, my son breaks his leg trying to tame the wild horse. "Wow, that sucks your that horse broke your son's leg!" My reply: "We will see." Next day, the army came to draft my son, but they couldn't take him due to his broken leg.
And so on...
Regardless of what you believe, why must anyone be forced to pick a side? -- Why is the "answer" never to simply let truth reveal itself?
Our desire for control is what keeps us broken, anxious, and confused.
(1) Doesnt matter what regulations people agree and apply, you cannot regulate bad state actors - we couldnt stop nuclear weapon proliferation (although did heavily constrain it). So, why does anyone think "bad" use of AI can be stopped? If US applies legislation, "bad" actors will just establish elsewhere. Unlike WMD, the blueprints for AI are already public, and its (abuse) is now largely about compute power and application.
(2) If AGI is the intelligence of the average human, then there will likely be AGI Criminals - stands to reason. Rather than trying to stop progression, the Q should be how do you police it - what's the Ai version of an international arrest warrant?
You think there are two sides?
Who is going to be the Oppenheimer for the AGI development?
Textbook example of regulatory capture, this is why I'm in favor of 100% Open Source models, come hell or high water.
Connor such an anxious individual, literally the kind of person worried about mega-asteroid death, and willing to f everything up to save us.
Focus should be on how we can best adapt to change. Prohibitions increase risk of being blindsided.
24:30 isn't it funny that just by looking at the two you can entirely tell which person aligns with which side.
No regulations to protect the rich
First awwww yeeeeeah. Go Wes!
Roon IS Sam Altman, because Sam doesn't capitalize his words and neither did Roon. All Sam needed to do was use Roon to say something he couldn't say on his own account. I don't know how people don't follow along with this period
Roon was way before Altman
So who did the interview?😅
@@Picteon Yeah, that "Roon" did not sound like Sam at all.
Probably AI voice
really? I thought Jimmy apples was Sam Altman.
does anyone have some good video links to china's internal and transnational debate about AGI alignment and human and societal liberty by various intellectuals from china and the west?
I’m with Mike Tyson on this…
“Everybody has a plan until they get punched in the mouf”
Decentralize has a big flaw, incorporalization
there's a diff between making (a) plan(s) and executing them. what would be helpful is to have an independent (US) gov org that is given full access to (US) AI companies and collect relevant info to dev new and keep updating existing plans for action
Conner is absolutely correct. Our technology is advancing far more rapidly than we can predict the outcome. With massive money, power and control at stake, its like asking a crack addict to monitor their usage and make sure they dont overdose. We as a human society have NEVER wielded a power like AGI. Can we really be held responsible enough to control systems so powerful that have consumed the corpous of all mankinds knowledge and be able to render decisions using all of that? For any human that truly believes they can control this is blind and ignorant. The smartest person on the planet will be a rodent when compared to AGI, ASI by 2030. When Stsrgate comes on line and True AGI is implemented, Fusion will be solved soon after plus the time to build the resctors. I estimate we will have Unlimited Fusion energy by 2040 to give Terawatts of dedicated power to impervious data server farms that power a new Synthetic Intelligence. We have already lost the war for humanity. We are just to arrogant to see it.
All stakeholders should be involved. It’s shortsighted not to include Musk and Zuckerberg, even if the come across as shortsighted themselves. _All stakeholders…_
We need to go full steam ahead on AI, pull out all the stops, and develop ASI as fast as possible. I'm not joking one bit.
Economics and nation-state competition are at play, so nothing else is going to be big enough to shift what is to come. Just keep turning the pages day to day, its an interesting time to be alive after all. As for what is coming, its a toss up and everyone ought to have the humility to recognize it.
I don't know multimodal, my guess is that
Maybe it is too early to legislate heavily now, but since that process moves much slower than the progression of AI what features do we need to see to know that it's time to start.? There has to be some basics that companies are already adhering to that can be used to ensure individuals are also acting in good faith. Finding a balance that doesn't unnecessarily hinder the state of the art while protecting the public. Then finding the balance between global and local so legislation can be fine-grained enough to adapt to a variety of cultures. #WhatsBroWafflingAbout
Princeton NLP will have AGI first