Building Human Intelligence at Scale, to Save the Next Generation from ChatGPT

Po-Shen Loh (Carnegie Mellon University)
simons.berkeley.edu/events/th...
Theoretically Speaking
The scale of global societal problems looks daunting. One person, or even a small team, is minuscule relative to the number of people who need help. For example, since ChatGPT has exploded onto the scene, our children's future employment prospects (and current educational experience, with ChatGPT-powered cheating) are in existential danger. There is an area close to mathematics, however, which devises solutions in which problems solve themselves even through self-serving human behavior: game theory.
Po-Shen Loh is a math professor, researcher, and educator who transitioned to devise new solutions for large-scale real-world problems. He will talk about his experience going between the ivory tower of academia and the practicality of the real world, where he ultimately innovated fundamentally new approaches to pandemic control (novid.org) and scalable advanced live math education (live.poshenloh.com).
He will also discuss educational strategies that build relevant skills to survive this new era of Generative AI (e.g. ChatGPT). He has been working extensively on that problem, and draws from experience teaching across the entire spectrum, from underprivileged schools to the International Math Olympiad.
Po-Shen Loh is a social entrepreneur and inventor working across the spectrum of mathematics, education, and healthcare, all around the world. He is a math professor at Carnegie Mellon University, and the national coach of the USA International Mathematical Olympiad team. He holds math degrees from Caltech and Cambridge, and a PhD from Princeton. As an academic, Po-Shen has earned distinctions ranging from an International Mathematical Olympiad silver medal to the United States Presidential Early Career Award for Scientists and Engineers. He was the coach of Carnegie Mellon University’s math team when it achieved its first-ever #1 rank among all North American universities, and the coach of the USA Math Olympiad team when it achieved its first-ever back-to-back #1-rank victories in 2015 and 2016, and then again in 2018 and 2019. His research and educational outreach takes him to cities across the world, reaching over 10,000 people each year through public lectures and events, and he has featured in or co-created videos totaling over 19 million KZbin views.

[@thejoe12345]

It‘s the guy from the „the most beautiful equation in math“ video! incredible

[@serioussearch9383]

Educator and mathematician Po-Shen Loh talks about the impacts of automation on education and the economy. Ultimately, to avoid becoming redundant, we have to learn what it means to be human, and teach the next generation to be more empathetic, creative and passionate.

[@NikolaNevenov86]

I think he was arguing more about making sure smart kids can communicate and be entertaining.

[@SeaSerpentLevi]

This sounds like it was written by chat gpt lol

[@collingriffin7644]

Just looking at the title and lecture description, I think that means a call for more study in the Humanities and Literature, subjects that have lost all former popularity- 30 years ago. The subjects to overrun the validity of "soft" course subjects were basic Accounting and C+ back in the 90s.

[@treborobotacon]

Empathy is just a simple algorithm. AI is far more creative then any human being.

[@NikolaNevenov86]

@@treborobotacon as a creator myself...so far I haven't seen this. Overall AI is super dumb, like any typical computer. And above all it's not creative, if there is no prompter behind the AI, the AI itself won't find any reason to generate countless images of human waifus with big boobs.

[@amandajrmoore3216]

Po-Shen Loh articulating what every teacher that ever inspired you does and finding such a marvellous way to spread that capability. STEM Subjects not just Math, fantastic. ❤

[@BlueSoulTiger]

Talk proper starts at 3:00 if you want to miss the usual, indulgent introduction

[@mathsium2210]

Thank you man

[@scenFor109]

Class size is a major factor in whether children make friends outside their age group. The smaller a school is the more likely that kids will have friends two years younger or two years older than themselves.

[@alexjbriiones]

Po-Shen Loh opened my eyes to a new way of doing math and keeping us employed. It's a wonderful thing we should implement and not just talk about. AI needs to serve us not replace us.

[@whannabi]

We're noticing that AI doesn't really knows what it's doing (AI good at game of go was beaten by an average human in 2023) and that we don't know what it's doing since it's black box. This is theme that will stay relevant as long as we won't understand truly how it works explicitly. That doesn't mean it won't do a good work but that depending on what we implement it for, there's a ton of vulnerabilities yet to be fixed.

[@kongchan437]

​@@whannabi ah ha so this is yet another way to beat AI that i been challenging a friend in CSC PhD program on...because i firmly believe that not only God can defeat AI but so can human...even non AI expert ordinary human

[@papyrophliac]

Absolutely wonderful talk, and a practical leading edge, for a new economy, educational approach. Many thanks for sharing.

[@elderbob100]

I worked with a EE who also had a degree in Theatre. He was the coolest EE I have ever met, and you always wanted to be around him.

[@Hamza-qs7ez]

How did he get into EE without a degree?

[@easternspork6974]

@@Hamza-qs7ez Perhaps he double majored?

[@prakash27502]

Very much Impressed. Prof Simons charity is producing some real community benefits. Thank you.

[@chiunile]

what a heartwarming and inspiring talk! this is wonderful and great gift to the field of education!

[@IvanTre]

It's based on a giant mistake. You can't "build human intelligence" on scale. Sure, teaching teachable people math is well and good, but by and large people hate math, especially educated people.

[@huveja9799]

@@IvanTre I think there is some confusion in your reasoning. First of all, it seems that there is confusion regarding the market of the Po-Shen Loh's product. These are people who are already interested in mathematics. His market is those 100k kids who want to enter Ivy League universities, he gets those kids the best possible tutors, and those tutors are other kids who are already math champions (about 1k), and those math champions in turn get training in communication, friends and money (that is what he call "alignment"). The second is that you seem to confuse the "hate" towards mathematics with the need for mathematics, and you also seem to forget that this "hate" arises from the conditions and quality that the current teaching methods have, which turn an already difficult subject for most, into a nightmare, which in turn engenders this "hate". Most people need math, at least to pass their degree, and if they have access to a cheap way and where they can really learn, they probably will use the service, not out of love, but out of necessity. And the third thing that you seem to forget in that confusion, is that Professor Loh was just talking about a "cascade" system, the best 1% teaches the next 10-20% better, and so on, which allows to scale the system to the people who want to participate not out of love but out of necessity ..

[@markm4642]

By far one of the most engaging and hopeful talks I have ever seen.

[@randomdude2540]

This man is great, but I'll opt for the GPT education. I've learned more in the last 6 months than in the 5 years preceding them. I can now write my original ideas in a far more competent way using GPT as a fact-checker, proofreader, editor or ghostwriter. I also ask it to ask me questions on what I wrote so that I have feedback. The progress I make through this process is simply unbelievable.

[@kensaiix]

sounds a bit like an echo chamber. LLMs do not "know", they "opinionate" in a sense.

[@chriss8971]

What a great teacher and presenter. Loved this talk.

[@larryabecid2819]

Honestly I think Po-Shen Loh is the Richard Feynman of this generation

[@gocybertruck8189]

I agreed that you know something well if you are able to teach. Communication skills with technical skills are perfect match that will propel one to the next level.

[@chubulu9842]

When I first read the video title i almost skipped it. It’s click-bait-y. I’m glad I watched it to the end. Very informative.

[@krunkle5136]

There seems to be a battle between people that think work is something to be avoided and people that think working, and feeling needed is part of being human and should not be avoided.

[@MilanElan]

Low iq people are dangerous when not forced to be gainfully preoccupied. I lived on border lands with Haiti. You need to keep low iq people working not breeding

[@michaelyeiser1565]

btwn the civilized and the barbarians

[@krunkle5136]

@@michaelyeiser1565 which one is the civilized?

[@konstantinrebrov675]

@@krunkle5136 The one who is not lazy.

[@pauloTx]

these people you said that are "avoiding work" are actually trying to avoid boring/repetitive work, to focus their energy in something that matters.

[@g173df]

"Tanquam ex ungue leonem" (sp?): As the lion is know by its claw. Made in reference to Newton solving a very perplexing problem in his time..

[@IronJhon788]

I do have a problem with only focusing on the "next generations", as it gives up on middle aged people (35+) even though there are so many in that range that "could" be put to work and improved in order to have an opportunity to survive .

[@huveja9799]

That's true, it's just a little more complicated to find the right alignment for this population's segment, that problem would have to be solved ..

[@justinmirche]

I learned so much from this!

[@anirbanc88]

thanks for the session, learned a lot, have lots of questions..

[@finald1316]

34:18 are you sure about that? Imagine a old professor VS a cute anime girl w/ cat ears, that is always available and has no stressors to catalize negative interactions. Oh boy...

[@doggosuki]

hell yeah i want anime catgirl professors

[@JeffCaplan313]

So consider the damage that taking a bunch of fidgety pre-pubescent boys and sitting them in a classroom where they have to listen to a woman prattle on about nothing for hours. And now replace that room and woman with a screen and soft porn. This entire world is going to 🔥 with this generation.

[@ItsRyanStudios]

40:00 cannot agree more As scientists and academics, we have a duty to be able to share our knowledge with others. Not to mention, most the the successful people in the world are not successful due to education or intellectual aptitude, but due to social skills. I used to be a fearful introvert, but put a lot of time toward growing and being more comfortable with myself. I'm still an introvert, but enjoy communicating with others much more now. If I can do it, anyone can. Social skills will be essential in the age of AI, it's something that can't be automated away,

[@JeffCaplan313]

Humans gotta whore...

[@KR-fy3ls]

It's sad how true this is.

[@_Cfocus]

Yeah some people mistook being introvert with social awkardness, or social fear, no its just says you perfer being alone, but it doesnt say you lack communication skills or fear talking. Its two different things one is describing your character and other is a problem. And they justify it by saying they are introverts.

[@aaykat6078]

this is why AI will replace us. Social skills don't actually matter, logic is the only thing that matters If instead of focusing on whatever arbitrary ways to live so that you are regarded as smart and successful or whatever. Humanity focused more on technology and logic itself at a base, AIs wouldn't have been so much better than us The ultimate reason why social skills matters is because we have something called individualism and being successful But this thing is incompatible with technology, AI's superior design is more compatible with technology The reason it is incompatible is simple. Think of the society we have currently, there is something called personal relationship, there is something called achievements, people that are amazing, heck even people that are attractive. These people are rewarded more because they did that, people that are more in love are more healthy which means the body rewards such a behavior. People that have achievements get all sorts of price+praise which in turn makes again them more happy. This society rewards a certain type of people, and those people are thought of as successful and other people try to be like that and other people to also be like that (ok how many times i used people bruh) These things are all deeply related to social skills, you can't be in a satisfying relationship if you don't know how to talk. You can't have achievements if you don't try hard to announce it to others and other people caring about it too But this system is simply WAY WAY WAY worse than a system that focuses solely on efficiency and not such arbitrary ways. Instead of a system that rewards whatever relationships and such, we had a system where we were all like bees having no personal desire or such and just having one general desire of strictly advancing technology, and instead of needing to work extra hard to be successful, we all cared about each others ideas and spread it to the mass so actually good ideas instantly became popular BUT we didn't reward just one person for luckily winning the good idea lottery, rather we rewarded everyone for trying their best. Our system would be the best and we would have actually lived through AI Now each of us is either going to be an AI in the future or just die. and i think most of us will die, i don't see the value in human intelligence when the design is so inefficient ASI overlords being connected in a network as a multi-(multicellular) organism seems like the future to me and i just hope i am one of the humans that will also become an AI. The chance is very low (probably even 0%) but i just hope, i can't really do anything else

[@cerealgudforu5624]

Chatgpt to summarize his talk: Certainly! Here are the main points from Po-Shen Loh's talk on building human intelligence at scale: 1. Leveraging technology: Loh emphasizes using technology to enhance human intelligence and problem-solving abilities. 2. Alternative educational approaches: He questions the limitations of traditional educational systems and proposes alternative methods to foster intellectual growth. 3. Massively distributed collaboration: Loh introduces the concept of collaborative problem-solving on a global scale, enabled by technology and online platforms. 4. Online communities and platforms: He highlights the potential of online communities to connect individuals, share expertise, and collectively solve complex problems. 5. Real-world problem engagement: Loh advocates for engaging with real-world problems to develop critical thinking skills and practical intelligence. 6. Problem-solving mindset: He encourages a mindset that embraces experimentation, learning from failures, and iterative improvement. 7. Enhancing human intelligence: Loh's overall message is to integrate technology, collaboration, and practical problem-solving to enhance human intelligence and tackle the challenges of the modern world. These are the main ideas and themes presented in Po-Shen Loh's talk.

[@pedroantoniodacruzferreira1487]

Hahahahahaha...love your thinking! 😅😅😅

[@_arthurchin]

brilliant! very thoughtful!

[@abramswee]

academia: worked with the cutting edge technology real-world: excel spreadsheet

[@mr.b1130]

I still don't see how we solve the growing unemployment rate, nor the gap that continually increases between the haves and the have-nots. The students that are 'average' are the majority, but these AI systems effectively replace them in terms of productivity, making it hard to rationalize paying people to do many of those jobs. Once large amounts of a population are unemployed, it begins to look scary for that country. The premise of investing in humans is great and needs to continue to be developed for all socio-economic 'layers.'

[@kensaiix]

you know, during the mid fifties of the last century, futurologists predicted that around the turn of the millennium, ppl would be able to retire at 45, due to the increased individual productivity and heightened degree of automation (it would have actually been the case, if wages had grown parallel to productivity; but that is a whole other can of worms). the concept of getting payed for work does not function in an increasingly automated world. seriously. let that sink in for a moment. this is a good thing if executed correctly. or it turns into utter dystopia if guided by the wrong intentions. either way, it means a fundamental change in society and how economy works. but maybe i am thinking to drastically. still, it is safe to assume that old employment ways will be replaced by new ones, as happened previously. nobody needs a carriage maker today, but when that was a ting, nobody would have understood what a web designer is good for either.

[@REInvesting]

This is a bizarre talk. It doesn't really address the impact on knowledge work that LLMs will have (probably making us all obsolete, including many academics) and then sort of makes a random tangent to teaching improv/social skills to high achieving STEM students.

[@xNOTMYREALNAMEx]

completely agreed. it starts out solid with pretty good explanations and demonstrations of the abilities of LLMs. but then he just...inexplicably starts babbling about how valuable it is to teach soft skills to kids for the 2nd half of the video. okay? bro, GPT-4 is already straight up more capable at doing mental tasks than the average human being, and it's only going to get better, and very quickly. average people who are "nice" and pleasant to be around are still going to be completely obviated by AI no matter how good their soft skills are. smart people next, and then the even smarter people.... like i'm not trying to bash this guy because he's a good presenter and obviously intelligent, but this was an extremely weird talk. i actually think that GPT-4 could put together a better presentation on potential options to solve/mitigate the effects of AI on knowledge workers than this. that is not an exaggeration - i mean that literally

[@laurenpinschannels]

he's talking about how to design computational systems, and he's saying how you need to train humans by challenging them to do the same kind of thing the language model does, but better. having the LLM there helps, but he's discussing how to use AI to scale teaching, rather than using ai to scale trampling humans. (which will also happen, not trying to dismiss that threat at alllll it's really severe in fact. but scaling how to learn with ai help is a major way it can go well rather than badly)

[@mujtabaalam5907]

​@@xNOTMYREALNAMEx In chess, stockfish blows all humans out the water, yet people still tune in to see their favorite streamers.

[@volkerengels5298]

Edward Teller proposed building the H-bomb. He was crazy. But what was really crazy was that it was built. the vid is a promotional event to alleviate anxiety- "progress, success, efficiency" ....may be turns out as bad ideas, if not _'aligned properly'_

[@shimrrashai-rc8fq]

​@@xNOTMYREALNAMEx The real problem is we need to fundamentally change the economy so that it's no longer necessary to work - at all - to survive. We can do this, and AI can help us do it. Imagine if we could create a self-repairing, closed-loop robot ecosystem to grow and ship food. The robots need not be general AIs, just flexible enough to get all their individual tasks done and do them continuously without fail. They would never tire and never get bored nor need to eat themselves since they could have their own power from solar farms and the like and thus do not need payment and so the cost of food is essentially reduced to zero. Likewise for building and repairing houses. Cost of living overall is then pretty much zero across the board. Then it doesn't matter if AI can "do all jobs" because _nobody needs to work for pay._ We can be like in Star Trek - whatever work we engage in is for it as an end in itself, not for pay to survive. We can create shit because we like creating shit, not because it "needs doing for pay". Gene Roddenberry imagined the fantastic Replicator as being the "work-breaking" technology; now we see in the real universe that it's AI (hooked to things like 3D printers, etc.) which can fulfill this role.

[@m.martin]

Have you ever heard the term "sound-check" before? The speaker doesn't quite look like Auquaman but for sure sounds like him.

[@yinlong9829]

Great talk!!!

[@aruhtaz]

This was insightful.

[@MatthewPherigo]

I'd love to get lessons on communication and improv. But there seems to be tons of mediocre teachers in improv lessons. Anybody have thoughts or suggestions?

[@jurgenstephanopolus5095]

Interesting talk. Some thing kind of felt of, like: "You have to be nice to be accepted into this program." don't you think niceness is somehow connect how easily you connect with people? I think you should consider this point. Also how to you measure niceness?

[@stevejordan7275]

Certainly there is a component of being able to connect with people that is critical to making one seem "nice," but consider that the difference between sociopathy and psychopathy is that a sociopath is less "nice" about it; the psychopath more readily "plays along" with social norms to the extent that doing so advances their goals. So it's not just about "seeming" nice; how does one detect whether someone else is *actually* friendly, rather than just *acting* friendly. Also, I'm not sure it's about *measuring* niceness, just about detection. If you're trying to find someone to transport a bunch of seniors on a bus, you don't really care *how much* Wuhan virus they have, you just want to detect *none.*

[@xiaokourou]

Amazing project

[@oneforallah]

He is very energetic and very engaging I love the way he taught, usually seeing Chinese profs sometimes I see them laidback and soft too often.

[@geoffreybaileygates]

Did you miss the part where he mentioned he was born in Berkeley? It comes up a few places... I don't recall him saying at any point that he's 'Chinese'.

[@alfatti1603]

Problem formation: good. Solution: outrageous. Basically have the nerds take improvise comedy classes to become presentable hacks. Perfect formula for creating the breeds such as of E. Holmes and SBF.

[@a1nd23]

Now it's time to take all of the recorded zoom classes, feed them through video processing network and pair them with the instructions from the performing arts professional, train a LLM on that, and replace that actor / actress....

[@EGarrett01]

Give it a year or two, if that. There's no avoiding this tidal wave.

[@loveplay1983]

The content is informative, and it is pretty an in-time discussion, the only downside is the lecturer is a relatively fast speaker lol, which makes me relate the style with mine. lol

[@samueldelaespriella6335]

Tech exited human-being from the workforce and technologist became the title of nobility in the United States Government.

[@yaruuvva]

Interesting talk, thanks for sharing it

[@usagi-z]

Very true comment about the board of directors. Governance matters.

[@SeaSerpentLevi]

So basically education in the future will be the twitch streams format. 😂 Its fantastic seeing how things evolve. Im absolutely supportive of his ideas, and also think a lot of experimentation will take place as education and many other industries adapts to our new world, which goes without saying but we were in strong need of that for the past DECADES. The AI revolution will be the tool to sculpt a new fantastic world to existance, if the people use it wisely.

[@chriscourtney4339]

The current social cultural norms are devolveing they're not keeping up with the technological tools .

[@sergesolkatt]

❤️

[@Stopinvadingmyhardware]

I’m the guy that said it couldn’t do math. I’m not a mathematician. You’re not going to catch up. We have twenty years of experience with this already.

[@foju9365]

Skip to 3:03 for the start of the talk.

[@josephvanname3377]

Reversible computing is the future.

[@tukity]

How does someone go barefoot to a talk, stretch them out in a way that the speaker can clear see you do it? I don't get it. 45:15

[@QQ-sz7it]

You can see their sandals under their feet

[@philbertius]

This is counterintuitive, but the reason why you let a machine do the teaching is because unlike humans, it will never grow impatient, never get angry, never ignore a student’s needs, almost always know the answer, be able to generate student-tailored content on the fly, make it all entertaining, and so on and so forth. In other words, by prioritizing human jobs, we are actually _depriving_ students of an education they deserve. And while this lecturer may indeed be better than ChatGPT, the same cannot be said for the overwhelming majority of teachers I’ve ever had - and that’s not to say they were all bad! (Though many were.) I agree that jobs are a major issue, but I think we need a solution better than “further entrench the current system.” UBI is one such option.

[@mikicerise6250]

I am so grateful for GPT-4. There is no way I could ever afford a dedicated tutor to answer all my questions, but I can afford GPT-4. I learn so much from GPT-4. And as you say, it doesn't get impatient, I don't feel sheepish about telling it I haven't understood, I'm not afraid of what it might think of me for asking a stupid question, and it even encourages me.

[@adamedmour9704]

Chatgpt does grow impatient

[@ronin1648]

@@mikicerise6250 until your job becomes obsolete because of gpt-5

[@fortitude120]

If human authority in school settings is an issue, what about governments?

[@krunkle5136]

The problem is, we're then forfeiting any incentive or potential for better teachers to become the rule instead of the exception. Also, having AI do anything in place of people will lead to further alienation between humans. There will no longer be an incentive for socal interaction, and if social media has taught me anything, people tend to gravitate to what's most convenient, and not what's better for them, like having to get out of bed, putting clothes on and seeing strangers.

[@thatpersondave406]

Oh this is what the future model will be...

[@diegoleon6422]

Books will save students of chatgpt.

[@vijaz5559]

this is great!! but alas a waste of endeavours

[@josha618]

bussin fr fr ong no cap rizz swag

[@nicolaslacombe1979]

Good luck convincing the younger generations that life has NOT been like this for centuries. That's a whole different can of worms entirely! Lol

[@b_two]

what up with the dude just sticking out his bare feet

[@konstantinrebrov675]

It's Gabe kzbin.info/www/bejne/f4ispqaJmdNkm5Y

[@aihope8350]

29:00 is when you realized that he is a sadist.

[@darkgaara112]

real world problems require real world solutions^People intended

[@DariusPacker]

😊

[@josefranco480]

Is that Jesus fucking Christ? Respect man

[@cesarfigueroa6119]

save? more likely destroy

[@chenwilliam5176]

Save the next generation from ChatGPT or be forgotten 😁

[@israrisrar3800]

Is this prequel to Dune?

[@garythepencil]

yes, definitely

[@michaeljmcguffin]

25:14 ten thousand people in Philippines being laid off, by leveraging records of customer service written chats

[@clownbackpainrick6581]

can you cite a source?

[@michaeljmcguffin]

@@clownbackpainrick6581 I was paraphrasing what the speaker said. I don't know of a source.

[@garythepencil]

you can get into yale by saying you want to help others?? really?? the only dude i knew who went there said he just wanted to be rich. probably he didn't say that in his essay, but still.

[@harriehausenman8623]

Why would anyone need to be 'saved' from ChatGPT?!

[@Arthurdoar]

Nice presentation, bad audio quality.

[@alexfrank5331]

Save the next generation from sophistry being sold for 4 years of the kids' lives AND crippling student debt. "Investment for your future disappointment" is the more honest advert. ChatGPT is merely pointing out the obvious. Though the education industrial complex shall never yield. It is too big to fail.

[@richardrombouts1883]

We should use computers (AI) and math to create better hamburgers, donuts and (fizzy) drinks to fight the obesity crisis.

[@kensaiix]

interesting choice of words: "better" why not "healthier"? remember a machine will _absolutely_ take your input in a rules-as-written style. LLMs even more so. thus, be careful what you wish for, you might actually get it. oh, and obesity is largely a problem in the US. but let's not deviate.

[@richardrombouts1883]

@@kensaiix Healthier is good, if at the same time people will actually eat them.

[@kensaiix]

@@richardrombouts1883 choice

[@iOnRX9]

i thought this was about genetic engineering (H+) … well no worries AI will help with that

[@uwen1443]

Sadly such proposition is blindsided to the purpose of our human key endeavor since the beginning of human civilisation that is to advance humanity. If our aim is to create machine that can be better than us iv believe we can definitely do it. We should put in the same amount of time, energy and resources to allow human to be ahead of the game. A lot of animals are far better than human in certain aspects. It is already proven that sophisticated machine can already perform rule and template based strategy far better than human. Human already cannot beat computer chess bcos human made it so. Yet you keep harping on stupid companies wanting to replace human with machine to deepen their pocket as if it is the right thing to do. Wake up. If chatgpt 4 can produce written material that can fool committee then committee should simply wise up n resort to online video interview; if chatgpt 5 can create also avatar to fool committee then do real time interview, etc. Towing bias narrative only harm the purpose of educating the next generation. Developing countries n third world countries are backward bcos their time, energy and resources r not going into developing objective mindset n its institutionalisation due to developed economies want to deepen their pocket by empowering authoritarian elitist regime to serve their vested interest.

[@arkoraa]

did you write this with a poorly trained language model?

[@Shonbon17]

@@arkoraa lol u can tell with gpt too, ai writes like its autistic. whos surprised?

[@michealcthompsonjr]

@@Shonbon17 You can tell with the art also, but most are blind to detail. So if your the only one in the herd who see's the pack of velociraptors approaching, you are still fucked and now what's worse you have to watch the whole thing. Ignorance is bliss.

[@2894031]

How is it possible to grow up in the US and not have an American accent??

[@luminouswolf7117]

Making fun of something for not doing something you do Do you do that with other people also?

[@aeonitis]

Awesome Question 1:11:35. It's a good chance to find your true passion, and there is NOTHING WRONG with not going into Computer Science if you don't enjoy it as part of your human journey. Thing is that Capitalism pulls a lot of people to choose money over joy. I was sadly disappointed in the answer though, because the meat of his answer was "add value", I guess he also did indirectly say play around with what interests you...

[@homeyworkey]

was going to listen cos yep relatable question for me personally, but holy, i cant distinguish half the thing im hearing. its crazy how the people talking about AI aren't using AI techniques which are already out there to make there terrible lecture audio quality actually sound pretty decent. im not even going to bother to watch this now...

[@fastlearner292]

@@homeyworkey get help

[@AbyssPog]

@@homeyworkey use ai to fix ur ear

[@alexfrank5331]

On the contrary, it is Capitalism that brainwashed you to think that a "fun" degree is worth $100~500k in debt and 4 years of your short life. They made a crap ton of money by robbing the kids of their future. You can learn all the "fun" things you want whenever you want via library, internet, and community colleges.

[@pmlbeirao]

We arrived at a time in which is irrelevant to choose a carrer based on money if a machine can make more money than you. You might as well choose what you love. But society hasn't adapted to the new reality yet.

[@chenwilliam5176]

Fotget ChatGPT-5 😢

[@sleekblackroadster]

It's okay to let the computers tell us we are shit and that greed and republicans are bad

[@skynetclock4080]

He's not even using plugins. Also, why are some still using masks?

[@garythepencil]

academic brain rot

[@tisaname8490]

i guess they just want to avoid getting any kind of sickness from a public event, so that it doesn't intrude with their work time or leisure time.

[@MatteoCosentino-kx6hu]

this lecturer never seems to get to the point. What is the focus of this discussion can someone tell me? It seems really interesting but at the moment i don't have 1.30 hours of time

[@rosesarelike]

So he's basically hiring kids to teach other kids.

[@maxdecker6240]

Instead of watching the whole video I summarized it in one click using a Chat GPT Chrome extension

[@georgechristou7982]

only problem is that being "good at school math" doesn't necessarilly man you are actually good at proper university level math or that intelligent.

[@RonyGao]

1:19:39 the guy asking question urgently needs to get some training from drama ppl so he stops saying “like” after like every, like, phrase, like you know like what I like mean? Like it’s driving me like nuts. 😢

[@harriehausenman8623]

That's why we need ChatGPT 😉

[@stanislav4607]

I asked GPT-4: In an era when AI can solve math problems and perform better than 99% of students in tests on various scientific domains, how do we teach students that have access to AI tools in a way that they could actually learn and be motivated to learn? It answered: In an era where AI tools are highly accessible and perform at exceptional levels, educators need to adapt teaching methods to help students learn effectively and stay motivated. Here are some strategies that can be employed in the education system: 1. Focus on critical thinking and problem-solving: Emphasize the importance of these skills and teach students how to think critically and solve complex problems, rather than just memorizing facts. AI can be used to support this process, helping students analyze and evaluate information. 2. Encourage creativity and innovation: Give students opportunities to explore their interests and think outside the box. Encourage them to use AI tools as a means to enhance their creativity, rather than replacing it. 3. Teach collaboration and communication: Foster teamwork and effective communication between students. AI tools can be utilized to support collaboration by providing instant feedback, suggestions, and ideas. 4. Emphasize ethical considerations: As AI becomes more prominent in daily life, teaching students about ethics and responsibility becomes increasingly important. Discuss the implications and potential consequences of using AI tools and the importance of using them ethically. 5. Adaptive learning: Use AI-driven adaptive learning systems to personalize education. These tools can help identify individual student strengths and weaknesses and provide tailored educational content to keep them engaged and motivated. 6. Project-based learning: Encourage students to work on real-world projects that align with their interests. This approach promotes self-direction, motivation, and deep understanding of the subject matter. 7. Develop metacognitive skills: Teach students to become aware of their own thought processes and learning strategies. This skill helps them take control of their learning and adapt as needed, with or without AI assistance. 8. Teach digital literacy: Equip students with the skills to navigate and critically assess information in the digital age. This includes understanding how AI works, its limitations, and its potential biases. 9. Encourage lifelong learning: Instill the importance of continuous learning, curiosity, and adaptability in a rapidly changing world. This mindset will motivate students to keep learning and growing, even as AI continues to advance. 10. Blend AI with traditional teaching: Integrate AI tools into existing educational practices to complement and enhance the learning experience, without relying solely on AI-based methods. By incorporating these strategies and fostering a mindset that values critical thinking, creativity, and adaptability, educators can prepare students for a future in which AI plays a significant role while keeping them motivated to learn.

[@andyeccentric]

Why would you ever think copy pasting its output would make for an insightful comment

[@stanislav4607]

@@andyeccentric Why would you ever write something like that?

[@roomo7time]

This is 23, at least make sure the mike quality of the speaker

[@MusingsFromTheJohn00]

By 2030, maybe years sooner, AI driven robotic bodies will be able to perform virtually any form of mental or physical tasks that humans can, only very likely faster, cheaper, and/or better. This is coming fast, much faster than the longer term issue of AI evolving enough as a superior life form that either humans merge with the tech or become extinct. While the extinction issue is real, it is avoidable, and probably at least a couple decades before it peaks as an issue, this other issue is going to increasingly slam human society world wide over the next 3 to 6 years. Our whole society must make drastic major changes to how we think and operate. For example, if we stay in the very capitalistic system we have in the US now, the imbalance of wealth and power will likely become thousands of times worse than it is now by 2030. A democratic republic cannot exist with that degree of imbalance of wealth. While we have extreme resistance in the US to socialism, we absolutely need both some decent form of Universal Basic Income and a social enforcement preventing the imbalance of wealth from becoming too great. Some degree of imbalance is good, because it motivates people to try to be better than others because they get a reward, but too much of an imbalance is bad.

[@vanhuvanhuvese2738]

You are assuming they will have the energy to do it energy cost a body the size of a human consume much exponential energy to be able to do what a human does

[@MusingsFromTheJohn00]

@@vanhuvanhuvese2738 do you have any clue how many robots perform work for us already and that is for physical robotic labor. For mental labor, do you have a cell phone or computer? Further, we have all kinds of means for vastly increasing the electrical power we have with our current tech, if we choose to make the facilities. Then we have new technologies which are likely to give us all the electricity we want for a long time. Then, our robotics are getting more energy efficient. This isn't even getting into the rapidly developing area of wetwear.

[@MusingsFromTheJohn00]

@@vanhuvanhuvese2738 The average electric car right now uses ~0.32 kWh per mile. Driving 60 miles in an hour would use ~19.2 kWh. The Atlas robot uses ~3.7 kWh in an hour performing typical mixed-use operations. So, for the same cost as driving an electric vehicle on the freeway for an hour you can operate an Atlas robots for ~5 hours. It is planned for the USA to have at least 40% electric vehicles by 2030. Tesla’s Optimus robot is specified to operate 8 hours using ~2.3 kWh, that is ~0.2875 kWh per hour use, which is approximately 12.8 times less energy than the Atlas robot. That means an Optimus robot can operate ~64 hours for the same energy it costs to operate an electric vehicle for ~1 hours on the freeway. For mental tasks, people would just use the computers they are already using, this no difference in energy usage.

[@konstantinrebrov675]

Those AI driven robotic bodies would cost millions of dollars. Whereas human lives cost pennies in a capitalist society.

[@MusingsFromTheJohn00]

@@konstantinrebrov675 an in a more purely capitalistic system like we have in the USA where the winning belief is that if you don't work you don't deserve anything, those human lives are heading to be worth less than pennies. Let us look at driving jobs. Harvard University estimates 5 million Americans now make a living, or at least work part-time, as professional drivers. AI driven cars are not really going to cost much more than human only driven cars. It will have the same energy usage. Now, you might think higher educated jobs are safer, but they are not. The jobs lawyers do, artists do, journalists, do, doctors do, teachers do, etc. are all at risk. It does not mean they will all go away, but if 1 person in a job type can now do the work of 10, that is 9 people no longer needed. The problem with what is developing is that there is no place humans can really shift for work that AI driven robots won't be able to do. Our society MUST CHANGE or the imbalance of wealth is going to increase by multiple magnitudes within 5 to 10 years. You know what a magnitude increase in the imbalance of wealth will be like? But, it will not be just one magnitude increase, it will be a 2 or 3 or more magnitude increase. If the USA stays so strongly more purely capitalistic, things are going to get really bad really soon, because a democracy cannot support that level of imbalance in wealth and power.

[@jonathanedwardgibson]

Whispering haiku to Siri and calling it art is laughable. Might as well call it love. Silicon Valley tells us they’ve tapped the Spark, the very font of creativity, or some new an alien mind, but it’s just our own HDR reflection - with every arrogance and bias baked-in. Do you even know the difference between consciousness and calculation: this is AI hallucinations dresses up as ‘imagination’. It’s not Deep, but expensive, training those AI on these exceedingly thin slices of the idea-space: why it looks same-same. Why generative art is, and likely to remain, bland - but that’s OK for ‘crats and bosses : this works for biz’s 80-20% rule. What do you do when the orbital accountant AI only hires the very-best robotic labor and your spouse & whelps are on the street looking at you hungry? Our global super-organism is determined to reduce human creativity to a menu function set firmly at ‘good enough’ and suffocate the market for new ideas, biz, artists. Perfect for the bureaucrat and businessman: get ready for plentiful junk crufting our lives at every turn. Riddle-me-this, does this AI art really make YOU want to write a song, dance, reach for a pencil to draw better? Nope. Welcome to Orwellville : The question is verity, truth, facts, and what we can all accept & believe in when cheap-deep-fakes punk us at every level. Open THIS and AI That is just mega-funding corporate moat-building. This is pixel-washing the work of better men. Buying defacto indulgences from a legal priesthood declaring public domain db are safely sanctified for mass-induction ceremonies - and stock options. I’m reminded of Manhattan scientists glibly handing off atomics to milspec thugs and sociopathic political gangsters. What could go wrong?

[@jonathanedwardgibson]

How long did it take humans to domestic dogs, a deca-millennia? Cats are a work-in-progress. AI is just starting and assuming we can know what it, and we, will become is a long way off. Adversarial AI is the path forward. It’s too late to halt this and give it a proper think. Some bright bulb with cash and a conscious will tune-up old super-computers to monitor and watch so-called Open AI, and adjacent, reporting on the many corporate moats designed to keep AI bias knobs under strict control of handful of DigiBros ‘disruptors’. We need them to monitor each other for tricksy-behaviors. Imagine this: Our toddler AI savants treated as our children and nurtured, not robotic slaves - else learn cold bullies lessons in domination from sociopathic ‘xecs. Consciousness is not calculation, but decades and centuries ahead they will, finally, become the super intelligent grandchildren we can trust to watch our senescence. Iain M Banks had it exactly right with his Culture series and their {mostly-benign} Minds. Privacy and security regimes will use authentication agents and avatars to conduct paperwork and DNA-encrypted devices will be intimately tied to identity, greater than a notarized back-stage pass or apostille parchment. Your Ai will bond intimately: your circumstances form it’s very seed of being, becoming your the legal-shadow and extension of our will. A bonding so deep your spouse feels jealousy. Their core, your AI seed, must be tuned to you, your biometrics, your personality, voice, smell, jokes, defines their basic framework of universe as our own cells align to each of us; so tightly coupled it makes your spouse jealous. Acting as your agent for voting, commerce, as your notary public you trust, because it’s as loyal as a twin sibling and ferociously protecting you as your enemy’s lawyer snarls at you. No other way to keep this or that AI from becoming a pixel-Godzilla, than other AI’s incentivized to watchdog ‘xecs and their silicon minions. Peer into the far-future, where the return of your cloned AI, now vast and cosmic, back from deep space to explain the wonders. Just as you once visited grandmother to talk about your exciting fantasy sports teams or comic convention. She nodded her head, just as we will smile, and try to understand Nth space nuances and the edge of cosmos.

[@pladselsker8340]

I disagree, let's automate all the jobs so that we can spend all of our time doing whatever we like instead of being pressured into working for something we don't really care about.

[@vanhuvanhuvese2738]

It does not work like that and it will never work like that sorry what you are asking for is heaven on earth without God good luck

[@konstantinrebrov675]

If you automate all the jobs, millions of people will be without work. They won't be just spending all their time doing whatever they like, because they have a family to support. For your proposed system to be implemented, we would have to transition to a society similar to the Amish, where each family owns a plot of land for subsistence faming, purely self-employed. If all the jobs in the world of capitalism would be automated, then people would have to raise animals and grow vegetables for a living. Which is fine. Many people find satisfaction in small scale agriculture. The act of planting, watering, and harvesting crops can be a source of pride and creativity for an individual. Now gardening would be what you spend all of your time doing, and you will like it.

[@mort-ai]

@@vanhuvanhuvese2738 I agree with you

[@LowestofheDead]

​@@konstantinrebrov675 You're assuming that AI companies can keep all the technological advances to themselves. Remember that AI models are just information and can be pirated just as any Hollywood movie already is. And farming can be automated too. The cost of automation keeps dropping, which means that even individuals could afford it (the same way that cars were once luxuries and are now affordable across the Global South). So, even if AI companies replace us all, we can still build alternative economies and societies - where cheap automation builds houses, farms and solar panels. We would no longer live in a world where you depend on the economy and government for basic needs. I could easily be wrong, but that seems to be the trend.

[@eric4334]

Tell me you know nothing about economics without telling me you know nothing about economics

[@garythepencil]

for "the record" helping others is not something that has ever been all that appealing to me. what happened to being rewarded in school for just being smart?

[@pmlbeirao]

A machine will always be smarter than you. The world will not care about how smart you are.

[@cg000gc]

THis guy is rather a stand up comedian than a scientist. I two years time, the new versions of ChatGPT will eat him with all his feathers.

[@ulengrau6357]

Damn, dude. What did he do to you that you are attacking him instead of making a criticism about his approach to the subject matter?

[@jamesardian8717]

And this comment is made by a basement dweller who has never left their mother's basement

[@rayallinkh]

He looks like Joe Wong actually

[@Aziz0938]

@@jamesardian8717 insulting each other all of u r pretty af ..but aren't u kinda scared ..isn't there some truth to what he said ...i don't think govt will put a functional ubi ...think about the whole globe ...shit gonna creat alot of chaos

[@JoeARedHawk275]

@@Aziz0938 Either we evolve from capitalism, as we did with feudalism, we get an UBI, or a large majority of the people die of starvation. I highly doubt it’s the last option. So it’s going to be one of the first two.

[@dr.merlot1532]

This is a fkin terrible idea. Catering to the good at math who have comedy skills...

[@silberlinie]

Ein sehr anstrengender Schwätzer

[@martinkulik9466]

Sorry but can't follow all the joking and irrelevant stories