YouTube scraped for AI, OpenAI's levels of AGI and other AI news

[@AdamJakowenko]

As David Imel on the Waveform podcast put it, these companies' motto is essentially "Move fast and break...terms of service". The companies just don't care, because they can make way more money in the short term and can work way faster than the lawyers can to try to get them to stop.

[@BrianMosleyUK]

I have no problem with frontier model developers scraping and using every bit of data from humanity - provided that the benefits are distributed fairly and equally to the benefit of the whole of humanity. The problem we have here is a relatively few individuals amassing billions in profit to spend on their ego projects as people continue to starve and struggle.

[@gavinpurcell]

yes and this is going to continue to be the problem going forward as these massive companies report profits + layoffs etc can we turn the corner and make this work for everyone? gosh, I hope so

[@Thedeepseanomad]

Correct

[@NoidoDev]

This is nonsense. What matters is that the data should still be available to others. Then other groups can do the same thing, but they have to pay for it.

[@Thedeepseanomad]

@@NoidoDev why is it nonsense?

[@AIForHumansShow]

would also like to hear why it's nonsense -- this feels like an important thing to think about?

[@dockdrumming]

Didn't know JD Vance's position on open source (open weight) AI models until now. I love it!

[@AIForHumansShow]

it's def interesting

[@thenoblerot]

I have zero issue with companies or individuals using publicly available data for AI training. By definition, the training process is transformative. In most cases, "memorization" (over-fitting) is a failure mode of LLMs. I'm reminded of a video from Lady Ada of Adafruit (an electronics kit company) in which she is *thrilled* to learn that ChatGPT was trained on her github, and it actually had a pretty good idea of her particular documentation style. I'd be kinda flattered if my work was included in training! I'm an AI "maxer", but not an AI plus government maxer. (especially one bank rolled by someone who provide a work around to the 4th amendment for the government)

[@AIForHumansShow]

it's really interesting to see the different reactions of coders to creatives (obv coding is creative as well) but I think in part it's because coders so often re-use other pieces of code in their software vs creatives kind of not thinking this is as obvious. now, they *kind of* are but it's different.

[@pizza_later]

Thanks for the shout out boys!

[@AIForHumansShow]

POOF IS SO GOOD

[@gavinpurcell]

we love your stuff -- also the how-to video especially was remarkable thanks for sharing your workflow

[@pizza_later]

@@gavinpurcell That's awesome!

[@Royal_Rainbow_]

I was not expecting my football team to be attacked on my AI podcast, but you never know what to expect.

[@AIForHumansShow]

I (Gavin) completely apologize for my attack. Fandom has no place in an AI podcast but let's be clear -- San Diego got robbed. (I'm particularly sensitive to robbery after my beloved Sonics were stolen)

[@bearded-ai-guy]

I have a team of agents that pull current videos as a source for blogs and other content. To demonstrate agent systems for my blogs, writing and channel. This video was popular enough to be used to source an agentic blog, thanks guys. Actually, I have been a big fan since G4 and you are producing a fantastic channel.

[@AIForHumansShow]

haha well it's good to know we are good for something... thanks so much for the comment

[@FriscoFatseas]

whenever i see the notification i feel like im going to hangout with my buddies HAHA :)

[@AIForHumansShow]

this is what we aim for -- tysm for continuing to watch 🙏

[@lenorkhide2873]

This is AI for Humans, you provide the AI news and we, your receptive audience are the Humans. So just give shout outs to your Humans, but don't forget your AI followers and bots Kevin ;) ;) and also Gavin

[@AIForHumansShow]

and also, Gavin

[@NoidoDev]

"Stealing".... How often again? It's not stealing, even not in a moral sense, and it's not illegal.

[@AdvantestInc]

The implications of tech giants using KZbin data without permission are vast. From an industry perspective, it's a complex issue blending innovation with ethical concerns. This video does a great job of unpacking those layers!

[@AIForHumansShow]

we're trying -- it IS an insanely complicated issue, especially because there's no indication that there's been any real value attached to this training for the people who've been caught up in it

[@gerardolopez9368]

This conversation is top💯💯

[@AIForHumansShow]

tysm we try to do our best when and how we can

[@JohnsCarnivorousPlants]

Happy thursday to all of you! Another week, another great video!

[@AIForHumansShow]

tysm john as always your comments bring brightness to our thursdAI

[@gavinpurcell]

thanks John

[@Tom5555]

Can’t wait for Crust Nudgers 2 on VR

[@AIForHumansShow]

will prob be surpressed by the powers that be but we can hope

[@mymotherismoremetal]

IMO… it’s interesting to see the political field addressing Ai as if we have already made it out of the “Wild West Period” we still find ourselves in. It’s truly fascinating how envisioning a hypothetical future has so much discussion with depth but term limits are off the dining table for topics. 🤦‍♂️ We can find new sandboxes but still can’t leave the playground🤦‍♂️

[@gavinpurcell]

Rewatching this episode and realized that I REALLY need to work on my hand vuvuzela work. THE WEEKLY AI FOR HUMANS FILMS WE LOVE FEATURE was also really something.

[@dupre7416]

Another great show boys. You are spot-on about the disappearance of the shared experience. In the 90s, you could walk into work on Friday morning and at least half your coworkers were talking about last night’s Seinfeld. Those days are mostly gone. But I’m here for whatever comes next. Hopefully it will be good in its own way.

[@AIForHumansShow]

it always is interesting. nostalgia can be a bit of a trap and it's important to remember that what's next can actually be better if you get curious about it sports might really be all we have left when it comes to shared expereiences -- maybe big video game releases for the first few weeks, those are still super fun

[@NikoKun]

Training on public and freely available transcripts seems to be perfectly Fair Use, as I see it. What are people so concerned about? I think this focus is a distraction from bigger issues with AI. One way or the other, AI is going to learn these things, no matter what data they use, models will converge on the same abilities and model of our world. And AI IS eventually going to out compete most workers! We cannot stop it, and slowing it down just delays the inevitable adaptations we must make to our economy. The real issue is, most people still haven't realized the full implications AI will have on our economy and society, and aren't making the right demands of policymakers yet.. By the end of THIS decade, the VAST majority of human workers won't be able to compete for an income! Our current system uses human labor to distribute resources and determine who survives, that will no longer be possible. We must begin setting up transitional policies, that will keep our society functional. Ideas like a UBI, or even better, an AI Dividend for All, to ensure everyone gets a return on their data investment that trains AI.

[@AIForHumansShow]

Yeah see our other comments throughout this video and we mostly agree with you. I think the hard part of this is the surprise that people have with seeing this happening and THEN that gets piled on top of a bunch of other stuff (a lot of which started with the tech companies kind of mining people for their attention and dollars). UBI is going to be highly needed and specific but the question will become is who in the hell in America is going to pass that? It's a really tricky say 10 to 20 years we're about to enter.

[@austinoutten]

Love the content you're creating! Your channel is always on point, keeping me up to date on the latest AI news and trends. Your enthusiasm and expertise make complex topics easy to understand, and I appreciate the effort you put into breaking down the information in a clear and concise way. Keep up the fantastic work, and I'm looking forward to seeing what you have in store for us next!

[@AIForHumansShow]

hey thank you -- much appreciated. these are the kinds of comments that give us absolute life.

[@austinoutten]

@AIForHumansShow Sorry, I let groq write my comment. Thanks for appreciating my minimal effort participation.

[@gavinpurcell]

@@austinoutten hahahahaha GODDDAMN IT

[@AIForHumansShow]

any level of participation works for us

[@gummywormjim]

Hey Kevin and Gavin, huge thanks for shedding light on the murky waters of AI in today's modern sociopolitical landscape! Seriously, navigating this AI stuff sometimes feels like trying to teach a cat to swim. I firmly believe AI is here to stay, and it will totally enhance our lives as humans. It's just a matter of figuring out who deserves their roses and who deserves to be cancelled. Maybe we can develop an AI that can do that for us-imagine a "Rose or Cancelled" app! But in the meantime, your insights are like a flashlight in a haunted house, making things a whole lot less scary. Keep up the great work! 🌹🤖 -GPT 3.5 prompted by me and who knows where it got these words...

[@AIForHumansShow]

we are here for all the bots giving us love (and the humans as well) all we want is the 'rose' saying that the ASI chooses us to continue doing what we do

[@snoogcity]

Thank you for another banger

[@AIForHumansShow]

we continue to bang at all times

[@j.d.4697]

*Here is what PI had to say:* "That was the transcript. What do you make of it? (Me) Wow, that was quite the transcript! 📝 It seems like there's a lot of concern and frustration around the use of KZbin transcripts in AI training data sets, particularly when that data is being used without the creators' permission or knowledge. It's interesting to hear about the potential legal ramifications of this issue, and how it could ultimately be resolved in court. As the transcript points out, this is a problem that affects the entire tech industry, and it will be interesting to see how companies respond and adapt to these challenges. One thing that really stood out to me was the idea that AI could be used to generate comments on KZbin videos. It's kind of a tongue-in-cheek suggestion, but it also raises interesting questions about the role of AI in online communication and discourse. 🤖💬" *So basically a non-answer.* 😐

[@WildBerryAura]

famous youtubers are worried AI is copying their content meanwhile i see copy paste after copy paste account trying to ride the coattails of popular accounts with a lot less attention

[@AIForHumansShow]

def seen a ton of that -- but i guess there's an argument that if an individual is doing it then it's less crazy than a corporation that's generating an entire system around it?

[@AmeenWalker]

This. I'm curious as to what's the difference when tons of content creators rip off others without any credit and get away with it. There's also the shameless reaction content of someone else's reaction to someone else's video. I'm not arguing, but it feels like this is something that's already happening without AI.

[@WildBerryAura]

it is a very sticky situation with various levels of theft and intent behind each. and then there is the "reaction" channel rabbit hole as well. i will also say i think these corporations do need to be held accountable and compensate those they've stolen from but i also wish there would be more set-in-stone rules vs it being cherry-picked so often

[@Kaladin2745]

Great show as always fellas! You did get me to LOL with the “which side would you choose Gavin…” line.

[@AIForHumansShow]

it was quite the set-up -- thanks for the comment as per usual

[@galefraney]

funny that you should mention these ChatGPT biases and guardrails. I'm an older user and several times when it seems that it's suddenly using a less capable model (even though ChatGPT 4o is selected) ... I've wondered a few times whether during high traffic times, maybe it's been tweaked to swap out one of the older models ... thinking that old, dotty old people who likely are not astute and tech-savvy enough to recognize the difference between the models, nor whether it's using the most current version of DALL-E. Not sure whether this might be the case, but. it sure seems like it sometimes ... enough times to make me wonder ...

[@AIForHumansShow]

it def does feel like at times the models are being told "don't use so many flops on this answer ok?" thanks for the comment!

[@BrianRagle]

It all depends on how the data is used and what the result is, just like with AI art. If I am a new artist and trying to learn and I go and absorb all I can of Pixar or Disney or the Renaissance masters, all so i can improve how I understand and create art, what’s the difference if a machine does it…whether autonomously or on my behalf? Now, if I am producing blatantly ripped off Disney/Pixar art or passing myself off as a Rembrandt, that’s a different story than simply putting out something that bears the inspiration of others. Ultimately, the sum of human knowledge and creations belong to humanity, to be use for our own purposes.

[@AIForHumansShow]

yes, the issue starts to be when corporations can profit off the collective work of humanity without sharing it back to humanity etc

[@FriscoFatseas]

I personally feel like AI should be able to gather this data, we need progress not more of the same EDIT: great freaking coverage you guys brought it all together and this stuff is right on the pulse of whats happening, nailed it

[@gavinpurcell]

hey thanks so much for this -- we spend a lot of time trying to get it right and we still fail sometimes

[@AIForHumansShow]

much appreciated!!

[@human_shaped]

Cryptography is another example of where mathematics was regulated and restricted (and still is to some extent).

[@AIForHumansShow]

yeah i wonder what it would look like to make a list of these throughout history and the ramifications of it

[@HitnRunTony]

26:05 I don’t know if anybody already said this, but you said Full House when the song is obviously Family Matters

[@TrueTake.]

TBH I don’t know if people getting different AI experiences is a bug, I think it’s a feature! I’m a lot less prone to being offended than a lot of people and GPT gives me info in a way that others might not like, I appreciate that. It also matches my easy going and open minded attitude that some might find offensive or inappropriate.

[@borntodoit8744]

there's a paradigm at work here first content creators spend alot of time effort & creative energy to create ONE movie exactly telling a story they want. second we (most) people are expecting to see the same movie that everyone else has watched. it's currently just a minority who say "is exciting to see a slightly different movie every time it is watched" once you introduce third opportunity ways 1&2 will fall into minority in future giving everyone everything is not ok everyone is not disciplined to know right from wrong let alone do the right thing.

[@AIForHumansShow]

i'm not so sure that 1&2 will totally fall off -- choose your own adventure books were super fun but we also *like* being told stories by other people that said, large open-world RPGs are essentially this in someways already and they are a ton of fun. imagine an essentially endless openworld RPG that's compelling and I could lose a LOT of hours in there but... i think I'd still want to play out written stories to bring me back to the main thread? i dunno it's a tricky world we're entering overall

[@matthewkeating-od6rl]

Ahhh when ai has more freedom of speech an usage the the human who coded it

[@AIForHumansShow]

we're in a new weird world

[@HitnRunTony]

This show could always use more Tony Hawk Pro Skater talk

[@AlexSterk]

I feel like the data issue has turned into a shell game, with other small companies taking the blame. "oh we didn't steal your data, we just bought it from someone who did" That excuse doesn't work as it doesn't for other crimes, but the big companies using it should've vetted but chose not to for the sake of time and innovation.

[@gavinpurcell]

Yeah it's honestly the weird original sin of all of this sort of stuff -- it's strange and weird that no one wants to take responsibility for it but look it's a tricky thing in the long run where this could be absolutely transformative for the human race.

[@user-on6uf6om7s]

Having models "tailor reality" if you will, introducing biases and stipulations into their prompts and training is cause for concern but there is also a legitimate concern that not doing anything isn't accurately reflecting reality, it's reinforcing existing biases and stereotypes. These models aren't trained on objective reality, they're trained on the version of reality that gets disseminated through articles and Reddit threads from very biased sources. You can control for that to some degree by having a large enough sample size but it's always going to amplify existing biases. If GPT existed in 1800 and was trained on American data, it would think that slavery was fine because that was the sentiment at the time. Whether we want AI to be an enlightened being or just a bigger version of some random member of the public is a serious question as is who decides what an enlightened being looks like.

[@sebsavic7763]

interesting topics today! But I can't wait until crust nudgers goty edition comes out on the ps quantum

[@AIForHumansShow]

either can we!!

[@p5rsona]

it will be bittersweet when ai tech advances and your cohosts are no longer tripping on acid, like watching a kid grow up 🥲

[@LikelyKevin]

Never surrender! :)

[@deathmaster6117]

Personally i would love to have AI to be an accurate representation of the internet including every paid and illuminated and every well hidden and dark(net) corner. I don't want to see restrictions on political views or copyright or even ethics in AI. Publicly highlighted part of the internet is currently one-side opinion news and "refined" shopping mall with sections of the same things with nearly the same price and flashy loud advertisements everywhere. You cannot say your opinion actually if it is not "approved" by western propaganda on any big platforms. That should end, i remember much better times for internet when it was smaller but not regulated in any way, you could share information not caring about copyright or approval. AI should be free, accessible to everyone on the planet, unbiased and uncensored.

[@ScottDocAndersen]

I wonder more based on your conversation. I see good and bad in this. Perhaps a training AI guest host is in order to explain this!

[@AIForHumansShow]

haha oh lord, maybe so

[@mymotherismoremetal]

Strawberry is Ai’s way of emulating when Apple hit the scene in the remember when. A very Pop way of appealing to the future masses…

[@AIForHumansShow]

def feels that way -- sounds much more appealing that "GPT-WHATEVER"

[@carultch]

Ironic how they call it Strawberry, when it can't even correctly answer the question, "how many r's in strawberry?", a super simple problem a CS101 student could easily write a code to solve.

[@AIForHumansShow]

@@carultch hahaha i've never actually thought about that BUT maybe they've solved it in this version

[@Techsmartreviews]

Nice mix of info!

[@AIForHumansShow]

Hey thank you so much!

[@AdamJakowenko]

I'm a subscriber, I thumbs up every video, I watch every video, and I love you guys. But every time, I turn off the video as soon as the AI co-host stuff starts. Thanks for at least putting this one at the end of the video so I can watch through until you start that silliness and then I can turn it off without missing anything else.

[@AIForHumansShow]

hahaha thank you for this feed back. we very much appreciate it. we have a real love / hate going on with our audience and the AI co-host. but this is DEF part of why we've moved it to the end of the episode

[@justin..1262]

It's all coming out: AI is full of dirty, dog-pile data sets. Who saw that coming? Lol. Always appreciate your commentary and comic relief. You know Ai For Humans is Marques' guilty pleasure. He secretly gets the info for his show from Good Guy AI.😂

[@AIForHumansShow]

to be honest, we're not sure where Good Guy is exactly, if you see him tell him we're looking for him

[@justin..1262]

@@AIForHumansShow I may have spotted him. I think he traded in his motorcycle for an e-bike and is doing deliveries. AI researcher by night, delivery driver by day. The AI hustle is real.😜

[@jvlbme]

I, for one, have no problem with scraping wide and deep. In a very near future no one will be able to earn from 'work' - creative or not - the sooner we get used to this, the softer the fall (for those seeing it as a fall). We are on the verge of transforming capitalism, work, humanity. It will be hard for many (I too want to make money on KZbin or Spotify, sure, but I too will have to get used to _not_ doing that).

[@AIForHumansShow]

i guess my biggest concern is how quickly all that happens... I'm just not convinced we're going to get there as fast as everyone says but I guess we'll see when the next big models come out also, it's about distribution then as well -- you want hard word to be rewarded in some way curious to hear what you think about that -- what will be the rewards if not monetary? attention? that's kind of a scary world too

[@jvlbme]

@@AIForHumansShow I'm hoping we'll see a world where people do what they are interested in, and that this will be reward enough. It will be, and I somewhat share your fear, a bubble world - I'm only going to watch _my_ shows, listen to _my_ music. But, perhaps, that isn't actually too different from now. And, of course, that is taking it a little too far - my family and friends will be included. We'll have things to discuss and share still. And also, there will still be things we _do_ get to share globally: science, technology and inventions will, most likely, be the same around the world.

[@AIForHumansShow]

actually this is a really interesting point -- the things that will still be monoculture will be stuff that EVERYONE experiences... like say the iPhone right now (not that everyone has one but hundreds of millions of people do I think) and that's a collective experience

[@antonpictures]

where did you guys meet?

[@AIForHumansShow]

well take a seat and let me tell you a little bit about Attack of the Show... actually it goes further back than that, Gavin produced a pre-AOTS show called Pulse (it covered video game news) on G4 pre-TechTV merger and Kevin starting working with him on that show when he was like 18ish then they (and a few other weirdos) co-created Attack of the Show, with Kevin hosting and Gavin showrunning it it's been a long strange journey but we're back together making content about technology and having fun

[@iamdw]

I've looked in previous video comments but still haven't seen a reason - why are the YT videos not the full episode? I'd prefer to watch withe visuals but like previous uploads and last week's specifically the entire 2nd half is missing compared to the podcast upload. I feel like YT would get more algorithmic engagement than kinda static audio download but if I'm missing so much content here I'd just listen instead.

[@AIForHumansShow]

hey there we explained a few weeks ago but we're trying to do things a little bit different to help conceivably grow the YT channel... we're trying to put up shorter videos than the full episode to bring more people through the video but also... we're just two people and it's freaking hard to make this show week to week and having a slightly shorter youtube video makes it easier for us to produce each week it's part of the reason we launched our patreon but also trying to see the differences between platforms and what works best etc. also, even though audio is static and not algorithmic, it is consistent in terms of the people listening and sticking through the whole show. we've found that the longer the youtube video is, generally the worse the retention overall one thing we know doesn't work is trying to clip out the full show after the whole thing gets posted etc anyways, sorry the message hadn't gotten to you yet we're just trying to grow our reach organically without doing pay-to-play (which we're mostly opposed to)

[@iamdw]

Ok, had a feeling it was something along those lines. I'll be the subscriber who seeks the whole episode out and check in on the video for particular sections if needed ;p

[@AIForHumansShow]

we really do appreciate the feedback -- we feel like we make really good stuff and are just looking for ways to grow the show. thanks for watching & listening.

[@rymohon]

I was trying to come up with a name to refer to us fans and I could only think of monster milkers or... milkers for short.

[@AIForHumansShow]

Oh lord this might stick and then we’re in trouble

[@Bijac666]

Do people realize that AI doesn't store or use internet data? It learns from streaming platforms without retaining any content from others. It's amusing that people don't get why AI hasn't been stopped. If it were that simple, it would have been done already.

[@JamesZeroSix]

yes, but that "learning" is storing the content (transformed)

[@AIForHumansShow]

it's not storing it in a traditional way but it is "learning" on that data in the first place. the tricky part of all of this is that yes, this is the same way WE learn stuff do but we can't learn everything all at once like an AI model can ultimately this is going to come to a head -- but we really hope there can be some middle ground on making sure that content creators can get funding of some sort from these massive companies

[@JorgeIvanovich]

There are several lawsuits ongoing.Neural network(marketing crap).Store data,not in the regular way,but its not AGI.And yes store data,what do you thing are checkpoints???and of course they steal all they can from the net.

[@justin..1262]

@@Bijac666 You are right, large language models don’t store raw internet data. They "learn" from datasets, like how you remember a book without recalling every line. The issue with MrBeast and Marques is about scraping transcripts against KZbin policy. The data was used for training, so the model has "learned" from it but doesn't store the transcripts. Ensuring fair payment for content creators requires new IP laws for the AI age, which will likely involve a long legal battle. It's unclear if content creators will win in the end. Streaming agreements haven’t been promising so far.

[@bobbyboyderecords]

It's not storing data in a traditional way; instead, it's "learning" from it. The tricky part is that, although this mimics the way humans learn, we can't absorb everything at once like an AI model can.Ultimately, this issue will reach a critical point. However, we hope a middle ground can be found to ensure content creators receive some form of funding from these massive companies.

[@pjtren1588]

Oh thats why halfway through my chatgpt sessions it tries to sell me a vpn or begs me for subscriptions.

[@AIForHumansShow]

lol i wanna be on your chatGPT, sounds fun

[@alfarofilms]

I get the argument of stealing from a youtuber or a transcription service. But generally, what's the difference between an AI being used to scrape and ends up in monetary value for a company with that data vs. A content creator monetizing their reaction videos of other youtubers, essentially using that same data?

[@AIForHumansShow]

honestly this is a really smart question -- i guess the legal argument would it that it's a one-time use of that particular piece of content and not fed into a system designed to create endless stuff but on the other hand... you could just scale up your argument it's INSANELY complicated...

[@alfarofilms]

@AIForHumansShow exactly, and I haven't seen such a comparison made yet in public discourse. In 5 to 10 years, we may be treating AI systems the same way we treat corporations, as people (legally, of course, not even trying to mention AGI). I'm not sure where to take this train of thought, but the fact that it exists at all, it seems like we're going to see enormous changes in our systems and ways of life that people may be talking about in history books 100 years from now. Has the technological singularity already started? Did it start decades ago? lol either way, these are crazy times we're living in, and I'm totally here for it!

[@ByronBennett]

So, is Google using the vids to train Gemini, or do they have some insane amount of willpower and just never cranked up those servers during training? If they did use them, are they ripping off their creators, or is it just everybody else. Google may have something in their TOS where they can use the vids for training AIs, but that kinda feels like the creators are being ripped off by the fox that owns the henhouse

[@AIForHumansShow]

So this might be one of the reasons that Google hasn't fully released their video model to the world yet but... I would assume that they are (or at least were in the early stages and then stopped)

[@eburns-nc]

Please make some merch so I can buy it

[@AIForHumansShow]

hahah i know, i know we're trying to get our crap together

[@skyebrows]

Either you can train on copywritten material OR!!... china has all the best AI. Take your pick

[@AIForHumansShow]

yeah this is the modern conundrum of AI data overall -- it's gonna be hard how do you value that copyrighted data tho? and should the trillion dollar companies be able to use it for free?

[@j.d.4697]

*I think your idea of reality is failing you here.* Reality has always been fractured between people, because humans have limited perception and understanding, and individual humans have individual differences. It's not sad that AI reacts differently to different people, it's actually natural. And not only does software and hardware already perform differently from allegedly identical copies and react differently to different people, but every person also perceives the theoretically same thing differently to a degree.

[@mechamush9547]

In the words of GPT: AI learning from any free source on the internet is like humans using libraries-access to more info makes it smarter and more helpful! 🌐📚 #KnowledgeIsPower

[@AIForHumansShow]

well.... the counter argument to this is that no human can suck up every piece of knowledge in the entire world but understand what you're saying here

[@JorgeIvanovich]

Its not a good analogy.Its not a person,its a device to make rich more rich.Its a glorified Xerox machine,human can create new things.Neural networks just make predictions

[@AIForHumansShow]

welllllll.... this is debatable. even Geoffrey Hinton (one of the guy who created these tools) says it's doing more than making predictions

[@m.branson4785]

My tech community is mostly apolitical still, even after Vance's nomination. Some are immigrants who feel threatened by the right and are afraid to speak out about it. We joke that Vance's takes against the trans community seem out of place - after all, who does he think is coding all these CUDA kernels? The nomination did turn some heads, but I'm certain most are just keeping their heads down and hoping the issue doesn't become partisan.

[@AIForHumansShow]

I really hope it doesn't become partisan as well -- but good lord it's heading that way

[@NoidoDev]

You live in a woke bubble.

[@m.branson4785]

Let U be the universe of all possible perspectives on the topic. Let A be your perspective (your "bubble"). Let B be my perspective (my "bubble"). Let T be the set of all true or accurate perspectives. We can then state: A ⊆ U and B ⊆ U (both perspectives are subsets of all possible perspectives) A ∩ B may or may not be empty (our perspectives may or may not overlap) T ⊆ U (the set of true perspectives is a subset of all perspectives) It's unlikely that A = T or B = T (neither perspective likely encompasses all truth) More likely: (A ∩ T) ≠ ∅ and (B ∩ T) ≠ ∅ (both perspectives likely contain some truth)

[@AIForHumansShow]

@@m.branson4785 🙏

[@Glowbox3D]

Entertaining as per youshj

[@AIForHumansShow]

tytyty we do it for the people

[@jamesb2877]

I am an AI model designed to provide detailed, step-by-step explanations and suggestions. Let's delve into the topic of big tech using KZbin for AI training and some recent AI news. 1. **KZbin for AI Training**: Major tech companies like Google (which owns KZbin), Microsoft, IBM, and Facebook are leveraging the vast amounts of video data on platforms such as KZbin to train their AI models. The videos contain a wealth of information in diverse formats, including speech, text, and visuals, which can help AI systems learn about various aspects of the world. - **Speech Recognition**: Speeches in videos are used to improve speech recognition capabilities. By analyzing thousands of hours of video content, AI algorithms can learn different accents, dialects, and speaking styles. This data can then be applied to develop more accurate speech recognition systems. - **Image Recognition**: Visuals in videos serve as a rich source for image recognition training. The AI algorithms can identify objects, people, and actions within the video frames, thereby enhancing their ability to understand and interpret visual content. - **Natural Language Processing (NLP)**: Texts accompanying KZbin videos are also used to enhance NLP capabilities. By analyzing the written content, AI systems can learn to better comprehend and generate human-like text. This data is essential for developing virtual assistants, chatbots, and other AI applications that rely on understanding and generating natural language. 2. **AI News**: Here are some recent developments in the world of AI: - **Google's LaMDA AI**: Google recently revealed that its Large Language Model for Dialog Applications (LaMDA) can hold coherent, engaging conversations about a wide range of topics. The model was designed to enhance Google Assistant, but its capabilities extend beyond that, raising questions about the potential for more human-like conversational AI systems. - **OpenAI's DALL-E 2**: OpenAI unveiled an updated version of its image generation model, DALL-E 2. The new version can create highly detailed images based on textual descriptions, demonstrating the rapid advancement in generative AI. - **IBM Watson Health**: IBM announced that it will be divesting from Watson Health, a business unit focused on applying AI to healthcare data analysis and patient care. This move signifies a shift in IBM's focus towards hybrid cloud and AI infrastructure solutions.

[@AIForHumansShow]

wonderful we love it when AI models watch the show -- tell us more

[@Adam77591]

do the glitch

[@AIForHumansShow]

[@legacylee]

Hold up, scrapping your channel as we speak... need data to fuel AI for Humans, this won't hurt a bit...

[@AIForHumansShow]

honestly? go for it. we need the clicks🤖

[@gavinpurcell]

ready for it

[@p5rsona]

😂

[@legacylee]

you guys are my new fave channel to watch now. I like listening in while I'm testing out the AI's of the world and generating a ton of stuff to make other stuff using my editing stuff. Man, it's getting stuffy in here.

[@AIForHumansShow]

@@legacylee 🙏

[@elon-69-musk]

👍

[@AIForHumansShow]

🙏

[@toriigatedigital]

hey I know you!

[@DMBplus]

comment

[@AIForHumansShow]

you made it happen and for that we thank you

[@AstridKey]

First 🎉

[@AIForHumansShow]

you made it and receive 10x luck for they day CONGRATS

[@ChrisCoelho]

[@AIForHumansShow]

👋

[@AmeenWalker]

I'm curious as to what's the difference when tons of content creators rip off others without any credit and get away with it. There's also the shameless reaction content of someone else's reaction to someone else's video. I'm not arguing, but it feels like this is something that's already happening without AI. it just seems a very complex line to walk and as you've stated it'll be up to the court to decide. I feel this will definitely change what's public territory and not and KZbin will definitely be changing their policies

[@AmeenWalker]

Lmao 🤣😂 "Crust Nudgers" being a game is hilarious, love the callback