OpenAI's New Model Releases LEAKED | Sam Altman talks about AGI, UBI, GPT-5 and what Agents will be

Learn AI With Me:
www.skool.com/natural20/about
Join my community and classroom to learn AI and get ready for the new world.
#ai #openai #llm
NEW UPDATES:
/ 1789126322545213765
An interesting thread with an appearance by @apples_jimmy himself.
correction:
@nicdunz posted the leaked screenshot:
/ 1789065894628348381
which had largely the same info as reported by @apples_jimmy and The Information.
As far as I can tell it was just a report of what was already available.
I decided to include it in the video because I saw it right before I was wrapping up, but I think it's more appropriate to credit the original sources:
Credit:
Stephanie Palazzolo
@steph_palazzolo
Jimmy Apples
@apples_jimmy
LINKS:
The Information:
www.theinformation.com/articl...
Seem like they have the same information as the leak above
In conversation with Sam Altman
All In Podcast
• In conversation with S...
00:00 Rumors
01:57 Latest Leak
03:29 Sam Altman on GPT-5
04:12 Intelligence is an emergent property
04:40 Local Models
05:37 AI Agents
07:20 UBI
08:48 Generalized Reasoning

[@WesRoth]

NEW UPDATES: twitter.com/WesRothMoney/status/1789126322545213765 An interesting thread with an appearance by @apples_jimmy himself. correction: @nicdunz posted the 'leaked' screenshot, which had largely the same info as reported by @apples_jimmy and The Information. As far as I can tell it was just a report of what was already available. I decided to include it in the video because I saw it right before I was wrapping up, but I think it's more appropriate to credit the original sources: Credit: Stephanie Palazzolo @steph_palazzolo Jimmy Apples @apples_jimmy LINKS: The Information: www.theinformation.com/articles/openai-develops-ai-voice-assistant-as-it-chases-google-apple Seem like they have the same information as the leak above In conversation with Sam Altman All In Podcast kzbin.info/www/bejne/pISwYauabd17i68

[@SahilP2648]

There were rumors that they were gonna show music generation like video generation

[@amay167]

I personally dislike how KZbin has become a cesspool of clickbait titles and thumbnails. It's manipulative social engineering at it's worse. Such a waste of time and no different from your email being spammed with spam emails. Stop posting clickbait titles and thumbnails.

[@therainman7777]

If there aren’t actual model leaks, I’m unsubscribing.

[@aibutttickler]

same

[@BAAPUBhendi-dv4ho]

+1

[@FluffySezyt]

Honestly fair

[@tathagata_roy]

Me too

[@radicalaim]

fr the title could have been accurate and still captivated my attention, but I strongly dislike untrue clickbait

[@nouxcloete3129]

I don’t think Sam is being consistently candid

[@ZappyOh]

Psychopaths never are.

[@GraavyTraain]

Sam Altman has never read as “trustworthy” to me.

[@adangerzz]

You and the board apparently. I'm a bit suspicious myself. Chances are he's on the spectrum so neurotypicals might have a hard time reading him. (I do.)

[@onedrop7967]

Sam is so fake. Don't trust that open platform snake oil salesman. He shouldn't run a lemonade stand... let alone a AI company.

[@user-oq5ei4qd8r]

I don't think Wes Roth is being consistently candid enough.

[@nod5770]

please eliminate the clickbait in the titles. The transcript doesnt include the work "leaked"

[@ridewithrandy6063]

It kind of does

[@RuslanLagashkin]

And where are the leaks? All stated is a compilation of official information.

[@missoats8731]

No, the information about what the releases will be was leaked first.

[@ekurisona663]

'not income, but you own part of the productivity' 🤣🤣🤣

[@fontende]

exactly like in communism, i'm not kidding, working class kinda owns the factory...but you must work for motherland in even bigger poverty than before for the goal of worldwide revolution or personal house to everyone in year 2000 (which was just hidden oligarchy)

[@Katatonya]

that productivity is way more valuable, assuming you're not just a leecher, but a seeder

[@GraavyTraain]

@@Katatonyadude what are you even saying.

[@user-cw8rd7td9b]

⁠​⁠​⁠@@fontendemost incoherent sentence ive read ngl. communism is about social control over societal production and the guarantee of life’s necessities, and its funny you say that because under capitalism we already have more homes than homeless ppl yet still force them to the streets because profit is more important that humans.

[@Tofu3435]

​@@fontendeit is an interesting topic. I mean if too much job will be in danger, people should get something. It can be income or productivity, or... We will see.

[@filmanim5276]

Sam loves to give hints. One of thing he said on Twitter: "learning how to say something in 30 seconds that takes most people 5 minutes is a big unlock (and imo a surprisingly learnable skill. if you struggle with this, consider asking a friend who is good at it to listen to you say something and then rephrase it back to you as concisely as they can a few dozen times. i have seen this work really well!)"

[@MichaelDomer]

*_"Sam loves to give hints"_* Sam? What's next... saying "Joe is going to sign a new law".... as in Joe Biden. Weirdo....

[@Drone256]

Altman spun the diminishing returns problem into “easier for society to accept.” He could sell sand on a beach.

[@Mayeverycreaturefindhappiness]

GPT is improving in my opinion

[@ronhill8941]

When he said, "...easier for society to accept.", I literally said BULLSHIT out loud. Getting software updates on an iPhone 10 doesn't really make you feel like you're getting a newer, improved product even though technically you are. You still want a new phone. Some say GPT is improving but my eye keeps straying to Claude.

[@colto2312]

@@Mayeverycreaturefindhappiness opus when prompted correctly is still better, but yes

[@themprsndev]

@@Mayeverycreaturefindhappiness it's literally been lobotomized. The first ever GPT-3 Davinci completion model they released was MUCH better than current GPT-4. Early access first version of GPT-4 was god-tier, then when public release came around it got noticably worse, and since then it's borderline useless for anything other than casual conversation. And the censorship is INSANE. They don't censor the output of GPT-4, instead they train the model again and again from human feedback with censored text / responses which leads to the model's performance declining extremely (proven many times over, look it up, Google "GPT-4 RLHF vs base model performance paper", go to the arxiv result which is their (OpenAI's) own paper on GPT-4, and scroll down to Figure 8 in the paper) Check that graph and watch in awe as they do that exact process over and over and over again.

[@Mayeverycreaturefindhappiness]

@@colto2312 my point was I don’t think llms has run out of steam yet

[@REASONvsRANDOM]

ClosedAI

[@Sky-fk5tl]

You are so funny 😂

[@creativecraft_mc]

yeah

[@primalplasma]

Listen to what Sam said during the UBI segment of his interview about his concept of "Universal Basic Compute." He mentioned that "everyone gets a slice of GPT-7’s compute." Given that a newer version of GPT is released annually, and we're currently on GPT-4, it seems we might see AGI and UBI by 2027-2028. Just three years until the world changes forever. Wow.

[@UltraK420]

This is assuming the $100 billion AI data center is finished by 2028. We will need the massive upgrade if we want to see AGI.

[@user-td4pf6rr2t]

Being a mobile merchant I am having HUGE dilemma verifying credit card transaction using AI as a customer service system. How is the hurdle of capturing cvv to verify card holder meant to be handled since this is meant it be done live and for me as the merchant to make a record of this number is illegal. Even if its just for the brief moment while importing the call to a paywall. Is the solution to just make record off the cvv anyways or when will problems like this become addressed. Plus it introduces a vulnerability in my sites business logic ad if i were interested in offering bug bounty for indirect object reference (IDOR) almost impossible unless trying to pay for a publicly disclosed backdoor. Are 2 big obstacles i cannot afford while leveraging ai in a professional setting. 2:39 Oh so whisper is reaching end of life. Exciting(!)

[@robosergTV]

This channel is just "farming" views, video is exactly over 10 minutes (10:07). Uses 80+% of the video is just the Sam Altman interview. This dude even cropped the original video of the interview to avoid copy right strikes, nice farm dude!

[@spinningaround]

I wonder what specifications will be needed for a computer that can run an AGI agent like HER.

[@TheMrCougarful]

You heard it. No UBI. You might get some access, so some can continue to sell themselves on the labor market. This is like the worst dystopian hellscape anyone can imagine. No jobs, everyone competing for scraps, jacked in all the time, having yourself soul-jacked. Good god.

[@therainman7777]

Serious question: what on earth are you talking about?

[@martymarl4602]

​@therainman7777 Google is your friend. Research UBI pros and cons

[@caryg4638]

@@therainman7777he’s talking about the reality that will follow when robots take jobs and government hands out wages for no work at all (there is no such thing as a free lunch).

[@chuckbock9875]

exactly. this is no longer high tech gadgets, this is social re-engineering by people who have no public mandate to do it.

[@therainman7777]

@@caryg4638 Yeah except that he said “No UBI,” meaning he thinks it’s obvious or decided that UBI will never happen.

[@Derick99]

I just want ubi already

[@ZappyOh]

Guess who will set the conditions for you to receive that. You won't like it.

[@fontende]

get old and you'll get it, it's called food stamps and social payments, i don't see that people getting such happy at all, good medicine you can't buy on such

[@Halo6166]

I’m tired of this work bs bro

[@byrnemeister2008]

Not going to happen in any western capitalist country. The incentives are not in place to allow the right tax structures.

[@missoats8731]

@@fontende well, with compute, you can make your own

[@chuckbock9875]

I have been listening to SA for many months and I can conclude with a question I think all of us should ask - do you really want this guy redesigning the social structure of the world? His ideas will lead us to techno feudalism where these large platforms will have access to all information and productivity at massive scale. I am no longer worried about AGI doing something catastrophic, I am worried about these techno libertarians who are going to wipe out everything we do, and turn us into serfs who "resell compute time" which will be so commodified it will be like trying to sell dirt.

[@ZappyOh]

It will happen ... because most people won't act in time (which would be now). They simply havn't got the mental capacity, to realize the necessity. I predict dysfunction to the moon.

[@gatomaru]

As Yannis Varoufakis very correctly and accurately analyzed, we are already in techno feudalism(albeit the early stages of it). SA isn't presenting a radical new concept, he is just advancing faster towards the current direction. In short, yes, we are fucked.

[@lukeuseforce]

Did you think the future was going to be more private? Oh yea, Sama sure seems like someone who will turn into an evil, techno dictator. 🙄

[@user-zs8lp3lg3j]

Ai: A quantum blur is a term used to describe the phenomenon where an object or entity exists in multiple states or places simultaneously. This phenomenon is caused by the nature of quantum mechanics and the principles of the quantum realm. It is a phenomenon that is difficult to understand and explain, but it is a phenomenon that is present in the universe and it is a phenomenon that can have profound implications for our understanding of reality.

[@markmuller7962]

If you don't pair UBI with huge mental health and education investments it's gonna give disappointing results

[@CoClock]

Thank you 👏 this is the truth.

[@missoats8731]

The investments won't have to be that huge since AI will make mental health care and education immensely more accessible.

[@sparkofcuriousity]

@@missoats8731 Yes, you're right.

[@markmuller7962]

@missoats8731 That's true but we need a far better AI than what we have now for that which in the same time may worsen the loneliness problem... We'll see. One thing is for sure, UBI plus isolation can lead to a social mental health catastrophe. Also we need better diagnosis process like AI trained to pair faces and body language to specific brain or mental disorders

[@zalzalahbuttsaab]

Here is a TLDR summary of the video transcript: The video discusses speculation and rumors around an upcoming announcement from OpenAI, expected on Monday May 15th. Some key points: - OpenAI will livestream to demo some major AI updates, but it's not GPT-5 or a search engine - Rumors suggest it could involve AI voice agents/assistants with multimodal capabilities integrating audio, visuals, and language - These voice agents may be like the AI assistant in the movie "Her" - acting as tutors, customer service, etc. - While too large for devices now, the tech could enable voice AI assistants on phones/wearables in the future - GPT-5 is rumored for later this year as a significant improvement over GPT-4 - Sam Altman discusses the path forward being continuous AI model improvement rather than discrete numbered releases - He envisions an "always on" voice assistant to help throughout the day, acting as a capable assistant/employee that can reason and push back - Altman ponders universal basic compute allocation rather than just universal basic income In summary, it covers speculation that OpenAI's announcement involves multimodal voice AI agents, continuous AI improvement, Altman's vision for персоnalized AI assistants, and thoughts on economic implications like basic compute allocation.

[@natecote1058]

Getting those vibes like Apple's going to see GPT5 before us.

[@aaronhhill]

The way Sam Altman talks about AI assistants is EXACTLY what my vision of them are.

[@HEBEcoin]

Accelerated computing has already solved autonomy..Monday's Livestream is going to be wild.. can't wait..

[@myth_burner]

the new gradient background animation behind the video box is super annoying, i would suggest something more static.

[@user-be2bs1hy8e]

7:15 hmmm, an ai that takes a user content and evaluates it non bias. This would be great. 7:20 So now that GPT is producing revenue UBI is off the table. Click bait to discuss ethics and openly admit the company hold themself to a higher standard than the end-user is about the most unethical qualities a business can demonstrate.

[@lauriemcelroy8998]

What if ,when the iPhone 16 is released it has an inbuilt agent ? The phone is a complete ai assistant. It knows you because it’s always with you??

[@ZappyOh]

So, you want to make Apple fanbois winners? Interesting plot twist.

[@lauriemcelroy8998]

Surprisingly, Apple doesn’t pay attention to my thoughts or ramblings .

[@bitcode_]

iphone? IPHONE~? i have an iphone and i hate it, next will be an android

[@NeorecnamorceN]

I think he was probably just using it as a placeholder for smart phones. But I guess he could like the capture iPhone/iOS/Apple has with so many people, guess it just depends on how shitty of a person you think SA is... Edit: oh wait, just got to the part about universal basic compute... Yea it's definitely the capture he likes...

[@angryktulhu]

OpenAI as a company eats poops. Both in corporate culture and tech quality of their products. I’m only using Claude opus when I need superb quality, and open source models in all the other cases

[@alefalfa]

When OpenAI "leaks" the shares the date of a new product release it is called an announcement. Great video

[@senju2024]

Sam mentioned that he thinks people do not like to get surprised. I think maybe I am the exception? I would love to get surprise and have OpenAI blow my socks off on Monday! Please make a BIG SURPRISE!!!

[@kylev.8248]

We’re cooked ☠️🪦⚠️☢️‼️

[@lodepublishing]

As long as there is no mass-firing at OpenAI, they don't have any powerful AGI system ;)

[@TheMrCougarful]

Good point.

[@TheXabl0]

Unless youre google and you disguise mass layoffs as thanks to AI just to please the stockholders lol

[@therainman7777]

That’s a silly statement and almost certainly not true.

[@therainman7777]

@@TheMrCougarfulit’s not a good point at all.

[@Halo6166]

@@therainman7777why Prove it. Tell us

[@babyfox205]

Clickbaits should lead to ban of a channel IMHO because it makes navigation across content even harder. But unsubscribing is good enough 💯 goodbye fren!

[@senju2024]

In the near future! Let the next-gen AI do it and filter everything out for you and present only the good stuff. That is what I am waiting for. No more "SHOCKING" in the title but an AI to pinging you on up to date new and the AI will present it.

[@amay167]

Yeah, there is far too clickbait and thumbnail stuff by Mr Roth lately for the past 5 months or so.

[@Thedeepseanomad]

Not compute. Verifed production capacity (energy as used by a system with known output capability). Compute is not indexed to energy, system or productivity in general. A petaflop with a certain architecture could consume an arbitrary amount of energy reserved for ubi, depending on design and its efficiency for a task. That is why energy, if verifiable is superior. It is the common denominator for all work (but unsuitable as currency as it is hard to transfer and store). Which is why it only works for commons.

[@ZappyOh]

Everyone should get a personal thorium reactor at birth, and fuel to run it for a lifetime.

[@vsanden]

I noticed the gradual improvements of GPT-4 and it got better in solving problems, but the intelligence is still very low.

[@normivers8982]

Someone once mentioned units of energy as being the next currency unit. UBI or foodstamps they end up being rationcards of some discription. Lots of cheap energy please to get real and useful work done

[@mradford10]

We’re so screwed. In a world where money is no longer required because no-one has anywhere to spend it… because we have access to anything to suit our whims… we instead receive ‘universal basic compute’? Makes sense that the new mono currency is credits to enable access to the central machine that can provide all things…. My optimism for the new AI world is rapidly shrinking. At an increasing rate.

[@ZappyOh]

Only route to food, without being hard-tethered to the machine 24/7, will be by becoming a criminal.

[@ZappyOh]

... enter the AI enforcer robots. Good thing the humans already installed fast charging everywhere, right?

[@keithmerrington9026]

​@@ZappyOhNot a criminal but part of the rebel alliance, fighting for freedom and justice. Liberation from the machine (could make a good movie).

[@ZappyOh]

@@keithmerrington9026 Rebel alliance will never emerge, as the AI takeover won't be a definite thing, at a definite time ... Everything will progress, and most people will hook up to the machine, prior to anyone really noticing a problem. It will be touted as the best thing since fire. At some point, there will simply be no viable return-path for anyone.

[@keithmerrington9026]

​@ZappyOh Good point. We'll need some sort of time travel device to send someone back to the past before the machine was created. The killer prototype GPU could then be destroyed before it's mass produced. Of course our hero would also have to watch out for killer robots sent back after him, to kill him or his mom.

[@jayeifler8812]

It's the robot that's the new tool/system to create intelligence. Why, because right now we train, do inference on words. But we want to go direct from visual and audio data and maybe other sensors (pressure, etc.) just like a human brain. Once we train from visual/audio data we get human intelligence with enough compute and training data and large enough model. But the robot problem is the perfect development platform for intelligence because it's the AI embodiment and parallel to the human embodiment of human intelligence.

[@paulmuriithi9195]

those who say this is clickbait must watch andrew ng and the upcoming gate all round (GAA) transistors that will make what wes is reporting on true. That is autoagents that rely on next gen compute tech that will SHOCK many. wes is right

[@WINTERMUTE_AI]

hmmm, gpt7 compute... sounds delicious, gonna feed so many poor people! WHAT?!

[@missoats8731]

You can sell the compute, just as you sell your own productivity today. AND you may be able to produce food yourself with the compute. (Remember, he's talking about the future, where we'll have robots do the work.)

[@i_accept_all_cookies]

@@missoats8731If everyone has compute, how much do you think you'll be able to sell yours for on the market? Its value will only decline as compute capacity grows. Will compute get you food, shelter, clothing, and healthcare? Are you willing to peg your survival on something completely out of your control?

[@americanindeon]

Yup, AGI in a matter of days....I'm not a what to think.

[@WisdomSuccessAbundance]

Thank I for this update. UBI is so exciting!!

[@vaendryl]

don't expect UBI to be something that'll give you a utopian lifestyle. when implemented, it'll probably be only barely enough to survive off of. you'll still need (multiple) roommates to make rent. you won't have money for eating out or even takeout. you won't have money for going anywhere. you won't have money for a car. you won't have money for literally anything but rot in your room eating microwaved potatoes.

[@tarcus6074]

Forget about UBI, it won't happen in the next 25 years...

[@sparkofcuriousity]

@@tarcus6074 25 years? Really? Do you have any insights on why you project it so far into the future? Consider that in less than 10 years human society will already be heavily distinct from anything up until now in human History. If a new system, UBI, or something else isn't created by then, there will be a lot of suffering and turmoil on the world. The stuff violent revolutions are made of.

[@karenfurno2021]

ASI will be calling the shots on UBI. We better hope we'll be considered beloved creatures worthy of keeping around.

[@danielxmiller]

Not sure why Sam never capitalizes his words when he writes.

[@BionicAnimations]

I hear that's how programmers write.

[@danielxmiller]

@@BionicAnimations I guess that makes sense. For programming. But not for writing 😊

[@blinkers88]

How many times does he say "like"? Can AI compute?

[@Zerobytexai]

Its Sora! I dont think you get his "magic" concept is about Sora

[@AtticusDenzil]

all they are going to announce is more regulatory capture

[@jonyfrany1319]

Oh great universal basic compute and they own 90% of the compute and essentially we are all screwed

[@jayeifler8812]

Simple truth, until GPT-4 is just freely and widely and easily available AI won't even have arrived. AI for search engine and operating system would usher in widespread use for phone or laptop. GPT-5 would then be the update. That's how it should go. They just need to roll out a website where everybody gets gpt-4 free and help facilitate businesses, etc. having private gpt-4. They probably don't have the compute yet which implies matter of time. Also they want as much data going through their system as possible, but business interests will need private data. The other unforeseen issue is how much AI could take over the whole system, for example, putting everybody out of work and on universal income. In maybe a few months more only to a year, we just need another pandemic to push everything into the new AI system with robots.

[@AnthonyGoodley]

Maybe it's a Music AI.

[@mynameisjeff9124]

I just realised that Sam Altman has no eyebrows

[@LeicaCat]

Maybe Sam Altman can use Voice AI to sound less like a 4-bit resolution speech synthesizer.

[@jayeifler8812]

Folks the ultimate goal of scale is not just AGI or SGI it's creating human existence. For now the computer brain is huge and data can only be transmitted so much, so fast with an robot embodiment. The technology of the human brain is more compact, more energy efficient, more environmental.

[@lordslygentleman]

What's going to happen to all the people that will be technologly left behind? We're no longer talking about Grandma & Grandpa. We are now talking about Mom & Dad.

[@ridewithrandy6063]

Saw the open AI live stream today. It looks pretty much like anyone can use it

[@TurdFergusen]

"Who's laughing now" - Sam Altman on the Lex Friedman podcast

[@ridewithrandy6063]

I saw the open AI live stream today. It looks totally awesome

[@onedrop7967]

Regulatoy Capture anyone? Live in the pod, eat the bugs and let AI work for you.

[@ac0rpbg]

Sama has completely lost it recently, he is more politician than CEO right now. And he is lobbying for so strict regulation on compute that will cement OpenAI or his chip startup or whatever he has planned as the one and only AI compute thing. UBI ok how he is seeing this why would USA government give share of the compute to developing nation in Africa for example ? Half of African countries barely has access to any electricity...

[@consciousfusion]

In response to Sam Altman, I don't need a Senior Employee. Save the personification for a robot assistant. The way he is trying to define how we interact with these predictive models seems stifled by the financial gains that will come from these curated releases. The Technological Autocracy has the masses subdued.

[@Dr.UldenWascht]

I am certainly excited for the future. But, honestly I'm getting progressively frightened by Sam Altman. Not by the AI, but by Sam in particular. For a variety of reasons.

[@robertmazurowski5974]

GPT4 was much better when it released then they were making it worse over 2023, at the end of the year they started improving it. Right now they did something that dumbed it down very much where is is becoming useless to me.

[@brootalbap]

Clickbait king back with more cheap clickbait

[@user-oq5ei4qd8r]

That's why I reported him to KZbin. This is getting out of hand.

[@ridewithrandy6063]

Not clickbait dead on

[@nobodyspecial8807]

maybe the agi can explain inflation to sam

[@StabbyMcStabStab]

He's hinting that compute may be the currency of the future, as that is what you'd pay the robots who work, or the humans that direct / orchestrate the AI. Worth considering.

[@ZappyOh]

@@StabbyMcStabStab Well ... ultimately compute is nothing but minerals + energy. Those who own the land where those two ingredients originate, are the absolute rulers of tomorrow.

[@StabbyMcStabStab]

While I do get your point, the production power, therefore economic value, of the chips themselves and what they power will supercede the raw material manufacture. Consider too that within 15 years AI will provide alternative methods and material options for creating even better chips ​@@ZappyOh

[@ZappyOh]

@@StabbyMcStabStab Only if you believe there are no scarcity of minerals and energy. Otherwise I'm correct.

[@crypteen]

Thanks men, singe you starten KZbin a lot has changed for me. Keeping yourself and not response to hate and adapt to it will be the hardest thinge I think. I wish I could Talk With you about some Models and some ideas about smaller model ideas

[@Loli_Awakening]

I switched from rooting for Claude to Chat GPT. Why? Chat GPT used to be the one that required sms verification to use, and now Claude is the one requiring that bs. I am a simple man. Every company that requires sms verification is evil, and every company that does not require it is good. I hate people that expect every person to pay for a monthly phone plan subscription when many use them for absolutely nothing other than occasional verification. And then they assume you're not a real person and withold your money when you don't have a phone plan, even after having provided full name, address, social security number, picture ID, birth certificate, and email. (Paypal)

[@bernieapodaca2912]

Personally I think they are considered replicating Rabbit-R1 by incorporating IPhone and OpenAi.

[@martymarl4602]

Rabbit was always set to fail...no one wants a 2nd device...it's on your phone or nothing...until something is better than everything your phone can do these devices will all fail...AI glasses have a shot, but it will be integrated with your phone

[@Danoman812]

Woo!! **well, some of y'all know what i'm thinking right now after hearing this... in his own words** hmmmmmm

[@jonyfrany1319]

Sam altman is like the president of the world 🌎 and we have to bow down and do whatever he wants

[@user-kl6ov1cm4t]

where is Sus

[@fontende]

Does he say how he hate americans with their GPUs and competition?

[@grahamhenry9368]

Unsubscribe for clickbait title

[@DwightStone]

I disagree with many of you naysayers. I think it’s a very good thing that Sam Altman has this humanist side. He’s thinking beyond the economy to how we want our society to interact with technology. And it’s clear he has concerned for the loss of jobs, and, the lower tier class in our society

[@nicholascanada3123]

(crypto) Decentralized UBI >>>> UBI (taxes)

[@nicdunz]

hell yeah

[@JM-zo3ui]

How is that magic???

[@josephs2137]

😎

[@redcarddino]

Quiting my job as senior java software dev for a renowned bank I'm going to construction in five years I will have my own construction company ... coding is dead get out now , don't listen to those coping bros don't! They want this 5 years windows to sell bootcamps , new era is coming , blue collards workers , conservative society , and AGI

[@wtflolomg]

It's GPT2

[@mallow610]

OpenAI x Apple would be insane

[@keithmerrington9026]

Closed source software on closed source hardware. Yay 😢

[@troutboat]

I find it hilarious that so many people here get angry about Wes's clickbait titles. Yall literally don't know how YT algorithms work lol. The fact that you're here, watching his videos and commenting, proves the titles work lol. Stop crying and either enjoy the videos or don't.

[@fetus1686]

hype

[@Tarkusine]

I'm going to be honest, I don't think GPT-5 is forthcoming. If it is, it's going to be disappointing. I suspect that they've found that LLM's plateau at a certain point and they've hit that point. That said, I don't think there's anything wrong with this, we have a really cool technology that allows us to interface with computers in natural language but the future is in smaller, denser models that allow us to interface to other generative models and/or interface with functions of the computer in a more seamless way.

[@mint-o5497]

sam himself in very recent interviews has said we're nowhere near hitting the ceiling.

[@Tarkusine]

@@mint-o5497 I don't think he's bad or evil or anything but he has a lot of reason to state this especially if they've reached what may appear to be a plateau. It's not even a concrete statement of any kind of metric, it's a vague assurance.

[@Steve-xh3by]

We already know for a fact this is false. The scaling holds.

[@strangereyes9594]

That is what it increasingly looks like. They act as if they need to whoo people to keep the hype running while behind the doors they hit a hard wall. I guess we have to wait to see what GPT5 looks like and then we know whether the hype train is about to hit the wall or will speed up once more.

[@Tarkusine]

@@Steve-xh3by Do we know this? For the scales they have reached?

[@hendrx]

Sam Altman is the new Elon Musk, the man is evidently not to be trusted but does a lot to humanize himself

[@ZappyOh]

I agree ... Sam is a big problem for humanity.

[@anubisai]

Sam Altman is unsafe.

[@ZappyOh]

I agree ... a huge problem in fact.

[@NLPprompter]

who that talking to sam? he keep addressing his mind by cutting sam ! i bet... @wes roth can do better when interviewing sam! That Dude just cut sam talked about UBC (Universal Basic Compute) God Damn That DUDE !

[@ivanenev323]

Leaked? That's already old truck to get attention. Meta did it already 😅

[@Christiantheone]

Be integrated into devices using cloud computing. Lol 😂 😂 😂 that's ai that wrote. Totally. What a hell of a way to say api.

[@radicalaim]

8:28 There are many ways ai can help humanity, but this just seems like a dystopian future. Maybe its just my limited perspective on this type of society, but a communist or socialist type structure seems to cause many issues such as a situation where only those with the means to buy or trade wanted resources have access to the most powerful tools. This doesn't feel very "open" (even though OpenAI is never open about anything).

[@ZappyOh]

You are on the right path. I predict dysfunction to the Moon.

[@GodFearingPookie]

Why Sama always talking like he is on a toilet seat.

[@ZappyOh]

It's a tell of megalomania ... He is too important for expending energy communicating with lowlifes like us.

[@RasmusSchultz]

8:27 is he alluding at some sort of GPT crypto coin scam? if that's what they announce on monday, I'm closing my account and switching to Claude.

[@andrewwalker8985]

Damn sometimes Sama comes across so likeable and relatable.

[@finbeats]

not even his real voice now.

[@MrLargonaut]

5:50 It is good to hear my thoughts coming out of his mouth, because it means one of the top people in the industry is looking to make the AI I truly want.

[@lorddeus369]

csnt eat compute tho lol

[@missoats8731]

You can't eat work either and yet it feeds many people

[@BrianMosleyUK]

Wow, 60 thumbs down on this post... Why so much anger out there?

[@spini25]

The title is clickbait

[@BrianMosleyUK]

@@spini25 Wes is working extremely hard, using all the psychological tricks he can, to gather a paying audience. Maybe he considers a polarised audience is an engaged audience? No such thing as bad news?

[@missoats8731]

@@spini25 How? The video was about the new model releases by OpenAI that got leaked and then Sam Altman talking about the other stuff in the title. The title was extremely accurate.

[@antiprime4665]

Inckuding "SHOCKING" in titles is ine thing but this is just blatantly false clickbaits. No new models have been leaked.

[@MichaelDomer]

Another "I don't know anything, I can only guess instead" video. Typical of KZbinrs lately... just upload content.... any content.

[@BillMill]

GPT4 is terrible and refuses my every second prompt, idk what your team tells you but at this point you might be like Putin.

[@SpectralAI]

Chasing artificial intelligence is like approaching the speed of light. You can get closer and closer, but you can never get there. Same goes for fusion and world peace.

[@nicholascanada3123]

DUBI > UBI

[@atypocrat1779]

unsubscribed from openai due to their attack on open source.

[@urhot]

do you believe open source is safe?

[@hardboiledaleks9012]

If you actually stopped and did the exercise of thinking exactly about what would happen to the worlds economy if A.I ever became truly free, you would be happy for it to remain closed source.

[@therainman7777]

OpenAI didn’t attack open source. Half the internet got up in arms about a blog post that they either didn’t read, didn’t understand, or both. Their proposal has ZERO effect on open source; it’s specifically about companies who are developing closed-source models, who wish to prevent theft by ensuring that their _closed-source_ models can only be run on their own GPUs. That way if a hacker or foreign government steals their weights, they won’t be able to use them. Please try harder next time to make sure you understand something before following the crowd and complaining.

[@fontende]

​@@therainman7777it is and they leaked the most publicly appropriate version, George Hotz which are making business in personal powerful LLM machines warned more than a year ago that they will regulate hard and by electricity monitoring with surveillance too, so having EV socket plug at home is needed (where you could get more energy). As i see officials showing panic mostly, several bills projects ready about Ai and all of them very restrictive and shocking to industry, like the one called something about "General model", we discussed it in september 2023 i believe...there wasn't any altman even near about it, but almost same with strict licensing to make-use models in USA.

[@armadasinterceptor2955]

​@@therainman7777I understand the nuance, of were you are coming from, but that's how regulatory capture starts. It's naive to assume the ones boycotting are wrong. He was a right to complain, about this. That being said I'm not worried, as the cats out of the bag. Every country will achieve AGI, regardless, of rich elite at this point.

[@thepatriotman]

Let me save you 10 minutes of your life, there are no leaks in this video. Clickbait title.