Q* Open AI's New GPT 5 May Be Reason For Sam Altman Firing?!

All right guys there's been a lot of breaking news in terms of new speculations that AGI may have been achieved or is right ahead of us let's dive right into [Music] it all right so to start this thing off if we look at Ilia suers Twitter he tweeted out there exists no sentence in any language that conveys how happy I am and that's a replied tweet to open ai's statement saying they've reached an agreement in principle for Sam ultman to return

To open AI as CEO and then you see Elon Musk tweeted below it saying what does this letter refer to and he's linking to a routers a rutters.com article and he tweeted it out again saying extremely concerning so let's look at this article you can see here open AI researchers warned board of AI breakthrough ahead of CEO ster Source says ahead of open AI CEO samman's 4-day Exile several staff researchers wrote a letter to the board of directors warning of a

Powerful artificial intelligence Discovery they said could threaten Humanity two people familiar with the matter told Rutters the previously unreported letter of AI algorithm were key developments before the board aler of mman the poster child of generative AI the two sources said prior to the triumphant return late Tuesday more than 700 employees Tred to quit and join backer at Microsoft which we covered a lot of those videos so you can check some of those out if you need to get up to

Date with all all that that drama the sources cited the letter as one factor among a longer list of grievances by the board leading to alman's firing among on which were concerns over commercializing advancements before understanding the consequences Rutter was unable to review a letter of uh a copy of the letter the staff who wrote The Letter did not respond to requests for comment after being contacted by rotor open aai which declined to comment acknowledged in an internal message to staffers

A project called qar and a letter to the board before the weekend's events one of people said an openingi spokesman said that the message sent by longtime executive Mir moradi alerted staff to certain media stories without commenting on their accuracy some of opening ey believe qar pronounced qar could be a breakthrough in the startup search for what's known as artificial general intelligence one of the people told routers open AI uh open AI defines AGI as autonomous systems that surpass humans in

Most economically valuable tasks given vast commuting resources the new model was able to solve certain mathematical problems the person said on condition of anony because the individual was not authorized to speak on behalf of the company though only performing math on the level of a grade school student eing such tests made researchers very optimistic about Q star's future success disorder said so let's kind of break some of this down um you know fill you fill you in on kind of what

This means and whatnot so you may just be thinking okay well it's basic math problems what's the deal with this these basically this you know essentially the way they're saying is a breakthrough because um they can kind of do um you know predictive scaling and see where you know things are going to scale with models so even if it's just starting to solve these new breakthroughs and in mathematical problems that it may not have been able to do before then you

Can kind of predict where it's going to scale to with these with these models so that's one thing if you can solve have those breakthroughs on a small level also two qar so if you're not familiar what qar is uh Q is is kind of the meaning of q learning so we'll get into that in just a second and give you kind of overview of what that is okay so here is a tweet from a P qar can solve grade school

Math problems so what what's the big deal why did the research send the letter why did Ilia freak out because open AI has already has a stack where they can predict intelligence based on compute and data these disclose this with the gp4 release and I'll show you that in a second qar is relatively is a relatively small model solving grade school math problems reliably and consistently using reasoning it's an algorithmic breakthrough and they have scaling graphs to predict the intelligence increase

From data and compute so they know they can probably solve Elite human and Beyond human problems they don't have AGI yet but they have a clear uh near-term road map now the question is whether to go down the path or not Sam's reaction to this was to raise tens of billions of dollars to get a Joon I device into people's hands and build data centers everywhere to deploy this Tech as soon as possible all right and here's an example of their

Predictive scaling so a large focus on the gp4 project has been building a deep learning stack that scales predictably the primary reason is that for very large training runs like gp4 it is not feasible to do extensive model specific tuning we developed infrastructure and optimization that have very predictable Behavior across multiple scales to verify this scalability we accurately predicted in advance GPT 4's final loss on our internal code base not part of the training set by extrapolating from Models trained using

The same methodology but using 10,000 times less compute open AI codebase next words prediction and then observed prediction and gp4 you can see here on the graph all right now qar comes from Q learning all right there's qar there's a star and now what is Q learning well let's break it down very simply using chat gp4 so what is Q learning imagine you're teaching a robot to navigate through a maze each time it makes a move it either gets closer to

The exit a positive outcome or bumps into a wall a negative outcome Q learning the Q learning is a method used to teach the robot which moves are the best in different parts of the maze key points of QQ learning learning through trial and error so the robot tries different paths in the main maze each time it tries a path it learns something new about the maze two rewards and penalties the robot gets rewards like points for good moves like getting

Closer to the exit and penalties for bad moves like hitting a wall three memory of past events or past actions the robot remembers the Rewards or penalties of each move at different spots in the Maze this memory helps make it bet make better decisions about where to go next and for it making decisions the robot uses its memory to decide the next move it usually chooses the move that it remembers having the best outcome and how is Q learning different from

Other forms of machine learning so one no need for a map so it's model free so unlike other AI methods in q-learning the robot doesn't need a map of the maze beforehand it learns by doing not by being told or shown a correct path and then two focus on actions and outcomes so Q learning is about finding the best actions in each situation it's different from say pattern recognition like identifying cats and photos where the focus is on learning from examples

And then three adaptability so Q learning is good for situations that change for example if the maz's wall move the robot can adjust because it's constantly learning from the environment and then four long-term strategy the robot learns to think about the long-term consequences of its actions like chess players thinking several moves ahead not just the immediate reward and then an easy analogy is like learning to cook think of hue learning like learning to cook without a recipe you try adding different

Ingredients actions to a dish some will make it taste better which is a positive reward and some will make it Ru and some will might ruin it which is a negative outcome over time you remember which ingredients work together well and start making better dishes even though no one gave you the recipe to follow in simple terms Q learning is like teaching a robot to solve a puzzle through practice it remembers what works and what doesn't gradually getting better at the

Task this is different from other AI methods that might require a set of instructions or examples to learn from Q learning is all about learning from experiences and making DEC decisions based on past successes and failures so sounds very similar to you know how you would lots of times May learn something and it's kind of leading towards it's kind of like the stepping stone people are thinking of you know this next level of autonomy getting towards AGI and all that good

Stuff so um yeah it will be interesting to see what's next all right and here's some different things from people in the AI Community here's someone saying if qar is an agentic algorithm breakthrough that can reason and goal seek via something like AAR search against an entire world model embedding it completely explains this weird statement from Sam last week is this a tool we've built or a creature and I'll play that video right now but it has been I think this

Is like definitely the biggest update year for people yet and maybe the biggest one we'll have because from here on like now people accept it powerful is going to happen and there will be incremental updates but there's like you know there was like the year the first iPhone came out and then there was like everyone since and at this point like you you really know the difference between this one and last year's one um so so this was like I think

A big moment one thing that I am happy about is uh I think people are viewing these viewing these systems as correctly as tools artists in particular but um other people as well and there was this I think there's a real moment of fear which is like is this a tool we have built or a creature we have built and what is what is that going to mean I think people now view this as a new thing in the toolbox of

Humanity and are doing like really remarkable things with it uh it's very all right and another thing that was interesting guys EMT Shear who is the ex CEO of twitch and was the interm CEO of open AI when Sam Alman was fired he was joining the talks and was the one of the turning points in the neg NE iation or actually was the Turning Point apparently and he asked the board to provide concrete evidence of Sam's wrongdoing when they did not

Do so he said he could not remain in the CEO Rule and helped both sides find Common Ground around getting Sam back reinstated so really cool I guess from EMT Shear to do that all right guys so a lot of people are speculating that either AGI has been achieved internally and that was the whole thing that made Ilia Sask really um you know lose his mind and really freak out or that we basically are on the path to achieving it because

Like you saw in that video Sam man also said that there's only been four times in open AI history where it's really seen that he's pushed past that veil of um you know having a breakthrough and he said the most recent one was a couple weeks ago and this was right before he got like four times now in the history of open ey the the most recent time was just in the last couple of weeks I've gotten to be in the

Room um when we sort of like push the front the sort of the veil of ignorance back and the frontier of Discovery forward and getting to do that is like the professional honor and you can see here he actually well not tweeted he posted on Reddit saying AGI has been achieved internally and then he edited that post and said edit obviously this is just meing yall have to chill when AGI is achieved it will not be announced with a Reddit comment

All right so he said that and he also said this right after Jimmy apples and if you're not familiar who Jimmy apples is he is basically a open AI Insider who's gotten uh information correct every single time and he tweeted out saying AGI has been achieved internally and and then Sam alond tweeted that out fairly after so some people are speculating that it has been achieved some people are well this kind of shows that Sam Alman probably knows of Jimmy G

Jimmy apples um and uh also too open a has quietly changed its core values and being thoughtful and audacious no longer makes the cut so this is actually a few weeks ago or um some time ago that this actually happened but you know this is kind of just referencing back um that open ai's Mission and their values and what they their whole purposes did change kind of quietly on their website and if we go to their website here you'll see that

Their core values are AGI Focus right so we are committed to building safe beneficial AGI that will have a massive impact positive impact on Humanity's future anything that that does not help with this is out of scope so anything if they're maybe building other products they that is no longer in their scope everything they want to do from now on is Agi focused all right and then I'm also going to insert a clip here from Shane lag who is actually the

Founder and chief AGI scientist of Google deep mind and you can kind of hear his thoughts on some of these new models and these new Frameworks and uh what he has to say these Foundation models are World models of a kind and to do really creative um problem solving you need to start searching so if I think about something like alphao in the move 37 famous move 37 where did that come from did that come from all its data that it's

Seen of human games or something like that no it didn't it came from it identifying a move as being quite unlikely but know possible and then via process of search coming to understand that the that was actually a very very good move so you need to you to get real creativity you need to search through spaces of possibilities and find these sort of hidden gems that's what creativity is I think current language models they don't really do that kind of a

Thing they really are mimicking the data they are mimicking all the human Ingenuity and everything which they have seen from all this data that's coming from the internet that's originally derived from humans if you want a system that can go be truly beyond that and not just generalize in novel ways so it can you know these models can blend things they can do you know Harry Potter in the style of a West r or something even though it's never happened they

Can blend things together but to do something that's truly creative that there is not just a blending of existing things that requires searching through a space of possibilities and finding these hidden gems that that that are sort of the hidden away in there somewhere and that requires search so I don't think we'll see systems that truly Step Beyond their training data until we have powerful search in the process all right guys so let me know what your thoughts are in the

Comments down below about this news with qar with you know all these this different framework that you know essentially that is kind of being said that they're they found or that they're using now this new model for GPT 5 and do you think they've already achieved AGI or do you think this will just basically be a huge breakthrough for them to achieve AGI or let me know any of your thoughts down below what this could possibly mean what you guys are

Thinking is the future for you know Humanity for AI for AGI for the whole Singularity what do you think is going to happen how fast are we on that trajectory um this is you know what we're seeing you know breakthrough after breakthrough we're having so um it's really only a matter of time before um things really start to get crazy here so if you want to stay up to date with all the news make sure to smash that subscribe button smash

That like button um let me know what your thoughts are about everything about all this in the comments down below also too guys make sure to join our stride Community it's a free community of entrepreneurs AI enthusiasts where you can Network share your thoughts and there will be a lot of free resources in there it's a free Facebook group and Discord Channel link for that will be in the description below and then you can opt in for our free 6-day accelerate

Your stride challenge where there's a bunch of AI resources and business growth tactics in there it's 100% free links to all that free stuff will be in the description below guys and I will see you in the next video keep hustling keep grinding and of course accelerate your stride take care