so open Ai and Sam Alman are up to something there's a flurry of activity we're still not 100% sure what's happening but let's take a look at what might be happening subscribe for more sweet AI content open AI has been relatively quiet they've announced a few things that some of which haven't even rolled out yet like the voice mode for Chad gbt and today we find out why they have been relatively quiet and also potentially why that's about to change so first and foremost Chinese developers scramble open a blocks access in China if you're an open AI developer if you use their API to do various projects and you provide access to those projects to the whole world you might have received an email from openi saying that they're limiting some countries from accessing those features now this isn't completely unexpected this isn't completely new so open AI last month announced plans to crack down a china-based AI developers Now by default you can't access open AI in China there's a firewall but a lot of Chinese AI developers are getting around that but these developers have come up with fish ways to get around the firewall by using well virtual private networks vpns as you might have expected now some of the AI researchers in China have said this move caused significant concerns within China's AI Community they're saying the decision raises questions about Equitable access to AI Technologies globally now the cynic and me is wondering do they use that word Equitable unless unless they're talking to Western reporters I mean is that a is that a thing that gets talked about or is that more kind of like PR for the Western World I'm genuinely curious because reading some of the Chinese AI research papers they're really good at I guess you would call it localization like certain things they publish doesn't just get translated word for word it also gets changed to appeal to that country or that population now the rumor mill is going crazy a lot of people are asking if this means that open ey is getting ready for another big release they're wondering if they're going to throw out a new model GPT 5 perhaps or maybe API updates of some sort there does seem to be some other changes looks like they're adding new domains EU API open that were added yesterday so tber is asking sounds like the rest of the world is getting new models or API updates soon let's come back to that in just a second because there might be a clue as to what opening eye is going for in terms of how they release their new products and new models there's some new data about that but let's come back to that in just a second so I'm sure most people are Unum level familiar with this person Ariana Huffington so she is the co-founder of The Huffington Post and more importantly for what we're talking about the CEO of Thrive Global but anyway she mentioned this article that is in Time magazine called AI driven Behavior change go transform Healthcare it's written by her and Sam Alton there they're talking about the problem that we're facing with health the cost of Health Care chronic disease Etc but they're saying Behavior change can be a miracle drug both for preventing disease and for optimizing the treatment of disease which is true with things like optimized sleep nutrition exercise there are tons of stuff that can be done without assisting without needing to go to the hospital or visiting a doctor but Behavior change is hard but through hyper personalization they say it's also something that AI is uniquely positioned to solve so the idea is to use AI large language models so something like Chad PT that knows a lot about you your habits maybe some of your medical files what you're trying to achieve and is able to coach you through how to do that and these are the ideas behind Thrive AI Health the open AI startup fund and Thrive Global are jointly funding to build a customized hyper personalized AI health coach that will be available as a mobile app and also I guess they're going to have some sort of an Enterprise side of it here's the thing I could see this working I could see this working really really well people that are tracking their calories and macronutrient profiles Etc tend to eat better just the idea of being aware of what you're eating and how many calories it has and how that fits with your within your daily macro allowance that by itself makes you a little bit more intelligent about what you eat and how what choices you make there was another study where they took a number of Hotel housekeepers or or part of the cleaning crew that kind of uh you know tiddies up your hotel room when you're done with it and half them were told about how good the activities or cleaning are for your weight loss and uh Health efforts that it's like a good exercise how many calories you burn per day just doing that and they found that that group of people started moving more burning more calories because it was just more intentional so just having somebody that as they put it offers real time nudges and recommendations yeah I genuinely could see this being a big deal for somebody that needs to take little micro actions little tiny steps towards Health I've been messing around with building my own kind of voice AI assistants they use various language models like CLA 3.5 or GPT 40 I've showcased them on this channel and I constantly think wouldn't it be nice to be able to use that in this particular scenario for example if you're doing a workout something that can log how many sets you're doing how many reps you're doing over particular exercise the weights that you're using you can just verbally say okay I did eight reps 40 lbs and that gets logged somewhere so that the next time that you're working out you can ask okay what did I do last time currently you're either using an app for that where you're like tracking your typing in the numbers or maybe you just use a pen and paper but both cases that takes you out of the exercise you have to stop and write it down cuz you might not remember every single rep and weight that you used at the end of the workout same thing for using something like AI Vision to where you take a snapshot of what you're eating and it automatically gets tracked as calories proteins carbs fats Etc it doesn't have to be 100% accurate just a rough estimate of what you're eating tracked every single day with no friction no frustrations it just gets easily tracked in your own personal app with all your personal details and then you being able to use it later you know say oh I felt really good the whole last week in the mornings felt great what was I eating and it's able to reference that data give you recommendations Etc now what's interesting about kind of what opening ey is doing is right now it seemed like they were being fairly quiet no big new releases but at the same time there's a lot of like rumors and whispers about all the stuff that they're doing across the board so for example here they're partnering up with a hell startup a lot of people have mentioned that they're talking to a lot of the Hollywood Studios about using something like sword to start developing movies Meanwhile we're hearing almost daily news of various other companies that are approaching or outperforming GPT 4 GPT 40 for example here Chinese AI company sense time takes the lead with sense NOA 5.5 Reports say that it has outperformed GPT 40 across key metrics so here's Jr KB saying they are the only ones to have grown significantly recently so they are no longer afraid to take their time before releasing their next month model general public is not aware of the benchmarks and even for those of us that are aware of the benchmarks and kind of know who's an elite or not I got to say I mean I I am kind of dismissive of the benchmarks yeah I'll take a look at them and if something's like dominating that to me is interesting I will look further to try to figure out how good that model is but I just don't take it at face value but the point of all this is this chart so this is desktop and mobile web visits world wide that blue line well that's chat GPT right approaching 3 billion visits and this rocket ship like growth this was around April 2024 can you guess what caused that why did it take off like that after being kind of a in a in a holding pattern for a while by the way these are the other models so we have Google's Gemini here in orange character AI a very interesting company we have to look at it more is kind of crazy for a lot of different reasons not only because of who the founders are but also how many visits it gets and also what people use it for we'll look at that in a different video and below here we have perplexity Ai and CLA AI now to me it's very strange looking at this chart because this to me is very surprising because being deep into the stuff every day I'm thinking oh CLA is nose to nose with Chad GPT Etc just because of that's what people are talking about but the real reality is the majority of the people out there probably can't name the Google AI model they're not going to say it's a Gemini most people have no idea what perplexity AI is right everybody knows Google people know Bing no one knows perplexity AI people know chadt they don't know claw as much it's a tiny fraction of the amount of people that have at least heard of Chad PT so what this is showing to me and take this with a grain of salt I don't know this for a fact but it sure seems like opening eyes feeling very comfortable in its position as the worldwide leading AI model also notice what happens whenever it releases the next best thing very quickly there are 10 different copy cats that are trying to get closer to it and succeeding the announcement of Sora was Earth shattering everybody was just Minds blown but now we're having the releases by cing AI Runway gen 3 Luma Ai and I got to say all of those are looking really good I don't know if they're as good as Sora or not we don't know if the examples we saw from Sora are cherry pick the best of the best or just your average outputs but it does seem like usually when opening ey releases something within 6 months you get a number of other companies that catch up and do either as good or almost as good so if I had to guess I think open ey is probably laying the groundwork to be everywhere all at once every single time the next big thing that they have gets released they want to have the groundwork for having that Health app the coach they want to have some deals with the Hollywood Studios they want their API to be everywhere they wanted to be and not showing up in all the places they don't want it to show up like China because that could potentially maybe cause problems because of the whole us China conflict thing that's happening on the AI front so once the next big thing rolls out they already wanted plugged into everywhere with users with applications with everything before you know basically on day one because as soon as they release it to the public that clock starts ticking where everybody else is able to for example train their own models on the outputs of open AI models create something similar or better so I think the goal is going to be here to release into an already existing infrastructure with everything plugged in now if you're wondering why there's this rapid growth I mean this was around the time they announced the voice engine but we still don't have the voice engine what that growth I think is is the fact that now you can access Chad GPT 40 you can access it for free you can access it without logging in if I open up an incognito window I can use Chad gbt without logging in without signing in without creating an account with the free plan you get limited access to GPT 4 the best model in the world is free to the whole world now so this to me feels like the Cal before the storm I don't think they're giving away everything for free because there's nothing bigger coming I think number one there's something really big that is being prepared for release number two I don't think they're going to be as focused on selling it to the consumer right so the $19 a month Chad gbt subscription that's not their main focus I think they want to sell it to Enterprise businesses to Hollywood to these AI Health companies so they can make apps for the end consumer or sell it on the Enterprise side to Apple for example right that was the big announcement that they're going to be integrated with iPhones and I think that's the game plan they want to be plugged into apple and Hollywood and all those places so when they roll out the next big model it just gets updated and it's live on all those devices and all those places at once here's gnome Brown he's researching reasoning at open aai he co-created the first superhuman No Limit poker Ai and he co-created Cicero we talked about cisero just a little bit on this channel so if you imagine a game like Risk where you're supposed to take over the world but imagine that there's much more diplomacy where you're able to negotiate with other players make alliances break those alliances and basically work together to try to survive or take over the world very strategic game very social game in the sense that you have to understand what your opponents are thinking CU they're not really your opponents they can be your allies your opponents however you kind of Define that relationship well Cicero co-created by this person became as good as the top World players of that game an agent that can play at the level of humans in a game as strategically complex as diplomacy is a true breakthrough for Cooperative AI Yan Lon so they're describing it as a dialogue aware strategy which is that's a great way of putting it like it's strategy game based on you have to be aware of the dialogue that you've had with all the opponents the point being is this is an accomplished person in the realm of AI and he's working at open aai and today he posted this when I joined openai a year ago I feared Chad gbt's success might shift Focus from long-term research to incremental product tweaks which that is kind of the perception by some from the outside of what open ey is doing right just put our product and tweak it here and there to improve it and that's the status quo now but he's saying but it quickly became clear that that was not the case openingi excels at place big bets on ambitious research directions driven by strong conviction which by the way this is kind of Sam alman's Mo if you look at how his investment career kind of went what he says is the secret to a lot of his success in terms of investing in startups is he focuses on long-term projects that require a lot of capital and he's saying there's not a lot of people competing in that space people want a quick return people don't want to be placing big bets on some future ambitious project or goal and so if you're able to do that and you're able to do it successfully the payoffs are huge and here gnome is kind of supporting that open AI is still going that direction that that is still happening they're still placing big bets on ambitious projects that are long-term that are far away and that they stand by them with conviction meanwhile it seems like the open AI team is busy behind the scenes I think because of how many people are trying to glimpse into what they're doing try to pull out any shred of evidence for something new coming out now they're actually attempting to hide their various new experiments by kind of masking them by not making them obvious looks like what used to be called chadt included in voice Alpha so I'm assuming that's The Voice demo the voice engine that's rolling out to a small select group of alpha testers right so maybe seems like it got renamed to whatever 3199 whatever some string of numbers to kind of like try to hide that a little bit not make it as obvious so as excited as I am about all the various new progress and new features that are being rolled out by all the other AI models the different tex2 video platforms anthropics new releases have been phenomenal but as exciting as all those developments are I do think that openi is cooking up something crazy and Secret in their underground lab and I think it's going to be scorched Earth as soon as it gets released cuz it will rapidly get applied to all the projects that they're building infrastructure for right now that's my guess by the end of the year I'll let you know if I'm right or wrong and I'll reference this video in the meantime this person nak Lea is saying that they're convinced they can automate all of factorio using language models and looks like they're going to be using uh the open- source llama 3 or their own kind of version their own fine tune version of it it seems like it's beginning to play factorio on its own I don't know about you but to me this is incred incredibly exciting and interesting and just it's fascinating to watch something about automating games with large language models something about that is just incredibly exciting to me I I I don't quite know why but I can't wait till this is more of a reality when it's more approachable for the average person I've experimented doing botting back in the World of Warcraft if you know what that is but they were just simple script based Bots and a lot of people didn't like the fact that people are Bing that game I get that but there is some strange Fascination and satisfaction for me and I know a lot of other people as well from creating little systems like that little automations little Bots that run around and do your bidding 24/7 that you can tweak and optimize and make better and of course large language models with vision reasoning skills and the ability to interact with these games that will be the next big thing on that front by the way if any of you are building something like this something similar something at the intersection of AI specifically large language models and things of that nature and games automation stuff like that let me know I want to help in whatever way I can happy to provide some coverage on the subject especially if you're making content with it like video or tutorials if you're making anything that's going to be useful for people either with a tech background but also non- techies if it's useful for them to get into this world if it helps them learn and kind of dip their toes in the waters of working with llms if you're doing anything like that let me know I'm happy to shine a light at it in whatever way I can that said my name is Wes rth and thank you for watching