China's AI Guru SLAMS Sam Altman: "He Isn't Doing Well!"

think about open ai's cost of $7 billion of operating cost in 2024 deep seek probably operated with 2% the operating expense so with that kind of formidable competitor I think uh Sam Alman is probably not sleeping well yeah well he so that was Kai fi and this guy is a prominent figure in Ai and Technology entrepreneurship who actually currently serves as the CEO of covation Ventures a venture capital firm he founded that manages 2. 5 billion in assets developing high-tech Chinese companies now this guy isn't just some incredible investor he's also the founder of 01 an AI startup launched in July 2023 and he's had leadership positions at Google Microsoft Apple and this guy has established himself as an influential voice in AI development often referred to as the Oracle of AI he has books including the AI superpowers AI 2041 which explore the future impact I and so this guy isn't just some random person giving an interview it's actually a very established businessman and well-respected entrepreneur talking about the state of artificial intelligence now the clip that you just saw was a short snippet from an interview in which he's talking on Bloomberg about the state of AI and of course you can see he says that Sam mman is probably not sleeping very well now I do understand exactly where he's coming from the deep seek drama has been absolutely incredible in this clip because it was something that I don't think anyone really saw coming now whether or not deep seek did lie about how expensive their model is I think that is besides the point but I want you guys to watch this full clip because the implications of what he's saying are certainly true and of course if you watched yesterday's video you'll know that it's already having you know certain impacts on how openi are changing their actual company structure so take a look at this full clip which I'm going to dissect in a moment in both us and China the pre-training of a giant model has Consolidated and is consolidating and it's become clear that open source will be the winner um there's still many that will not conceed right open AI anthropic who build their businesses believing they can build a better closed model than everyone else I think they got shocked when they saw a model as good if you think about open AI cost of seven billion dollars of operating cost in 2024 deep seek probably operated with 2% the operating expense so the issue really isn't whose model is 1% better I think they're all very good but the issue is open ai's model even sustainable right I mean you're spending 87 billion a year making a massive loss and here you have a competitor coming in with an open- Source model that's for free and that company also is infinitely uh lasting because this founder has enough money to fund it at a current level and has reduced the cost of computing by a factor of 5 to 10 so with that kind of formidable competitor I think uh Sam Alman is probably not sleeping well yeah so it was definitely a rather interesting clip you can see at the end of the interview there's actually a smirk on his face because of course he makes the comment that Sam Alman is probably not sleeping well and honestly I think it's really interesting to see some of the moves open ey has pulled if we take a look at what they've done recently we can see that open ey called Deep seek a state controlled chatbot and it actually called for a ban on PRC produced models now this is something that I will say is 50/50 in terms of it being true and of course in terms of there being a massive incentive for openi to do this because of course we know that if deep seek does get banned that helps out open AI a ton you can see that they actually called the Chinese lab deep seek a state subsidized and state controlled and recommend that the US consider Banning these models from the outfit and similar PRC supported operations now it isn't just a completely baseless claim because deep seek does require under Chinese law to you know submit certain data when they do have requests for certain demands for that user data but I think what will probably happen if anything does happen is that deep seek will probably only be banned on government devices and probably not anything else so this move from open AI to say that look this is Chinese owned operated and controlled whilst it does have some kind of grounds I think a lot of people are looking at this move as a rather strategic move to sort of Wipe Out the competition and honestly you can't blame them for doing that now the craziest thing about all of this is that this isn't the only problem

that open ey are facing deep seek of course is coming in with cheaper prices and at the end of this interview one of the things that we can actually see he stated was the fact that the models are actually all consolidating and this is something that is evident if you pay attention to the Benchmark so recently there was a model called Deep seek R V3 and deep seek V3 actually got a significant Improvement so this is the artificial analysis lead leaderboard intelligence index for non reasoning models that do not think before answering and this index incorporates seven evaluations spanning reasoning knowledge math and coding and this shows us that at the time of recording this video it probably may change when you are watching this because the AI industry does move quite rapidly but deep seek V3 had an update to where it pushed it to be the state-of-the-art model in terms of the non reasoning model now this was I think around a week ago and since then several models have updated to surpass it just a little bit but the point is that this is a serious problem if you are opening eye from just purely a model standpoint because not only are they managing to out compete in terms of the non- reasoning model it's also probably something that is a lot less cheaper than the compe iors which of course is going to factor in into how people decide which models they use to deploy in various applications and API usage actually makes up a significant portion of opening eyes Revenue so it will be interesting to see you know in next year's financials how the company has managed to change that so I do think that this you know for me at least was probably bigger news than deeps V3 because this was you know better than Gro 3 DPT 4. 5 Gemini 2 3. 7 Sonic in terms of this AI Benchmark which is an independent evaluation now I personally don't use deep seek that much I've tried it and among many other chat Bots it does seem a bit generic for me I haven't managed to put it into any workflow but I do know many individuals that do actually use the model on a day-to-day basis now like I said before this model was more impressive than R1 because you can see right here when they spoke about this you know someone tweeted that this is arguably more impressive than deep seek R1 and potentially indicates that R2 is going to be a significant giant leap forward so the reason that this is so crazy is because a lot of people are going to think that you know if deep seek fe3 is this good what is R2 going to look like in terms of that base model and if you don't know already deep seek have discussed actually getting the base model out by April and they are really trying to rush those models out because they want to really dominate the AI space so they didn't realize that they were going to capture the AI mind share but now that they have they've spoken about releasing this model even sooner than they anticipated so if we're seeing jumps from Deep seek V3 all the way to the frontier some people argue that you know deep seek R2 and the next base models could be state-of-the-art in terms of the application so some people are stating that this deep seek V3 version is going to be the base for R1 and R2 when released could be better than o03 mini and other reasoning models that currently exist so it will be very interesting to see where this space is at the end of the year because with multiple contenders now trying to be the state of the art model across these benchmarks and dropping the prices of course that leads these companies to implementing some changes now like I said I spoke about this in yesterday's video opening ey didn't really abandon the AI race but they revealed some major changes with as to how they're structuring the company if you missed this video I'm going to give you guys a quick summary on what I said basically in this interview samman you know he acknowledges the fact that like there are tons and tons of models and they're all kind of consolidating around the same point the interviewer spoke about how LG released a new model and they spoke about there's a lot of state-of-the-art models so what are you going to do and he basically says look it's like the trans transistor when the transistor was you know initially created it was expensive only a few companies could make them and now they're everywhere pretty much in every phone and he says that it's probably going to be the same with llm so what they're focusing on now at openai is not really state-of-the-art models they of course will have state-of-the-art models so that they can deliver incredible products but there is a lot to be said about delivering a good customer experience take a look at what he says when discussing what will be more valuable in 5 years you can see it says

what is going to be more valuable in 5 years A 1 billion daily active user destination site that doesn't have to do customer acquisition or the state-ofthe-art model and Sam Alman says the 1 billion user site now you have to remember that openi they've shifted their focus if they were just a AI research lab that is focused on Frontier AI research then sure you could say that this is a big issue but they are now a tech company and tech companies have two main objectives number one to make money that's the bottom line growth and money that is the bottom line these companies want to grow year on year and of course number two they want to have as many users as possible so of course Frontier Air models I don't think openi would care that much if they had froner a models provided that they were making the line share of income and they had a billion users on their site every single day and I think over time we're probably going to see that is probably going to be a lot of other companies Focus too as they start to deploy these a systems but either way I think it is very interesting that he made this statement so if you guys did enjoy this video do not forget to leave a like subscribe all that good stuff and let me know what you think about the AI race are things shaping up as you thought or are things getting a bit too hectic