OpenAI Seems Like It's Imploding!

More and more every day, it's starting to feel like OpenAI might actually be imploding. I feel like almost every week I'm sharing some new drama that came out of OpenAI and this week it seems like there's just been a pile of more drama. So in this video, I want to quickly speed run you through where we're at right now and kind of how we got here. It all started back in November of 2023 when the OpenAI board booted Sam Altman from the company. At the time he was fired, the board consisted of Ilya Suskever, one of the co founders of OpenAI with Sam Altman, Greg Brockman, another co founder with Sam, Helen Toner, Tasha McCauley and David D', Angelo, who is most known for Quora and its chatbot po. When all of this went down, it came out that the people that voted to boot Sam Altman included Ilia, one of his co founders, as well as David, Helen and Tasha. At the time, it seemed like Greg Brockman, the other board member, was the only one on his side and he was just as blindsided by the firing as Sam was. Fast forward a few months later and there was some major brain drain from the company. In February of 2024, Andrej Karpathy decided to leave OpenAI. Now this is one of the pioneers in AI advancement. He was instrumental in the creation of convolutional neural networks or innovation image recognition. His technology was actually used in Tesla's self driving cars and things like that. And he was a prominent member of OpenAI. In February, he decided to step away from OpenAI to focus on his own projects. A month later, Logan Kilpatrick, the head of developer relations, the person that was kind of the face of OpenAI next to Sam Altman on places like Twitter, he decided to step away. And when he stepped away, he actually went and joined Google to to help with their developer relations for their AI stuff. Then not too long after that, Ilya Suskever, who again as we just mentioned, was part of the coup to boot out Sam Altman, decided that it was his turn to step away. Now Ilya was also one of the co founders. He had nothing but good things to say about OpenAI, as did Andre Karpathy, as did Logan Kilpatrick. They all just wanted to step away to focus on their own projects. But anybody who's been paying attention to the drama at OpenAI would know that Ilya probably had some weird tensions with Sam Altman and the company because, well, he was One of the people that actually voted to get rid of Sam Altman. But that was sort of the end of the non confrontational era of engineers and leads over at OpenAI. Just stepping away from the company because just a few days after Ilya left, Jan Laika left and well, he didn't step away on the greatest terms. In fact, he wrote up a giant Twitter thread that got over 6 million views here about all of the issues that he was seeing over at OpenAI. A few posts into his thread he starts talking about how he started disagreeing with OpenAI's leadership, about the company's priorities, how he believed that more bandwidth should be spent getting ready for the next generation of models on security monitoring, preparedness, safety, adversarial robustness, super alignment, etc. And he was concerned that these problems weren't going to be fixed. He goes on to say that over the past years safety culture and processes have taken a backseat to shiny products and sort of spilled the beans that he didn't think OpenAI was taking safety seriously. But here's the thing, a whole bunch of people were stepping away from OpenAI but nobody was talking negatively about OpenAI's practices until Yan came out and started talking negatively about them. Why is he the only one that had anything bad to say about the company? Well, it turns out that OpenAI made their employees sign an agreement that that they would not disparage the company if they decided to leave. Otherwise they would lose their vested equity. Although a lot of people seem to be leaving OpenAI, none of them were talking poorly about it, but that was because they were afraid that any equity that they had vested in their time working at OpenAI would just be taken away from them if they decided to speak out against OpenAI. OpenAI changed a bunch of these policies. They said that nobody was going to lose equity and they've started having people sign different agreements now that don't have these non disparagement agreements in them. And it seems like things sort of calm down a little bit over at OpenAI. Fast forward to today, August 5, 2024 and this article comes out over on the information trio of leaders leave OpenAI with a giant picture of Greg Brockman, the guy that was Sam Altman's only ally during the time he was getting fired, who was on the board. Of course Greg Brockman is also a co founder and was the President of OpenAI underneath Sam Altman. Him leaving seems like insane. Huge news. But to be honest, it's a little bit clickbaity now. Two people did leave OpenAI and it is a big deal. First, we have John Schulman here who was another co founder of OpenAI, and he decided to step away to go join a competitor over at Anthropic. He says, I've made the difficult decision to leave OpenAI. The choice stems from my desire to deepen my focus on AI alignment. So very similar reasons to why Jan decided to leave. He decided to pursue this goal at Anthropic, where he believes he can gain new perspectives and do research along people deeply engaged with the topics he's most interested in. He does go on to say that the company leaders have been very committed to investing in this area, but his decision is a personal one based on how he wanted to focus his efforts. So he's not trying to claim that OpenAI isn't focused on alignment. Sort of like Jan was. He pretty much had nothing bad to say about OpenAI and Sam Altman even replied to his tweet saying thank you for everything you've done for OpenAI, you're a brilliant researcher, et cetera, et cetera. We also learned from this article that Peter Dang, a product leader who joined the company last year after working at Meta, Uber and airtable, has also left today. This article here is really the only confirmation we have. He doesn't really do Twitter very much. I haven't seen anything public of him posting that he's stepping away, but he is another person that did leave today. But what about Greg Brockman? The article says a trio of leaders leave OpenAI and they're clearly talking about John Schulman, Peter Dang, and Greg Brockman himself. Here it acts like in past tense. He he used to work at OpenAI. Brockman held a critical role turning AI researchers breakthroughs into large scale AI models and products such as ChatGPT has also been important ally to Altman, including during Altman's fight to return to the company. But then there's one little sentence here at the very end. Brockman told staff he planned to return after his extended vacation. Brockman didn't leave. He's going on vacation. The rest of the article kind of goes on to break down all of this stuff that I just kind of told you. But the fact that they're talking about a trio of leaders leaving with a giant picture of Greg Brockman strongly implies that Greg Brockman just quit OpenAI. Here's Greg's actual statement over on X I'm taking a sabbatical through the end of the year. First time to relax since Co founding OpenAI nine years ago. The mission is far from complete. We still have safe AGI to build. He's not going anywhere. Now. Why he's stepping away for such a long time, that's unclear. I mean, we still have August, September, October, November, December. We still have five months left in the year. So that seems like a long time to be sort of stepping away and not involved in a company that you co found and were acting as president of. So it does seem kind of weird that he's stepping away for that long of a time. But people are making it out like he quit, he's done. Just like Ilya and Andre and Logan and Jan, he's gone, he's not coming back. But that does not appear to be the case. It sounds like he's probably just really, really burnt out. There's a lot of drama over at OpenAI, seemingly all the time. And it seems like OpenAI has sort of backed off on shipping as frequently as they used to. Now they are in this trend of announcing something really cool, getting people really excited about it, and then not actually releasing it for anybody to use except for a few, select few. So I think between the lawsuits and the drama and the lack of shipping and more and more people just sort of calling out OpenAI for these things, the dude just needs a break. I'm sure more details and more info will come out over the coming days about what actually is going on. But as of right now, in this moment, Greg Brockman has not left OpenAI. He's just going on a really long vacation, saying that this is all happening at a time where it seems like there's still a lot of turmoil over at OpenAI. For instance, there's been articles circulating in the past week about how OpenAI could be on the brink of bankruptcy in under 12 months with a projection of $5 billion in losses. Apparently the startup spends $7 billion on training its models and another 1.5 billion on staffing. So OpenAI is burning through cash right now. There's also a handful of lawsuits happening. A YouTuber just filed a class action suit over the fact that OpenAI is scraping transcripts from YouTube channels. Marques Brownlee has started to become more and more outspoken about how frustrated he is by this practice. All signs point to the fact that Sora was probably also trained on YouTube videos. Elon Musk, who recently sold OpenAI because he believed it moved too far from its mission, but then dropped the suit, has now resued OpenAI. So there's a brand new lawsuit against OpenAI alleging that the company breached its founding principles by shifting from a nonprofit to a for profit model, prioritizing commercial interest over the public good. And then there's also the case of OpenAI really, really buddying up with the US government. So OpenAI is endorsing a handful of Senate bills, for instance. They're endorsing the Future of AI Innovation act, which forms a new regulatory board called the United States AI Safety Institute. He's all for that because he probably wants to be on that board that gets to decide what constitutes safe AI and what doesn't. OpenAI also recently pledged to give any new models that they come out with to the US Government early. And there's also that thing where, you know, he hired a board member who is an ex NSA member. So Back in June, OpenAI appointed a retired U.S. army General Paul M. Nakasone, to board of directors. If we look at his bio here, in his career as an army officer, he was pivotal in the creation of the U.S. cyber Command. He was the longest serving leader of U.S. cybercom and also led the National Security Agency where he was charged with safeguarding the United States digital infrastructure and advancing the country's cyber defense capabilities. So they brought in a ex member of the NSA as one of the board members in OpenAI. So now we have a lot of government involvement, a lot of government hands messing with OpenAI. They get to see the models early. They have an NSA member on the board. They're trying to create a new regulatory board which Sam Altman is endorsing. I don't know what's going on. All I'm doing is putting a lot of puzzle pieces together and just sharing them with you so that you can make your own opinions about what is happening over at OpenAI. But we've seen a lot of people leave the company. We've seen multiple people leave and basically say that they don't believe OpenAI is focused on safety. They had these practices where members weren't allowed to say anything negative about OpenAI. If they left otherwise, they would lose their vested equity. We're losing more and more early members and co founders of OpenAI. It seems by the month they're under pressure from a whole bunch of lawsuits. They're burning $5 billion a year. They've buddied up with the US government and have started working much closer with them. And to top it all off, they keep on showing really interesting cool things and talking about things like AGI, but then never actually putting any of the stuff that they're showing off into people's hands while other companies just seem to be shipping right past them. We saw Sora, but we didn't get access to. That opened the door for Luma Dream Machine Runway Gen 3 Cling, all of these other tools to sort of step in and create something similar. We saw GPT4O, which looked really, really good. But now Google's model beats it. Claude's models beat it. There's other large language models out there that seem to be doing it better. Dall E3 was really, really good when it came out, but as pretty much everybody would probably agree on. Midjourney is better at most things. Stable diffusion and the new flux is better at most things. Ideogram is pretty good. And there's just more and more AI image generation models that are popping up that are doing better than OpenAI. Is OpenAI imploding? I don't know. There's definitely a lot of drama, there's definitely a lot of turmoil and they're definitely under a lot of pressure. So it's going to be interesting over the next few weeks or months to see how they deal with that pressure and how they deal with all this sort of brain drain that's happening over at OpenAI right now. It seems like a lot of people that have worked there are agreeing less and less with the decisions of the higher ups, and more and more people are getting vocal about it and even stepping away from the company to go work with other companies that they feel are doing a better job at the things that they joined OpenAI to do. It's all just fascinating to me. This is my reality tv. This is my drama fix. I love reading about the latest tech news, the latest AI news, keeping up with the drama and then turning around and trying to loop you in. It's just fun for me. I don't know anybody at OpenAI. I have no insider information. I probably don't even know what I'm talking about and just wasted your time. So I apologize for that. But I have fun watching this drama play out and turning around and sharing what I find interesting with you. So if you like stuff like this, maybe like this video, subscribe to this channel. I'll keep you in the loop with all the latest AI news. I'll make tutorials, show you where the latest cool tech and research is going to and just talk about really fun, nerdy stuff like AI and emerging tech. And again, if you like that stuff like this video, subscribe to this channel. I'll make sure more of this shows up for you. Really appreciate you. Thanks for nerding out with me. See you in the next one. Bye. Bye.