OpenAI's Biggest Fear Is Coming True (AGI BY 2027 But NOT by OpenAI)

There's a fresh player in the race to develop ultra intelligent AI safely, a startup called Safe Superintelligence. SSI, founded by some big names from OpenAI, who departed under controversial circumstances. So Ilya Sutskever, who you might know as one of the brains behind OpenAI, left the company just last month and now he's already jumping into a new venture, like almost immediately after leaving this new company. SSI is all about creating safe superintelligence, basically making super smart AI that won't turn against us humans. Sutskever made this announcement on Twitter, or X as it's called now after the whole Elon Musk thing, saying, and I quote, we will pursue safe superintelligence in a straight shot with one focus, one goal and one product. But it's not just a solo gig for Sutskever. He's teamed up with a couple of other heavy hitters in the AI world. There's Daniel Gross, who, who used to lead Apple's AI and search efforts, and Daniel Levy, another former OpenAI engineer. So it's like a little supergroup of AI experts, if you will. Together they're setting up shop in Palo Alto, California and Tel Aviv, Israel. I mean, it looks like these guys are serious about their mission to keep AI safe and secure, and they're not letting any short term business pressures get in the way, which is really cool. Now let me give you a little backstory on Sutskever's departure from OpenAI. It actually came after some pretty dramatic events. He was part of a group that tried to oust Sam Altman, the CEO of OpenAI, last November. Yeah, like an actual attempted coup or something. That whole episode caused quite a stir in the AI community, and Sutskever even publicly apologized later, saying he regretted his part in the attempt. It's clear that there were some deep disagreements about how to handle AI safety, which likely contributed to Sutskever and some others leaving the company. So now Sutskever's new venture, ssi, is all about making sure AI remains safe as it gets smarter and more advanced. The company's message is clear. Our singular focus means no distraction by management, overhead or product cycles. And our business model means safety, security and progress are all insulated from short term commercial pressures. But let's take a step back for a second and look at what happened with OpenAI's super alignment team, which Sutskever co led. This team was focused on steering and controlling AI systems to ensure they didn't go off the rails, which is obviously super important. But after Sutskever and his colleague Jan Leika left. The team was dissolved. Laika, by the way, went on to join another AI firm, Anthropic, which is also working on AI safety. In a blog post from 2023, Sutskever and Laika talked about the potential dangers of superintelligent AI and predicted that we could see AI with intelligence superior to humans within the decade. Can you imagine that? Like AI that's smarter than us humans, they emphasize the need for research to control and restrict such powerful AI systems to this has been a consistent theme in Sutskever's work, and it's clearly something he's passionate about. Alright, so now SSI's approach is to advance AI capabilities as quickly as possible while making sure safety measures stay ahead. They believe this strategy will allow them to scale their technology without running into major issues, which is obviously super important. Sutskever and his team are recruiting top talent for their offices in Palo Alto and Tel Aviv, aiming to bring together the best minds in the field to to tackle this huge challenge.

As researchers and engineers continue to work on AI, the day will come when the digital brains that live inside our computers will become as good and even better than our own biological brains.

There's an interesting bit about how SSI differs from OpenAI in terms of its business model. Unlike OpenAI, which started as a nonprofit and later restructured to accommodate the massive funding needs for its projects, SSI is being set up as a for profit entity from the get go. Daniel Gross, one of the co founders, mentioned that raising capital won't be an issue for them thanks to the booming interest in AI and the impressive credentials of their team. So they're not really worried about the money side of things. Alright, now let's talk about something that was revealed in a recent interview with an OpenAI employee. This interview was on the Hard Fork podcast by the New York Times and it featured Daniel Kokotello who who shared some pretty surprising insights about the internal workings of OpenAI. So get this. Kokatello mentioned that Microsoft, which has a big partnership with OpenAI, went ahead and deployed GPT4 in India without waiting for approval from the Internal Safety Board. This board was supposed to ensure that any major deployments of AI were safe and well considered, but apparently Microsoft just kind of jumped the gun, which caused a lot of disappointment and concern within OpenAI. The this incident shows that even big companies can sometimes bypass important safety protocols, which is pretty alarming if you ask me. Cocotelo also talked about the culture at OpenAI after the whole Sam Altman ouster attempt that I mentioned earlier. He said there was a lot of anger directed at the safety team, which included himself. People at OpenAI felt that the safety folks were slowing down progress and that the old board wasn't honest about why they wanted to fire Altman. This created a pretty hostile environment for those focused on AI safety, which is just not cool. This, this tension might have been a factor in why John Leake and some other researchers decided to leave OpenAI. You remember Jan Leake, right? The guy who went to Anthropic after the Super Alignment Team, which was crucial for AI safety, was dissolved. It's clear that there were significant internal conflicts about how to balance rapid AI development with ensuring safety, which is a tough balance to strike, for sure. Now, there was a pretty bold prediction that came out of this interview with Cocotylos. He mentioned that many OpenAI employees, including himself, believe that AGI could be here by 2027.

I think that publicly available information about the capabilities of this model and the progress that's been happening over the last four years is already enough to make extrapolations and say, holy cow, it seems like we could get to AGI, you know, in 2027 or so or sooner.

That's just a few years away, folks. This is a big deal because AGI would mean AI systems that are as smart as, or even smarter than humans. If this prediction holds true, it could transform our world in ways we can't even fully imagine. Yet. Kokotello pointed out that even publicly available information about AI's progress suggests we're on track for AGI by 2027. He said, you don't need any secret info to see this coming, which is pretty mind blowing if you think about it. This timeline is consistent with what other AI experts, including Sam Altman himself, have been saying. They've been predicting AGI within the next decade, and it looks like the consensus is leaning towards it happening sooner rather than later. So what does all this mean for the future? Right. Well, if we do see AGI by 2027, it's going to have a massive impact on everything from how we work and live to how we address global challenges. But it also means we need to be super careful about how we develop and deploy these powerful AI systems. That's where companies like SSI come in for focusing on making sure AI remains safe and beneficial for humanity, which is obviously super important. To wrap things up, let's reflect on the bigger picture here for a second. Ilya Sutskever's move to start SSI shows a strong commitment to AI safety. Despite all the drama at OpenAI, he's staying focused on what he believes is the most important challenge of our time creating safe superintelligence. His new company, with its clear mission and talented team is poor poised to make significant strides in this area, which is really exciting to think about. The insights from Daniel Cocatajlo's interview also highlight the complexities and tensions within the AI industry. You know, it's not just about making smart machines, it's about doing so responsibly and ethically, which is easier said than done. Of course, the next few years are going to be crucial in shaping the future of AI and it's going to be fascinating to see how it all unfolds. Alright folks, that's all for today's update. I I hope you found this as interesting as I did. If you did, make sure to smash that like button, hit subscribe and ring that notification bell so you don't miss any future updates. Thanks for watching and I'll see you in the next one.