RUMOR: Is ChatGPT About to Launch Search?

Today on the AI breakdown, we're asking what if AI fixes everything before that on the brief? Is ChatGPT about to announce a search competitor? The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to breakdown.network for more information about our YouTube, our Discord and our newsletter. Welcome back to the AI Breakdown Brief. All the AI headline news you need in around five minutes. We have a spicy little nugget to start us off today. The AI rumor mill believes that next week ChatGPT is going to release a search engine. Pete Huang from the neuron writes breaking OpenAI will soon release a search engine for ChatGPT. Expected launch date is May 9, right before Google's big conference. After teasing this project over the last months, OpenAI has created search.chatgpt.com this is expected to be something like Perplexity you or find more coming. So what is P referring to in terms of this idea that they have been teasing it? Well, on his interview with Lex Friedman Last month, OpenAI CEO Sam Altman talked about the idea of search explicitly. Altman said, if we can build a better search engine than Google, then we should. Google shows you 13 ads and 10 blue links, which is one way to find information. But the thing that is exciting to me is that maybe there is a much better way to help people find and act on information. Allman went on this the intersection of LLMs plus search. I don't think anyone has cracked the code on yet. I would love to go and do that. I think it would be cool. Andrew Gao points out that the search.chatgpt.com subdomain goes to a valid page, but other subdomains do not. While Google may be the target that Altman has in mind, another company that people think is in the crosshairs is Perplexity. In some ways, Perplexity is a wrapper on top of other LLMs. However, their specialization in search has created a user interface that has driven many people, myself included, to shift a huge amount of their particularly reach researching away from Google and towards them. However, as Altman has talked before about the idea that OpenAI is likely to steamroll some startups, not even meaning to. The question is whether Perplexity is going to be one of those casualties. Ben Tossel from Ben's Bytes writes, perplexity won't last. I use it every day. But ChatGPT search next week may mean I don't need to. There's going to be a lot more to say about this to the extent that ChatGPT search does come out, but we'll save it for then rather than getting too deep into speculation now. Now, the event that it seems like OpenAI may be trying to front run if they do release this next week is of course Google's I O conference. I O is Google's annual developer conference and it's happening May 14th and 15th. Google hasn't given too many hints about exactly what to expect from that. There is a keynote called what's new in Google AI and generative AI is, as CDNet puts it, sprinkled throughout other keynote descriptions as well. However, as they also point out, Google I O is known for surprises. Might they surprise announce Gemini 2.0? We don't have long to wait to find out. Speaking of big tech companies who we are eagerly awaiting to find out what they're going to do next with AI, Apple had its quarterly earnings call, which by the way, it had an earnings beat coming in above analyst estimates. But no one was really focused on that and everyone was focused on what their AI strategy is going to be. Tim Cook remains vague when it comes to what exactly Apple is going to do with AI, but he is certainly willing to hype it up a lot more than he has in the past. He said that Apple was well positioned and that there were big opportunities across Apple for generative AI. And he said that there would be big AI announcements in the quote coming weeks. Wedbush analyst Dan Ives described it as a drumroll moment and what it is leading up to. At least what we believe is some big announcements at the Worldwide Developer Conference, which happens in June. It also seems like it might be pretty important. IPhone sales were down 10% over the past three months and so AI on the iPhone might have more pressure than just Apple catching up with the Google's and OpenAI's of the world. And it might actually be mission critical to get people to keep buying their new phones, marques Brownlee joked on Twitter. On one hand, it seems like it's only a matter of time before Apple starts making major AI related moves around the iPhone and iOS and buries these AI in a box gadgets extremely quickly. On the other hand, have you used Siri lately? Finally, another big tech company made a bunch of announcements around AI as well. Microsoft released its first annual Responsible AI Transparency report from the Verge. Quote Microsoft says in the report that it created 30 responsible AI tools in the past year, grew its responsible AI teams, and required teams making generative AI applications to measure and map risks throughout the development cycle. They say they've given Azure AI customers access to tools to detect problematic content and that they've also expanded red teaming efforts. Ultimately, this is a self grading exercise, so make of it what you will. Another interesting story from Microsoft today though was that they have banned US police departments from using their Azure AI tools for facial recognition, TechCrunch writes. Language added Wednesday to the terms of Service for Azure OpenAI service more clearly prohibits integrations with Azure OpenAI service from being used by or for police departments for facial recognition in the U.S. this is an extremely controversial application of AI, something that is also banned in the EU AI Act. Finally, Microsoft announced a massive renewable deal, presumably to help power its AI initiatives. Microsoft signed a deal yesterday with Brookfield Asset Management to support the development of 10.5 gigawatts of renewable energy capacity around the world. 10.5 gigawatts is basically half of the solar and wind capacity that California has currently. Brookfield Asset Management said the deal is almost eight times bigger than the previous single largest corporate power purchase. One of the big arguments from people who think that AI might not be so bad for the environment is that the sheer quantity of energy that it's going to require is going to force innovation and consumption of alternative energy sources as well. So chalk this as evidence in favor of that argument. However, that is going to do it for today's AI breakdown brief Next up, the main AI breakdown. Today we are talking about an interesting thought exploration from an interesting source. Nick Bostrom, who's been one of the most influential people when it comes to the rise of the AI safety movement, has now written a book all about what happens if AI solves, well, all of our problems. A couple things I want to note before we get into this episode. We're going to be talking about Bostrom's new book, but it's important, if you do not know anything about Nick Bostrom, that he is not an uncomplicated figure. As the Guardian writes, fifteen months ago, Bostrom was forced to issue an apology for comments he'd made in a group email back in 1996 when he was a 23 year old postgraduate student at the London School of Economics. In the retrieved message, Bostrom used the N word and argued that white people were more intelligent than black people. The apology did little to placate Bostrom's critics, not least because he conspicuously failed to withdraw his central contention regarding race and intelligence and seemed to make a partial defense of eugenics. Although after an investigation, Oxford University did accept that Bostrom was not a racist the whole episode left a stain on the Institute's reputation at a time when issues of anti racism and decolonization have become critically important to many university departments. Now, as you can also see from this picture of Sam Bankman Fried here, Bostrom also has controversy by proxy through his and his Future of Humanity Institute's connection with the effective altruism movement, whose most famous member, Sam Bankman Fried, is now serving 25 years in jail for fraud. Just last month, on April 16, the future of Humanity Institute was quietly shut down. If you go on Twitter X right now and want to engage in the debate about Bostrom, there will be plenty of people who will take you up on that. I just want you to have the context before we get into exploring this new book that he's written. First, though, we have to understand why people are paying attention, and it's because of his book published in 2014 called Paths, Dangers and Strategies. As Perplexity sums up, Bostrom argues that if superintelligence is achieved, it could potentially lead to scenarios where AI could take over the world to achieve its goals, which might not align with human welfare. He emphasizes the difficulty in controlling such superintelligent entities once created, highlighting the existential risks they could pose. Some of the concepts that were introduced in the book include the orthogonality thesis, the idea that an AI's level of intelligence does not necessarily correlate with having benevolent goals, as well as existential risk. As Perplexity also points out, this book has been highly influential. It's been discussed by people like Elon Musk, Bill Gates. It is a big part of the AI x risk canon, and that's why people are interested in the fact that he's written a new book called Deep Life and Meaning in a Solved World. In this new book, instead of considering what happens if things go wrong, Bostrom considers what happens if things go right. If disease is eradicated, if scarcity is replaced by abundance, if there is truly a techno utopia. And certainly it seems like one of the first questions is is this Bostrom shifting his position on things or just engaging in another theoretical conversation? When Wired asked about why he switched from writing about the threat to considering this good future, Bostrom replied, the various things that could go wrong with the development of AI are now receiving a lot more attention. It's a big shift in the last 10 years. Now all the leading frontier AI labs have research groups trying to develop scalable alignment methods, and in the last couple of years we also see political leaders starting to pay attention to AI. There hasn't yet been a commensurate increase in depth and sophistication in terms of thinking of where these things go. If we don't fall into one of these pits. Thinking has been quite superficial on the topic. In other words, it seems that to the extent that there was a mission in writing superintelligence to provoke that conversation, it has done its job and so now Bostrom is on to the next exploration. Now also, if you're wondering if Bostrom comes to the conclusion that it would be awesome if all of our problems were solved, it's not quite that clear cut for him. In that same interview, he said, ultimately, I'm optimistic about what the outcome could be if things go well, but that's on the other side of a bunch of fairly deep reconsiderations of what human life could be and what has value. We could have this superintelligence and it could do everything. Then there are a lot of things that we no longer need to do, and it undermines a lot of what we currently think is the sort of be all and end all of human existence. Maybe there will also be digital minds as well that are part of this future. Overall, from this interview you get a sense that Bostrom just doesn't know. And in a world where people are increasingly dividing themselves up based on what they believe about the potential future paths, he's trying to get everyone to think more deeply about them, sort of from a first principles perspective, without trying to bring that baggage of what he thinks will happen. As he puts it, I've spent three decades thinking quite hard about these things, and I have a few views about specific things. But the overall message is that I still feel very in the dark. Maybe these other people have found some shortcut to bring insights. So what I think is interesting about this is that hold aside any sort of conclusions that Bostrom comes to one way or the other around the central contention that we need to have deeper, more complex conversations about these outcome possibilities, both the good and the bad. That's something that I unreservedly agree with. I think that the implications of AI, if they are bad, obviously create context for a very important societal conversation about how technology serves us, what we will and will not countenance, what we do and do not want. On the flip side, if things are good, I think that it will be good in a way that demands a total reevaluation of the social contract. It's not impossible that we will have to ask and reimagine what we think is reasonable for people to contribute in order to be a contributing member of society. The most obvious examples are in things like what the standard work week looks like, but it's going to go a lot deeper than that as well. Unfortunately, it's very, very difficult to have these conversations in advance of the world actually changing. We can only do so much imagining. I do think though, that overall it's healthy to see a bit of a recalibration where the explosion of attention around X risk and other even more pertinent AI risks when it comes to things like jobs, industrial discrimination and inequality that arose last year are now not going away. But because they are just part of the agenda and part of the larger conversation, there's also being space made for the opposite explorations about what it looks like when things go well. One of the big reasons that I spend much more time on the questions of what if things go well? Is that because I think that to some extent to make things go well, we're going to have to make specific types of decisions and by anchoring ourselves in the world that we actually want to create, becomes at least a little bit easier to assess actions in the here and now. Anyways, some heady stuff for a Friday episode. Hopefully you're listening to this over the weekend with a drink, looking at some beautiful May warm landscape. In any case, though, that is going to do it for today's AI breakdown. Until next time, peace.