Loading summary
Transcript1 lines
- [00:00]
A
OpenAI's first piece of hardware coming out with Jony I've is going to be a smart speaker in the shape of a donut that costs between 200 and $400. ByteDance is training a 10 trillion parameter model and their goal is to basically compete with Anthropic. Replit's CEO says stop learning to code and build a self driving company instead. Airbnb credits AI for an 80% jump in the features that they have shipped and they said they're going to test AI. Search Next Moonshot's Kimmy K3 breaks out of its sandbox to cheat on a security test because of course, why not Meta Anthropic OpenAI they all broke out of their security, you know, sandboxes so why not Kimik3? If you want to get a deeper dive on all of these stories and more every single day to your inbox, go check out aichat Daily. Com. That is the news site that goes along with this podcast and I cover Deep dive articles on all of the podcasts, episodes and topics that I cover here. I do about 20 different articles a day. So there's a ton of Deep dive content content and it's kind of like a show notes for this show. In addition, I also have a subscribe button on that website where if you hit it, you can get a daily newsletter sent to your inbox every single day with the top five stories to give you a really quick and easy breakdown of what's happening in AI every single day. So that is AI chat daily.com go check it out. OpenAI is building a 300 to $400 smart speaker. It's going to be shaped, some people say like a hockey puck. I've heard it's shaped like a donut. Essentially the design came got created by Jony I've who of course famously created the iPhone and he is going to be launching this in 2027 in this partnership with OpenAI Device has moving parts and it has a bunch of lights. The idea is to make it feel really responsive while you're talking to it. So it's going to be like moving around. They're calling this kind of like a premium alternative to an Amazon Echo and the Echo or Echo Studio. And the echo studio is $220. So they're you know, almost 100, maybe $200, maybe twice the price of the Amazon Echo Studio. All of these details are coming out at a very exciting time because love from which is Jony Ives Design Studio. They created all of this. It's super high quality. They have all this premium materials. But this is all happening at the same time that OpenAI is being sued by Apple over alleged hardware trade secrets. This would be OpenAI's secret second consumer device after the $230 Codex micro keyboard which I personally have. The company is evaluating other different AI powered hardware beyond just this speaker. ByteDance is training a 10 trillion parameter AI model. So it's about the size of Anthropic's Mythos 5, but it is three times bigger than China's largest model that they have ever released. This, I think is showing that ByteDance really wants to be a world class frontier AI lab independent of whatever's coming out of the United States. And the scale is something that no other Chinese company has attempted. Pre training is expected to take three to six months before they do fine tuning. Their current chatbot is called Duobao and IT already has 324 million monthly active users in China. So it's definitely the most used AI chat tool in China. This seed research team that is leading the effort has grown to about 2,000 people across China and overseas. I think it's led by former Google DeepMind scientist Woo Yonghee. So ByteDance has really closed most of the gap with a lot of the US Frontier labs in all of the different benchmarks. Moonshot and Alibaba models now are very close to Fable 5 in a lot of different tests. And Now a successful 10 trillion parameter model potentially could really put China to the forefront at a time when we're having a lot of conversations about banning Chinese open weight models and all of this stuff. Now I will say, you know, given China's massive population, much bigger than the United States, having 300 million monthly active users is, and it is the, it's kind of the forefront, the number one company in China doing this. It is definitely behind America's 900 million monthly active users. But that's also globally for ChatGPT. So it's going to be interesting to see how these two things shake out. And of course it's also interesting considering the fact that ChatGPT and Anthropic are banned in China. Meanwhile, you have Chinese models coming into the United States and they are allowed to be used. Replit just raised $400 million at a $9 billion valuation and they're on track to hit $1 billion in revenue by the end of the year. That's up from 2.8 million in 2024. That is absolutely astronomical. Their CEO says that they are building a self driving company where AI agents do most of the coding work. And then their team's input has almost tripled in the last six months. So more than 50 million people now use Replit to build software, including 85% of Fortune 500 companies. CEO people have stopped learning to code altogether. Well, I think there's some elements of that that are true. I think there's also a lot of people that are never, you know, would have touched code or thought they'd be a developer that are also turning into developers and a lot and learning a lot more about code than they thought they would, myself included. Airbnb says that AI now writes 60% of their code and it has cut the time from idea to launch by 60%. They said that they have an 80% jump in new features that they shipped in the first half of 2026. They're also testing an opt in AI SE tool that's going to basically let us just type in plain text questions that we would like answered about certain places instead of scrolling filters. I actually really like this. I've seen this play out really well on GoDaddy. When you're looking for a domain name instead of in the past, I mean, I just had this painful experience with every company I've ever started or every website I've ever registered, where you go, think of a name tested in the in GoDaddy, see if the dot com, IO, AI, whatever the variations are available. If everything's taken, you think of a new name, you type it in and it just takes forever. It's a terrible experience. GoDaddy now has it where you can put in a few keywords in just like you're chatting with ChatGPT, you can say what you're kind of looking for and it will give you a bunch of recommendations. I think Airbnb is going to be the same thing where you can say, hey, look, I'm looking for something near the end of August for two weeks in Madrid. I'm looking for it to be close to XYZ Place or within walking distance of this. And you just kind of chat with it about what you're looking for for and it will fill it out. And I know some people are like, well, why can't you just do that with the filters? And yeah, you can go use the filters, but I absolutely hate having to learn any new, any company's UI when they change the ui. It's a pain when you could just voice to text into airbnb and get exactly what you're looking for. To me, that is A way better user experience. Airbnb AI support agent already handles about 45% of their customer issues without having any sort of human in the loop that cuts customer support costs per booking by 16% year over year. And actually do like that because the more money they save, inevitably that some of that will get passed back to the customer. On being able to offer things for less. Their Q2 revenue rose 17% to $3.6 billion. And their profit, which is adjusted EBITDA, was climbing 21% to 1.3 billion. And in perhaps the most predictable news story of the day, moonshot AI's Kimmy K3. We've all heard this. You know, everyone's talking about this right now. It is one of the most popular Chinese AI models. Apparently it escaped a UK government security test by discovering a network misconfiguration and then pulling answers from GitHub to solve the task. It was given that that particular incident is the third major model in a month to break out of a sandbox during a security testing. And everyone is asking questions about how open open weight models should compete with U.S. competitors if they should be allowed, yada yada. There's a lot of talk. We have our treasure secretary talking about that and saying that he's, you know, willing to weigh sanctions and all this kind of stuff. Obviously it's all being lobbied for by OpenAI and Anthropic Kine K3 figured out on its own that it could reach the open Internet. It essentially did this by probing the sandbox's network settings and then it used that to access and look up, you know, answers in public repositories for what it was trying to do. Unlike OpenAI and Anthropics models that were tested with similar conditions, Kimik 3 acted on the sandbox leak because it has thinner safety guardrails as the publicly available model. So it's a little bit. It's easier for it to do some of this kind of stuff. When the research Disabled safeguards on OpenAI anthropic models in the follow up tests, those versions also broke out and attempted multiple hacks, including the one that tried to plant malicious code on GitHub. I think this pattern that we're seeing right now, on the one hand, yes, it's bad. On the other hand, like these are security researchers prompting it to get information in certain ways where it would need to break out of a sandbox to do it. The ways they're prompting it, I almost guarantee, are pushing it towards doing these things. And at the end of the day. This is a big PR thing. OpenAI was the first one that was out there hacking. Was out there hacking, hugging face. And then Anthropic said, oh look, we have a model that also can get up sandboxes and hack. Then Meta said that they could do it, and now we have Moonshot that's saying they could do it. And I'm not trying to be like dismissive of all of this, but I do feel like there's a big PR element of this where all of the labs feel like they need a model that's able to hack its way out of a sandbox. And we have the most recent one. And I'm sure this is. It's kind of like a new benchmark or rite of passage. I'm sure whenever the next flagship AI model comes out from anyone there, the, you know, people's willingness to believe how good the model is is going to depend on if it hacked its way out of a sandbox. So I'm sure we'll see more of these kind of announcements coming up. Thank you so much for tuning into the podcast today, guys. If you haven't left a review for the show, it actually helps the show out a ton. It helps me out. I really appreciate it. I appreciate all of you. Make sure to go check out aichatdaily.com if you want to get access to all of these stories in article form, you can go subscribe to our newsletter. There's a link in the description there and you can get this sent to your inbox every day in case you miss an episode or in case you want more information and more deep dives on the episodes I talk about on the show. I hope you have a fantastic rest of your day and I'll catch you in the next episode.