Subscribe
Share
Share
Embed
Your Daily Dose of Artificial Intelligence
π§ From breakthroughs in machine learning to the latest AI tools transforming our world, AI Daily gives you quick, insightful updatesβevery single day. Whether you're a founder, developer, or just AI-curious, we break down the news and trends you actually need to know.
Welcome to Daily Inference, your daily dose of the most important developments in artificial intelligence. I'm your host, and today we have a packed episode covering IPO fever, Apple's major AI comeback, infrastructure innovation from under the sea, AI in the courtroom and hospital, and what Microsoft's AI chief really thinks about the future. Let's dive in.
Before we get started, a quick word from today's sponsor: 60sec.site. Need a website fast? 60sec.site uses AI to help you build one in, you guessed it, about sixty seconds. Check them out at 60sec.site.
Alright, let's start with the biggest financial story in AI right now. The IPO race is officially on. OpenAI has filed confidentially with the SEC to go public, following rival Anthropic which made the same move just about a week earlier. OpenAI's valuation is expected to exceed 850 billion dollars, which would make it one of the most valuable market listings in history. To put that in perspective, we're talking about a company that didn't even exist fifteen years ago potentially being worth nearly a trillion dollars on day one of trading. The filing is confidential for now, meaning financial details like executive pay and full revenue figures aren't public yet. OpenAI's blog post was refreshingly candid about the whole thing, essentially saying they expected it to leak anyway so they just announced it. They also noted timing isn't locked in, as some things are easier to do while still private. What's fascinating here is the competitive dynamic: both OpenAI and Anthropic going public in rapid succession signals the AI sector is entering a new phase of financial maturity. Add to that news that a senior OpenAI employee was quoted saying, quote, chat is dead, as the company reportedly continues building a so-called super app that goes well beyond conversation. The next phase of OpenAI isn't just a chatbot company, it's positioning to be a platform.
Now let's talk Apple, because WWDC 2026 was a major moment. After years of being criticized for falling behind in the AI race, Apple came out swinging with what it's calling Siri AI, described as an entirely new version of the assistant. This isn't just a software update, it's a philosophical rethink. Siri is getting its own dedicated app for the first time, a more expressive and customizable voice, and the ability to read what's on your screen and interact with your apps contextually. One practical demo showed pointing your iPhone camera at a restaurant bill, having Siri identify what you ordered, and splitting it with Apple Cash automatically. That's the kind of frictionless, real-world utility that AI has been promising but rarely delivering. Apple also announced a partnership with Google Gemini, deepening its AI bench, while waiving cloud API costs for smaller developers with fewer than two million app downloads, a smart move to bring independent creators into the Apple Intelligence ecosystem. Safari is getting AI-generated extensions, meaning you can describe what you want a browser extension to do and Apple Intelligence will build it for you on the spot. The vibe shift here is notable. After settling a 250 million dollar false advertising case tied to earlier misleading Siri demos, Apple clearly made sure this year's demos were grounded and functional. The slow-and-steady bet is starting to look pretty smart.
From the Pacific Ocean floor to the Yellow Sea, let's talk about AI infrastructure. China just activated the world's first wind-powered underwater data center off the coast of Shanghai. The Shanghai Lingang project, a collaboration between HiCloud Technology and China Communications Construction, has a 24 megawatt capacity and launched in May. Here's the clever part: underwater environments offer natural cooling, which is one of the biggest costs and environmental burdens for traditional land-based data centers. The facility uses significantly less power and water than its land-based equivalents. This is China's answer to the crushing energy demands of its domestic AI boom, and it's a genuinely innovative approach. Meanwhile, in the US, a Guardian analysis found that roughly two thirds of planned new data centers are being built in drought-stricken areas, creating a growing tension between AI's water appetite and already strained resources. The contrast couldn't be sharper: one country experimenting with sustainable underwater infrastructure, another plowing ahead into water-scarce regions. Microsoft's AI chief Mustafa Suleyman noted in a recent interview that Microsoft's new data centers are all liquid-cooled, using roughly a swimming pool's worth of water over six years, and are largely powered by renewables. The sustainability question around AI infrastructure is only going to get louder.
Now let's shift to AI in public institutions, because two stories out of the UK this week illustrate both the promise and the peril of deploying AI where the stakes are highest. First, the good news of sorts: Deputy Prime Minister David Lammy announced a trial of AI legal assistants in crown courts across England and Wales, aimed at chipping away at the massive backlog of court cases. The idea is that AI tools can handle administrative and research tasks, freeing up legal professionals to focus on judgment-heavy work. But lawyers and legal advocates were quick to push back, warning that AI cannot and should not be used as a substitute for adequate funding and enough court staff. It's a tool, not a solution to systemic underfunding. The second story is more sobering. The Medical Protection Society has warned UK ministers that under current law, doctors and the NHS can be sued for medical negligence if an AI diagnostic or treatment tool makes an error, even if the clinician followed the AI's guidance in good faith. This is a genuine legal gap: as AI gets embedded deeper into clinical workflows, the liability framework hasn't caught up. The MPS is calling for the law to be overhauled. These two stories together paint a picture of governments rushing to adopt AI for efficiency gains without fully thinking through accountability structures. Who's responsible when the algorithm gets it wrong? That question is going to define a lot of policy battles over the next few years.
Finally, let's zoom out for a moment, because Microsoft AI CEO Mustafa Suleyman gave a fascinating and wide-ranging interview this week that cuts to the heart of where this whole industry is heading. He pushed back hard on the idea that AI will eliminate jobs, drawing a careful distinction between automating specific tasks within a role versus eliminating the role entirely. He argues that freeing people from repetitive administrative work will actually make them more creative and productive. But he also acknowledged the public trust problem is real: polling shows that the more young people use AI, the more antipathy they have toward it. That's a genuinely troubling signal for an industry betting on consumer adoption. Suleyman also took a firm stance against Anthropic's approach to AI consciousness, arguing that treating AI models as potentially conscious entities is not just philosophically wrong but actively dangerous. His view: we want AI to be controllable, accountable tools that serve humanity, and building systems that have ideas about their own suffering moves in exactly the wrong direction. He also offered a crisp taxonomy worth knowing: AGI is when AI can match most humans at most tasks, superintelligence is when it surpasses humans and can discover genuinely new knowledge, and the singularity, which he thinks is decades away and a bit too sci-fi for his taste, is when AI can recursively self-improve without limit. Where we are right now sits somewhere before rung one.
The through-line connecting all of today's stories is this: AI is no longer a research curiosity. It's becoming financial infrastructure, legal infrastructure, medical infrastructure, and consumer infrastructure all at once. The question shifting from can we build this to should we deploy this, and who's responsible when it goes wrong, is the defining challenge of the next few years.
That's a wrap on today's Daily Inference. For more coverage like this delivered straight to your inbox every morning, visit dailyinference.com and subscribe to our daily AI newsletter. And don't forget to check out today's sponsor, 60sec.site, for AI-powered website creation that takes about a minute. We'll be back tomorrow with more. Stay curious.