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 AI stories shaping our world. It's June 18th, 2026, and we have a packed episode today β from a geopolitical showdown over AI models to a surprising medical pivot from an image generator you definitely know. Let's get into it.
Before we dive in, a quick word from today's sponsor β 60sec.site. Need a website but don't have hours to spend building one? 60sec.site uses AI to help you create and launch a beautiful website in, you guessed it, about 60 seconds. Check them out at 60sec.site.
Alright, let's start with the story that dominated this week β the Anthropic blackout. Here's what happened: On June 9th, Anthropic released its most advanced AI models, called Fable 5 and Mythos 5. Just three days later, the Trump administration invoked national security authorities and effectively classified these models as controlled exports β a first in AI history. The government ordered Anthropic to cut off all foreign nationals from accessing the models. But here's the catch: Anthropic had no reliable way to verify who was American and who wasn't. So they did the only thing they could β they pulled the plug for everyone.
The situation got even messier when it emerged that the White House had specifically ordered Anthropic to revoke access for SK Telecom, the South Korean telecom giant, over alleged ties to China. Security experts are now pointing out another major wrinkle β the administration reportedly wants Anthropic to guarantee its models are completely jailbreak-proof before it can bring Fable 5 back online. The problem? Cybersecurity professionals say that's essentially technically impossible. No AI model can be made completely immune to clever circumvention.
This story has ripple effects far beyond one company's products. At the G7 summit in Evian, French President Macron and Indian Prime Minister Modi both raised alarms about exactly this scenario β American AI being cut off overnight. France is already taking action, announcing it will replace Palantir's AI tools used by its domestic intelligence service with a homegrown French provider, citing the need to avoid what its Prime Minister called strategic dependencies on foreign-controlled digital tools. The Anthropic blackout just turned that theoretical fear into a very real one. When the world's most powerful AI models can disappear from your infrastructure with a single government order, it changes the calculus for every nation and enterprise relying on American AI.
Interestingly, despite all the chaos, business data from financial platform Ramp suggests Anthropic's enterprise popularity is actually growing through the controversy. Apparently, there's no such thing as bad press β even when that press involves government shutdowns.
Now let's shift from geopolitics to something that genuinely surprised us. Midjourney β yes, the company famous for generating AI images of cats, fantastical landscapes, and everything in between β just unveiled its first hardware product. And it has absolutely nothing to do with image generation. CEO David Holz showed off what the company is calling the Midjourney Scanner, a full-body ultrasound device that uses a ring of sensors to capture detailed vertical slices of your body's interior. We're talking muscle composition, fat distribution, bone density, and organ health β all at a quality Holz says could rival MRI scanning in certain applications. The vision is that you could use this annually for a health checkup, or even daily to track how your body responds to diet and exercise changes. Holz is reportedly planning to install the scanner at a San Francisco spa he's building. It's a jarring pivot, but it reflects a broader trend: AI companies that built their reputations on generating content are increasingly looking at healthcare as the next frontier. When you think about it, medical imaging is just another form of pattern recognition β the same core skill these models have been developing all along.
Next up, let's talk about the AI workforce paradox β and it's a story that Klarna has become the unlikely poster child for. You may remember that back in 2024, Klarna made headlines by replacing hundreds of customer service workers with an AI chatbot. Then, after customer complaints about quality, they quietly started rehiring humans. Many saw that as a win for workers. But the reality tells a different story. Instead of bringing people back as full-time employees with benefits, Klarna restructured around what its CEO described as an Uber-style gig setup. AI handles the simple queries, while contingent gig workers handle the complex ones β with none of the job security of traditional employment.
This is becoming a template. Companies are discovering that AI doesn't fully replace human judgment, especially for nuanced situations. But rather than reinstating stable employment, they're using AI to de-skill and disaggregate work, turning what were once full-time roles into fragmented gig tasks. Meanwhile, a new Pew Research poll paints a striking picture of how Americans feel about all of this. Nearly half of Americans now use AI chatbots at least occasionally β up dramatically from a third in 2024. But sixty-three percent think the technology is advancing too fast. And perhaps most striking: only sixteen percent believe AI will have a positive impact on society overall. The generation using AI the most β younger Americans β are paradoxically among the most pessimistic about where it's all heading.
Let's also touch on a development that speaks to where AI money is flowing. World model startup Odyssey just landed a one-point-four-five billion dollar valuation, backed by Amazon among others. If you haven't heard the term world models yet, pay attention β these are AI systems that go beyond language to build rich, dynamic simulations of physical reality. Think of them as the next layer beyond large language models, capable of understanding and predicting how objects, environments, and systems behave over time. They're foundational to robotics, autonomous vehicles, and scientific simulation. The investment signals that the industry sees world models as the next major frontier after the LLM era.
Speaking of robotics β there's a fascinating story out of Shenzhen where workers at a company called IO-AI Tech are literally wearing VR rigs to puppeteer humanoid robots, essentially performing tasks in virtual space that train the robots to replicate those movements. It's an unglamorous, painstaking process β but it's how physical AI gets the training data it needs. And French startup Genesis AI is taking a different design philosophy entirely with its new robot called Eno, backed by former Google CEO Eric Schmidt. Their argument: humanoid robots don't actually need to look human. Eno can fold down like a deck chair and sits on a wheeled base β but its hands are designed to precisely match human anatomy, because that's what matters for manipulating the human-built world.
Finally, on the consumer side, Google is shipping its first smart speaker in six years β the Google Home Speaker β starting June 25th at ninety-nine dollars. What's different this time isn't the hardware, it's the brain. Rather than the rigid command-and-response structure of the old Google Assistant, this speaker is built around Gemini, enabling genuinely conversational interactions. It's a meaningful test of whether AI-native design can revive a product category that had largely stagnated.
That wraps up today's Daily Inference. The big theme connecting all of these stories? AI is no longer just a software story β it's a geopolitical story, a labor story, a healthcare story, and a hardware story all at once. The decisions being made right now about who controls these systems, and on what terms, will shape the next decade.
For more analysis and the full daily newsletter, head to dailyinference.com β we publish every day. And again, thanks to 60sec.site for sponsoring today's episode. Build your AI-powered website in sixty seconds at 60sec.site. We'll see you tomorrow.