🤖 Meta's Secret Model, UK's £8B AI Scandal & AI Weapons Hitting 850K Targets
Welcome to Daily Inference, your daily dose of the most important AI stories shaping our world. I'm glad you're here, because today we have a packed show covering everything from Meta's next big model to a Scottish infrastructure scandal that's raising serious questions about AI's promised future. Let's dive in.
Before we get started, a quick word from today's sponsor. If you've ever wanted to launch a website in under a minute, check out 60sec.site. It's an AI-powered tool that lets you build a beautiful, functional website in literally sixty seconds. Perfect for side projects, portfolios, or that business idea you've been sitting on. Head to 60sec.site and give it a try.
Alright, let's talk AI news.
First up, the AI model race is heating up again. Reports are circulating that Meta is developing a model internally codenamed Watermelon, and by all accounts it's being positioned to go head to head with OpenAI's GPT-5.5. We don't have all the technical specs yet, but the fact that Meta is naming and sizing up a competitor model so explicitly tells you something important about where the frontier AI race is right now. These companies are watching each other incredibly closely, and the gap between the leaders is narrowing. Meanwhile, on the developer tooling side, Cursor has launched a mobile app that lets you go from a screenshot of a bug straight to a fix. That's a genuinely interesting workflow shift — imagine catching a UI glitch on your phone and having it resolved before you even sit down at your desk.
Now let's talk about a story that deserves way more attention than it's getting. A Guardian investigation has pulled back the curtain on what may be the UK's most overpromised AI infrastructure project. The Lanarkshire datacentre complex in Scotland — an eight-point-two billion pound development by US firm CoreWeave and Scottish company DataVita — was announced in January with bold promises. It would be fully powered by on-site renewables. It would create thousands of jobs. It would be operational by 2030. Sounds great, right? Except the investigation found that government officials and developers privately acknowledged there was a power provision, quote, issue, with the site. The renewables promise appears to be unachievable. The jobs haven't materialized. And in the village of Newarthill, residents say they were misled — some were approached about selling their properties or giving up green belt land, all for a development that local people are now calling smoke and mirrors. This is part of a broader pattern with Britain's AI growth zones — a government initiative to attract massive datacentre investment. One source described some of these plans as, at best, unclear, and at worst, complete bunk. When you combine this with the separate story about OpenAI's Stargate UK project — where twenty billion of the touted thirty billion pound investment appears to have been largely hypothetical, and OpenAI reportedly never even visited one of the key proposed sites — you start to see a troubling trend. Governments are being handed ambitious AI investment announcements that look spectacular in a press release but don't always hold up under scrutiny. The infrastructure gap is real, and the energy demands of these facilities are enormous — we're talking about needing the equivalent output of a nuclear reactor to power a single major datacentre campus. That's not a detail you can paper over.
Sticking with the theme of AI and geopolitics, UK Foreign Secretary Yvette Cooper made headlines this week with a striking warning. She compared the potential risks of unregulated AI development to the scale of impact the Hiroshima bomb had on humanity. Strong words. She's calling for international agreements — involving the US, China, and other major powers — to establish global rules around AI development. Cooper believes AI governance will dominate foreign policy conversations for the next two years. And when you look at the other stories in today's news, it's hard to argue with her. We have an Israeli defense contractor, Elbit Systems, revealing that its AI-powered targeting system called Tzayad identified approximately eight hundred and fifty thousand targets across the Gaza and Lebanon conflicts between October 2023 and the end of 2025. That's roughly a thousand potential targets identified per day. Whether you view that through a military efficiency lens or an ethical one, the scale of AI-assisted targeting in active warfare is staggering, and it underscores exactly why Cooper and others are pushing for international frameworks before this technology proliferates further.
Let's shift to something that might hit closer to home for many of our listeners. The NHS in England is rolling out AI-powered triage on its app, with the goal of reaching around two hundred thousand patients over the next year. The system will assess your symptoms and direct you to the right service — whether that's a GP appointment, a pharmacist, or A&E. It's part of a ten billion pound overhaul of NHS systems. Now, here's an interesting parallel: across the world in Australia, the government is sounding the alarm about a different AI health application. AI scribes — tools that sit in on doctor-patient conversations and automatically generate clinical notes — have exploded in popularity among GPs over the past eighteen months. The Australian federal health department has raised privacy concerns, and regulators are now considering what safeguards need to be in place. Both stories point to the same underlying reality: AI is moving fast into healthcare, and the regulatory frameworks are genuinely struggling to keep pace.
Finally, let's end on a story that captures the cultural friction AI is creating right now. China is making significant advances in what researchers call embodied AI — specifically, the incredibly hard problem of building robotic hands that can actually replicate human dexterity. Tying a shoelace, buttoning a shirt — these seem simple to us but require extraordinarily complex neurological coordination. Chinese companies are reportedly closer than anyone else to cracking this. And if they do, it transforms humanoid robots from expensive novelties into genuinely useful tools. At the same time, the fanfiction community online is essentially at war with AI, with writers and readers trying to root out authors who are using tools like Claude or ChatGPT to generate stories. The detection methods being used are questionable at best, and there are real concerns about false positives ruining the reputations of human writers. It's a microcosm of a much bigger cultural debate about authenticity, creativity, and what we value in human-made work — a debate that Australian author Anna Funder put plainly this week when she wrote that her books were ingested into AI systems without her consent and without payment, and that US tech companies are essentially helping themselves to decades of creative output for free.
That's your Daily Inference for today, July 6th, 2026. The big themes tying everything together? Infrastructure promises versus reality. AI in warfare demanding urgent global governance. Healthcare AI moving faster than regulation. And creative communities fighting back. These aren't isolated stories — they're all facets of the same transformation happening in real time.
If you want to go deeper on any of these stories, head over to dailyinference.com for our daily newsletter with links, analysis, and more context. And once again, thanks to 60sec.site for sponsoring today's episode. Build your next website in sixty seconds flat. We'll see you tomorrow.
