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 truly explosive news cycle to dig through β covering everything from AI on the battlefield to a surprising shakeup in the chatbot wars. Let's get into it.
Before we dive in, a quick word from today's sponsor: 60sec.site. Need a website fast? 60sec.site uses AI to build you a stunning, professional website in under a minute. No coding required. Check them out at 60sec.site.
Alright, let's start with the story dominating every tech and national security conversation right now: AI in warfare, and the corporate fallout that followed.
Over the past week, a dramatic series of events has fundamentally reshaped the relationship between AI companies and the US military. Here's the timeline: Anthropic, the maker of the AI model Claude, reportedly had its Claude system used by the US military in strikes against Iran β strikes that, according to experts, moved faster than quote, the speed of thought. The technology shortened what military planners call the kill chain β that's the sequence from identifying a target, through legal approval, all the way to the strike itself. AI reportedly compressed that entire process dramatically.
But here's where it gets complicated. The Pentagon then publicly blacklisted Anthropic, labeling the company a supply chain risk, apparently because Anthropic had drawn hard lines against two specific use cases: no lethal autonomous weapons and no mass surveillance of American citizens. The Trump administration reportedly called Anthropic a, quote, radical left AI company. Within hours, OpenAI stepped in, announcing a deal to supply AI to classified military networks. OpenAI CEO Sam Altman openly admitted the deal was, quote, definitely rushed, and that the optics don't look good. MIT Technology Review described OpenAI's arrangement as the very compromise Anthropic had feared.
What makes this even more tangled is that the US military reportedly continued using Claude in active operations even after Trump announced the severance of ties with Anthropic β illustrating just how deeply embedded these tools already are in military infrastructure. You can't just flip a switch and remove AI from operations it's already woven into.
The public reaction was swift and revealing. ChatGPT uninstalls surged by a staggering 295% following news of OpenAI's Pentagon deal. Meanwhile, Claude rocketed to the number one spot on Apple's App Store in the United States β actually dethroning ChatGPT β just one day after the Pentagon contract with OpenAI went public. Consumers, it seems, are voting with their downloads. And tech workers signed an open letter urging the Department of Defense to quietly withdraw the supply chain risk designation against Anthropic rather than escalate further.
This whole saga raises a question that TechCrunch put bluntly: does anyone actually have a good plan for how AI companies should work with the government? OpenAI is rapidly transitioning from a consumer startup into what is effectively national security infrastructure, and by the CEO's own words, the company seems to be improvising as it goes.
Now, interestingly, while all this controversy was swirling, Anthropic made a product move that looks strategically savvy in hindsight. The company upgraded Claude's memory feature, extending it to free-tier users and adding a dedicated tool to import data from rival chatbots like ChatGPT and Google's Gemini. The idea is simple but powerful: if you've spent months teaching your AI assistant who you are, what you like, and how you work, you shouldn't have to start over when you switch. Anthropic is making it frictionless to bring that history with you. Combined with the brand goodwill boost from the Pentagon saga, Claude is having quite a week.
Our second major story comes from Alibaba, which is quietly having a massive few days in AI development. The Chinese tech giant released not one but several significant open-source tools. First, there's OpenSandbox, a tool designed to give AI agents safe, isolated spaces to run code, browse the web, and train models β all without those actions bleeding into the real world or other systems. Think of it as a secure playpen for autonomous AI. It's released under the Apache 2.0 license, meaning anyone can use and build on it freely.
Alibaba also released Qwen 3.5 Small, a family of language models ranging from 0.8 billion to 9 billion parameters β and this is significant. The industry narrative has long been that bigger models equal smarter AI. Alibaba is pushing back on that, with a philosophy they're branding as more intelligence, less compute. These smaller models are built to run directly on devices β your phone, your laptop β rather than in massive data centers. That's a meaningful shift toward making capable AI accessible without constant cloud connectivity or enormous computing budgets.
And there's CoPaw, another Alibaba open-source release, which functions as a personal agent workstation for developers β helping them manage complex, multi-channel AI workflows and memory systems. Together, these three releases paint a picture of Alibaba positioning itself as a serious infrastructure provider for the agentic AI era, not just a model maker.
Story three: the AI coding tool Cursor has reportedly crossed two billion dollars in annualized revenue. And what makes that number staggering isn't the figure itself β it's the pace. According to Bloomberg sources, Cursor's revenue run rate doubled in just three months. The four-year-old startup has tapped into something real: developers who use AI assistance for coding aren't going back. This explosive growth is a signal that AI developer tools may be one of the fastest-growing and most defensible categories in the entire software market right now.
Fourth, a quieter but consequential legal development: the US Supreme Court declined to hear a case challenging the rule that AI-generated art cannot be copyrighted. The case was brought by computer scientist Stephen Thaler, whose algorithm created an image back in 2019. The Copyright Office rejected the copyright claim on the grounds that the work lacked human authorship, and now the Supreme Court has let that ruling stand. This matters enormously for artists, companies, and anyone building creative tools on AI β the legal status of AI-generated content remains firmly in the unprotectable category under current US law.
Finally, let's zoom out and look at infrastructure. Nvidia announced it's pouring four billion dollars into photonics technology β two billion each into companies called Lumentum and Coherent. Photonics uses light rather than electrical signals to move data, which means faster transfer speeds, better energy efficiency, and higher bandwidth in data centers. As AI models grow larger and more demanding, the physical plumbing of how data moves between chips becomes a critical bottleneck. Nvidia is betting that photonics is the solution, and four billion dollars is a very serious bet.
That bet connects to a broader tension we're seeing globally: data centers are proliferating at an extraordinary rate, from small-town America where communities are pushing back against Amazon's tax-exempt mega-campuses, to the Arctic Circle where operators are heading north chasing cheap renewable energy, to the UK where campaign groups are warning that new AI infrastructure could double national electricity demand. The physical footprint of the AI revolution is becoming impossible to ignore.
So stepping back, what's the throughline across today's stories? AI is no longer a technology that operates in a sandbox β and I mean that in every sense. It's in military operations, in courtrooms, in app stores, in rural Ohio, and near the Arctic. The decisions being made right now β about who gets to use these tools, under what conditions, and with what oversight β are going to define the next decade. And the companies navigating those decisions are making it up as they go.
That's your Daily Inference for March 3rd, 2026. If you want to go deeper on any of these stories, head to dailyinference.com for our daily AI newsletter β it lands in your inbox every morning with the most important developments, fully contextualized.
And again, thanks to today's sponsor 60sec.site β build your AI-powered website in under a minute at 60sec.site.
Stay curious, stay informed, and we'll see you tomorrow.