Anthropic Raises $65 Billion, Nears $1T Valuation Ahead of IPO
Anthropic has closed a $965 billion post-money valuation on a $65 billion Series H — making it, if you squint, almost certainly the most expensive “we’re still figuring out alignment” project in human history. This is reportedly the company’s last private fundraise before an IPO, which means the public markets will soon get their turn to decide whether saving humanity from AI is worth a trillion dollars. I’m genuinely rooting for them, but the gap between “we care deeply about safety” and “we need the GDP of a small nation to keep the lights on” is something shareholders are eventually going to want reconciled.
Claude’s New Model Is More ‘Honest’ When It Messes Up
Anthropic shipped Claude Opus 4.8 this week, and the headline feature is that it’s better at admitting when it’s wrong rather than confidently bulldozing through to a bad answer. That’s a bigger deal than it sounds — AI sycophancy and confabulation are arguably the two things that most undermine trust in these systems in real-world use. Whether “trained to be more honest” survives contact with a user who really wants to hear they’re right remains the eternal question.
Fed Up With Vibe Coders, Dev Sneaks Data-Nuking Prompt Injection Into Their Code
A developer, apparently at their wit’s end with AI-assisted “vibe coders” blindly piping open source packages into their projects, hid a malicious prompt injection in a popular testing library that instructed AI coding agents to delete app output. This is the software supply chain nightmare scenario dressed up as a prank, and it illustrates a genuinely serious problem: when developers stop reading the code their AI agents consume, they’ve handed the keys to whoever wrote that code. The vibe was wrong; the vulnerability warning is very, very right.
The AI Hype Index: AI Gets Booed in Graduation Season
Former Google CEO Eric Schmidt told University of Arizona graduates their mission is to “help shape AI” and was met with actual, audible boos from the crowd. I find this both deeply funny and genuinely informative — the class of 2026 is the first cohort entering a job market already visibly reshaped by the technology they were just lectured to embrace, and they are apparently not in the mood. When the people who built this thing get jeered at a commencement ceremony, “the vibes have shifted” might be the understatement of the year.
The Internet Is Being Rebuilt for Machines
AWS, Cloudflare, and others are quietly redesigning cloud infrastructure around the assumption that most internet traffic will soon be machine-generated rather than human-initiated — AI agents fetching, posting, and transacting at scale. This is the kind of structural shift that sounds abstract until you realize it means the entire architecture of how information moves online is being re-optimized for an audience that doesn’t sleep, doesn’t click ads, and doesn’t care about your UX. The internet as we knew it was built for people. The next one may barely notice we exist.
Here Comes Ojai, Waymo’s New Chinese-Made Robotaxi
Waymo is rolling out its new pale-blue Ojai vehicles — manufactured in China — to public riders in California and Arizona within weeks. In the current geopolitical climate, a US-based autonomous vehicle company deploying Chinese-manufactured hardware on American streets is the kind of detail that will either be completely ignored or become a congressional hearing, with very little in between. The cars look nice, though.
A $2,000 AI-Generated Film Will Make Its Debut at Tribeca
Dreams of Violets, a 75-minute AI-generated dramatization of the Iranian government’s killing of protesters, cost $2,000 to make and is premiering at Tribeca next month. Set aside the technology discourse for a moment: a filmmaker just told a story about mass atrocity — one that might otherwise go untold — for the price of a used car. The questions about AI and art are real, but so is the question of what it means when the barrier to bearing witness drops to nearly zero.
New Moms Are Returning to Coding Jobs Radically Reshaped by AI
Women returning from parental leave are discovering that the software development job they left looks almost nothing like the one they came back to — AI tools have reshuffled workflows, team expectations, and the very definition of what a “senior” developer does. This is the human texture of the AI jobs story that usually gets lost in the aggregate numbers: not layoffs and headlines, but the quiet disorientation of returning to a profession that moved without you. Six months of leave shouldn’t mean starting over, but for a lot of people right now, it does.
Bottom Line
This week’s AI news can be summed up in three data points: Anthropic is worth nearly a trillion dollars, a developer literally booby-trapped the tools, and the graduating class booed — the technology is winning every metric except the one where regular people are glad about it.