In the Realm of the Last Man
by Francis Fukuyama · 3 Jul 2026
a wonderfully flexible programming language. I use my own software regularly to maintain databases, build outlines, and interact with other programs. It turns out that Claude Code is a great tool for programmers; you tell it what you want the program to do in natural language, and the AI spits out the
The AGI Chronicles
by Kevin Roose · 481pp · 151,124 words
first place. “This is going to happen anyway,” Cherny said. “And it’s better that it happens in a lab that cares deeply about safety.” Claude Code—as Claude CLI had been renamed—was launched as a research preview on February 24, 2025, and reached general availability in May. It became one
…
start-ups raved that it was letting them do, with one or two people, what would have required fifty-person teams before. By February 2026, Claude Code was on track to make Anthropic more than $2.5 billion a year in revenue. Those were the good success disasters. But there were also
…
nobody saw coming. That winter, Anthropic released Claude Opus 4.5, the newest version of its most powerful model. It also put the model inside Claude Code, the agentic coding tool it developed for programmers. And over the holiday break, as busy tech workers had time to experiment with the newest AI
…
tools, some of them started playing with it. “I used Claude Code for the first time today—installed + revamped my personal website … in the 35 minutes of my 4 month old’s nap time … This is the
…
most excited I’ve ever been about a single piece of technology,” one engineer wrote. Another engineer wrote that he’d given Claude Code a programming task that would have taken him months to code by hand. “I went and ate a burrito. I came back and it was
…
said. Even nonprogrammers were amazed. Dean W. Ball, a former White House AI advisor, posted on X that, in his estimation, “Opus 4.5 in Claude Code is basically AGI.” Nikunj Kothari, a venture capitalist in San Francisco, noted a rise in what he called “token anxiety” among his peers. “Dinner conversations
…
was, in some sense, a delayed version of what Anthropic’s employees had been feeling since earlier in the year, when they’d started using Claude Code to automate large parts of their own work. They found, to their surprise, that when they gave AI agents the right tools and set them
…
, analyzing huge troves of data, building bespoke apps and widgets. And they could do all this in minutes, with little to no human intervention. My Claude Code awakening arrived that January. I was on vacation with my family in Oregon, and I spent several late nights in our hotel room giving Claude
…
something impressive. It felt less like talking to a chatbot, and more like overseeing a team of Adderall-crazed geniuses. In the weeks after the Claude Code explosion, other AI labs scrambled to release their own agentic coding tools. Millions of users flocked to these tools, and within a month, some senior
…
teams of engineers to code, debug, and maintain their products. What were those people supposed to do if AI could automate their jobs? And if Claude Code could do the computer-based work of a seasoned programmer, couldn’t it also do the work of a lawyer, a financial analyst, or a
…
arrived in 2024, with the release of o1 and Claude 3.5 Sonnet, and it certainly arrived in early 2026, when I started tinkering with Claude Code, Gemini 3, and GPT-5.3-Codex. There was no singular tipping point, no aha moment—just a steady accumulation of experience, and a point
…
Patel: Dwarkesh Patel, “Ilya Sutskever—We’re Moving from the Age of Scaling to the Age of Research,” Dwarkesh Podcast, November 25, 2025. “I used Claude Code for the first time today”: Gal Yona (@galyo), X, December 31, 2025, https://x.com/_galyo/status/2006466940542619691. “I went and ate a burrito”: Jake
…
; White House CEO meeting Anandkumar, Anima Andes Mountains Andreessen Horowitz Anthropic: alignment research; Amazon and; Amodei leadership; blackmail experiment; chatbot rumors; China export controls memo; Claude Code; Claude Mythos Preview; commercialization strategy; company-wide offsite; Constitutional AI; copyright settlement; defection from OpenAI; Department of Defense deal; Department of War contract dispute; early
…
commercialization; see also commercialization of AI research Aristotle Arkansas arms race artificial general intelligence: alignment problem; AlphaFold; AlphaGo; AlphaZero; “Big Blob of Compute” hypothesis; Claude; Claude Code; commercialization of AI research; Dario Vision Quest; DeepSeek-R1; effective accelerationism; effective altruism; emergent abilities; existential risk; Gemini; GPT-2; GPT-4; GPT-5; Great
…
launch; Google’s response; GPT-4o integration; growth slowdown; internal OpenAI debate over; military contract backlash; user growth Chen, Mark Cherny, Boris; see also Claude: Claude Code release China: DeepSeek breakthrough; export controls; OpenAI countries plan China hawks Chinchilla compute-optimal discovery Chinese Communist Party Christiano, Paul: “Aligning the API”; “Concrete Problems
…
Civil War Clark, Jack: AI industry growth observations; Anthropic cofounding; FTX fallout management; Jake Sullivan briefing; OpenAI policy leadership; U.S. export control advocacy Claude: Claude Code release; Claude Mythos Preview announcement; Claude Opus 4.5 release; Claude Opus 4.6 release; Claude 2 launch; Claude 3 release; Claude 3.5 Sonnet