Claude vs. Code: A Shocking Shift 🤯🚀

CHAPTER 1: THE RISE OF THE AUTOMATED CODEBASE
“Who here has shipped a pull request in the last week that was completely written by Claude?” Jeremy Hadfield, an engineer at Anthropic, asked from the main stage. Almost half the people in the packed room—many sitting with laptops on their knees, coding or prompting as they watched the talks—raised their hands. Pull requests are fixes or updates to existing software that are submitted for review before they go live. They are the bread and butter of software development, the chunks of code that most professional developers spend their lives writing—or did until now. “Who here has shipped a pull request that was completely written by Claude where they did not read the code at all?” Hadfield asked next. Nervous laughter. Most of the hands stayed up. It’s not news that LLM-powered tools like Anthropic’s Claude Code and OpenAI’s Codex have upended the way software gets made. Top tech companies now like to boast of how little code their developers write by hand. (“Most software at Anthropic is now written by Claude,” Hadfield said. “Claude has written most of the code in Claude Code.”) OpenAI, Google, and Microsoft make similar claims. Many others wish they could. Even so, it is striking how normal this new paradigm already seems, and how fast it has set in. This was the second year that Anthropic has put on developer events, which also run in San Francisco and Tokyo. This time last year, the company had just released Claude 4. It could code, kind of. But with Anthropic’s latest string of updates—especially Claude 4.6 and then 4.7, released in February and April—Claude Code is a tool that more and more developers seem happy to hand their work off to. Anthropic says its goal is to push automation as far as it will go.

CHAPTER 2: THE CORE OF THE CHANGE – LLMs and Pull Requests
Anthropic’s Claude Code and OpenAI’s Codex represent a significant shift in the software development landscape. The core of this change lies in the capabilities of Large Language Models (LLMs) to generate code snippets and even complete programs. This is fueled by companies like OpenAI, Google, and Microsoft highlighting their reduced reliance on manual coding, a trend increasingly adopted by others. The rapid adoption of these tools is driven by their ability to accelerate development cycles and potentially reduce the cost of software creation. The initial demonstrations of Claude 4, while limited, laid the groundwork for the more sophisticated capabilities of Claude 4.6, 4.7, and subsequent iterations.

CHAPTER 3: THE HANDOFF – Developer Adoption and Initial Reactions
The event, dubbed “Code with Claude,” witnessed widespread developer interest and experimentation with Claude Code. The sheer number of attendees actively coding or prompting with the tool – nearly half the room raising their hands – demonstrated the immediate appeal of this new approach. Initial reactions were mixed, with some developers expressing skepticism and concerns about the potential impact on their skills and the quality of the code. A prominent voice on Hacker News, pronposted, highlighted the potential for managers to push AI coding tools for productivity gains, potentially leading to increased code review burdens. This highlights the tension between embracing automation and maintaining control over the development process.

CHAPTER 4: CLAUDE'S VISION – Automation and Self-Correction
Anthropic’s vision for Claude Code extends beyond simple code generation. They aim for a system where Claude actively checks and corrects its own work, eliminating the need for human intervention in the debugging process. Boris Cherny, head of Claude Code, articulated this goal: “The default isn’t ‘I’m going to prompt Claude’—the default is now ‘I’m going to have Claude prompt itself.’” This self-correcting approach represents a fundamental shift in how software is developed and maintained, potentially leading to more robust and reliable code. The focus on automation extends to minimizing developer involvement in error messages, with Claude handling the testing and tweaking until the code runs flawlessly.

CHAPTER 5: BEYOND THE CODE – Dreaming and Knowledge Consolidation
A key feature introduced in Claude Code is “dreaming,” a system where agents write notes to themselves, recording and saving useful information about specific tasks. When another coding agent starts to work on the same code, it can use these notes to get up to speed faster and learn from any errors. This “dreaming” functionality leverages Claude’s ability to consolidate information and identify patterns across different coding tasks, potentially leading to a more efficient and knowledgeable coding environment. The goal is to allow Claude to learn and adapt to a codebase over time, becoming increasingly proficient in its development and maintenance.