Something Big Is Happening

What It Covers

Matt Schumer's viral essay "Something Big Is Happening" argues AI has already transformed tech work and will soon disrupt all knowledge professions within one to five years. The piece sparked massive debate about AI capabilities, adoption timelines, and whether current models represent genuine productivity gains or elaborate productivity theater.

Key Questions Answered

• •Coding autonomy breakthrough: GPT-5.3 and Opus 4.5 released February 2025 enable developers to describe desired applications in plain English, then walk away for hours while AI writes tens of thousands of lines of code, tests functionality, iterates on design, and delivers production-ready work without human correction or guidance needed.
• •Self-improving AI acceleration: AI labs deliberately focused on coding capabilities first because building better AI requires code. Once AI writes code well enough to build the next version of itself, it creates a compounding improvement cycle. OpenAI confirmed GPT-5.3 was instrumental in creating itself, marking this inflection point as already achieved.
• •Early adopter advantage window: The single biggest career advantage available now is being early to adopt AI seriously. Professionals who demonstrate AI-enabled productivity gains like completing three-day analyses in one hour become the most valuable team members. This advantage disappears once everyone adopts, creating urgent pressure to act immediately rather than wait.
• •Adaptability over mastery: Building the habit of rapid adaptation matters more than mastering specific AI tools. Current models will become obsolete within a year, requiring continuous workflow rebuilding. Professionals who develop comfort with perpetual beginner status and experimental mindset will outperform those who perfect single-tool expertise in this environment of constant change.
• •Tool-shaped objects critique: Will Munidis argues many AI implementations produce work-shaped objects rather than actual value, with elaborate agent workflows generating memos nobody reads. However, this critique reveals more about the nature of modern knowledge work itself than AI limitations. The real question becomes whether AI eliminates genuinely valuable work or merely automates already-questionable corporate busywork.