They lied to us about AI
In a Nutshell
OpenAI is in chaos—hiring a Facebook manager for a "code red" on project overload, shipping distractions like Sora and a browser while losing to focused Anthropic in coding and enterprise. AI like LLMs boosts code volume by 41% but slows senior developers 19% due to trash output and review burdens, failing to replace jobs. AGI hype sells fear of imaginary superintelligence, while real tools can't handle proprietary work or learn on the job.
These notes were generated by AI and may contain inaccuracies.
In February 2026, an engineer at OpenAI (makers of ChatGPT) faces a mess: strategy in disarray, projects scattered, losing market share to Anthropic in coding and enterprise. Instead of building an AI agent or training a new model, they Google "good manager" and hire Fiji Simo from Facebook.
Last week, Fiji Simo declares a code red to the entire OpenAI company for having too many projects. Code red implies something extreme like AI becoming sentient and asking for equity, but it's just project overload. OpenAI's pitch is AI makes one person as productive as 10, yet they can't manage one company. They solve it with meetings, a lady with PowerPoint, and bad news instead of using their own AI tool.
I have never been more bullish on middle management in my entire life.
Sam Altman, OpenAI CEO, has been on a product rampage: shipped Sora (TikTok-like app for bot videos to share with humans), Atlas (web browser that MIT Technology Review called incomprehensible), bolted shopping on ChatGPT (hallucinating spending advice). Joked about launching an airline. He's distracted from preventing human extinction by building fake news video feeds.
Sora team crammed into research department, like putting marketing under legal (ads sound like cease-and-desist letters).
Anthropic (nerdy, fewer people/money/compute, boring products) focuses on coding tools and enterprise, skips browser/TikTok/image gen, kicks OpenAI's ass. Claude Code so popular developers binge-used it over Christmas (Wall Street Journal called it a "Claude bender").
If you're spending Christmas with an AI agent instead of your wife, that is not early adoption. That's a divorce.
AI isn't replacing workers or developers yet. Meter study: 16 senior developers on real codebases/tasks using AI were 19% slower. 41% of code now AI-written, but productivity isn't increasing—code goes to trash, burdens senior devs like Mike reviewing pull requests at 11 p.m., crashing out, quitting for a bar or golf sim.
AI is not replacing developers. It's creating a second job for them.
CEOs promising job replacement mean AGI (artificial general intelligence), not current AI like ChatGPT—AGI doesn't exist, no one demonstrated it's possible. Selling fear of unreal tech.
LLMs released, hyped as AI, but it's autocomplete: math function predicting next word from internet patterns, not thinking. Industry pivots: "Real AI is AGI later." Raised $300 billion calling fast generator "AI," then defer to future.
My girlfriend goes to a different school.
LLMs can't learn on the job, only generic tasks trained on internet data. Companies have proprietary, undocumented internal tools/systems requiring deep judgment/company context—not on internet.
The thing that could replace your job, AGI, is imaginary. And the thing that's real can't replace your job. And the entire industry is praying you don't notice this difference.
Keep Mo Bitar in your library
Save the videos and channels worth coming back to, and find them again in one place.
