Ai Will Take Our Jobs

Content interview preparation vs actual work O'REILLY Designing Data-Intensive Applications THE BIG IDEAS BEHIND RELIABLE, SCALABLE, AND MAINTAINABLE SYSTEMS Martin Kleppmann

The modern coding triangle of dependency! Students and ChatGPT walk hand-in-hand down the path of enlightenment (or cheating, depending on who you ask), while Stack Overflow watches from the shadows like a disappointed parent who knows they'll come crawling back eventually. Remember the good old days when we actually had to understand error messages? Now it's just "Hey ChatGPT, fix this garbage code" followed by "Actually, let me check Stack Overflow because this AI hallucinated a function that doesn't exist." The circle of developer life continues...

Content 0000000077770000 0000000000000000 000000000T071000 TOT000000TO7000T TOTOOTTITOOT000T 000TOOTOTOTOTOTO [andut] 'Te dss 'dqu) das) dou auf! ama лоШ ЛоШ ysnd Dos 45 (Ртол) итеш 74? арод // 09 == InduT) 1 (PTo)u aUF) I 'e [induri dsa 'doJ don dou au! duo (ЛоШ) ЛоШ usna эроз // і (I == indut) IT 3 (Ртол)итеш 74т (ртол)итеш 7ит. { эрод и/ } (I es andut) sI } (ртол)итеш 7ит эроз // { 3 (I E= andur) UT • (PIOn)uJeW Zug

Google engineer spends a year building distributed agent orchestrators, probably through countless architecture meetings, design docs, code reviews, and debugging sessions. Then Claude Code recreates it in an hour because someone finally knew how to describe what they actually wanted. The brutal truth: AI coding assistants are incredible when you already know the solution architecture. It's like having a junior dev who codes at 10x speed but needs crystal-clear requirements. The year-long project? That was figuring out what to build. The one-hour recreation? That was just typing it out with extra steps. Turns out the hard part of software engineering was never the coding—it was always the "what the hell are we actually building and why" part. AI just made that painfully obvious.

From debugging complex algorithms to flipping burgers at McCode's. The great tech replacement didn't quite pan out as expected, did it? After years of training AI to "automate all the things," it finally mastered the art of stealing your job while leaving you with an apron and a drive-thru window. The irony is delicious—much like the fries you're now serving to the engineers who built your digital replacement. At least you've still got job security... until they build a robot that can wave goodbye better than you can.

The brutal honesty of this hits harder than a production outage on Friday afternoon. Nobody's writing code from scratch anymore—not ChatGPT, not developers, not even that one colleague who claims they never use Stack Overflow. Modern programming is basically sophisticated copy-paste with extra steps. We're all just standing on the shoulders of GitHub repos and praying the dependencies don't break. The real innovation is knowing which code to steal and how to make it look like you didn't.