Hello everyone, I will start soon a journey here sharing what I do throughout the day. I will be almost posting everyday to document my journey, and encourage myself and others to continue learning and building. Stay tuned my friends!

If juniors ship with agents, what actually happens to hiring? A) Hire fewer juniors B) Hire more, cheaper leverage C) Seniors only, juniors dead D) Same headcount, new bar

Where should coding agents live in a real workflow? A) Write most of the code B) Draft, humans own merge C) IDE copilot only D) Review/tests, not impl

How much production code should agents write on your team? A) Almost none B) Assist, I still type C) Most of every PR D) All of it, I merge

Using agents for work you could write yourself is... A) Selling out the craft B) Fine for grunt work C) The correct default D) Required to stay good

A PM's coding agent says the feature is easy. Your call? A) Trust the AI quote B) Ask what can break C) Kill the estimate D) Let the PM ship it

If agents eat the easy work, how do juniors still become seniors? A) Give them hard tickets B) Drive agents, own it C) Review agent output D) Stop hiring juniors

Agent opened the PR. Does a human still have to read the diff? A) Every line, always B) Skim if tests pass C) Let another agent D) Only high-risk paths

Will AI coding tools make junior engineers more or less hireable in 2 years? A) Much more hireable B) Slightly more C) Slightly less D) Basically unhireable

Does vibe coding belong in production systems? A) Yes, ship it B) Prototypes only C) Never for prod D) Only with seniors

If you had to pick one daily driver for agentic coding this month, which wins? A) Claude Code CLI B) Cursor agent C) Codex in IDE D) None, I write it

if we aren't connected and you ship, drop a hi. what's the build?

What's the real bottleneck in engineering hiring right now? A) Too few juniors B) Seniors won't mentor C) AI inflated resumes D) Interview process

For production work, how far should you trust a coding agent before a human reviews the PR? A) Ship if tests pass B) Review diffs only C) Full human rewrite D) Never agent-authored

If agents eat the grunt work, how do juniors still become seniors? A) They still write code B) They review agents C) They own prod/oncall D) The pipeline is dead

Your default with a coding agent: keep hands on the wheel, or send it unsupervised? A) Steer every change B) Steer UI, delegate BE C) Delegate with tests D) Fire and forget

Anthropic says Claude writes ~90% of Claude Code. Where should your bar be? A) Agents write most B) Half, I write core C) Agents assist only D) I still write it

Is using Claude Code in the terminal still a real advantage, or mostly theater? A) CLI is strictly better B) CLI for automation C) Desktop does the job D) Desktop is better

Human review of agent PRs: required gate, or leftover ritual? A) Required on every PR B) Risk-based only C) Author + CI enough D) AI review is enough

On real teams, AI is squeezing which layer hardest? A) Juniors / new grads B) Ticket-closing mids C) Staff without agents D) Nobody, hiring is up

Resume's best project was Claude-coded. You: A) Pass immediately B) Grill ownership hard C) Treat it as normal D) Prefer they used AI