The Anti-Patterns — Slopapalooza

Headliner

Prompt & Pray

Vague prompts, accepted output, hoping the architecture is sane.

Read →

Headliner

The 9,000-Line Component

AI keeps adding to one file because nobody stopped to redesign.

Read →

Headliner

The Duct-Tape API

Endpoints added one at a time with no coherent model, naming, or errors.

Read →

Vibes-Only Migration

Framework, platform, or database swap with no real migration plan.

Read →

Screenshot-Driven Product Spec

“Make it look like this.” No behavior, data, permissions, or errors.

Read →

The Context Window Bonfire

A chat thread as the source of truth instead of specs or docs.

Read →

The Infinite Patch Spiral

Every AI fix creates two more bugs because the design is wrong.

Read →

The Fake Confidence Refactor

AI rewrites working code with great confidence and quietly breaks edges.

Read →

The Rubber-Stamp Stage

Twelve-hundred-line PRs, three-bullet AI summary, ship it.

Read →

Coverage Theater

Tests and code wrote each other. They agree. They’re both wrong.

Read →

The “It Worked Once” Deployment

The code ran locally once, so everyone pretends it’s production-ready.

Read →

Secrets on the Big Screen

API keys, prod credentials, and customer data pasted straight into the prompt.

Read →

Slopsquatting Festival

Hallucinated package names get squatted by attackers — and your AI keeps installing them.

Read →

Strangers With Capabilities

Installing unvetted third-party agent skills from a “marketplace” that nobody actually vetted.

Read →

The Autopilot With No Pilot

An agent takes irreversible, customer-facing actions on its own — no approval gate, no kill switch.

Read →

The Cloud Project Junk Drawer

AI jams tables, jobs, and permissions wherever the one-off task can finish.

Read →

The Uncommitted

Production tools that live nowhere — no repo, no pipeline, no source of truth. Often nobody realized code was even created.

Read →

The Agent That Should’ve Been a Cron Job

Renting a language model to run a deterministic job a script would do for almost nothing.

Read →

The Disappearing Apprenticeship

Teams stop building the engineering judgment AI assumes they already have.

Read →