shanecode.org / fadstack

Fadstack

A visual history of languages, architectures, and frameworks — what endured, what faded, and which fads burned real money.

Scroll to excavate decades

Catalogued

141

languages, frameworks, patterns

Stuck

69

still carrying real weight

Costly fads

20

invoice scars, not just memes

1970–1989

Foundations

Machines learn to share a language

Unix, C, and the first relational databases set the floor. Most of what still runs the world was sketched here — quietly, without a conference hashtag.

This era built the substrate. Process models, pipes, files-as-interfaces, and relational algebra were not fashion — they were constraints that happened to scale. C and Unix made portability a real goal; SQL made data queryable without rewriting the world. Later decades mostly decorate or redistribute these ideas. When something from this era still looks “boring,” that usually means it won.

Full era essay →

1990–1999

Object Wave

Everything became an object — until it didn't

Java, CORBA, and enterprise frameworks promised reuse at scale. Some of it stuck. A lot of it became consulting fuel.

Objects promised reuse; enterprises bought the promise in bulk. Java gave a portable runtime with real teeth. CORBA and UML-heavy processes gave vendors a decade of workshops. The survivors were the languages and platforms that outlived the methodology theater. The lesson — complexity sold as architecture burns cash — returns every decade with new branding.

Full era essay →

Language · 1995–now

Java

Stuck

Write once, run everywhere — then enterprise forever. Still the backbone of serious backend estates.

Language · 2000–now

C# / .NET

Stuck

Microsoft's managed stack grew up — cross-platform, open source, and still dominant in enterprise Windows shops.

Practice · 1995–now

Heavy SAP customization

Costly fad

Customize the ERP until upgrades are impossible. The enterprise tax that keeps consultants in villas.

$ ABAP extensions and bespoke modules that made every upgrade a multi-year program. Millions spent aligning ERP to process instead of process to ERP.

Language · 2004–now

Scala

Stuck

JVM language with functional teeth. Spark and Kafka culture kept it relevant long after the Twitter hype cooled.

Language · 1986–now

Erlang / OTP

Stuck

Let it crash, supervision trees, and nine-nines uptime. The telecom substrate that outlived every "modern" backend pitch.

Language · 1998–2010

Visual Basic 6

Stuck

Drag-and-drop business logic that powered SMB back offices worldwide. Microsoft killed it; the apps refused to die.

Framework · 1999–now

J2EE / EE frameworks

Mutated

EJB-heavy stacks taught a generation that XML config was architecture. Spring and lighter runtimes absorbed the survivors.

Practice · 1990–now

Big Design Up Front

Mutated

Spec first, code later. Agile didn't kill planning — it killed pretending the Gantt chart was reality.

Language · 1995–2015

Delphi / Object Pascal

Faded

RAD Windows client development before .NET. Still runs vertical-market apps whose source lives on a burned CD.

Architecture · 1991–2008

CORBA

Costly fad

Language-agnostic distributed objects. Complexity scaled faster than value for most shops.

$ Enterprises spent years on IDL, ORBs, and vendor lock-in for distributed objects that SOAP and REST later made look quaint — after the invoices cleared.

Practice · 1998–2014

XML-as-architecture

Faded

When every wiring decision lived in 4,000-line XML files. Configuration became a second programming language nobody wanted.

Practice · 1997–2012

UML Everywhere

Faded

Diagrams as deliverables. Useful as a sketch tool; catastrophic when treated as source of truth.

2000–2009

Web Boom

The browser ate the desktop

LAMP, Ajax, Rails, and the first serious front-end stacks. Shipping speed became a competitive weapon — and a career treadmill.

The web became the default delivery surface. LAMP and Rails made shipping cheap; Ajax made pages feel like apps; SOAP/SOA tried to enterprise-ify the wire. REST won by being debuggable with curl. This decade invented the modern product cadence — and the habit of treating every new framework as destiny.

Full era essay →

Language · 1991–now

Python

Stuck

From scripting glue to the default language of data and AI. Readability as a growth strategy.

Architecture · 2000–now

REST / HTTP APIs

Stuck

Resources over RPC theater. Won the API wars by being boring enough to debug with curl.

Platform · 1991–now

HTTP / HTTPS

Stuck

The application protocol of the internet. HTTP/2 and HTTP/3 are evolutions, not replacements.

Platform · 2005–now

Git

Stuck

Distributed version control that ate the industry. Merge pain remains; centralized SVN does not.

Framework · 2005–now

Django

Stuck

Batteries-included Python web. Still the adult choice when you want an admin, auth, and ORM that behave.

Architecture · 2004–now

MVC web frameworks

Stuck

Model-view-controller as the default web shape. Still how most server-rendered apps think.

Framework · 2003–now

Spring

Stuck

Java's practical operating system for apps. DI and Boot made enterprise Java shippable again.

Architecture · 1998–now

LAMP

Mutated

Linux, Apache, MySQL, PHP — the cheap, fast path to a website. Mutated into countless managed stacks.

Language · 1995–now

PHP

Stuck

Dismissed by snobs, powering WordPress and half the web. Ugly pragmatism beats elegant irrelevance.

Framework · 2004–now

Ruby on Rails

Stuck

Convention over configuration. Taught a decade of frameworks how to ship MVPs — and how to hit scaling walls.

Framework · 2010–now

Flask

Stuck

Small Python core, extensions for everything else. The pragmatic counterweight to Django's batteries.

Architecture · 2005–now

Ajax SPAs (early)

Mutated

Partial page updates without full reloads. Became the mental model for every modern front-end framework.

Practice · 2001–now

Agile (as practiced)

Mutated

Individuals and interactions — then SAFe slideware. The manifesto stuck; the ceremony industrial complex is the fad layer.

Framework · 2011–now

Bootstrap

Mutated

Opinionated CSS that made every site look like a startup for a while. Still the fastest path to a decent admin UI.

Language · 1987–2010

Perl

Faded

The duct tape of the early web and sysadmin era. Regex culture in a language that read like line noise.

Framework · 2007–now

Sinatra

Faded

Minimal Ruby routing for apps that did not need all of Rails. DSL elegance in a few hundred lines.

Framework · 2006–now

jQuery

Faded

DOM peace treaty for a fractured browser world. Mission accomplished — then frameworks ate its job.

Platform · 1995–2015

ColdFusion

Faded

Tag-based rapid web apps in CFML. Government and agency sites that outlived every contractor who knew the syntax.

Architecture · 2000–2016

SOAP / WS-*

Costly fad

Enterprise web services with ceremony. Interoperability theater with XML envelopes.

$ Banks and governments sunk fortunes into WS-Security, WSDL toolchains, and vendor stacks — then quietly rebuilt on REST/JSON while still paying maintenance on the old bus.

Architecture · 2003–2015

Enterprise SOA

Costly fad

Reuse via shared services — in theory. In practice: central buses and political ownership fights.

$ ESBs, governance boards, and "service portfolios" burned seven-figure programs. Many orgs got distributed monoliths with worse latency and a permanent integration team.

Framework · 2010–2016

Backbone.js

Faded

Minimal structure for early client MVC. Taught the browser to think in models before React rewrote the rules.

Framework · 2010–2016

Knockout.js

Faded

Observable view models for declarative UI. MVVM before the term became enterprise cosplay.

Framework · 2006–2017

Google Web Toolkit (GWT)

Faded

Write Java, compile to JavaScript. Enterprise AJAX before anyone admitted they hated JavaScript.

Framework · 2007–2018

Ext JS / Sencha

Mutated

Enterprise desktop UI in the browser — grids, trees, and megabytes of widget chrome.

Architecture · 2002–2016

Enterprise Service Bus (ESB)

Costly fad

Message bus in the middle of everything. SOA's plumbing invoice with vendor licensing.

$ Middleware vendors sold central buses that became single points of failure and political bottlenecks. Integration teams grew; end-to-end latency did too.

Platform · 1996–2020

Adobe Flash

Faded

Animations, games, and video in a plugin the web learned to hate. Steve Jobs wrote the obituary; HTML5 dug the grave.

Framework · 2007–2013

Silverlight

Costly fad

.NET in the browser via plugin. Netflix used it; the web moved on anyway.

$ Enterprises built internal dashboards and streaming pilots on Silverlight just as HTML5 and iOS said no to plugins. Migration projects burned budgets Flash already warned them about.

2010–2016

Cloud Scale

Split everything, then pay for the glue

Microservices, NoSQL, containers, and mobile-first. The patterns that fit Netflix were sold to every mid-market CRUD app.

Hyperscale patterns escaped the companies that needed them. Containers unified packaging; Kubernetes became the cloud OS; microservices and NoSQL were sold as defaults. Mobile-first stuck because screens changed. Cargo-cult distributed systems stuck around as invoices. The durable move was packaging and ops maturity — not rewriting every app into a mesh.

Full era essay →

Platform · 2013–now

Docker / Containers

Stuck

Ship the environment with the code. The packaging unit of modern compute.

Practice · 2010–now

Mobile-first / Responsive

Stuck

Design for the small screen first. Became table stakes, not a differentiator.

Architecture · 1970–now

Modular monolith

Stuck

One deployable with clear boundaries. The unfashionable default that keeps shipping while microservices argue about ownership.

Platform · 2014–now

Kubernetes

Stuck

Container orchestration that became the cloud's OS. Overkill for many; unavoidable for many others.

Platform · 2014–now

Apache Spark

Stuck

In-memory data processing that made Hadoop feel overnight. Still the batch/stream workhorse under many lakehouses.

Platform · 2011–now

Apache Kafka

Stuck

Durable log as nervous system. Event-driven architecture with real operational teeth.

Platform · 2009–now

Redis

Stuck

Fast structures in memory with persistence options. Cache, queue, session store — pick your footgun.

Framework · 2013–now

React

Stuck

Components and unidirectional data. Won the front-end by being good enough and having a gravitational ecosystem.

Language · 2009–now

Go

Stuck

Small language spec, fast compiles, great concurrency story. Cloud infra's favorite hammer.

Platform · 2014–now

Terraform

Stuck

Declarative infra with state files everyone learns to fear. The closest thing to versioned datacenters.

Language · 2014–now

Swift

Stuck

Apple's successor to Objective-C. The price of admission to serious iOS and macOS development.

Language · 2011–now

Kotlin

Stuck

Java without the boilerplate ceremony. Google's Android pick and Spring's polite nod kept it climbing.

Platform · 2012–now

Webpack

Mutated

The bundler that taught JavaScript developers what suffering means. Still under legacy apps; Vite owns greenfield.

Platform · 2012–now

Ansible

Stuck

YAML over SSH that sysadmins could actually read. The config-management winner by approachability.

Platform · 2009–now

MongoDB

Mutated

Document DB that became a punchline, then a grown-up product. Survived by getting serious about transactions and ops.

Platform · 2010–now

Elasticsearch

Stuck

Full-text search and log analytics at scale. Powerful, hungry, and everywhere in observability stacks.

Framework · 2014–now

Vue

Stuck

Progressive framework with a friendlier learning curve. Steady share without the cult volume.

Framework · 2016–now

Svelte / SvelteKit

Stuck

Compile away the framework. Less runtime theater, more shipped HTML — including this site.

Framework · 2010–now

Express.js

Stuck

Thin routing layer on Node that became the default API sketchpad. Middleware culture in a single file.

Architecture · 2015–now

gRPC

Mutated

Binary RPC with contracts. Fast internal glue that should never face the public internet without a gateway.

Language · 2011–now

Elixir

Stuck

BEAM reliability with modern syntax. Phoenix and LiveView keep the torch for pragmatic functional web.

Platform · 2015–now

Bazel

Stuck

Hermetic builds for monorepos that hurt until they save you. Google brainworm or serious reproducibility — pick one.

Framework · 2014–now

Phoenix / LiveView

Stuck

Real-time Elixir on the BEAM with HTML-over-the-wire before it was cool. Small but fiercely loyal community.

Platform · 2008–now

Apache Cassandra

Stuck

Wide-column store built for Netflix-scale write throughput. Resume-driven clusters optional.

Platform · 2007–now

Neo4j

Stuck

Property graph database that finally made graph queries a product. Niche that pays when the graph is the domain.

Architecture · 2012–now

Microservices Everywhere

Costly fad

Independent deployability as religion. Correct for some; ruinous cargo-cult for most.

$ Teams split 4-person products into 40 services, then hired SREs to babysit the mesh. Many later consolidated — after years of distributed tracing invoices and on-call trauma. See case studies on premature microservices and big-bang cutovers for named failure modes.

Practice · 2014–now

SPA for Every Page

Costly fad

Client-side apps as the only architecture. Great for apps; absurd for brochures.

$ Marketing sites and docs portals shipped multi-megabyte JS bundles for content that wanted HTML. SEO and performance teams spent years unwinding "modern" defaults.

Practice · 2005–now

ORM-as-database

Mutated

Hide SQL until N+1 queries own your night. Good ORMs survived; "never write SQL" didn't.

Platform · 2011–now

Jenkins

Mutated

Plugin palace CI that defined pipeline culture — and Groovy nightmares. Still running somewhere in your org.

Framework · 2011–2020

Ember.js

Faded

Convention-over-configuration SPA framework. LinkedIn and Discourse kept the faith; everyone else went React.

Framework · 2013–now

Ionic

Mutated

Mobile UI kit on Cordova then Capacitor. Still ships hybrid apps for teams that prioritize one codebase over polish.

Platform · 2007–2023

Heroku

Faded

git push deploy and twelve-factor bliss. The PaaS that taught a generation what deploy should feel like — then priced them out.

$ Free-tier sunsetting and Salesforce neglect sent hobbyists and startups shopping mid-project. Dyno pricing that made sense at demo scale became a line item at traction.

Practice · 2011–now

SAFe / scaled agile industrial complex

Costly fad

Agile at enterprise scale with certifications and train metaphors. Process theater with a subscription model.

$ Certified release trains, PI planning offsites, and agile coaches billing seven figures to wrap waterfall in sticky notes. Velocity went up on slides; cycle time often did not.

Practice · 2009–2019

NoSQL for Everything

Costly fad

Schema-optional databases sold as a lifestyle. Right tool for some jobs; default for none.

$ Startups ditched Postgres for document stores "because scale," then rebuilt relational integrity in application code — and paid twice when joins came back as a product requirement.

Platform · 2009–2020

Apache Mesos

Faded

Datacenter kernel before Kubernetes won the orchestration war. Twitter survived on it; most did not need it.

Platform · 2006–2018

Hadoop Everywhere

Costly fad

MapReduce as lifestyle. The elephant in rooms that needed a spreadsheet.

$ Enterprises stood up Hadoop clusters for gigabytes of data that fit on one Postgres instance. Hadoop admins, ZooKeeper nightmares, and ETL rewrites burned millions before Spark and cloud warehouses retired the elephant.

Framework · 2010–2021

AngularJS (1.x)

Faded

Two-way binding magic that became a migration tax. Angular 2+ was a rewrite; the 1.x apps lingered for years.

Framework · 2012–2018

Meteor

Faded

Full-stack JS with live data sync before anyone agreed on modules. Magic demo, fragile production.

Framework · 2011–2022

Xamarin

Faded

C# for iOS and Android before Flutter and React Native won the cross-platform narrative.

Framework · 2013–2020

Polymer / Web Components hype

Faded

Google's web components push before the platform was ready. Shadow DOM slides; React won the component war.

Framework · 2009–2019

Apache Cordova / PhoneGap

Mutated

HTML in a native WebView with camera plugins. The first cross-platform mobile bet before React Native ate lunch.

$ Teams shipped WebView shells as "native apps," then paid twice for performance fixes, plugin rot, and App Store rejections when users noticed the jank.

Platform · 2005–2018

CouchDB

Faded

Document store with replication and offline sync as the headline. MongoDB took the NoSQL spotlight.

Platform · 2009–2020

Chef

Faded

Ruby DSL infrastructure as code with cookbooks and a community site. Ansible ate the simpler path.

Platform · 2005–2019

Puppet

Faded

Declarative config management with a heavy enterprise sales motion. Still haunts datacenters that virtualized early.

Practice · 2010–2020

CQRS everywhere

Costly fad

Command/query separation as default architecture. Right for high-scale event domains; absurd for to-do lists.

$ Teams split read and write models for CRUD apps that never needed separate projections. Double the schemas, double the bugs, double the on-call surface — for a blog admin.

Practice · 2011–now

Event sourcing everywhere

Costly fad

Store every state change as an eternal log. Audit gold for some domains; operational nightmare as a default.

$ Immutable event logs for domains that needed UPDATE statements. Replay debugging sounded elegant until production needed GDPR deletes and schema migrations on millions of events.

Architecture · 2017–2024

Service mesh everywhere

Costly fad

Sidecar proxies for every microservice. Netflix problems with startup headcount.

$ Istio sidecars on three-service startups burned SRE headcount on mTLS ceremonies nobody could debug. Many teams ripped mesh out before the observability dashboard paid rent.

Platform · 2012–2018

Grunt

Faded

The first JavaScript task runner that made config files a lifestyle. Gulp and webpack ate its lunch.

Platform · 2013–2019

Gulp

Faded

Streaming build pipelines with code-over-config aesthetics. Briefly cool; webpack became the sun.

Platform · 2012–2017

Bower

Faded

Front-end package manager before npm figured out flat dependency trees. jQuery plugins as a service.

Platform · 2009–2018

Riak

Faded

Dynamo-inspired KV store from Basho. Bet on eventual consistency; lost to operational simplicity elsewhere.

Platform · 2011–2017

Parse

Faded

BaaS before Firebase owned the category. Open-sourced on the way out; trust never recovered.

$ Facebook shut down Parse with a year's notice, forcing thousands of mobile backends into emergency migrations or self-hosted forks.

2017–2021

Platform Wars

Frameworks fought for the runtime

React vs the world, serverless, Jamstack, blockchain-for-everything. Some platforms matured. Others left invoice scars.

Front-end frameworks consolidated power; TypeScript became the floor; serverless and Jamstack rebranded old ideas with better CDNs. Blockchain-for-CRUD and big-bang rewrites collected the decade’s dumbest budgets. The winners were boring platforms with ecosystems. The losers were narratives that skipped product-market fit.

Full era essay →

Language · 2012–now

TypeScript

Stuck

JS with a typechecker that conquered the ecosystem. The rare fad that became the floor.

Language · 1995–now

JavaScript

Stuck

The only language that runs in every browser. Accidental monopoly turned deliberate platform.

Platform · 2009–now

Node.js

Stuck

JS on the server. Unified the stack and created a package ecosystem that is both miracle and hazard.

Language · 2015–now

Rust

Stuck

Memory safety without GC pauses. Climbing from systems niches into cloud infra and WASM.

Framework · 2016–now

Next.js

Stuck

React with routing, SSR, and deployment opinions. The default answer when someone says "we need a web app."

Framework · 2017–now

Tailwind CSS

Stuck

Utility classes in markup upset purists and accelerated shipping. Design systems without meeting paralysis.

Framework · 2018–now

FastAPI

Stuck

Python APIs with types and OpenAPI for free. The default backend for ML demos that became the default for a lot else.

Framework · 2017–now

NestJS

Stuck

Angular-style structure for Node backends. Enterprise comfort food with decorators and modules.

Platform · 2020–now

Vite

Stuck

ESM-native dev server that made webpack feel instantaneously obsolete for new projects.

Architecture · 2015–now

Serverless-first

Mutated

Functions as the unit of deploy. Brilliant for spikes and glue; awkward as a whole product architecture.

Architecture · 2016–now

Jamstack

Mutated

Pre-render + CDN + APIs. The brand cooled; the pattern (static where possible) stuck.

Framework · 2020–now

Hypermedia / htmx

Stuck

HTML as the engine of application state — a deliberate counter to SPA maximalism.

Framework · 2016–now

Angular (2+)

Stuck

Batteries-included SPA platform for teams that want opinions and CLI scaffolding.

Framework · 2013–now

Electron

Mutated

Chromium in a trench coat. Slack, VS Code, and RAM usage everywhere you look.

Framework · 2017–now

Flutter

Stuck

Paint your own pixels on iOS and Android from one codebase. Google's bet against native UI fragmentation.

Framework · 2015–now

React Native

Stuck

React components on native views. Share logic with web until platform differences ruin your sprint.

Platform · 2014–now

Firebase

Mutated

Auth, realtime DB, and hosting as a product. Prototype paradise; vendor lock purgatory at scale.

Platform · 2017–now

WebAssembly

Stuck

Portable bytecode beyond the browser. Slow burn with real teeth in edge and plugin runtimes.

Framework · 2020–2025

Remix

Mutated

Web standards–first React framework with nested routes and forms. Merged into React Router; the ideas survived the brand.

Language · 2011–now

Dart

Stuck

Google's language that found purpose as Flutter's syntax. Rare case where the UI framework rescued the language.

Architecture · 2015–now

GraphQL Everywhere

Mutated

Clients ask for exactly what they need — and sometimes invent a second backend team to make that safe.

Practice · 2018–now

Low-code will replace eng

Mutated

Drag-and-drop apps for the enterprise backlog. Useful for internal tools; fantasy as a full replacement for engineering.

Language · 2016–now

Zig

Stuck

C without the footguns — in theory. A systems language beloved in blog posts and absent from most job boards.

Platform · 2020–now

Deno

Stuck

Node reimagined with TypeScript native and permissions. Correct instincts; wrong moment against npm gravity.

Framework · 2021–now

SolidJS

Stuck

Fine-grained reactivity without a virtual DOM. The framework perf nerds cite when explaining why React rerenders hurt.

Framework · 2021–now

Qwik

Stuck

Resumability and zero-JS-first marketing. Correct diagnosis of hydration pain; crowded prescription aisle.

Practice · 1995–now

Big-bang rewrite

Costly fad

Replace the working system instead of strangling it. A perennial money pit with new branding each decade.

$ "We'll freeze features and rewrite in the hot language." Multi-year projects shipped late, incomplete, or not at all — while competitors iterated on the ugly system that made money.

Practice · 2017–2023

Blockchain for CRUD

Costly fad

Distributed consensus for problems that needed a database and an audit log.

$ Enterprises funded "innovation labs" to put loyalty points and supply-chain spreadsheets on ledgers. Most pilots ended as slide decks; the gas fees and audits did not.

Framework · 2018–2024

Web3 frontends

Faded

Wallet popups as UX. A subculture with real tech under a mountain of token incentives.

2022–now

AI Present

The stack gets a co-pilot

LLMs, vector DBs, and agent frameworks arrive faster than governance. The question isn't whether AI sticks — it's which abstractions survive the hype tax.

AI pair programming is already changing how code is written. Agent frameworks and vector stores are still sorting winners from demos. The durable layer will look familiar: evals, retrieval quality, product UX, and ownership. Autopilot rewrites without tests are just big-bang migrations with better slides.

Full era essay →

Practice · 2021–now

AI pair programming

Stuck

Autocomplete that actually understands the file. Changing how code is written faster than how it's reviewed.

Practice · 2023–now

LLM eval pipelines

Stuck

Regression tests for nondeterministic models. The unglamorous infrastructure that separates demos from products.

Platform · 2023–now

Postgres + pgvector

Stuck

Embeddings back in the database you already run. The quiet winner of the vector DB gold rush.

Platform · 2022–now

Bun

Stuck

All-in-one JS runtime that actually feels fast. Winning scripts and side projects while production waits for edge cases.

Practice · 2023–now

RAG as default architecture

Mutated

Retrieval-augmented generation as the answer to every knowledge problem. Often right; often Postgres with pgvector would suffice.

$ Teams stood up vector pipelines, chunking strategies, and rerankers before asking if fine-tuning or a SQL query would answer the question. Retrieval infra became the product.

Practice · 2024–now

AI stack consolidation

Mutated

The shakeout from seventeen LLM wrappers to three vendors and one observability bill. Hype cycle entering boring procurement.

Practice · 2023–now

Agent ops / LLM observability

Stuck

Tracing, cost caps, and prompt versioning for production LLM features. Datadog for tokens.

Framework · 2023–now

LLM app frameworks

Mutated

Chains, agents, and prompt wrappers shipping weekly. The durable pieces will be boring: evals, retrieval, and product UX.

Platform · 2022–now

Vector DB gold rush

Mutated

Specialized stores for embeddings. Many workloads are folding back into Postgres extensions — after the vendor tour.

Practice · 2023–now

AI will rewrite the codebase

Costly fad

Big-bang rewrite, now with a chatbot. Same failure mode; prettier slides.

$ Exec decks promised autonomous migrations "by next quarter." Teams that skipped tests, ownership, and incremental strangler patterns bought expensive demos and fragile diffs.

Language · 2019–now

Gleam

Stuck

Erlang semantics with friendly syntax on the BEAM. The language your functional meetup loves and your employer has never heard of.

Framework · 2023–now

Multi-agent orchestration hype

Costly fad

LangGraph, CrewAI, AutoGPT cosplay — agents delegating to agents for tasks one function call could do.

$ Autonomous agent swarms burned API budgets on coordination loops that a single prompt and a cron job would have handled. Debugging "which agent lied" became the new on-call sport.

Practice · 2024–now

Vibe coding

Costly fad

Accept AI output on feel, skip the diff, hope tests exist. Fast demos; fragile ownership.

$ Ship-to-prod without reading the diff became a flex until incidents revealed nobody knew which prompt wrote the auth middleware. Review debt arrived in one weekend demo.

Practice · 2023–2024

AutoGPT-class autonomous agents

Costly fad

Give the model a goal and a credit card. Viral in 2023; quietly shelved when the loop never stopped spending.

$ Open-ended agent loops burned API credits chasing goals nobody specified well enough to finish. Demo videos outpaced production deployments by orders of magnitude.

Map

Relationships, stickiness, and outcomes in one constellation.

Open graph →

Invoice scars

Named cases and costly fads — the hype tax, itemized.

Case studies →

Archive

Catalog, signals, patterns, and essays.

Open catalog →

© 2026 Fadstack · Shane Code

Opinionated history · not a ranking