Framework · 2023–now · AI Present

Multi-agent orchestration hype

LangGraph, CrewAI, AutoGPT cosplay — then vendor harnesses shipped the same loop as an API. Open-source swarms faded; leased orchestration stuck.

How Multi-agent orchestration hype mutated

Open-source multi-agent frameworks failed as defaults because non-determinism compounds with every hop and observability lagged behind demo videos. They mutated into first-party session APIs — OpenAI Agents, Anthropic managed agents, Bedrock AgentCore — where the swarm is a product SKU. The costly fad was autonomous headcount replacement; the stuck residue is a vendor-owned loop with MCP tools and a human somewhere in the interrupt path.

Cost of the fad

What Multi-agent orchestration hype cost

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.

Case studies

Patterns

2022–now

AI Present

Part of the AI Present era (2022–now).

Read the AI Present era →

Compare with

Related

Framework · 2023–now

LLM app frameworks

Mutated

LangChain-class glue mutated into MCP clients and thin wrappers around vendor APIs. The durable pieces are still boring: evals, retrieval, and product UX.

Practice · 2023–now

Agent ops / LLM observability

Stuck

Tracing, cost caps, and prompt versioning for production LLM features — mostly constrained tool loops, not autonomous agents. Datadog for tokens.

Practice · 2024–now

Vibe coding

Costly fad

Accept AI output on feel, skip the diff, hope tests exist. Fast demos; CVEs and nobody owns the middleware.

$ 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.

Framework · 2023–now

LLM orchestration frameworks

Mutated

Chains, tools, memory, and agents as a framework — LangChain-class glue sold as architecture. Fatigue was real; the repos became MCP clients, not tombstones.

$ Teams adopted chain/agent frameworks before they had a single reliable tool call. Abstraction layers multiplied while prompt quality stayed flat.

Platform · 2025–now

Vendor agent APIs

Mutated

The Codex/Claude loop as a product SKU — durable sessions, compaction, subagents, and a sandbox you do not own.

$ Teams replaced a weekend of LangGraph glue with a managed session API, then discovered residency, recovery, and kill switches all live on the vendor side of the harness.

Architecture · 2024–now

Context / tool protocols

Stuck

Standard ways for models to talk to tools and data — MCP after the 2026 stateless rewrite, not one-off plugin snowflakes.

© 2026 Fadstack · Shane Code · Privacy

Opinionated history · not a ranking

Site updates

Occasional Fadstack notes. Confirm by email — this list stays off the book and advisory lists.

I use the address for Fadstack updates only. Privacy.