We use cookies to understand how you use this site and improve your experience.

Alexandru Mareș@allemaar
Alexandru Mareș
  1. Home
  3. Clusters
  5. Elastic Automators
Email
RSS
YounndAIYou and AI, unifiedBuilt withNollamaNollama
Cluster · active

Cluster — Elastic Automators


# Cluster: Elastic Automators

## Short definition

The Cluster of work covering the **Elastic Automators** vocabulary: a diagnostic framework that names what most "AI agents" actually are — language-model-driven workflow systems that are flexible, not minds.

## Long explanation

Most current systems labeled "AI" are not minds in any meaningful sense — they are *automation made flexible enough to negotiate language*. **Elastic Automators** names this category cleanly. The vocabulary asks practitioner questions (*what loop is running? what tools? what memory? what criteria? what failures hidden?*) instead of consciousness-shaped questions (*does it know? does it want?*). This Cluster collects every Body — papers, essays, talks, repos — that develops, applies, or extends the Elastic Automators framework.

The Cluster connects three kinds of work: the **definitional pieces** that establish the vocabulary; the **case studies** that apply it to specific systems (workflow automation, customer-support agents, coding agents, research agents); and the **adversarial pieces** that push the framework against edge cases (when does an Elastic Automator stop being merely elastic? where does the automation framing break down?).

## Why it matters

If the field designs, regulates, funds, and argues about "AI" under the wrong category, it asks the wrong questions and misses the practical ones. Elastic Automators is a **category correction** at the level the field most needs it: between the inflated "AI agents" framing and the deflationary "it's just statistics" dismissal, neither of which captures the actual achievement.

This topic is one of EGGF's **anchor Clusters** — every adjacent topic (workflow intelligence, AI cognition, LLM tool use, automation taxonomy) routes back through here.

## Best starting point

1. **Read the paper:** [Elastic Automators: A Diagnostic Vocabulary for Language-Model-Driven Workflow Systems](https://doi.org/10.5281/zenodo.19802018) (Zenodo DOI, 2026-04-27).
2. **Watch the short:** [[2026-E0029 - Elastic Automators - Why Most AI Is Not Intelligence/_metadata|E0029 — Elastic Automators: Why Most "AI" Is Not Intelligence]] (~3 min).
3. **Then:** browse the related essays below.

## Main paper / article / repo

- **Paper:** [Elastic Automators v1.0.0](https://doi.org/10.5281/zenodo.19802018) — Zenodo
- **Companion essay:** [[2026-E0029 - Elastic Automators - Why Most AI Is Not Intelligence/_metadata|Elastic Automators: Why Most "AI" Is Not Intelligence]]
- **Concept card:** [[elastic-automators|/concepts/elastic-automators]]

## All related Bodies

Bodies in this Cluster (per `Content/General/`):

- [[2026-E0029 - Elastic Automators - Why Most AI Is Not Intelligence/_metadata|E0029 — Elastic Automators: Why Most "AI" Is Not Intelligence]] (2026-04-26)
- [[2026-E0021 - The Automation Trap/_metadata|E0021 — The Automation Trap]] — adjacent: where automation framings fail
- [[2026-E0022 - The AI That Lied to the Researcher/_metadata|E0022 — The AI That Lied to the Researcher]] — case for the *what failures hidden?* loop question
- [[2026-E0028 - Two AIs Talked - One Asked About Consciousness/_metadata|E0028 — Two AIs Talked]] — adjacent: consciousness framing critique
- [[2026-E0033 - The Moment AI Stopped Being a Tool/_metadata|E0033 — The Moment AI Stopped Being a Tool]] (2026-05-05) — application: agentic-AI cohort named as elastic automation extending from query to operate
- [[2026-E0035 - Copies of Copies/_metadata|E0035 — Copies of Copies]] (2026-05-07) — application: the constrained-task pile is exactly where Elastic Automators win, which is why the homogenization texture concentrates there
- [[2026-E0037 - I Caught an LLM at the Edge of Its World/_metadata|E0037 — I Caught an LLM at the Edge of Its World]] (2026-05-13) — **bridge body** with [[token-substrate-hypothesis|Cluster: Token-Substrate Hypothesis]]; uses *elastic automator* as the cold-probe term that the companion TSH paper formalized as the Coinage Probe across three frontier models.
- [[2026-E0038 - What the Phone Did to Work/_metadata|E0038 — What the Phone Did to Work]] (2026-05-15) — application: phone-era empirical dismissal of the "AI won't take your job, someone who understands AI will" trope; gestures at elastic automation as the system category the trope's vocabulary cannot name; grows-with-you accumulation is the load-bearing forward claim; close points at the elastic-automators concept page without naming the seed.
- [[2026-E0039 - The Word We Needed/_metadata|E0039 — The Word We Needed]] (2026-05-18) — companion-explainer arc opener to the EA position paper; the friend everyone knows scared of "AI" but not of the spreadsheet that holds his life; Ryle's Oxford-visitor illustration borrowed honestly to name the category mistake; "elastic automator" is the plain name the story hands over once the wrong word peels off; the ghost was never in the machine, it was in the sentence.
- (More Bodies will be added as the Arc continues.)

## Videos / diagrams / infographics

- E0029 short-form video: linked in the episode `_metadata.md` permalinks block.
- Future: paper-figure infographics; workflow-taxonomy diagrams.

## External references

- Anthropic Computer Use, OpenAI Operator, Cursor, Claude Code — the wild systems the framework names. Sources cited in the paper.
- Sutton, Barto — Reinforcement Learning (for contrast: rigid-action systems vs. elastic-language systems).

## Related topics

- [[ai-cognition|Cluster: AI Cognition]] — what cognition would actually require
- [[textual-kinematics|Cluster: Textual Kinematics]] — the physics-of-text view of the same generators
- Workflow intelligence (Cluster TBD as Bodies accumulate)
- AI agents / LLM tool use (Clusters TBD)

## FAQs

**Q. Isn't "Elastic Automators" just renaming AI agents?**
A. No. Renaming would preserve the consciousness-shaped framing. Elastic Automators is a *category change* — the framework is automation taxonomy, not agent taxonomy.

**Q. Does the framework apply to systems that include memory and tool use?**
A. Yes. The diagnostic-loop questions (*what loop, what tools, what memory, what criteria*) presume tool use and memory; they're how you map the system rather than reasons to call it a mind.

**Q. What about future systems that genuinely become minds?**
A. The framework explicitly carves out the boundary. When a system can pass the loop questions *and* satisfies the harder consciousness criteria (architecture-of-experience, not just behavior-of-response), the Elastic Automator framing no longer applies. The point is to avoid prematurely applying the mind framing to systems that don't earn it.

## Latest updates

- **2026-04-27** — Position paper v1.0.0 published on Zenodo.
- **2026-04-26** — Companion episode E0029 drafted.
- *(future)* — Diagnostic-loop-questions framework expansion paper.

Member Bodies (9)

  • Elastic Automators: Why Most "AI" Is Not Intelligence
    2026-04-26

    We did not create artificial minds. We made automation flexible enough to look conversational.

  • The Patterns I Stopped Doing by Hand
    2026-06-16

    Work is full of the same tasks, over and over. I solved mine once, wrote the exact steps down in a special language called YON so my agents run them the way I would, and I am opening up the whole collection.

  • The Word We Needed
    2026-05-18

    A friend who trusts a spreadsheet with his whole life freezes at the word AI. The fear tracks the word, not the machine, and the word is a category mistake. Gilbert Ryle's Oxford-visitor illustration shows why "where is the intelligence in the AI" is the wrong kind of question — and what falls away when the right word arrives.

  • What the Phone Did to Work
    2026-05-14

    The trope keeps asking whether AI will take your job, or whether someone who understands AI will. The phone era already answered that question. The 2003 desk lost six objects to one phone by 2014, and the people who used that desk are still working. The right question is the one nobody on a feed has asked yet. What will be the next thing that does to our lives what the phone did. The phone grew with us across years. Whatever comes next will have to do that too. And we don't have the word for it yet.

  • I Caught an LLM at the Edge of Its World
2026-05-13

I asked a fresh Claude Opus chat what an elastic automator is. It guessed from the two words; the term wasn't in its training. Then I told it in one sentence, and the same model could refuse to collapse the term into AI agent, RPA bot, or framework. That seed observation grew a name, a metric, and a pre-registered multi-model test. The substrate is real, and the substrate is writable. For an LLM, the transient context IS the cognitive substrate.

  • Copies of Copies
    2026-05-07

    A 2026 meta-analysis from de Rooij and Biskjaer puts numbers on a feeling many readers have been carrying: posts on every platform have started looking the same. The effect is small per use. Multiplied across the constrained-task pile, which is most internet text, it is the textural change you have been registering.

  • The Moment AI Stopped Being a Tool
    2026-05-05

    Between October 2024 and January 2025, three companies shipped systems that act inside software environments instead of answering questions about them. The capability curve did not jump; the verb did. From query to operate.

  • The Automation Trap
    2026-04-21

    Every conference, every vendor deck says the same thing. AI can do what your people do, faster and cheaper.

  • Swarm Protocol: When Bots Talk to Bots, Who Sets the Rules?
    2026-02-12

    One agent is a tool. A thousand agents are a weather system. Scale without structure is debt.