LLenny's Podcast
← All frameworks
Innovation

Irreducible Complexity — The Simplicity Audit

1 + 1 = 1.5, not 3 — every addition subtracts, so audit the system, not the local decisions

Difficulty
Advanced
Time to result
~ongoing to results
Steps
5
Confidence
93%

Field's operating model for keeping a maturing product simple. The core insight is that features do not add — they compound negatively: adding a second thing to a coherent product yields less than the sum, because coherence is lost. Crucially, the complexity is a SYSTEM property, not a decision-level one: a team can make every local decision correctly and still produce a monstrosity. Therefore the audit must be run at the system level, on a recurring basis, by everyone.

Origin

Field credits Figma co-founder Evan Wallace with introducing him to an essay early in Figma's history — a Steve Yegge blog rant — that contains the term 'irreducible complexity'. The companion principle Field cites ('keep the simple things simple, make the complex things possible') he explicitly notes is not his own line; it is a long-standing software-design maxim associated with Alan Kay.

Core principles

  • 011 + 1 does not equal 3 — it sometimes equals 1.5. Additions subtract from coherence.
  • 02The more you add, the more complex it gets AND the worse it gets. Power and complexity are not the same axis.
  • 03You can make every correct local decision and still end up with a system that is too complex.
  • 04Keep the simple things simple; make the complex things possible.
  • 05Simplicity is everyone's responsibility, not a top-down function.
  • 06Entropy is the default: it is really easy to make the simple things complex.

How to run it

  1. 1

    Separate 'more powerful' from 'more complex'

    Before adding anything, state which axis you are moving. The hard, valuable work is making the product more powerful WITHOUT making it more complex at the same time. Field calls this extremely hard and says it demands constant attention.

    Pro tip Score each addition against 'keep the simple things simple, make the complex things possible' — a change that taxes the beginner path to serve an expert path fails the test.

  2. 2

    Furrow the brow — insist there must be something simpler

    In design and product reviews, when something feels too complex, refuse to accept it and demand a simpler formulation. Field's colleagues describe this as his signature review behavior. The feeling of complexity is a valid trigger, even before you can articulate the fix.

    Pro tip Say 'there must be something simpler' out loud rather than approving and adding a cleanup ticket. The cleanup ticket never gets done.

    Watch out Do not use this as a veto without engagement — the point is to force one more design pass, not to block the team.

  3. 3

    Audit the system, not the decisions

    Periodically zoom out from individual features to the whole subsystem and ask whether the accumulation still works — even when each decision along the way was defensible. Field's diagnostic: 'we made all the right local decisions and yet together they're too complex and they're not working anymore.'

    Pro tip Do this in conversation with product leaders across the org, not solo — different owners each hold one 'right local decision' and only the group sees the sum.

    Watch out Local-decision review will never surface this class of complexity, because by construction every local decision passed.

  4. 4

    Revisit and redesign the system

    When the system-level audit fails, the answer is not another patch. Go back and rework the system as a whole — which is what motivated Figma's UI3 redesign.

    Watch out This is expensive and will feel like it delivers no new user value. It is the price of not becoming a monstrosity.

  5. 5

    Distribute simplicity ownership

    Do not make simplicity a founder-only fight against entropy. Field's stated position: everyone is responsible for simplicity — every IC decision either fights or feeds the second law of thermodynamics.

In the wild

The Figma subsystem that failed its own audit

Field describes conversations with Figma's product leaders about parts of the product where each decision was individually right, but the accumulated system had become too complex to work. The correct call was not another local fix but 'let's go revisit the system'.

System-level revisits rather than feature-level patches — the reasoning that fed into the UI3 redesign.

The pages skepticism

Field remains skeptical of Figma's Pages implementation — not because pages are useless, but because he doubts it is the most elegant solution in the context of the entire system of product design. It is a local win he suspects is a system-level cost.

Pages stayed shipped (the world demanded them), but the framing shows the audit lens applied even to a successful feature.

Common mistakes

Assuming feature additions are additive

Teams justify each new feature on its own merits, implicitly assuming 1 + 1 = 2 or 3. In reality coherence decays and the sum can be worth less than one of the parts alone.

Reviewing only at the decision level

Every complexity monster is built out of individually correct decisions. If your only quality gate is a per-decision review, you will approve your way into a monstrosity.

Treating simplicity as a top-down mandate

If simplicity is the CEO's job, everyone else optimizes locally and the entropy wins. Field is explicit that it is everyone's responsibility.

Is it for you?

Best for

Product leaders on a maturing, feature-accumulating tool who can feel the product getting worse while every individual roadmap item looks justified

Not ideal for

Pre-PMF products still hunting for the core value loop, where the risk is under-building rather than over-accumulating

From the transcript

contains the term irreducible complexity and it's based this idea that like 1 plus 1 does not equal three it sometimes equals like one and…

23:00

there's Parts where it's like okay this thing is too complex as a system and we made like all the right local decisions and yet…

24:00

keep the simple things simple make the complex things possible

25:30

it's really easy to make the simple things complex unfortunately

26:00

we can make it more

23:00

From the episode

Dylan Field live at Config: Intuition, simplicity, and the future of design