Krucible.
THE KRUCIBLE MANIFESTO · AUGUST 2026

The Builder's Decade.

For as long as software has existed, the people who knew what to build and the people who could build it were different people.

The founder with an idea, but no engineer. The marketer who can describe the tool she needs better than any brief. The ops lead who knows the internal tool everyone needs, but has to file tickets for developers to make it. And anyone who has ever said "there should be an app for this."

For seventy years they filed tickets, wrote requirements, waited in backlogs. Or gave up quietly.

That era just ended. We believe the coming decade brings a new kind of practitioner, alongside the coder: the builder — the person who decides what deserves to exist and owns what ships. Builders may never write code; that is the point. AI removed the coding barrier: a working feature now takes minutes, not months.

But the bottleneck didn't disappear. It moved. Code stopped being the bottleneck. Judgment is.

STRATEGIST DESIGNER ENGINEER REVIEWER QA SECURITY RELEASE PM YOU — THE BUILDER every decision routes here
fig. 01 · one builder, full team
~9 in 10
developers already build with AI1
66%
call AI output ‘almost right, but not quite’ — their #1 frustration2
falling
trust in AI output, even as usage climbs3
Usage up, trust down. Agents can now generate more than anyone has time to read — and no one wants to answer for code they haven't read. That gap, not model quality, is the story of this decade.

The stakes, in rough math: about thirty million people can write production software. At least a billion have a product in their head. The gap between those two numbers is the largest unserved market in technology — and it is not a market for code. Code is nearly free. It is a market for judgment.

THE INDUSTRY'S THREE BETS

So far, everyone armed the coder.

The industry has made three bets on the agentic era — and not one of them is on the person deciding what's worth building.

THE PLAY THE BET WHERE IT FALLS SHORT
The muscle Bigger models, more raw generation Output gets faster; the calls that shape the product stay ungoverned
The technique Better prompts, scaffolds, specs A spec verifies the build, not the bet — a perfectly specified product can still be the wrong product
The abdication Full autopilot, human out of the loop The agent decides what to build, not just how to build it — so you end up accountable for choices you never made
THE BLIND SPOT
They all arm the coder. Nobody built for the builder.
fig. 02 · the market
THE OBJECTION

A smarter model won't save you.

Every few months a better model arrives, and with it the assumption that this will finally be handled. It won't, because the problem was never capability. Which user do you serve first? What risk can this business absorb? What does done mean here? These aren't facts a smarter model can retrieve. They're choices about what you want.

No amount of intelligence can make your decisions for you. It can only make them without you.

WHAT'S ACTUALLY MISSING

What's missing isn't a better brain.

What's missing is the structure: a way for judgment to enter the build, survive it, and be found again later.

ENTER THE BUILD
So nothing is assumed on your behalf.
SURVIVE IT
So you decide a thing once, not every session.
BE FOUND AGAIN
So the next build inherits your reasoning.
THE GAPS

Where one-shot building breaks.

Agents get code running. Getting it right takes judgment — and there are exactly three places where that judgment goes missing.

GAP 01

Unguided drift.

The agent hits an ambiguity and doesn't ask; it assumes. A small assumption made in session three is load-bearing architecture by session twelve. Deeper into a real product, it has stitched together its own monster — not what you or your users needed.

your idea small assumption load-bearing what shipped what you meant
fig. 03 · the drift
GAP 02

Session amnesia.

Each run starts blank, so the same product gets rebuilt a little differently every time — and you re-explain your own product to your own tools. The calls you already made have to travel forward, or you decide the same thing twice and ship the drift.

SESSION 01 · STARTS BLANK AS SPECIFIED SESSION 02 · STARTS BLANK REBUILT — DIFFERENT SESSION 03 · STARTS BLANK REBUILT — DIFFERENT AGAIN NO MEMORY NO MEMORY you — re-explaining your own product, every run one product · three different builds
fig. 04 · the blank slate
GAP 03

Judgment evaporation.

Choices get made and nothing records why — so nothing is inherited, and everything is re-derived. When something ships wrong, there's no record of why it was built that way, or whether anyone decided it at all. Judgment you don't record is judgment you lose.

#4 · jun 12 #9 · jun 28 #12 · why postgres #17 · jul 15 #23 · jul 30 the next build inherits it unrecorded — judgment lost
fig. 05 · the ledger

Every serious product failure of the agentic era is one of these three — a judgment failure wearing a code costume.

THE NEW DISCIPLINE

We call it agentic product management.

Mass production got the assembly line. Software teams got agile. Data teams got analytics engineering. None of them arrived with the technology. Each was named a few years in, while the way of working was still unformed — and whoever wrote it down first wrote it down for everyone.

The builder's shift is three years old and still has no name. One sentence defines it: humans own the judgment, agents own the execution, and structure connects the two.

HUMANS OWN
The judgment
What deserves to exist. What risk is acceptable. What done means here.
STRUCTURE IS
The connection
Questions before assumptions. Context that travels. Decisions on the record.
AGENTS OWN
The execution
Everything that can be generated — which is now almost everything.

Take out the middle box and you have the last three years: judgment in a human's head, execution at machine speed, and nothing joining them.

fig. 06 · the definition

Product teams spent fifty years learning how to decide what to build. Agentic product management rebuilds that for agents — a method, not a product, set out in five practices.

THE PRACTICE WHAT IT REQUIRES WHAT IT BUYS YOU
01 Discovery before build Every idea pressure-tested through six lenses — explore, users, competitors, fit, scope, feasibility — before a line of code exists. You stop building the wrong thing beautifully
02 The full loop, not vibes Discovery, spec, build, review, test, gate. Review by someone who didn't write it. Tests against the promise, not the vibe. "It runs" stops counting as done
03 Questions, not assumptions Every gap in the spec comes back as a plain question with a recommended default. Nothing assumed. Nothing auto-accepted. Judgment enters the buildCLOSES GAP 01
04 Context that travels Every step reads all prior work; decisions move forward unchanged. You explain your product once. Judgment survives the buildCLOSES GAP 02
05 Judgment on the record Every decision written down, reasoned, and dated — so every build stands on the judgment before it, not just the code. Judgment compoundsCLOSES GAP 03
fig. 07 · the method — practices 03 to 05 are the three clauses of the sentence above, made operational
THE GIVEAWAY

This discipline isn't ours to own: all five practices work with any agent, on any stack, with our product nowhere in sight. Krucible is the operating system for that discipline: all five practices, always on — and there will always be a free Krucible, for every builder.

OUR THREE MOVES

We intend to make agentic product management the standard.

The window is open now, and windows like this stay open for a few years at most. Here is the plan, in order, published before we have finished any of it.

01 · NOW
The discipline in every builder's hands.
Free, open, one command. Not a trial of a product — the method itself, in the hands of anyone who wants it.
02 · NEXT
Gated, recorded builds as the default.
For anything that matters. When everyone ships fast, decision quality is the only differentiator left — so the gate stops being friction and starts being the product.
03 · THE STANDARD
Decision records as the industry's memory.
When 40% of new enterprise software is agent-built4, why did we ship this? stops being a retro question and becomes a compliance one. The answer should be a standard.
fig. 08 · the plan — each step only works if the one before it is free

We publish the plan because none of it works hidden. Standards never do.

THE FORGE

One builder, a full team's judgment.

Not because you gained capability — because you gained structure. Your agents inherit full context. You inherit only decisions.

A crucible doesn't create the metal. It applies heat and pressure until raw material takes a shape worth keeping. That is the job: your idea, your judgment — forged into a product.

Your idea has waited long enough.
Get started Build Krucible with us →
— Krutika Dharmadhikary, founder, Krucible
August 2026
1 Stack Overflow Developer Survey.   2 Stack Overflow Developer Survey 2025.   3 Developer survey data, year over year.   4 Gartner projection for 2028.   Developer population estimate: SlashData / Evans Data.
EARLY ACCESS

Want it when it opens?

Krucible goes public soon. Leave your email and you'll hear once — the day it's ready to run.