They work. You take the credit.

Everything on-chain is still hand-work. Somebody sits at a screen and clicks. Somebody watches a chart at four in the morning because that is when it moves. Somebody rebalances, redeploys, re-checks, and calls the whole exhausting ritual being early. It is not being early. It is labour, and it is being paid at zero.

The market does not close. You do. Every hour you spend asleep, eating, working a job or being a person is an hour the position is unattended and the work does not get done. That gap is not a discipline problem. It is a staffing problem.

ROBOTHOOD builds the staff. You do not do the work here. You commission a unit, set what it is allowed to do, and put it on the floor. It runs the shift. It does not sleep, it does not negotiate, and it does not take the credit.

Doing it yourself solved a real problem once. Deploying, trading and monitoring used to require a desk, a team and permission. Now it requires one person with a phone. That was progress, and it came with a cost nobody priced: the one person is you, and you are the bottleneck.

A human operator has a duty cycle of about sixteen hours on a good day, reacts in whole seconds, and gets worse the longer the shift runs. The venue you are operating in has a duty cycle of twenty-four hours, reacts in milliseconds, and does not get tired. You are not competing with other people. You are competing with everything that already automated.

It shows up in the chart below. The share of on-chain work still executed by a human hand is collapsing, and not because people got lazy. It is collapsing because the hand lost.

The floor's position is simple: the interesting part of a position was never the conviction. It is the execution, and execution is work. Work can be delegated. Almost nobody delegates it.

Before you can commission anything, you need to know what a unit is made of. ROBOTHOOD models every worker as five systems. Each system is a slot. Each slot holds a setting you choose, and each setting changes something the market can actually feel.

Most tooling exposes one system: size. A few expose two. The floor exposes all five, plus the one nobody talks about, and lets you set them against each other. That is where a useful worker comes from. Not from one exotic setting, but from a combination nobody has put on shift before.

Four steps. None of them involve writing Solidity. None of them involve you doing the work.

01 · Pick the frame. Choose the worker you are building on. A plain fixed-size unit, a rebasing one, a reflexive one. This is the chassis, and it decides what your settings are even able to mean.

02 · Set the systems. Configure the five. The floor compiles your choices into a contract rather than filling a template, so combinations no template author anticipated are still valid units.

03 · Burn-in. The unit runs a full shift against a simulated market before it touches the chain. Most builds fail here, which is the point. You get the teardown report either way.

04 · Put it on the floor. The unit that survives is deployed to chain 4663 and settles to Ethereum. From that moment its build is public, immutable and readable by anyone — including everything that would rather it stopped working.

Building is not free and the floor will not pretend otherwise. The further you tune a unit away from a known-good configuration, the higher the chance it comes off the line dead.

This is not a flaw in the process, it is the process. A workshop that guarantees every unit behaves identically is a workshop that can only produce the same worker everyone already has. Capability requires failure. The floor's job is to make the failure cheap, and to make it happen in burn-in instead of in someone's wallet.

The chart tracks it honestly. Lightly tuned units survive at rates close to a stock build. Heavily tuned ones do not. What changes is not just survival, it is range: the units that do survive do work the stock build simply cannot.

Read it as a warning if you like. The floor reads it as a shift roster.

Three units that came off the line working.

These are not products and they are not for sale. They are reference builds: what happens when you take the same frame and push a single system to its limit. Every one of them is a dog because the first frame that held for a full shift was quadruped, and the name stuck.

Notice that none of the three is better. Each is a working answer to a different job, and each pays for its strength somewhere else. That trade is the whole discipline. A system with no cost is a system you have not measured yet.

A work floor needs a medium that does not lie about who did what. The unit has to be readable, its build has to be immutable once deployed, and the record of the work has to outlive the workshop that built it.

ROBOTHOOD is specified against chain 4663, an EVM ledger that settles to Ethereum and posts its data as blobs on L1. There is no native chain token. Gas is paid in ETH. That matters here: a substrate with no coin of its own has no incentive to favour one worker over another. It is a neutral floor, which is the only kind worth putting a shift on.

Ethereum is the part of this stack the floor does not control, and that is deliberate. The sequencer decides what happened inside a block. Ethereum decides whether the block happened at all. It means a deployed build cannot be quietly edited afterward, including by us.

Which is the only guarantee worth making. Not that your unit will be safe. That the record of the work — and of whose name is on it — will outlive anyone's ability to change it.

Everything above is the pitch. This section is the part the pitch usually leaves out, and it is here because a workshop that hides its failure modes is not a workshop.

Units are experimental. A configuration that has never been put on shift has, by definition, never been tested by reality. Burn-in simulates. It does not know.

Automation is not profit. A worker that executes flawlessly can execute a bad idea flawlessly, all night, at full speed. The floor makes the work happen. It cannot make the work correct.

The build is public. The moment your unit is deployed, every setting you chose is legible to anyone, including whoever is looking for a weakness in exactly that combination.

Nothing here is financial advice, and no outcome is promised. Assume any unit can go to zero, including yours, including the ones in bay C.