Ship it internally on Monday. Even if it's bad.

The internal beta is not a marketing move. It is the cheapest instrument for measuring the gap between your head and the world. Skip it and you will polish the wrong thing for six months.

The hook

I watched a team lead demo an internal tool last quarter. Nine months of work. Beautiful architecture. The prototype was behind a feature flag that nobody outside the team had ever seen. In the one-on-one afterwards, he said the thing I have learned to dread: "We are close. We just need to fix onboarding." The onboarding had been close to fixed for four months. I have been that person. The tool was not the problem. The unshipped-ness of it was.

कर कटावरी तुळसीच्या माळा · Hands on Hips, Tulsi Beads Around

कर कटावरी तुळसीच्या माळा । ऐसें रूप डोळां दावीं हरी ॥ १ ॥ ॥ ध्रु. ॥
ठेविले चरण दोन्ही विटेवरी । ऐसें रूप हरी दावीं डोळां ॥
कटीं पीतांबर कास मिरवली । दाखवीं वहिली ऐसी मूर्ती ॥ २ ॥
गरुडपारावरी उभा राहिलासी । आठवे मानसीं तेंचि रूप ॥ ३ ॥
झुरोनी पांजरा होऊं पाहें आतां । येईं पंढरीनाथा भेटावया ॥ ४ ॥
तुका म्हणे माझी पुरवावी आस । विनंती उदास करूं नये ॥ ५ ॥
This abhanga names the ache a person feels when the thing they can clearly see inside their head is not yet in the world. The word "झुरोनी पांजरा" means wearing away into a cage of bones.

What I keep seeing

Every sprint, one more piece added to the polish column. Every standup, a little quieter than the last. The team lead can describe, in exquisite detail, what the tool will do and why users will love it. They just cannot let anyone touch it yet. Not because it is fragile. Because letting someone touch it would mean finding out it is wrong in ways they have not anticipated, and nine months of certainty is a terrible thing to trade for two weeks of embarrassment.

I am not above this instinct. I have nursed a side project for a year before letting a colleague poke at it. The colleague found the core problem in eleven minutes.

The mechanics

Every serious AI lab this year shipped their coding-agent CLI in a rough state. Terminal UX was thin, session management was awkward, docs lagged features. Every one of these tools has been iterated on hard since ship, and the iterations are informed by things nobody on the building teams could have predicted from the inside.

There is a mechanical reason. A tool inside a company is used by roughly seven people, all of whom understand its intent. A tool shipped internally to 200 people is used by people who read the tooltip wrong, mount the wrong volume, expect the wrong keybinding, and file the same misunderstanding three times in a week. That distribution of misunderstandings is the signal you cannot get any other way. You cannot A/B test your way to it, design-review your way to it, or user-interview your way to it. You can only get it by shipping.

An internal beta is not "marketing to your own company." It is the cheapest instrument for measuring the gap between what you think the tool does and what a competent stranger will experience it doing.

Where Tuka comes in

The abhanga has a specific word for what happens when a person holds a clear picture inside their head for too long without letting it meet the world. "झुरोनी पांजरा": wearing away into a cage of bones. Not a metaphor for sadness. A specific state of unmet longing that turns physical.

An engineering team in month nine of an unshipped internal tool is in exactly this state. Not because the team is undisciplined. Because holding a clear vision protected from reality is expensive. It costs the vision itself. By month twelve, the team can no longer see what the tool was supposed to be, only the specific compromises they made against a spec that has drifted. Tuka asks for the meeting to happen urgently: "विनंती उदास करूं नये" (do not treat this request with indifference).

What I would actually do

Set a rule for every internal tool project: ship to 20 non-building colleagues within eight weeks of starting, in whatever state it is in. Write a note that says: this will be rough for two weeks, then it will be much better because of you. If the team lead pushes back with "we cannot ship this yet, it is embarrassing," that is the signal, not the objection. The embarrassment is what you are buying. Twenty ruffled colleagues, some of whom will forgive you and some of whom will become your best product informants. A week of discomfort buys you a year of building the right thing.

Chetan Dhandal