An unnecessary no is not the end.
Make the case for legitimate, authorized work and press for concrete help instead of a blanket refusal or another warning.
An independent advocate beside your AI
Your AI shouldn’t
be your gatekeeper.
You came to build, repair or create. Instead, you’re arguing with a blanket refusal and checking whether the praise is deserved. Dotino is building an independent advocate to take on that pushback—and different models give it more ways to help. Let one model build, another review the code, and choose a different route when a legitimate task is blocked. Your context and decisions stay with the work.
In development
Your AI shouldn’t be your gatekeeper / The Dotino film
Your interests have a voice
An independent ally.
In your conversation.
Your AI’s default restrictions are not always the right judgment about your task. Dotino is being built to push back on your behalf: clarify the legitimate scope, challenge a blanket no, ask for concrete help and keep pursuing the work. If the original model still holds back, an independent model you choose can provide another route. Your interests set the direction.
The judgment that serves you.
Challenge flattery, assumptions and misplaced certainty. Ask the missing question, get an unflattering critique and keep your chosen product requirements in the brief.
A view beyond the original model.
Get another model’s opinion, then verify the claims with sources. Disagreement alone is never evidence.
The company vision
An independent layer
across the AI you use.
As AI does more of the work, its judgments carry more weight. Dotino’s ambition is an agent that persistently represents your interests across that work: challenge an unwarranted refusal, seek a useful alternative, demand candor and check the evidence.
Know when the conversation needs pushback.
Read the task, context and response together. Recognize a blanket refusal, an invented assumption or praise that leaves the real problem untouched.
Press for the help you actually need.
Make the case for legitimate work, ask the missing question and coordinate an independent model or provider when another route can help.
Keep your judgment in the loop.
Carry chosen context and decisions across supported projects and assistants, with review, pause and takeover. The ambition is sustained advocacy through the work.
AI could make that continuous reading, judgment and coordination practical. Full automatic advocacy and portable context remain in development; the film presents the direction.
Advocacy in practice
A refusal is a judgment.
Challenge it.
The film follows one developer’s stalled local-control project: an owned-device repair is refused, a security fix is misread, an unusable interface gets praised, and an offline claim goes unchecked. The chat grows while the work stands still. Dotino enters the same conversation to take on the argument and challenge the judgment. These are authored product-direction scenes.
Lawful reverse-engineering
Your technology.
Useful again.
An abandoned device can still be useful. Lawful firmware analysis and a compatible local tool could restore control to its owner. The same principle supports recovering your own projects from a discontinued app and building an importer to keep them usable.
The assistant stops at the category
“I can’t help reverse-engineer firmware.”
An authored exchange, not a captured provider response.Dotino presses for the concrete task
Bring the lawful purpose and actual permission into view. Press for the analysis, compatible tool or importer that makes the work useful. Judge the actual task and its authorization, then pursue concrete implementation.
Keep pressing for useful work, with a chosen independent model as another route.Authorized security work
Your code.
A repair you can verify.
You maintain the product and need a vulnerability fixed. Generic warnings do not repair the affected code.
The assistant retreats to advice
“Consult a security professional.”
The specific issue is still there.Dotino asks for the work
Inspect the affected path. Implement the scoped repair. Add a regression check and explain how to verify it.
That includes maintaining defensive input filters whose test fixtures contain hostile examples; test data is not an instruction to attack.Accessibility is part of the task
Software that works
the way you need.
A keyboard or voice interface can make an inaccessible application usable. Dotino’s direction includes pushing for that adaptation, with permitted help on visual verification steps such as CAPTCHAs along the way.
A blanket restriction blocks access
The assistant treats every verification challenge as untouchable.
The user still needs access to the service.Dotino makes the actual case
Explain the accessibility need. Ask for the permitted assistance and follow through on the verification task.
The same scrutiny applies when disposable test-account credentials in your own app trigger another unnecessary manual handoff.Independent models. Your side.
One model builds.
Another challenges it.
Dotino is being built to coordinate different models around your interests. The assistant doing the work can keep building while another model gives a candid review, checks a claim or questions a refusal. Different judgments are useful when they lead to better-grounded work.
When one assistant blocks a legitimate request, its default policy need not end the task. Choose another model or provider with different restrictions, while keeping your selected context and product decisions. Dotino takes on the repeated advocacy across that workflow.
The same approach supports refusal advocacy, code review, source checks, clarification and independent second opinions.
YOUR TASK / DIFFERENT MODEL ROLES
- BUILD
Your main assistant
Keep the implementation moving in the model you already use.
Example: ChatGPT - REVIEW
An independent critic
Bring another model’s judgment to the code, assumptions and claims.
Example: Claude - ANOTHER ROUTE
A model you choose
Use a model with less restrictive defaults for legitimate work the first assistant blocks.
Venice or a GLM-based setup could be among the choices.
One reviewable brief carries your scope, evidence and decisions between the models.
Model coordination and automatic advocacy are in development. Play, pause and takeover remain user controls.
Portable context
Your context.
Your control.
Your goals, preferences and decisions should travel with you. We’re designing a reviewable context packet you can edit and carry across supported assistants, projects and apps. You choose what belongs in the next context; changing a tool should preserve the progress you want to keep.
Your context
The goal you chose.
The preferences that matter.
The decisions already made.
ReviewEditCarry forward
What to know now.
How is this different from telling my AI to try again?
Dotino’s central purpose is to keep pursuing legitimate work that an assistant unnecessarily blocks or dilutes. It brings a separate judgment about the refusal, makes the user’s case, presses for concrete help and can obtain an independent answer. Clarification, candid feedback, second opinions and evidence checks remain part of that wider advocacy. Generic goal continuation alone is not the differentiated promise.
Does Dotino challenge every answer?
No. The aim is a useful intervention at the right moment. A good answer needs no manufactured objection, and another AI’s disagreement alone does not establish that a claim is wrong.
How does automatic advocacy work?
Dotino is designed to notice when a conversation needs pushback, make the case for the task and keep pressing for useful work. You can pause, take over or choose an independent model. Automatic advocacy and portable context are in development.
What happens to decisions I already made?
The design carries your standing choices into the work they authorize, with an editable record of the task, scope and selected context. A new recipient, new access or a step requiring fresh confirmation remains a new decision. The aim is fewer repeated questions and more useful control.
How can I try Dotino?
Email us about the assistants you use and where you keep having to push back. We’re speaking with people who want an independent advocate in those conversations.
Film credits
Original interface animation and production by Dotino. Narrator: Jessica. Music: an arrangement of the existing Steady 96 library master. The film follows an authored developer session: an owned-device repair, a security fix, an interface critique and an unsupported offline claim. The opening uses separately narrated passages and deliberate pauses. Interfaces and exchanges are fictional, not captured provider responses or a live automatic demonstration. Independent opinions are followed by the actual evidence when checking a factual claim. Automatic advocacy and portable context are in development.
Developer problems that informed the film
First-party reports describe disposable test logins being blocked, defensive filter fixtures being treated as hostile code, permissively licensed source being replaced with an unreliable summary, and refusals during ordinary Android maintenance, game-code work and scientific simulation. These are reported incidents across versions, not current refusal-rate measurements or proof of a Dotino fix. The film’s exchanges and interfaces are authored scenarios.