I use AI at work
Start with a small task and a clear review step.
Open Technology / Learning library
AI Control Layer
Use AI with a clear purpose, the right access, and evidence you can check.
We build OpenTechnologyApp with AI assistance. This library connects what we learn about building with AI to practical guidance for using it, integrating it and choosing open-source tools.
English guides · Free to read · No account needed
AI Learning Series · 4 lessons
Most AI results are decided before the first prompt.
Pick the model. Write the prompt and set up the project. Structure the repository. Then spend less on all of it — without losing quality. Four lessons, in order, each built on the last.
Lesson 01
Model Usage
Which model does this task need?
Model tiers, context windows, tokens, and a decision checklist you can keep.
6 min read
Lesson 02
Prompting, Project, and Context
What goes in front of the model?
Prompt anatomy, acceptance criteria, and standing project context.
6 min read
Lesson 03
Documentation, Rules, and Skills
How does a repo stay consistent across sessions?
Three layers of guidance — and which one each rule belongs in.
7 min read
Lesson 04
Cutting Cost With an AI Workflow
How do you spend less without losing quality?
Baselines, one lever at a time, and guardrails on anything that loops.
7 min read
Status: outlines. Each lesson gets its on-screen walkthrough once it airs.
Start with your work
I build with AI
Make each change traceable from intent to evidence.
I choose and run tools
Check operating cost, access and the exit path.
Six tracks. One continuing practice.
Read a guide, try a bounded task, and keep what you can verify. Existing articles provide background; revisit version-sensitive details before following them.
Use AI well
Define the job, limit the data, and check the result.
Guide · 6 min
A practical routine for using AI wellA repeatable brief, review and feedback process.
Exercise · 5 min
Review an AI Summary When the Data Is IncompleteFour synthetic records, a flawed summary and a supported answer.
Exercise · 6 min
Build a Small AI Evaluation Set You Can ReuseFive reusable cases with a separate reviewer answer key.
Exercise · 5 min
Share Less Data for AI-Assisted WritingPrepare a useful writing brief with less sensitive context.
Build with AI
Connect instructions, proposals, review and release evidence.
Guide · 6 min
Build with AI: keep the instructions connectedFrom README to proposal to tested change.
Exercise · 6 min
Write Tests That Catch Mistakes in AI-Assisted CodeRun six local tests and deliberately break a date boundary.
Exercise · 6 min
Turn Verified AI Findings into Reviewable ProposalsConnect evidence to an existing work item and release checks.
Existing article
A beginner’s development workflowBackground reading. Recheck current versions and deployment instructions.
Control app integrations
Understand setup, permissions, context and available tools.
Guide · 7 min
Controlling App Integrations: What's Implemented, What's PlannedThe blog-post version of the app-control and setup-and-api pages, in one place.
App documentation
OpenTechnologyApp AI controlsCurrent source capabilities, setup checks and planned work.
Setup guide
Setup and API accessResponsibilities by permission level and secure test setup.
Choose open source
Evaluate freedom, fit, maintenance and an exit plan.
Guide · 6 min
Open source: benefits you can evaluateA practical pilot and maintenance decision.
Run local and hybrid AI
Choose where work runs and measure the whole workflow.
Existing article
Start learning about hybrid AIBackground reading. Model, pricing and setup details need a fresh check.
Existing article
Explore a hybrid AI workflowA previous workflow example, not a current price or quality guarantee.
Follow AI changes
Turn announcements into checked, useful decisions.
Guide · 5 min
Read AI news as an operatorAn evidence checklist for future AI and company-integration coverage.
Try the practice pack
Review an incomplete summary, reuse five synthetic evaluation cases, or run a local code exercise. The cases include a reviewer answer key; send only the input to an approved AI tool. No provider calls or model scores are included.
The code exercise uses Node.js. Read the instructions and expected results first.
Inside OpenTechnologyApp
Documentation is part of the control layer.
README and setup guidance orient the work. Agent instructions guide proposal and implementation work. Tests and release evidence support the docs people use. Server permissions govern what the app can execute.
Read the app control guideBuilt to keep learning
Future coverage will examine AI changes and company integrations: what was announced, what is available, who it affects and what needs testing. The news track starts with an evaluation guide; dated news coverage will be added after source review.
Open-source choices are one part of this library. OpenTechnologyApp has its own product licensing terms; using open-source tools does not make the app itself open source.
See product and self-hosting options