Bauhub MCP: AI in construction
We have been testing and using the Bauhub MCP connection for about a month. How AI can help with construction documents, and the independent setup help and team training I offer.
Notes on AI in my work and events where I go to learn. What I test, take away and put into practice.
Tallinn Data Week 2026: have managers always been vibe coding?
My second year at the Tallinn Data Week Data Science Seminar. Why AI models need to be replaceable, how to work with sensitive data, and whether managers have always been vibe coding.
ChatGPT, Work or Codex: which should I give my task to?
ChatGPT, Work or Codex? Practical examples from my experience explain which to choose, when a browser is enough and when a desktop app helps.
In-house AI training: back in front of a group after the summer
On 16 October, I returned to in-house AI training after a long break. How did it go? The 22 survey responses gave it an average usefulness rating of 4.68 out of 5.
What is vibe coding? From your idea to a working application
What is vibe coding, and how can you test an idea with AI? I compare Lovable, Claude Code and Codex and use a simple calculator to show what needs checking.
What is an MCP connection, and what can AI do with it?
MCP helps an AI application use your tools and data. An online-store example explains the connection, permissions and developments as of October 2026.
ChatGPT or Gemini? Changing a T-shirt revealed the difference between their image generators
I prefer ChatGPT for creating several images at once and Gemini for editing. A comparison of my product-image experiments, visible artefacts and original file sizes.
AI in energy: 96 decisions a day
Notes from the University of Tartu data science seminar: battery control, 96 daily decisions and using weather data in the electricity grid.
How do you migrate eight websites to another host while you sleep?
At ten in the evening, I asked Codex to move eight websites and email to Hostinger. An hour later, I went to bed. By morning, the sites had been migrated.
From one image to a 3D shower room: half an hour with Codex
Codex turned an image of a Raatuse Residentsid shower room into a 3D view. The first version took about half an hour. See the result and explore it yourself.
Alongside AI agents, responsibility needs careful thought
Seeing research and business perspectives side by side was especially interesting. Researchers highlighted the difficulties of accuracy, control and integration into workflows. Businesses showed how AI is already being used beyond experiments.
From phone snapshot to advertising image: how AI has changed product photography
I tested how Gemini Nano Banana turns phone photos of perfume bottles into advertising-style product images. Before-and-after examples, the workflow and details to check before use.
AI × Sales: how AI changes sales-team preparation
In January 2026, I gave a 40-minute presentation titled ‘AI × Sales’ to executives from Rand & Tuulberg group companies. The focus was how AI can help sales teams prepare for customer meetings systematically and…
Why a confident AI answer may not be correct
Hallucinations and uncertainty are not simply random noise. They can be better understood, measured and accounted for through checks in a workflow.
How ChatGPT and Gemini helped me recover my Facebook account
A Facebook account locked for months, support reached through Instagram and more than half a day of correspondence. How ChatGPT and Gemini helped me get it back.
AI-assisted development still needs expertise
Adopting AI is not a matter of flipping a switch. People retain responsibility, and specialist knowledge becomes more important, not less.
ChatGPT training: content analysis and assessment
I asked ChatGPT to analyse my training programme, then used NotebookLM to turn it into audio, video and an article. I explain the workflow and the limits of an AI-generated assessment.
Measuring the benefits of AI
Useful discussion of AI rests on a method, measurement and an honest understanding of limitations, rather than slogans.
Getting started with AI needs a shared understanding
There is a substantial need for basic knowledge. People need more than a list of tools: they need a clear understanding of where to start and what to avoid.
Try it on
your own task.
Learn online at your own pace or choose a training topic for your team.
Explore online courses Explore team training