ChatGPT interrupts work: What is behind the recent disruptions?
Documented ChatGPT disruptions, capacity problems and original evidence while using GPT-6.1 Sol: understand the limits and protect work in progress.
Read articleThe TNSystem journal
Good to Know
Understanding IT. Making sense of AI development.
Latest articles
26 articles
The AI sprint in autumn 2026: From chatbot to agent
AI agents in autumn 2026: GPT-6, Claude 5.5 and Gemini, computer use, API costs, application connections and responsible human oversight in business.
Read article
Is AI really getting worse? Why ChatGPT and Claude sometimes lose quality
Repeated mistakes, forgotten requirements and inconsistent results: what evidence shows, and why a poor answer does not prove secret throttling.
Read article
The AI productivity paradox: When faster code creates more work
Why the first output tells us little about real time savings, and how developers can assess AI assistance sensibly.
Read article
AI pricing in autumn 2026: What really costs more?
New AI subscription tiers, API prices and changing allowances compared. Which costs have risen or fallen since September 2026?
Read article
Same bill, less room to work? What users criticise about AI pricing
Why AI users debate prices, allowances and predictability. Personal experiences, documented changes and the cost of completed work.
Read article
Where is AI heading? What the major providers are planning next
A clear look at OpenAI, Anthropic, Google and Microsoft’s AI plans, distinguishing announcements and previews from availability and predictions.
Read article
The future of AI in everyday work: What could change for users and businesses
From AI agents to software development: How emerging AI plans could affect workflows, costs, oversight and the skills people need.
Read article
Software updates without unpleasant surprises
Prepare software updates around real workflows, data changes and a workable recovery plan—not just a successful installation.
Read article
Desktop, web or cloud: Which software actually fits?
Compare software around offline tasks, device access, collaboration and operating responsibilities—not labels or blanket cost promises.
Read article
When software fails: Why monitoring alone is not enough
A green status does not rule out every failure. Respond to software outages with meaningful checks, useful logs and clear responsibilities.
Read article
Software without vendor lock-in: Can we really take our data with us?
Why data exports, open interfaces and a planned supplier change matter when choosing software.
Read article