AI
Engineering
Can You Trust AI's Answers? Building a Maintainable AI Eval Set
Learn how to build and consistently maintain evaluation sets to ensure your AI models are working correctly.
Tag: Prompt Engineering
Prompt Engineering
AI Utilization
Talking to AI: Is Writing Longer Actually Worse? The Secret of 'Short and Clear' Prompts
We explain in simple terms why short and concise commands yield better results when talking to AI, exploring the language efficiency of prompt engineering and the latest research findings.
If You Give an AI 100 Books at Once, Will It Read Them All? The Trap of the 'Large Context Window'
AI companies boast about massive context windows of a million tokens, but AIs actually suffer from a 'lost in the middle' problem where they forget information in the middle. We explore why small, accurate information is important.
Can We Set a 'Thinking Time' Before AI Answers? The Mechanics and Hidden Costs
An easy-to-understand explanation of the principles behind the 'Thinking Effort' adjustment feature in the latest AI models, the concept of chain-of-thought, and the resulting correlation between wait times and costs.
AI
Learning Methods
In the Age of AI Homework, Why Are Some People Using AI for 'Real Learning'?
Explore how to use AI like ChatGPT as a 1:1 tutor rather than a 'vending machine' for answers, turning new information into true personal knowledge.
AI
Agentic Patterns
AI Planning and Choosing Tools on Its Own? The Secret of 'Agentic Patterns'
Going beyond simple chatbots, we explain in an easy-to-understand way what 'Agentic Patterns' are—the core of autonomous AI that selects tools and performs tasks on its own.
Is the Way AI Writes Documents Changing? The Era of Chatting via 'Web Pages' Instead of Text
Lately, it's becoming a trend among AI developers to receive AI responses in HTML instead of plain text or Markdown. We explain this interesting shift sparked by an Anthropic engineer and the reasons behind it in an easy-to-understand way.