Give the task
a better beginning.
Practical prompts, careful research and useful drafts.
Start with a real job. Leave with a way forward.
Prompt Craft
The Four-Part Prompt Brief: Turn a Vague Request into a Useful Draft
A useful prompt describes a job, its context, its boundaries and the shape of a usable answer. This reusable brief helps you supply all four without writing an essay.
Read the guide ↗
Research Notes
NotebookLM Source Selection: Build a Smaller, Better Research Packet
A focused collection of sources makes an AI research session easier to audit. Build the packet around one question rather than importing everything you have.
Read the guide ↗
Research Notes
Use NotebookLM Discover Sources Without Outsourcing Source Selection
Source discovery can speed up the start of a research task. The useful result is a shortlist to inspect, not a collection to trust automatically.
Read the guide ↗Notes for the next piece of work.
New guides join this collection as they are published.
According to Yonhap News Agency, Erdogan stated that South Korea might participate in Turkey’s Sinop nuclear power project.
Seoul, Nov 2 (Reuters) – South Korea could become a partner in the Sinop nuclear power plant project that Turkey plans to build, Turkish President…
Get More Useful AI Feedback on a Presentation Outline
A presentation outline needs a clear audience, one purpose and a sequence that makes sense. Ask for targeted feedback before asking AI to decorate the…
Anthropic adds intelligent reporting, employee surveys, and shared budgeting features to Claude Enterprise
Anthropic has introduced several new features for Claude Enterprise, designed to help organizations leverage AI tools more effectively while keeping costs under control. These updates…
Turn a Long Reading List into a Research Plan
A reading backlog becomes manageable when every source has a role. Sort the question, evidence and remaining uncertainty before asking AI for another summary.
A startup founded by a former Dogecoin employee reportedly claims to have used AI agents to acquire sensitive anthropological data.
Cathedral, a cybersecurity startup founded by former employees of the U.S. Department of Government Efficiency (DOGE), is facing new scrutiny following allegations that it breached…
OpenAI has resumed training of frontier AI models following upgrades to its safety systems, and has disclosed a vulnerability in a widely used piece of software.
OpenAI stated on Friday that it has resumed training its most advanced artificial intelligence models, following the implementation of strengthened security measures and the disabling…
Ask AI for a Counterargument Without Creating False Balance
A counterargument can expose a weak assumption. It becomes less useful when the model invents evidence or treats every position as equally supported.
An ergonomics/human factors engineering firm has proposed an AI safety bill that could prevent the deployment of a model capable of causing more than 50 deaths.
Over the past year, concerns regarding the potential dangers of artificial intelligence have grown, and debates over the direction of safe AI development have intensified.…
Examples you can question.
Our guides distinguish documented features from suggested workflows, and leave room for uncertainty.
How we put a guide together →