-
ChatGPT at Work 9: GPT-Live Can Listen While It Talks
GPT-Live can listen while it speaks, making AI conversation more natural. A classroom interview explores the promise and the need to check its claims.
-
ChatGPT at Work 8: To Chat or To Work? That Is The Question
Use Chat to think through a problem and Work to delegate a deliverable. The difference involves the tools and process around the model.
-
ChatGPT at Work 7: Editable Writing Blocks
The newer inline writing block makes collaborative editing faster, more natural, and far less clunky than the older Canvas interface.
-
ChatGPT at Work 6: The First Response Is Slop
Useful iteration means blunt feedback, multiple attempts, strategic restarts, and knowing when to try another model—or stop.
-
ChatGPT at Work 5: Verify Before You Act
ChatGPT can assist with consequential IT work, but people must still verify evidence, protect data, and approve the action.
-
ChatGPT at Work 4: Turn Good Prompts Into Workflows
The value of ChatGPT grows when a successful prompt becomes a documented, measurable workflow another person can run.
-
ChatGPT at Work 3: Start With the Business Question
Useful AI analysis starts with a business question, clean data, and a deliverable someone can act on.
-
ChatGPT at Work 2: Make It Show Its Sources
Business research with ChatGPT depends on current sources, visible limitations, and a clear line from evidence to decision.
-
ChatGPT at Work 1: Define the Job
A useful ChatGPT prompt defines the job, context, constraints, and form of the finished work before generation begins.
-
Before and After Networks 10: The Telephone, From Circuits to Packets
The phone call’s quiet revolution: from a reserved circuit to best-effort packets, with dates.