I've been experimenting for the last three weeks with Wispr Flow and I have to be honest I don't think I could do without it now. As you can see from the attached screen shots, I use it a hell of a lot for writing prompts and I find it way better than either Claude's microphone or ChatGPT's microphone. It's particularly good at recognising technical words to do with hosting or domain names, things that normally would drive you crazy because after you've done speech-to-text you have to go back through and then edit it all correctly. As you can see from the attached screenshot it looks like I might be doing too much work and my poor friends don't get to hear from me too often. 😄 Anyway if you want to write detailed prompts extremely quickly and with minimal editing (I'm talking almost never), then I highly recommend trying out Wispr Flow. Cheers Jason

Most people use ChatGPT to generate ideas. That's the easy part. The more useful play is to use it as a second opinion on a decision you've already half-made. When you're about to raise a price, ship a new offer, or send a long email you're nervous about, you don't need more options. You need someone to poke holes in the one you've picked. ChatGPT is good at this if you set it up right. Here's the basic shape. Step 1. State the decision you've already made. Don't ask "what should I do", write out the choice in plain terms. "I'm planning to raise my price from $49 to $79 starting next month. Here's why…" Step 2. Give it the context. Customer base, recent feedback, what your competitors charge, why the timing feels right. Two or three short paragraphs is enough. Step 3. Ask it to argue against the decision. Try this line: "Argue against this decision as if you were a sceptical advisor who has seen plenty of similar plans fail. Be specific." That last word matters. Without "be specific" you get generic risk-talk. With it, you get the actual second opinion you came for. Step 4. Decide what to do with what comes back. Some of the pushback will be useless. Some of it will land. The point isn't to follow the AI's advice, it's to surface the angles you hadn't considered before you commit. This pattern earns its keep most on pricing changes, hiring decisions, big emails, and anything where you've already mostly made up your mind. The "give me 10 options" pattern is better suited to early-stage brainstorming. What's the next decision you'd run this on? Drop it below.

Hi everyone! 👋 I’m looking for some advice on optimizing my OpenAI API workflow and would appreciate your insights on two specific issues. Question 1: Tracking Token Usage per API Key How can I accurately track token consumption for each specific API key? Is there a built-in dashboard feature to see exactly how many tokens a single key is burning daily, or does this require a custom logging setup? Question 2: Dynamic File Updates for Assistants I’m using the Assistants API (platform.openai.com/assistants/) but facing a bottleneck with File Search. I frequently need to make minor, real-time edits to the reference documents. What is the current best practice/pipeline for this? Are there any out-of-the-box tools or integrations for seamless syncing (e.g., editing a Google Doc and having it auto-update in the Vector Store)? Or is the only viable route to write a custom script that triggers the deletion of the old Document ID and pushes the new one? Thanks in advance for the help! Pavel

Most prompting advice tells you to write the perfect instruction up front. For everyday business tasks, I think that's backwards. You're guessing at what matters before you've really thought it through, and you end up with a vague answer because you handed it a vague brief. Here's a simpler move: tell ChatGPT to interview you first. Instead of asking for the finished thing, add one line to your prompt: "Before you answer, ask me up to five questions that would help you give a better response." Now ChatGPT does the hard part. It surfaces the gaps: who the audience is, what the goal is, what tone you want, what you've already tried. Then you just answer in plain language. The final output gets built on real context instead of assumptions. This works best on the messy, higher-stakes stuff: a sales email, an awkward client reply, pricing for a new offer, a job description. Anything where the quality depends on details only you know. Two tips. Cap the questions. Five is plenty, or it spirals. And if a question doesn't matter, just say "skip that one." You're steering, not filling in a form. It feels slower for about ten seconds. Then the answer comes back sharper than anything a one-shot prompt would have produced. What's the last thing you asked ChatGPT for that came back generic? That's exactly the kind of task to try this on. Drop it below. Jason 🙌

You can now allow users to upload files directly in your chatbot window. This has been requested a lot, especially for support and onboarding use cases. Here’s what’s supported: JPEG (.jpg, .jpeg)PNG (.png)WebP (.webp)HEIC (.heic)PDF (.pdf)Word (.doc, .docx)CSV (.csv) Maximum file size: 10MB per file. This means your users can now: • Send screenshots when something isn’t working • Upload contracts or forms• Share product spreadsheets • Send mobile photos straight from iPhone • Provide documents for analysis You’ll find the toggle inside your chatbots appearance settings (paid plans only). Enjoy! Jason