Hey kids! 👋 Do you like coloring books? 🖍️ What if I told you that you could get a printable coloring book page—of almost anything you can dream up—delivered to your inbox every single day? Well, that's my latest AI workflow project. And those are my em dashes, so deal with it.

You: someone who builds AI workflows (in n8n, Zapier, or Make) AND you have more than 2,000 followers. Me: about to roll out subscription and monetization on @useplumb and I need a few people to kick the tires. DM me if you're interested. If selected, I will pay you $1,000 (that's right, one rack) to build a flow and promote it to your followers.

On this week's episode of AI Builders Club, we're tackling proposal generation. If you're a consultant or freelancer, you know what a pain proposals can be. What if we built a flow that would automate it?

Went to a coffee shop. Barista introduced himself. Nice guy. Smiles easily. Went back. He said hello. Called me by my name. Now I’ve gone seven days in a row because the barista feels like a neighborhood friend, and that’s how low the bar for service is these days.

Gemini 2.5 Pro is probably the best AI model for speech-to-text transcription right now I'll tell you why you should be doing a lot more voice recordings and getting the transcripts But first let me show you exactly how to do this for free If you have a voice note from your phone and you want to get a better transcript than Siri does, do this: 1. AirDrop or send the audio file to your computer 2. Go to AI Studio on Google on the web and upload the file into the chat. AI Studio is free btw. 3. Select Gemini 2.5 Pro from the drop-down menu where you pick a model 4. If you want speaker labels, which I find EXTREMELY helpful, give it a prompt like this: "Make me a transcript in plaintext from this audio. Assign speaker labels. Nick is me and I talk the most or ask the questions, the other person is Neville." 5. Send the chat and wait Results usually come back with a long transcript in about 1 minute Now let me tell you why you should be doing this (1) The native iPhone transcript on voice memos is usually garbage. It's fine in a pinch but usually you want the better accuracy and speaker labels (2) You want to give the AIs as much context as possible for making decisions. So giving long text transcripts I find is really helpful: they have huge context windows now and can process the data from long conversations Finally here are other options for your transcripts: (1) AssemblyAI was my previous favorite. They have their new Slam-1 model and you can play with it for free, with speaker labels, on the Playground section of their website But Gemini is better. I compared the two yesterday from a doctor's office recording that I made and Gemini easily came out on top, with more accurate speaker labels and better context clues for improved transcription (2) ElevenLabs is great especially for lots of speakers- like transcribing a large Q&A- and they let you add speaker labels names in a better way But their model is just too damn expensive. You can use Gemini 2.5 Pro to do your transcripts inside Google AI Studio for free now, so just do that It's still not easy enough for a normal person. But if you've read this far then trust me, you can easily do it And when you start giving AI long transcripts- of meetings, conversations, or even recording doctor visits- you'll be amazed at how helpful the additional context is to help you get better answers to your questions