Hi everyone, I wanted to post this because I asked a question about setting up guardrails for AI and found @Bas Rosario’s quite interesting. I was wondering if you could share it again, so others who weren’t at the member onboarding today can also learn from it. Also, I want to ask others to add to this idea of setting up guardrails. How do you go about it? My way of doing it was to write ‘don’t do X because of Y’ in a folder/file. Maybe even create a separate file for that? (Not sure if that’s a good idea because it sounds inefficient, but I have no proof on it). I reasoned that doing it this way adds another layer that prevents the LLM from accidentally doing X, because it’s coupled with an explanation, which might work better for reasoning models that are known to take action based on what they think is the best option. If you provide an explanation, the reasoning models might view that as a more valid reason for not acting upon it. However, I haven’t read much into AI research, so this is all hypothetical and speculative. Would love to hear everyone’s ideas around creating guardrails for AI.

📦 EVERY ENTRANT GETS A FEEDBACK FILE 📦 🔍 HOW WE READ THESE We cloned every repo and pinned it to its last commit before the deadline, so nobody got judged on late work. Only two repos had any post-deadline commits at all. Cleanest field this competition has had. Then we read file by file. Identity, rules, examples, the reference layer, the code. We ran every checker in the field on our own machine. Three would not run. For eleven entries we went further. We cloned the territory, or the entrant shipped it, and checked the cards against the source. Eleven is all the field allowed. Everyone else pointed at something private, unpinned, or gone. Two of those eleven were public codebases nobody in this community controls. On one gate we planted a bad citation to see if it would fire. Comments got you through the door. The judging happened inside the folders. 📚 WHAT THE FIELD TAUGHT Three lines split forty builds: ✅ The one rule held. The brief was blunt: a README that tells people to add every file has failed. Not one entry made that mistake. Every search hit was either a rule against it or a fixture broken on purpose. Name a failure precisely and this community stops doing it. ✅ Enforcement is now the baseline. Sixteen entries ship a checker that runs offline with no API key. Last cycle that was the headline. This cycle the checkers got better. One broken fixture per gate is now normal, and the best gate also ships a clean fixture to prove it stays quiet on honest prose. ✅ The cold walk decided the round. About half the field shipped a receipt of a reader entering a catalog, opening one card, and stopping. The top of the field is almost exactly that set. A map can be accurate and still hard to walk. This week we scored the walk. 🥇 THE WINNER @Adam James The Cartographer, a map of a live GoHighLevel account, built so a VA can change one message without guessing which of a dozen lookalike workflows owns it. He wrote down his test method before any test ran. Then he did what almost nobody does. He handed his work to people who could make him look bad, and let them.

I've written a draft paper and I'd like feedback on it. It came out of the learnings & engagements in this group, so this is the right place to put it first. Some context on where I'm coming from. I hold a PhD in anthropology and sociology, a post-doc from the London Business School and after I was not able to pursue my academic career (no position, family life), I pvioted into AI and spent the last four years working in AI, currently leading AI governance inside an international humanitarian organisation. Between the fieldwork and the AI work there has been a fair amount of drift, and I have nothing published on any of this yet. Writing is how I think, so this got written alongside the practice it describes. The argument in one sentence: agentic capability is a property of the arrangement of model, task and environment, and the half no vendor can sell you is the environment, which for organisational work is a description of your own organisation. How work actually gets done, which rules apply, where the knowledge sits, what must never happen. Most of that has never been written down, and it can't be commissioned from outside, because the people who hold it are the people doing the work. ICM is in there, named and engaged with. It's 26 pages. Read whatever is useful. Feel free to drop the PDF into your own workspace and have Claude go at it, which is a fair test of the thing, and I'd rather see a hard read than a kind one. Thank you. Also open to any future collaborations.

This was the first community I joined, and my first real go at AI beyond just talking to a chat window. I found it pretty much by chance, lurked for a few months, and High Tea 6 was my first call. I didn't realise at the time how much that would shape things. The picture I got here was a healthy one. Build the thing, watch how it actually behaves, be honest about the bits that don't work. I've spoken to enough people since who came in through paid courses to know that isn't the norm. Nobody here sold me a shortcut, and I got pretty lucky landing here first rather than somewhere else. Something happened recently that made me take a step back. I met someone at an open networking event who runs corporate training and is writing a textbook alongside it. What she needs is her course turned into per-module walkthrough videos - her own slides on screen, voiceover over the top, one video per module. I built it as ICM, because that's just how I build now. Numbered stage folders, one agent reading the right files at the right moment, markdown carrying the context, and local scripts doing all the mechanical work that never needed a model in the first place. Her voice cloned from a module she'd already recorded, the deck walked and exported as frames, one script and one audio file per slide. The durations of those audio files become the cue sheet that stitches the whole thing together, so there's no separate timing step to get wrong. One module's through it end to end. 21 slides, 21 scripts, 21 audio files, one finished MP4. And there's around ten hours of modules to put through the same pipeline. As an 18-year-old that's a really good opportunity, and it's going to generate me some serious income. I wouldn't have built any of it without what I picked up in here first, so thank you to everyone who's answered a question of mine along the way. Nothing's finished yet. I've still got a lot to come and plenty I haven't cracked. It just felt worth stopping on rather than quietly getting on with the next thing.

Hi everyone! Finally, I feel like I can share what I’ve been building behind the scenes. AND I AM FUCKING PROUD OF THIS ONE! 🔥 What do you think something like this is worth? I’ve built an end-to-end contract automation pipeline that runs 100% inside the Microsoft ecosystem. The best part? There is ZERO AI in this flow. Just pure, deterministic, enterprise-grade automation that works flawlessly every single time—and costs next to nothing to run! ⚙️ How the Flow Works: Drop & Trigger: You drop a .pdf file into a specific SharePoint folder. This instantly triggers a Power Automate flow that extracts the data and generates a draft contract. Interactive Teams Approval: The person who uploaded the file receives an automated notification in Microsoft Teams. Right inside Teams, they can: - View the uploaded source file. - Review the generated draft contract. - Approve or deny the contract. - Edit key values directly (edits apply to the final executed contract). E-Signing (No Third-Party Fees): Once approved, the contract is automatically emailed to the customer with a unique link to review and e-sign. - Customers can sign using a mouse, touch screen, stylus, or by uploading a signature image. - Note: You can plug in Adobe Sign or DocuSign, but why pay per signature during testing when you can build it yourself? 😉 Legally Sound Audit Trail: The system captures and logs full verification parameters: [Signed By], [Signed At], [Signer Email], [Signer IP], [Signer Device], [Signature Method], [Document Hash], [Reference], [Approved By], [Approved At]. Final Delivery & Auto-Cleanup: Once signed, the finalized contract is emailed to both parties. The flow then automatically deletes all temporary input and output files, leaving a clean footprint where only the two emailed contracts exist. 🛡️ Built for Enterprise & Clients Bulletproof Reliability: Built-in safety features prevent duplicate entries, invalid file formats, or missing data fields.