Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
5 contributions to AI Automation Society
Day 2 done ✅ #AISChallenge
What I built: instead of Firecrawl I used an MCP server I already had connected, Claude in Chrome, and mapped the whole AIS Classroom into a CSV. 17 courses (15 I can open, 2 are AIS+ only), 258 lessons, with section, video length, done or not, and a direct link to each lesson. Same pattern as the lesson: a short CLAUDE file, a separate cheat sheet for the MCP tools, plan first, then the extraction, then a check that every course has exactly the lesson count the Classroom shows. One thing I learned: the self-correction from the video happened for real. The first approach got its output cut off, the second was blocked by the tool's own safety filter, the third worked. Each time it changed the method instead of retrying the same thing. One thing I'd improve: video length only shows up for 14 lessons in the page data, so the rest of the column stays empty instead of guessed. I'd fill it by opening each lesson only when we actually plan to watch it. One use case: for a client who imports solar pumps, pulling the specs of each model from the supplier's site into a table, so the grant paperwork matches the product data word for word instead of retyping it from PDFs.
Day 2 done ✅ #AISChallenge
1 like • 3h
@Jason Elam Good call. One of those pump entries came back from review today because a single line referenced the supplier's offer. Now each entry's text comes straight from the signed translation of the price list, and gets checked against it before it goes in.
0 likes • 1h
@Wasilis Paliakoudis It does stop it. The check exits with an error, and the render only runs if it passes. Earlier today a | tail I had added to the command hid that exit code, and the render went ahead anyway. Now the two steps are chained with && and nothing pipes the check's output.
Day 3 done ✅ #AISChallenge
What I built: an infographic mode inside my own video skill (reeler). I trigger it with "make an infographic from report X" or /reeler. It reads a markdown report, writes the numbers and short texts into a small JSON, and renders a PNG from HTML in local Chrome. Same idea as Nate's infographic builder, with one change from the start. I need the numbers exact, so the image comes from code instead of an image generator. The AI-generated version is still there on request, but every digit and letter gets checked before it's used. One optimization after watching it work: a guard. Before rendering, a script checks that every number in the JSON exists in the source report. On the first real run it stopped the two totals, because they don't appear in the report as such. Now a total only passes if the sum of its column actually adds up. That check runs as a script, so it doesn't depend on the model getting it right. Six runs so far, and watching them showed me what to fix: an orphan card in the calendar, a sliver on zero-value bars, long names wrapping, and the bottom line cut off because the size was measured before the font loaded. For Dive Deeper I added the MCP part: the same JSON goes to ChatGPT through Claude in Chrome, and the generated image comes back to be checked against it. First try, GPT also chose to build it from code. When I asked for the image generator, every number and word came out right, but the bars were only roughly proportional and the size was 1024×1536 instead of what I asked for. Both images below, same data: code on the left, GPT on the right.
Day 3 done ✅ #AISChallenge
1 like • 2h
@Jason Elam In the code version each bar's width is computed from the same number as its label, and the guard rejects a bar value that isn't one of the row's numbers. The GPT version is where your check is needed, since its bars came out only roughly proportional. Measuring each bar in that image against its value is the next thing I'd add.
Day 1 Newsletter build.
Day 1 done ✅ #AISChallenge What I built: a newsletter automation on the WAT framework. One workflow file, three small deterministic tools: an infographic rendered from HTML with headless Chrome, an inline-styled HTML email, and a Gmail sender. The content comes from our own study notes, so the newsletter is a weekly "what we're learning here, and what we already applied". Screenshot below. I swapped a few parts of the stack: no Perplexity (the content is our notes), images generated from code instead of an image API, and normal permissions instead of bypass. One thing I learned: WAT gave a name to something I was already doing with Claude Code skills (instructions in Markdown + scripts that do the work). What clicked: when something breaks, fix the tool and write the lesson into the workflow, not just rerun it. It happened twice today: the app password pasted with quotes and spaces made Gmail drop the connection, and the inline image showed up as a "noname" attachment. Both are fixed in the tool now, and both are written down in the workflow. One thing I'd improve: host the image instead of attaching it (Gmail still shows inline images as attachment chips), and add a small set of test inputs before letting it run on a schedule. On to Day 2 🚀
Day 1 Newsletter build.
1 like • 22h
@Wasilis Paliakoudis Great point, thanks. That's exactly the trap: the email looked complete while a number could be off. I'll add a check that compares the values in the rendered email with the source notes, not just that every section is there.
0 likes • 6h
@Tariq Osmani Good one, thanks. A connection that drops silently between runs is exactly the failure I'd only notice from a missing send. I'll add a daily check that just logs in, sends nothing, and pings me only if the login fails.
Nice 2 meet u all
Hi everyone! Marius here, from Romania 👋 I'm a web developer and I work with Claude Code every day. Each client project has its own small network of AI agents (site, CRM, reception app, mobile app) that share a mailbox, plus a central "HQ" that keeps memory, rules and decisions for all of them. Nate's kits were a big influence. I studied AIS-OS, the edit kit and scroll-craft, and adapted the ideas into my own setup (decision logs, one improvement per cycle, gated processes). Career goal: let AI agents handle the repetitive work, so I can focus on design and the decisions that matter. Happy to be here and learn from you all!
0 likes • 9h
@Glen Hutsell Thanks Glen! Agreed on the decision logs. I added mine after explaining the same "why" twice, and now the agent reads it before touching anything. One folder per area sounds like the same idea, does each folder keep its own log too?
0 likes • 9h
@Anthony Manganiello Thanks Anthony, nice to meet you too! 🙂
🏆 Weekly Wins Recap | October 24 – November 13
Three weeks. A lot happened. From 11 × $5.5K clients and first recurring retainers to products shipped, businesses transformed, and members using AI to solve problems they once would have paid thousands to outsource - AIS+ kept moving. 🚀 Standout Wins of the Catch-Up 👉 @Franklin Martinez signed 11 of 23 contractor leads at $5,500 each, using a repeatable CRM setup that now takes just 1–2 hours per client. 👉 @Brett Gavaghan locked in his first month of recurring retainer revenue and followed it up by sending proposals across logistics, construction, and building services. 👉 @Mike Thomson built a video-production pipeline that turned an estimated 104 hours of work into around 5 — shipping 24 avatar lessons plus a YouTube video in one day. 👉 @Sachin Mishra went from café owner to builder in about two weeks, launching his own website, rewards app, vouchers, and local SEO systems. 👉 @Nicole Jolie avoided a $10K+ accounting quote by using AI to handle the books across two companies — then pushed her own sites to 92+ AEO scores. ⸻ 🎥 Super Win Spotlight | @Erfan Atoofi Erfan’s story has moved far beyond “learning AI.” He walked into a 60–70-year-old Japanese company still processing handwritten fax orders and scoped an entire department from scratch. The company had already been quoted around $75K just to update its existing software. Erfan showed them a different path. He mapped the workflows, captured the tribal knowledge, demonstrated the ROI, and built enough trust that the client wired him $52K on day one — before the system was even built. For perspective, the overall deal was worth roughly three years of his salary. And now he’s preparing to transition from employee to full-time business owner. His biggest realization? The skills matter. But trust, understanding the business deeply, and being willing to stand behind your value are what turn those skills into life-changing opportunities.
🏆 Weekly Wins Recap | October 24 – November 13
8 likes • 3d
Grats!
1-5 of 5
Danila Marius
3
17 points to level up
@danila-marius-4321
ready steady go

Active 23m ago
Joined Nov 15, 2026
Powered by