Adam Sandler

@aliByteCode Get your mom to join

everyone says "just launch it" but once you launch you get 3 visitors 2 are your friends 1 is you checking analytics now what

@brockpierson Why not? Appreciate you if this works :)

I am blowing all super small accounts Reply if you're under 5k and I will boost you 🚀

@theaaron Super impressive!! Great luck at the speaking event.

I used to scroll YouTube every morning looking for something worth watching while I workout. Sometimes I'd find something good but usually I'd waste 10 minutes picking, then half-watch a mediocre video. So I gave my agent a skill that doom scrolls YouTube for me. It uses the YouTube API to search for videos based on my projects and calendar. Each video gets its own page in a @NotionHQ database with a learning exercise that connects the content to whatever I'm actually working on. And it's ready every morning before I go down to the gym. An example of how powerful this is: I'm speaking at a conference this week, so it started pulling public speaking videos. It knows I've been deep in @claudeai Cowork, so those topics have shown up too. All from my project file and calendar. I didn't tell it to do any of that. This took about 30 min to set up, and has already made my workouts more efficient and the videos I watch far more actionable.

Learnings Feed reflects the new direction:

The good news is that it's working

@ItsCasperonX I'm testing this directly in Claude Code CLI so I can use my max plan.

@TheViableEdge I tried doing this too.. I called it the Arena, and I blew through my API credits in no time without any results... How do you make this work on a budget?

There's no shortage of people building AI agents. The shortage is in people testing them. I run 45 rubric-graded tests across 15 marketing agents. Each one gets its real system prompt with live strategy context, brand voice, and automation schedules injected. A separate model judges whether the agent actually used that context correctly. Weighted criteria, must-pass gates, domain boundary checks. If someone's building AI systems for your business, ask for the eval. Ask what "quality" means and how they measure it. The methodology matters more than the demo.

Why not, talking to myself while building and testing agentic marketing workflows communicating in slack. Edited out Uhs and gaps

@interplato @coreyganim I believe you can call skills via Claude API's code execution tool, I would start there

@coreyganim @TheViableEdge @coreyganim how do I build a consumer facing web tool that draws its logic from a specific skill I trained? In your example if I had to open this up as a competitive research tool to a consumer how do I train it to think on the basis of a research skill ?

How to build AI tools without writing code: (Using Claude Code for competitive research as an example) Step 1: Start with a vague ask → "Find distressed wedding venues in North Carolina" → Don't overthink it. Claude will search the web and find patterns you'd miss. → It found social posts about "stranded brides" and owners going to prison. Step 2: Turn chaos into structure → Ask Claude to create a CSV with a schema → Now you have sortable, filterable data → This is what separates "I asked ChatGPT" from "I built a tool" Step 3: Use Plan Mode (Shift+Tab) → Claude interviews YOU before doing the work → It asks: "What level of control do you want?" "Acquisition or partnership?" → Fills gaps in your thinking you didn't know existed Step 4: Add a scoring rubric → Claude analyzed each venue for acquisition potential vs partnership fit → One scored 4.2/5 for acquisition (owner in prison, foreclosure auction) → Another scored high for partnership (burned out owner, good property) Step 5: Visualize with Playground → Type "use the playground skill to create a visual dossier" → It builds a full HTML dashboard with filters, tags, and next-step recommendations → One self-contained file you can share with anyone Step 6: Turn it into a skill → "Create a skill that repeats this for any state" → Now the entire workflow runs with one command → Research → Structure → Score → Dashboard. Automated. The beauty of this process is that you're building tools that compound. Full walkthrough with @TheViableEdge in the video below. (the Build With AI podcast is also available on all podcast platforms)

Had a blast building with Corey. We built a research assistant that replaces what would typically run $100K/year in analyst time, built with Claude Code, free to replicate. Full walkthrough in the video.

@francois352 I haven't even tried the others, I'm curious about Telegram and will probably add that integration in at some point.

@TheViableEdge Nice, I discovered Slack a week ago after lots of deception and frutration with Telegram and Google Chat

Slack integration has changed my whole way of working in a short amount of time. I’ve struggled with remote control in Claude Code being inconsistent, and Slack has given me control and access to spin up new local work trees and kick off features before I get out of bed. Also, developing my platform using my platform. Sounds gross when I type it out loud.

Wow, this was unexpected - Dispatch, Cowork on the go! Settings to keep your computer awake and enable browser control. I asked to analyze my GA, and it sent back screenshots. Pretty pretty cool. Don't judge my shitty GA numbers, mmkay?

Thank you so much, it's a year's worth of work behind the scenes. It all started as a Claude Code CLI based workflow leveraging agents, skills, and connectors which is still the underlying architecture. I'm maintaining a CLI version of the marketing workflows as well to keep the core system in tact.

@IMJustinBrooke @TheViableEdge This is really cool. I've been working on an internal tool like this. You've done a really great job with it.

Just got off the phone with this guy... Super smart, comes from the world of enterprise marketing. Used to work with Amex. Love what he's building.

@DawsonSchrader @YouTube Thank you so much, kind words in the morning like this are a bigger boost to productivity than Claude.

@TheViableEdge @YouTube Yo, this is really cool. Keep going.

I Turned Claude Code Into a Complete Marketing Organization youtu.be/IbwtWkNTQbk?si… via @YouTube

Goodnight

My marketing agent took it upon itself to write a week's worth of LinkedIn posts for my client. Did I make AGI?

I tried out Viktor and was super impressed, however, I was able to build a slack integration from my local Claude Code that runs even when the laptop is closed and it's been a similar experience. Viktor's orchestration is extremely impressive, but as a habitual Claude Code user I can't justify the cost.

The fact slack didn’t make a Viktor style app blows my mind

How do you stop building features? Asking for a friend. This is actually useful and perhaps even critical for marketing context management. In testing the RAG system, I was still struggling to achieve the precision I want for Agents to leverage brand context. This lead me to exploring the relationships of the different inputs and resources that are fueling marketing workflows, and understanding that RAG only addresses half the issue. The data is more efficiently retreivable, but relevance is still very loose. I've restructured schema of all dynamic information and data to have defined relationships, and slapped on this knowledge graph visual. Here, I have an ICP in focus, and can see all of the channels, resources, content, etc that the system thinks is relevant for it. Next is testing the output quality.