I completed Day 7 of Nate Hark’s AIS Challenge and built an AI assistant to support my Instagram marketing content creation. Since I usually work on Instagram marketing as a side business, I created a system that connects with the Instagram API and analyzes which posts are performing well through hashtag research. Based on that analysis, it transforms the content angle into one that fits my target persona and helps generate carousel posts. I also prepared files containing the interview steps needed for content creation, primary information, and catchphrase examples. The assistant can refer to those materials and create posts based on them. It can also analyze what kinds of angles are likely to perform well. On top of that, I created around seven different skills, including idea research, post creation, rewriting, updating existing posts, design planning, and creating AI image assets. I made it so these can be used with just one tap. I also added a skill called BrainDump. With BrainDump, I can put my daily thoughts, ideas, concerns, and mental state into the AI assistant. Then the assistant automatically responds based on my current needs and emotional condition. For example, when I have a request like “I want the assistant to work like this,” BrainDump helps plan how to improve the assistant itself. If it detects that my mental state is getting weaker, it can create a recovery project to help me get back on track. It can also manage my daily tasks and support my day-to-day execution. In the future, I want to expand this into something much bigger: a system that can automate and support many different areas of my life and work. Ultimately, I want this AI assistant to become a personal operating system that helps me think, plan, recover, and execute better every day. By applying what I learned from Day 7, I was able to build an AI assistant that I can actually use for my own Instagram operations and personal productivity. Thank you, Nate.

Which one is actually at the top? In this video, we look at Claude Code vs Google Antigravity to see which is better. I spent 100 hours testing Claude Code vs Antigravity to give you my honest results. These are the most honest results for Claude Code vs Google Antigravity. Hope you enjoy!

There's one free plugin that's completely changed how I use Claude Code. It's called Superpowers, and it turns Claude Code into a disciplined developer that clarifies, designs, plans, codes, and verifies before shipping anything. In this video I walk through all 14 skills, show real brainstorming demos, and share the results from a 12-run experiment comparing token usage and code quality with and without the plugin installed. SUPERPOWERS

Hey everyone, I built a Google Sheet that has every video I've published in 2026 along with the links to all resources, tools, and files mentioned in each one. If you're ever looking for something I referenced in a video, start here: 📌 YouTube Video Database This will get updated as new videos drop. Bookmark it. - Nate

What's up, everyone! Just finished Day 6, and I wanted to share some raw insights from today's implementation. I set up a scheduled task to pull KPI data from Supabase, which led me to a cold reality check and a major breakthrough. - The Reality Check: The PC Constraint Amidst all the AI hype, we often hear that "everything is automated." But the reality with the current Claude Code setup is that it still depends on your local PC being connected. It's a reminder that true 24/7 cloud autonomy still requires a bit more infrastructure or a different approach. - Execution Order: Stale Data is Meaningless The biggest concern during implementation was data synchronization. If the KPI retrieval happens before Supabase is updated, the entire automation is useless. It taught me that in automation, ensuring the "order of operations" is just as critical as the logic itself. - The Breakthrough: The "Self-Improving Loop" The highlight of today was discovering the power of the status.md loop. By having the agent read a status file at the start and overwrite it with "lessons learned" at the end, the automation evolves. It's no longer just a repetitive task; it's a process that gets smarter and more refined with every execution. - Moving Forward: Evolving the "Quality of Skills" I'm now applying this loop concept not just to schedules, but to the quality of the individual skills my agent uses. We're moving away from "set it and forget it" to "set it and watch it improve."