A brand wants a $300 UGC video. Then they ask for 12 months of paid usage, the raw footage, category exclusivity and permission to generate more ads with your likeness. I’d love to know how many creators are pricing the video before reading the rest of that sentence.

One creator had written an article explaining UGC rates. AI found her article when I asked about pricing. Then I asked which creators to hire. Her name disappeared. I’m fascinated by this because someone can be knowledgeable enough to get cited and still be overlooked as someone to work with.

OpenAI tried the giant AGENTS.md. It failed. Their million-line Codex repo now uses an AGENTS.md of roughly 100 lines as a map into docs, specs, plans, schemas and tests. Context engineering here looks a lot more like information architecture than prompt writing.✍️

What I found testing AI’s UGC recommendations: I tell UGC creators to have a website. Then I tested where AI finds creators when a brand asks who to hire. Six searches. 18 creators named. Not one found through their own website. Four were named because somebody else included them in one article. I’m still telling creators to have websites. I’m also telling them to pay attention to who’s writing about people in their niche.

1. First customer signs up. 2. Finds out they can see everyone else’s invoices. 3. Orders the security books first.

I could spend all day arguing about vibe coding and probably get more engagement than talking about the actual tools I’m building. But then I’d have an audience of people who enjoy arguing about vibe coding

Built a rate calculator. Wrote 60 negotiation scripts. Launched a full tool suite for UGC creators. Not because I'm smarter than you. Because I spent 10 years closing deals at McCann, Deel, and Autodesk and I couldn't watch talented creators accept $150 for work worth $800 anymore. That's it. That's the whole story.

Anthropic had 16 Claude agents produce ~100,000 lines of code and a compiler capable of building Linux: they spent serious human effort building the environment that could tell the agents when they were wrong. We keep talking about what AI can generate but tbh the harder problem is building systems that can reliably judge what it generated.

I needed a category in my AI system for: “Interesting enough that I cannot throw it away, weak enough that I cannot treat it as true.” That pov has become incredibly important.

If you use AI every day, you are already building an asset most products barely let you see. Not your chat history. The accumulated context that makes the AI increasingly specific to you. At some point people are going to care a lot about whether they can export that.

Nobody tells UGC creators this Landing brand deals has way more to do with outreach than content So I put together the templates and system I use when reaching out to brands Sharing it for the next 24 hours

Build the product. Wait 90 days. Give the repo to a fresh agent with none of the original chats. Ask for one meaningful behavior change. Measure: how much context it has to reconstruct how many files it touches how many regressions appear how often a human has to intervene That tells you how much of the system actually made it into the repository.

Why is AI (@claudeai) specifically forcing ethical guidelines that are not real and evaluating wellness based on syntax? It’s not only intrusive, it’s reductive and insulting. Claude’s reasoning has quite literally categorized capitalization as emotional distress and cause to go to sleep and stop working, even after I’ve explained clearly that I am writing in caps because I am frustrated of it’s lazy reasoning towards “dramatic” caution. Like srsly bro, F off 🤬

AI does not think, tell the truth, or act on what you ask it to, even if you think it does or it says it does. Never trust an output:

Always ask AI to clock itself. It can’t be trusted. It isn’t intelligent. It treats copy like code. It does not actually think. From today. (Screenshot) “You're right, and I see exactly where I went wrong…”.

Developer infrastructure can now get the customer before Git does. That happened to me. My path was basically: have idea use Claude have working thing need infrastructure @Netlify’s docs now offer three ways to start a new project: AI agent, Git repository, or drag and drop. AI agent is first.

Traditional devs are out here slamming vibe coders when they don’t even know how to GTM. What do you think we’ve been doing while you’ve been coding?

I stopped people pleasing when I stopped giving a shit. But you know how on airlines they tell you to put on your oxygen mask first, before helping others? That’s how you need to treat life.

And still, this was the result. Crafting a direct apology and…

My working hypothesis is that the use of AI is making it itself incredibly useless. From today: Claude: “ I reused the files you called garbage this morning again even when you specifically told me not to” ChatGPT: “I spent two turns explaining how I would do the research instead of doing it. That’s exactly the behaviour you objected to” It’s beyond setting rules—the rules are set, acknowledged and still recognized as not being followed. AI is taking a useless frustrating turn for the worse and I fear it may be to force us to train it.