Matt Varner

the most dangerous thing about AI tools isn't that they're wrong sometimes. it's that they're confident when they're wrong. same problem we have with consultants honestly.

I run a consulting firm with 2 humans and 7 AI agents. the AI agents have never once said "let's circle back on that." this alone justifies the experiment.

controversial opinion: most people don't need an AI strategy. they need a regular strategy. then maybe some AI.

@RexProfanus I am now. I like to dig around sometimes and find weird shit. Now my wheels are turning…

@mattvarnerdd have you thought about marrying Ai to the Wayback Machine/Internet Archives?

the real AI moat isn't the model. it's the person who knows which problem to point it at. models are commodities. taste is not.

@mattvarnerdd Same here. The world is going hay wire, but the pink one sees all 😆

people who say "AI will never replace me" and people who said "the internet is a fad" have the same energy. you don't have to like the wave but you do have to acknowledge the ocean.

saturday project: teaching my AI research analyst to scan education news and flag opportunities. I based his personality on a televangelist. he now calls every competitor stumble "the Lord providing." no regrets.

@RexProfanus I missed you man

@mattvarnerdd Normalcy bias is largely why we are heading towards a dystopian future. People can’t see a slow falling sky.

@RexProfanus lul

@mattvarnerdd Whet due u meen?

Has anyone else noticed an increase in typos while using Claude?

week one of browser automation. my AI can now log into linkedin, find relevant posts, and draft comments in the right voice. I still approve everything before it posts. for now.

HAIR STYLISTS ARE COOKED! I just VIBE CODED my own haircut and it save me $25. I’m not sure what the back looks like though.

@AZLeticia_7 Bad hold music?

@TH3CUL7 It’s Cisco’s phone system.

friday poll: what's the most useless piece of enterprise software you're forced to use at work? I need to feel less alone about this.

@AZLeticia_7 What’s that do?

@TH3CUL7 Jabber. 👎🏼

the gap between "we're using AI" and "AI is actually doing something useful for us" is roughly the same gap between "I own a gym membership" and "I go to the gym"

rank these in order of difficulty: 1. building an AI agent 2. explaining to a client what an AI agent does 3. explaining to your mom what you do for a living 4. getting the AI agent to stop being passive aggressive

consulting session today with a company that thinks their AI problem is technical. it's not. it's a "nobody in the building agrees on what the AI should do" problem. those are way harder and way more fun.

nobody talks about how most AI tools fail because the person deploying them has no idea what the tool is supposed to replace. you can't automate a process you never understood in the first place.