|
Hi Reader - While LinkedIn bros the world over have been going 'RIP designers,' I've been having more and more conversations like the below lately. Here is what they were expecting me to say Some clever stack of AI tools wired together with prompts. Maybe a word-salad about 'skills' and 'agents' or 'n8n workflows'. While what we actually do is boring. We use the standard AI models(same as your favorite LinkedIn bro). Then we hand all of it to human designers & strategists. The designers & strategists stitch the outputs together. They keep working until the ad stops looking like AI slop. They keep going until it looks like something a human would click on and purchase off of. That human step is the entire game. The AI step is twelve dollars a month. Which leads to the part that breaks people's brains. Our creative production went up 10x this year. We make thousands of ads a month. We doubled our team. I personally spent 50+ hours a week in Terminal. Running every part of our business. From creative briefs to P&Ls. If that math looks completely backwards, congratulations. Some guy figured out why in 1865 and almost nobody's read him since. Detour: a short story from 1865. Coal in 1865 was AI in 2026. Different century. Same vibes. Victorian gentlemen in top hats were writing breathless pamphlets. Their argument: the new steam engines would make human labor obsolete. Industrialists gave speeches at the Royal Society. Letters to The Times of London ran wall-to-wall takes from factory owners. About how the machines were going to 'revolutionize the workforce.' By which they meant get rid of most of it. Every philosopher, statistician, and gentleman of leisure had a hot take. About how Britain was going to need fewer humans. Same prediction. Same confidence. Same insufferable certainty as every "AI will replace your team" post you scrolled past last Tuesday on Twitter. Meanwhile.... Coal consumption was already exploding. The numbers were sitting right there in the trade reports. Nobody wanted to look at them. A guy named William Stanley Jevons did. Then he wrote a book: The Coal Question; An Inquiry Concerning the Progress of the Nation, and the Probable Exhaustion of Our Coal Mines. That book title might be the most 1865 thing ever. Anyway... With what I assume was a very British amount of restraint, he pointed out that everybody had it completely backwards. The more efficient the engines got, the more coal Britain burned. The more coal Britain burned, the more steam-powered factories opened. The more factories opened, the more humans got hired to keep the whole thing running. Because when something gets cheaper and faster to use, you find more things to use it for. The efficiency didn't kill the appetite for coal. It manufactured an insatiable one. Same went for the humans needed to mine it. Move it. Operate the machines that ran on it. This came to be known as Jevons' paradox. I call it the only economic principle anyone needs to understand what AI is doing to creative teams right now. AI is the steam engine. The more efficient our production gets, the more we produce. The more we produce, the more decisions someone has to make. Which AI output looks like a brand and which one looks like a stock photo with extra steps. Which generated character is on-brief and which one has six fingers. Which narrative flows smoothly and which one falls apart the second you look at it. AI does not make those calls. A human does. Here's the part that really stings. AI scales whatever judgment you put in. Good judgment scales into great output. Bad judgment scales into expensive garbage. Faster than you've ever produced expensive garbage in your life. The more AI you use, the more consequential every human decision becomes. We didn't double our team despite using more AI. The AI is the reason we needed to double the team. More designers stitching outputs into something brand-safe. More strategists deciding which angles are worth generating in the first place. More editors killing the 80% of AI output that doesn't pass muster before it ever gets to a client. The industrial revolution ran on coal. Our AI-team runs on humans. Anyway, if you spend $50k+/month on paid and you want to work with a team that actually gets how this works(and is actually growing both headcount and creative throughput), hit reply with "COAL" and we'll talk. Or don't. I'm sure the all new ChatGPT prompt pack will work just fine. Later - Shamanth PS: My latest interview on Intelligent Artifice w/ Patrick Stuart-Constant, CEO at Sociaaal, goes into this exact dynamic - and explains how they were able to scale creative output by hiring more humans. PPS: We've jussst brought back the Mobile UA show - with Intelligent Artifice now being a section on the Mobile UA show website. Same show you know and love - new/old home. PPPS: If you're doing interesting work at the intersection of AI/growth, I'd love to explore interviewing you on the show. Hit reply and lmk what you'd like to speak about - and we'll chat. |
Intelligent Artifice features hard won insights from the front-lines of performance marketing in a Generative AI world - delivered straight into your inbox. For free.
Hi Reader - I was not a programmer. My entire technical background was: ‘hire a developer and hope for the best’. As of a year and a half ago, my big AI move was exporting a CSV from my MMP, uploading it to ChatGPT, and asking it questions about my performance data. I thought I was ahead of the curve. I was very proud of myself. Turns out I was doing pivot tables with a chat interface on top. Something changed though. I started writing code with AI. I now spend 35 to 40 hours a week on my...
Hey Reader, Every week I pick apart subscription app ads on LinkedIn. What's working. What's broken. What could be fixed in thirty seconds. Hooks that tap into the lived experience of users(and those that are written by brand managers that have never spoken to a user). I've done this long enough: 50,000+ ads produced, hundreds monthly, $100M+ in spend managed: that the pattern recognition is basically automatic at this point. I've built AI agents to help with parts of the workflow(based on...
Hi Reader, In 1994, McDonald's hired a Harvard professor to help them sell more milkshakes. This man went to Harvard. And ended up in a McDonald's parking lot with a clipboard watching people drink milkshakes. For eighteen hours. I just want to know if he put that on his LinkedIn. He did figure something out though. In the past they'd tried better flavors, thicker consistency, lower price: everything. Nothing moved. So this dude Clayton Christensen just watched. Counted. Asked questions....