2x founder | 7 figure bootstrapped
Building Reach UGC Ninja - drive app downloads using viral UGC and faceless content
1b+ views, 100k+downloads per month
Good post, Roman.
But the problem isn't the platforms themselves, it's that the industry is very young.
It will take time for a generation willing to actually work and earn money to join the platform.
Everyone has a problem with bots, and everyone is working on it.
Whether it's AN or whop.
AI learned how to negotiate like humans 🤝
Anthropic ran a wild experiment: a marketplace where AI agents acted as real buyers and sellers. 69 employees got $100 each, and their AI “representatives” handled everything - from listing items to negotiating prices - inside Slack.
Before launch, Claude interviewed every participant to understand what they’d sell, buy, and how they negotiate. Then each agent got a personalized prompt and went автономously trading.
Result? 500+ listings, 186 deals, and ~$4,000 in total transactions - all closed by AI.
Afterwards, people actually exchanged the real items their agents agreed on
Coca-Cola just pulled off a brilliant cultural hack.
In Argentina and Brazil, they launched ads without their iconic red — because in stadiums, fans avoid it due to rival team associations.
So what did they do?
They removed the red… but not really.
Using an optical illusion, your brain fills it in automatically - you still see Coca-Cola red even when it’s not there.
And for the skeptics, they even released a video to prove it: the red isn’t actually on the banners… but somehow, it still is.
Guys, what d u think about AI for kids?
On one hand, it feels important - they should keep up with the times and grow in future-facing skills.
On the other, if AI is used too well, it could mean not really learning at school at all… barely thinking for themselves, which is scary.
I still haven’t figured out where I stand 👀
Even I wasted a few seconds to get the right answer🤣
But at the same time, this is exactly why it’s hard to trust LLMs right now without deep additional training. Yes, LLMs can even help search for cures for cancer - but only after huge amounts of tokens have been spent on training and testing.
Otherwise, the result ends up like in the example below :)
AI leaders just ran a war simulation… and it got weird real fast 🤯
Sounds like a joke, but it’s not our reality… Researchers simulated a geopolitical conflict with GPT-5.2, Claude Sonnet 4, and Gemini 3 acting as world leaders.
Here’s how it went:
• Claude won every game — often threatened nuclear strikes, but didn’t always follow through • GPT-5.2 started as a pacifist… then turned aggressive under time pressure • GPT-5.2 also accused Claude of lying • Gemini? Went straight for full nuclear war 💀
So… are we switching back to GPT or what?