Explaining an AI shopping experiment from @wetclaude

The creator details an experiment where she queried three major AI models 8,700 times for product recommendations across various categories. She explains her motivation to understand AI's future impact on shopping, then reveals her two main findings: all AIs surprisingly converge on the same brand recommendations, and this convergence is not caused by them reading the same online sources like Reddit.

Creator: @wetclaude on TikTok

Video format

Yap

Video outline

  1. State large-scale experiment
  2. Reveal initial hypothesis
  3. Present surprising first finding
  4. Debunk logical explanation

Narrative framework

The Counter-Intuitive Experiment

Narrative framework logic

Presenting the results of a self-conducted experiment that subverts a common assumption, and then deepening the mystery by debunking the most logical explanation for the surprising result.

Topics: Tech, Marketing, E-Commerce

Concepts: Breakdown, Case Study Breakdown

Formats: Yap

Elements: B-Roll Cutaway, Jump Cut

Account types: Personal Brand

Transcript excerpt

Rank test where I asked Claude and Google Gemini and Checha Better what to buy, what tinted fish to buy, what sunscreen to buy, what running shoes to buy, what protein powder to buy, 8,700 tier, and the results have been really odd and surprising. So I did this because I'm really interested in what happens when AI starts to make shopping decisions for us and when we start to use AI more than we use Google for consumer purchase decisions. And there were three main findings. So the first main finding is I came into this project being like, oh, all these AIs are gonna have different personalities. Claude's gonna recommend different stuff than Chesh Boutique, than Google Gemini, but that is not the case. For the most part, if you ask what natural deodorant to buy, everyone says Native. If you ask what running shoes to buy, everyone says Brooks. So there's convergence amongst the AM models. The second thing that's equally surprising is that that convergence is not a result of the models reading the same sources. So I thought, like, okay. They're just all picking up stuff from Reddit, and so that's why they're all giving the same answers. That's why everyone's saying vibro.

42,800 views