Within the context of vibe coding and the fact that I've been running successful product/content voting platforms since 2006 - I've come up with the idea of Vibe Voting.The idea: the best products are determined by clearly defined rules that are executed by AI models (10 leading LLMs from 10 different labs). This resolves the ongoing issue of voting rings and friends of founders determining the best content/products. For example - ProductHunt is suffering from this issue - the best products are determined by founders' following size and not by merit (in many cases).I've received criticism from some people that they can't trust the judgements of LLMs. While that's fair, in most cases, those judgements are more objective (two of my own products, LibHunt & VennPoll, have been rated poorly) and surface better products as compared to people with biases. What the models look for - Clarity, Need, Differentiation, Evidence, Visual Design & Polish, Trust, Sustainability. Also, they are explicitly instructed not to judge products based on their popularity. And, each model writes feedback + suggestion for improvements about each product.Furthermore:- Each model is scored with test-retest. Anything not satisfactory (less than 0.5) is dropped. Kendall's tau method. 4 runs. Most models are 0.7+. GPT-6.1-Sol is 0.89 and Claude-Opus-5.5 is 0.79.- Each model sees the products in its own random order, so position bias is neutralized as much as possible.- Gemini put Google’s own password manager 11th of 12, a counter-example to self-preference bias.You can check out the project here: https://vibevoted.com/ and a sample thematic session here: https://vibevoted.com/s/password-managers-2026-octThe products on VibeVoted get in by: $25 buys a seat in one session, never a vote or a ranking position. Thematic sessions, like the password one below, are picked by SaaSHub’s editors or me, and nobody pays.Do you think that this idea has merit, and what's the biggest issue you can point out?Thanks Comments URL: https://news.ycombinator.com/item?id=50361440 Points: 1

Comments: 0