Screening 1,000 ambassador applications with GPT-6 Astra
The author pulled a year of each applicant's posts via the X API and had GPT-6 Astra analyze them, finding 300 accounts with practical AI-use signals and 8 high-intensity users.
I’m reviewing the +1,000 applications we received so far for the Antseed ambassador program For me, the most important thing is that creators are actually using AI and there’s evidence of it I’m using AI to help with the user research and first filters. Sharing how I did it in case it’s useful! Simple workflow: download the posts, analyse them with AI, summarise the findings 1. Used the X API to pull each applicant’s posts from the past year 2. Asked ChatGPT (GPT 6 Astra) to analyse them and look for evidence of practical AI use 3. Grouped accounts by those signals, from basic use to people regularly sharing what they build with AI I wanted to separate people mostly talking about or promoting AI from those showing how they actually use it. Going through a year of posts from every person manually would take forever haha 300 accounts showed practical AI-use signals. 8 are high-intensity AI users Now I can focus my manual review on those profiles and look more closely at their work. Obviously, someone’s tweets don’t tell you everything, so I need now to study deeper, but first filter helped a lot Pretty cool that we can now do this kind of research fast without a whole team thanks to AI!!