The host reveals why Gemini AI left him disappointed compared to ChatGPT after testing both over a weekend.
The episode begins with the host sharing his initial excitement about testing Gemini AI over the weekend, followed by disappointment as he finds it less intelligent than ChatGPT. He describes spending Saturday and Sunday on automations using Telegram bots, N8N, and servers like Render and Google Cloud instances. The host compares how both AIs handle specific tasks, noting Gemini's tendency to switch to English and repeat errors, while ChatGPT performs better but can falter with lengthy queries. He provides concrete examples: Gemini fails with banking file modifications and gives outdated or incomplete instructions for N8N and Google Cloud, whereas ChatGPT admits mistakes and offers clearer guidance. A key insight emerges when the host discovers simple workarounds, like using screenshots of Google Calendar for ChatGPT to parse events, which Gemini handles poorly. He details setting up Casa OS on a small Google Cloud instance with ChatGPT's help, highlighting how AI assists his non-technical learning. The host discusses his current server setup, including a Xeon machine running Docker containers for N8N, Trillium, torrents, and file services at low resource usage. He mentions using N8N and Zapier for automations, such as notifying Telegram channels about new podcasts and syncing blog posts to Google Drive. Throughout, he records on a second-hand Lenovo Yoga Book, noting its low CPU usage after optimization tips from a listener. The episode blends AI comparison with personal tech experiments, emphasizing practical applications over hype.
