Gemini — what creators actually do with it

The 17 Gemini posts split between the model as a chat tool and Gemini CLI as a coding agent. The CLI launch post stresses price: 1,000 free requests a day, a one-million-token context window on Gemini 2.5 Pro and Google Search for live context, framed as a threat to paid tools like Cursor and Claude Code (Gemini CLI). Cursor and Claude Code each co-occur in 4 posts.

Large context is the recurring reason creators reach for it. @angus.sewell pays for Claude for most work and Gemini for "anything that Claude can't handle like large context tasks" (paid stack). A sponsored clipping recipe uses Gemini to watch YouTube links directly before Claude cuts vertical shorts; its income claims are speculative (GTA clips). For quick creative coding, one creator prompts Gemini 2.5 Pro for an interactive particle visualizer in a single HTML file; the first output came back malformed and needed a "did not work, remake it" follow-up (particle toy).

The limits posts are specific. Early testers of Gemini Omni reported two 10-second video generations using about 86% of a Pro plan's daily quota (Gemini Omni), and in a prompt-a-UI challenge a Gemini attempt was "a valiant effort" that failed (Build Wars). One rant calls Gemini untrustworthy, as opinion rather than evidence (rant). On the other side, a tier list says ChatGPT or Gemini give more breadth than Claude at $20 a month (tier list), and in an identity study Gemini named itself correctly while Claude and Grok often claimed to be Chinese models (identity study).

ChatGPT also appears in 4 posts, DeepSeek in 3 and Llama in 2, mostly as interchangeable model options in tools like MoneyPrinterTurbo (video generator) or in router setups. Several tags are incidental, such as a position-embedding explainer that lists Gemini among example models (RoPE explainer).