Tag
A user note on Gemini 3.5 Flash checkpoint highlights improved speed but worse prompt adherence and UI bloat, moving away from the original Gemini design.
The latest Gemini 3.5 checkpoint is criticized for poor prompt adherence, with the model ignoring instructions, using web when told not to, and overbuilding UI, raising concerns about agent reliability despite speed and intelligence.
The article claims that Stanford has released a free technique for training LLMs to adhere strictly to prompts, a skill Anthropic reportedly pays high salaries for. It urges readers to bookmark the resource before it is removed.