Gravar e-mail: Large language models provide discordant information compared to ophthalmology guidelines