Anthropic Engineer Shares ELI5 Prompting Method
Thariq Shihipar describes using slash commands to generate simple visual explanations.
TLDR
Thariq Shihipar, an engineer on Anthropic's Claude Code team, posted that colleagues use an ELI5 prompting technique. The command /eli5 followed by a topic directs the model to explain for someone with no prior knowledge, outputting an HTML artifact with large images and minimal text. His post included a video demonstration. Hiten Shah amplified the message on X. Replies on X split between praise for the visual approach and criticism that the need for such prompts highlights overly verbose model output.
Combined views
1.5M
3 Sources, first seen ago
Sources
BE 
Ben@Benuoa@trq212 Gave a go. I have to say, this is fantastic! Thanks for sharing! If all models treat us human beings like foolish kids and newbies, life is much more enjoyable.:-)
DI dinos@din0s_@trq212 I'd rather get castrated than have it Claudesplain stuff to me in that tone
BU busybody77@primaryNV@trq212 I still hate the way it writes. Very not human.
AJ Ajinkya@ajinkya_ml@trq212 Much needed release!!
FI Fish@fullquid@trq212 So useful!
LI Liz@Liz_LizLi@trq212 Nice！