Prompt Engineering for Beginners - Tutorial 3 - Having a Continuous Conversation

Рет қаралды 30,177

8 ай бұрын

OpenAI Docs: platform.openai.com/docs/intr...
Anthropic Docs: docs.anthropic.com/claude/docs
PromptLayer: promptlayer.com/
DevSprout: / devsprout
Source Code: github.com/buckyroberts/AI-Pl...

Пікірлер: 6

@billgordon6954 8 ай бұрын

Thank you. Very helpful videos! When appending the conversation so the history is retained, will that require that the entire previous conversation be resubmitted with each new user response? Does this then require larger and larger amounts of tokens each iteration (eg after g’ma gives us the recipe, are we returning that recipe back in the next API call and paying tokens for that each time?) Thank you again for the great videos.

@DevSprout 8 ай бұрын

That's right. Eventually you will run into a problem where you reach the max tokens allowed for the given model that you are using. There are strategies for avoiding this, like summarizing the history of the conversation and providing it as context, or storing the history in a database and retrieving various parts of it as needed. These are all things to consider when building out a production application with these technologies.

@grandparick3176 8 ай бұрын

wait who is this? This isn't bucky.

@AtomMachinerule 8 ай бұрын

😂.. he must be lost in the Adirondacks. 🏞🏕

@shivu.sonwane4429 8 ай бұрын

❤❤❤

@ex0ja 8 ай бұрын

Kookaburras also aren't native to WA and are a pest here. Most Western Australians don't know this and think they are native (same as Rainbow Lorikeets)