In production, Opus-level AI quality at a low cost, with near-instant results, is a game changer. This notebook makes it possible for anyone to implement this quickly. So how does it work?
AI
-
Opus generates diverse examples from single task descriptions
By
–
You give Opus a description of your task, along with one input/output example. That's it. From there, Opus will generate seven other great, diverse examples that are similar in structure to your example. It'll then use those + the task description to generate a system prompt.
-

Automated Prompt Generation and Testing Workflow for Claude
By
–
It'll then put the examples in the right prompt format for Claude, and test the generated prompt against your initial input example, using Haiku. Lastly, it'll save both the system prompt it generated and the Opus-created examples in a Python file, pre-formatted for generation.
-
Claude Opus teaches Haiku: 60x cheaper AI quality
By
–
Introducing `claude-opus-to-haiku` ✍️
— Matt Shumer (@mattshumer_) 21 mars 2024
Get the quality of Claude 3 Opus, at a fraction of the cost and latency.
Give one example of your task, and Claude 3 Opus will teach Haiku (60x cheaper!!) how to do the task perfectly.
And it's open-source: https://t.co/nrivU2BWmn pic.twitter.com/0HW6NTZ1HNIntroducing `claude-opus-to-haiku` Get the quality of Claude 3 Opus, at a fraction of the cost and latency. Give one example of your task, and Claude 3 Opus will teach Haiku (60x cheaper!!) how to do the task perfectly. And it's open-source: https://
github.com/mshumer/gpt-pr
ompt-engineer
… -
Claude Haiku matches Opus quality at 60x lower cost
By
–
This repo was inspired by this tweet that went viral. Claude 3 Haiku is *60x* cheaper than Opus, and 10x faster. I discovered that if you prompt Haiku with a number of great examples, it can match Opus' quality.
-
Training AI Agents on Open-World Game Partnership
By
–
Thanks Sean, it has been amazing to partner with you and your team on this, and to train the agents on your amazing open-world game!!
-
GPT-4 struggles with consistent Google Scholar browsing
By
–
GPT-4 can’t browse Google Scholar consistently
-

UN Adopts First Global AI Resolution; Policy Stocktaking Ahead
By
–
The @UN just adopted the first-ever global resolution on AI – the latest in a series of regulatory/policy initiatives around the world. @UNTechEnvoy Amandeep Gill will join us at the AI+Policy Symposium on Apr 16 for a global stocktaking. @StanfordCyber https://
stanford.io/4cFbIKb -
Market failures and regulatory tradeoffs in AI policy
By
–
One market failure is okay because there is also another one?
-
GPT Impact on Academic Publishing: Capability Implications
By
–
If / when it does work, I think academic publishing is going to be in for a wild time. (Not because of the GPT, but because or the capability it represents)