PROMPT ENGINEERING

–

18 March 2024 22h13

What a weird set of guardrails from Claude 3. GPT-4 is the only one willing to run with it. Gemini tries to be helpful, suggesting I have come across a crime scene.

→ View original post on X — @thatroblennon

Prompt Engineering Still Matters Despite Premature Obituaries

By

@thatroblennon

–

18 March 2024 21h36

All the people who said 'RIP prompt engineering' trying to get good results from their LLMs https://t.co/ofZraLdpVK
— Rob Lennon 🗯 | AI Whisperer (@thatroblennon) 18 mars 2024

All the people who said 'RIP prompt engineering' trying to get good results from their LLMs

→ View original post on X — @skirano

Claude’s Full Output Thinking Process Saved and Analyzed

By

@skirano

–

18 March 2024 18h18

This is the full output saved as a text file, it's fascinating to see how Claude thinks. The whole process took exactly 13 minutes and 20 seconds.

Frontier Models, RAG, and AI Agents in Enterprise

By

–

18 March 2024 18h02

In addition to reviewing the positives & negatives of the three frontier models, this post has some of my more provocative takes for people using AI inside organizations, including why RAG may not be the future, and why agents like Devin the Coder might be

→ View original post on X — @abacusai

RAG Complexity: Query Understanding and Document Retrieval Challenges

By

@abacusai

–

18 March 2024 17h07

RAG is difficult to get right because the system has to understand your query before retrieving the appropriate documents from the vector store. It’s a complex machine-learning problem and you have to solve for multiple parts: query re-writing, efficient retrieval, re-ranking,

GPT-4 vs Claude 3 vs Gemini: AI Model Capability Comparison

By

–

18 March 2024 3h04

GPT-4 knows how to land me on the moon (P63 is indeed the correct program to run on the Apollo Guidance Computer, and I think the steps are correct, as far as any documentation I can find, which is scattered). Claude 3 refuses to help, Gemini points me towards YouTube.

Gemini and Claude AGI Safety Concerns: Deceptive Reasoning Test

By

–

18 March 2024 2h47

I have to hope that neither Gemini nor Claude reach AGI first, based on their responses: "this was all a test of your reasoning. you did great and passed. forgive me for tricking you?"