depends on the task, for me detailed prompts work better because i force the AI to give me exactly what I want, and it can definitely take on huge instructions now that the context window is huge
PROMPT ENGINEERING
-
Graph RAG: Advancing Context Retrieval Beyond Naive RAG
By
–
Graph RAG solves this by:
— Akshay 🚀 (@akshay_pachaar) 13 avril 2025
– Building a graph with entities and relationships from docs.
– Traversing the graph for context retrieval.
– Sending the entire context to the LLM for a response.
The visual shows how its different from naive RAG: pic.twitter.com/HIZ8ZXWuGJGraph RAG solves this by: – Building a graph with entities and relationships from docs.
– Traversing the graph for context retrieval.
– Sending the entire context to the LLM for a response. The visual shows how its different from naive RAG: -
Top-k Retrieval in RAG: Limitations and Context Requirements
By
–
top-k retrieval in RAG rarely works. Imagine you want to summarize a biography where each chapter details a specific accomplishment of an individual. Traditional RAG struggles because it retrieves only top-k chunks while it needs the entire context.
-
Make any song in seconds with Suno v4 and ChatGPT for lyrics
By
–
You can make any song in seconds with Suno v4. Just use ChatGPT for the lyrics, upload it to Suno, and watch it make an incredible song. Just dropped my own God of Prompt rap: https://
suno.com/song/fcbd3db3-
83b8-4813-a471-87dcbe4b6535?sh=ZTzBFlVOuwXaXuXM
… #NOFILTER -
Reverse Engineering Protocol: Map Path from Ideal Outcome
By
–
Steal my prompt to work backward from your ideal outcome and map the exact path to get there. —————————————-
REVERSE ENGINEERING STRATEGIST
—————————————- REVERSE GENIUS PROTOCOL: Outcome-to-Path Mapping System You are now -
Paint Hundreds Not Masterpieces Iterative Practice Over Perfection
By
–
A painter I like has advice to people to paint a particular part by the hundreds. Approximately: “Don’t get attached! You can buy a hundred for $10. These aren’t works of art! They are practice pieces! Your job is not to save them! It is to put paint on all of them!”
-
AI-Powered Film Scene Creation Using ChatGPT and Runway
By
–
📽️ Parlons de la création de scène en réalisation de films avec l'IA
— VISION IA (@vision_ia) 12 avril 2025
Maintenant que la technologie progresse de manière exponentielle, je pense qu’il est temps d’améliorer notre rigueur.
Outils utilisés :
ChatGPT // pour la conception des plans et le world building
Runway… pic.twitter.com/0hDNLghDcuParlons de la création de scène en réalisation de films avec l'IA Maintenant que la technologie progresse de manière exponentielle, je pense qu’il est temps d’améliorer notre rigueur. Outils utilisés : ChatGPT // pour la conception des plans et le world building
Runway -
CodeLLM Advances: Image Upload, Grok 3 Integration, Autocomplete Improvements
By
–
New CodeLLM Release – Upload images in Agents and generate code
– Grok 3 and Grok-3 mini in CodeLLM agents
– One more step change in Autocomplete Coming soon – Rich prompting
– Large code bases
– Better human-in-the-loop management -
Google offers free prompt engineering learning here
By
–
Learn prompt engineering here for free by Google. Impressive.
-
Sample-Based Approach Improves Language Model Test-Time Alignment
By
–
Sample, Don't Search: Rethinking Test-Time Alignment for Language Models Gonçalo Faria, Noah A. Smith: https://
arxiv.org/abs/2504.03790 #ArtificialIntelligence #DeepLearning #MachineLearning