GenAI resources
Some curated GenAI resources.
My Top 5
As of 2025-01-10
- How does ChatGPT work? Explained by Deep-Fake Ryan Gosling (YouTube, April 2024)
- Mastering LLMs (Parlance Labs, 2024)
- What We Learned from a Year of Building with LLMs part I part II part III (O’Reilly Media, May 2024)
- Evaluating the World Model Implicit in a Generative Model (June 2024)
- python llm
- Msty app, the easiest way to use local and online AI models
Articles I enjoyed
- The fine art of human prompt engineering: How to talk to a person like ChatGPT (Ars Technica, April 2024)
- What We Learned from a Year of Building with LLMs part I part II part III (O’Reilly Media, May 2024)
- Tokens are a big reason today’s generative AI falls short (Techcrunch, July 2024)
- How I use AI by Nicholas Carlini, research scientist at Google DeepMind (August 2024)
Research papers I enjoyed
- Alpha Geometry (DeepMind, January 2024)
- Financial Statement Analysis with Large Language Models (May 2024)
- Evaluating the World Model Implicit in a Generative Model (June 2024)
- Artificial Intelligence, Scientific Discovery, and Product Innovation (November 2024)
Tools
Tools I am using
- python llm
- NotebookML by Google
CursorContinue- GenAIScript
- Msty app
Tools to look at
Prompt Engineering
WARNING: shelf-life is short as new models get better every day.
- Prompt Engineering Guide
- Impressive collection of prompt patterns (from Fabric project)
- Non-Obvious Prompt Engineering Guide (Tech-sistence, July 2024)
Performance evaluation
WARNING: this is a hard problem; most approaches are too generic and do not answer the question “which is model is best for you”.
Relevant quotes
“They [LLMs] are dream machines. We direct their dreams with prompts.”
Source: https://x.com/karpathy/status/1733299213503787018?lang=en
“As an analogy, artificial flight does not involve the kinds of things that are used to achieve flight by animals in non-human ecosystems. The goals of aeronautical engineers are not those of zoologists. Their methods and aims diverge accordingly.”
Source: https://arxiv.org/pdf/2407.08790
Electric motors are a general-purpose technology, but you don’t buy a box of electric motors from Home Depot - you buy a drill, a washing machine and a blender. A general-purpose technology is instantiated as use-cases.
Source: https://www.ben-evans.com/benedictevans/2024/6/8/building-ai-products
”[…] the problem with Alchemy of yore was not that Chemistry is useless, but that people wanted to delude themselves that Chemistry – a pretty amazing discipline on its own merits – can be Nuclear Physics if you prompt it just so.”
Source: https://arxiv.org/pdf/2402.01817
People I follow (alphabetic order)
My modest contributions to the conversation
- A wishlist for Generative AI technology (May 2023)
- WANTED: Chief AI Officer (March 2024)
- LLM-as-intern: revisiting the analogy (June 2024)
- Prompt like a hipster (July 2024)
- Some github experiments about LLM performance (work in progress)
- Coding on steroids (September 2024)