In this article I will share resources I use to stay up-to-date with AI research and AI-relevant technology. I’ll try to avoid the usual suspects like HuggingFace and arXiv, and r/LocalLaMA. Instead, I’ll focus on (perhaps) some lesser-known resources that I’ve personally found to be valuable. 

TL;DR

Survey Papers

A survey paper is typically a comprehensive showcase of SOTA research in a specific area. These papers are often provide a detailed overview of the current state of research, key findings, and future directions. They are an incredibly valuable resource to me when exploring new interests or whenever I’m stuck on a problem or need a fresh perspective.

I primarily use X (formerly Twitter) to run an advanced search for survey papers. Here is a typical search query that will return decent results. Note the min_retweets:10 parameter, which filters out papers that have not been shared much:

survey (paper OR research OR arxiv OR preprint) (llm OR ai OR "language model" OR "artificial intelligence") min_retweets:10 lang:en -filter:replies -filter:nativeretweets since:2023-01-01 filter:links

A few gems that I’ve found in 2024 include:

• Understanding the planning of LLM agents: A survey
• The Prompt Report: A Systematic Survey of Prompting Techniques
• A Survey of Reasoning with Foundation Models
• Knowledge Mechanisms in Large Language Models: A Survey and Perspective
• A Survey on Data Selection for Language Models

Research Papers & Search

• HuggingFace Daily Papers is a daily feed of AI-related research.
• Papers with Code is a platform designed to bridge the gap between academic research and practical implementation and is a decent place to look for current research with demonstration implementations. Personally I find it difficult to follow the abstract method representations in some papers however if there is code available I can usually follow along.
• arxiv-sanity: Allows for community tagging of recent papers hosted on arXiv.
• SearXNG is a meta-search engine that can aggregate results from multiple sources. Settings can be personalized and stored in the browser to query science/research databases. Moreover, it supports multiple output formats such as CSV and JSON so tools like Open WebUI can be used to automatically interact with it.

Cloud Native Landscape

The CNCF (Cloud Native Computing Foundation) is a nonprofit organization that promotes the adoption of cloud-native technologies. Basically it operates under the Linux Foundation and serves as a home for open-source projects that enable scalable, resilient, and portable applications in modern cloud environments.

The Cloud Native Landscape is a curated ecosystem of technologies, tools, and practices designed to support cloud-native development. While not targeted at AI research specifically, I often use the landscape to research the technology stack of candidate tools before I start using them.

GitHub

Github trending Is a good place place to check regularly to see what is popular. I try to take a look here once a week to see what is trending in the AI space. Additionally, you can also view developers are trending on Github which is an excellent way to find active community members who share your common interests. A great way to make new contacts in the AI community is to find folks contributing to repositories you use or are interested in and begin contributing yourself, join their Discord, etc.

Github Ranking is a general-purpose tool that ranks popular repositories from Github. While not specifically targeted at AI, the Python top 100 section has many high-quality projects relevant to AI research and development.

LibHunt

LibHunt is a platform designed to help developers discover open-source libraries, tools, and frameworks for their projects. It provides a categorized and searchable index of open-source projects, highlighting alternatives, trends, and recommendations for various programming languages and technologies. Typically, I use LubHunt to find alternatives to popular libraries or tools that I’m already using or have read about.

• Top