Leaders and Resources · Chapter 4

Resources and Directories

Where to find AI tools, models, code, research, courses, conferences, news, market evidence, and governance guidance.

The AI information problem has two sides: there is too little reliable long-term knowledge and far too much daily noise. Good resource use requires knowing whether you need a tool, a model, a research paper, a course, a market report, a policy source, or current product news.

Regional resource guides

The Global AI chapter collects official portals, laboratories, model projects, national programs, and learning resources outside the United States.

Open Global AI

AI-tool search and discovery websites

These are the best place to search when the question is, “Is there already an AI for this task?” Listings may include sponsorships, affiliate links, self-submitted descriptions, or rapidly changing product information. After discovering a tool, verify it at the developer's official website.

AI-tool directory

There's An AI For That

A task-oriented search engine and large directory of AI products. Particularly useful when you can describe the exact job you want the tool to perform.

AI-tool directory

Futurepedia

A curated AI-tool directory with categories, reviews, tutorials, courses, and practical business-oriented learning material.

AI-tool directory

Toolify

A very large multilingual directory covering websites, applications, browser extensions, models, categories, rankings, and newly added tools.

AI-tool directory

FutureTools

A searchable AI-tool database with categories, pricing filters, curated picks, collections, reviews, news, and workflow guidance.

Internal directory

AI Atlas Apps & Agents Directory

The Atlas's own structured directory of representative assistants, coding agents, creative products, enterprise agents, and vertical applications.

Models, code, datasets, and benchmarks

Models and datasets

Hugging Face

Models, datasets, demonstrations, libraries, documentation, and community collaboration.

Code

GitHub

Open-source code, model implementations, agent frameworks, evaluation tools, and software projects.

Research and code

Papers with Code

Research papers connected to implementations, datasets, benchmarks, and task leaderboards.

Datasets and practice

Kaggle

Datasets, notebooks, competitions, courses, and a large data-science community.

Benchmarks

MLCommons

Industry benchmarks and measurement efforts for training, inference, safety, and AI systems.

Research papers and scientific discovery

Academic search

Semantic Scholar

Academic search with citation networks, summaries, related papers, and author profiles.

Academic search

Google Scholar

Broad scholarly search across disciplines, publishers, repositories, books, and citations.

Literature mapping

Connected Papers

Visual exploration of related papers around a starting publication.

Annual report

Stanford AI Index

Annual evidence on technical progress, adoption, investment, labor, policy, education, and public attitudes.

AI trends

Epoch AI

Research and data on compute, scaling, frontier models, algorithmic progress, and AI forecasting.

News, analysis, and practical commentary

Newsletter

The Batch

DeepLearning.AI's newsletter summarizing important research, products, business developments, and policy.

Business analysis

Stratechery

Ben Thompson's analysis of technology strategy, business models, distribution, platforms, and AI industry structure.

Applied AI

One Useful Thing

Ethan Mollick's practical analysis of AI in work, education, management, and everyday organizational use.

Technical analysis

Simon Willison's Weblog

Close technical observation of models, tools, APIs, coding agents, security issues, and new product behavior.

Newsletter

Import AI

Jack Clark's long-running newsletter on AI research, policy, security, and international developments.

Courses and learning

Courses

DeepLearning.AI

Short courses and longer programs covering generative AI, agents, machine learning, data, and applied workflows.

Courses

fast.ai

Practical deep-learning courses designed to make powerful methods accessible to working programmers and practitioners.

Courses

Hugging Face Learn

Free courses on transformers, agents, audio, computer vision, diffusion, reinforcement learning, and open-source tools.

Courses

Microsoft Learn — AI

Role-based learning paths for AI concepts, Azure AI, agents, Copilot, development, and responsible AI.

Policy, governance, safety, and incidents

Policy

OECD.AI

Global AI policies, principles, country initiatives, data, analysis, incidents, and governance resources.

Regulation

European AI Office

Official information on European AI governance, implementation, general-purpose AI, standards, and adoption.

Incidents

AI Incident Database

A public database documenting harms and near harms involving deployed AI systems.

Responsible AI

Partnership on AI

Multi-stakeholder research and guidance on responsible practices, labor, media, safety, and social impact.

Major research conferences

Conference

NeurIPS

Machine learning, neural computation, systems, theory, applications, and broader impacts.

Conference

ICML

A leading international conference on machine learning research.

Conference

ICLR

Representation learning, deep learning, optimization, and related methods.

Conference

ACL

Natural-language processing and computational linguistics.

Conference

CVPR

Computer vision and pattern recognition.

Conference

AAAI

Broad artificial-intelligence research, applications, education, and community activity.

How to evaluate an AI resource

  1. Check the date: AI products and model capabilities change quickly.
  2. Find the primary source: confirm features, pricing, privacy, and availability on the developer's own site.
  3. Identify commercial incentives: sponsored placement, affiliate revenue, vendor funding, and paid submissions can affect visibility.
  4. Test with your own task: a high ranking may not predict performance on your documents, language, workflow, or risk level.
  5. Protect sensitive information: review data retention, training use, access controls, and deletion options before uploading private material.
  6. Separate popularity from reliability: a widely discussed tool may still be inaccurate, insecure, or unsuitable for consequential work.
  7. Keep alternatives: dependence on one provider creates pricing, availability, and data-portability risk.