RESOURCES

A curated collection of the tools, platforms, and reading that matter in 2026 — from self-hosted AI gateways to agentic frameworks, cloud infrastructure to cybersecurity. Updated regularly.

43 RESOURCES | 8 CATEGORIES | LAST UPDATED MAR 2026

AI Models & Platforms Self-Hosted AI Agentic Frameworks Developer Tools Cloud & Infrastructure Cybersecurity Reading List Communities & Newsletters

AI Models & Platforms

Claude (Anthropic) — Best-in-class reasoning and coding — Claude 4.5 and 4.6 lead on complex tasks Best-in-class reasoning and coding — Claude 4.5 and 4.6 lead on complex tasks GPT-4o (OpenAI) — Multimodal powerhouse with broad tool and plugin ecosystem Multimodal powerhouse with broad tool and plugin ecosystem Gemini 2.5 (Google) — Deep integration with Google Workspace and strong multimodal performance Deep integration with Google Workspace and strong multimodal performance Llama 4 (Meta) — Leading open-weight model — run it anywhere, fine-tune it for anything Leading open-weight model — run it anywhere, fine-tune it for anything Mistral Large (Mistral AI) — Europe's top open-source contender with strong multilingual support Europe's top open-source contender with strong multilingual support Qwen 2.5 (Alibaba) — Punches well above its weight at smaller parameter counts — ideal for local inference Punches well above its weight at smaller parameter counts — ideal for local inference

Self-Hosted AI

OpenClaw — Self-hosted AI gateway — route between providers, manage keys, and keep data in-house Self-hosted AI gateway — route between providers, manage keys, and keep data in-house Ollama — Run LLMs locally with one command — Metal-accelerated on Apple Silicon Run LLMs locally with one command — Metal-accelerated on Apple Silicon vLLM — High-throughput LLM serving engine with PagedAttention — production-grade inference High-throughput LLM serving engine with PagedAttention — production-grade inference LocalAI — OpenAI-compatible API for local models — drop-in replacement for any OpenAI SDK app OpenAI-compatible API for local models — drop-in replacement for any OpenAI SDK app LM Studio — Desktop app for running and testing open-source models with a clean chat interface Desktop app for running and testing open-source models with a clean chat interface

Agentic Frameworks

browser-use — AI browser automation via accessibility trees — agents that can navigate real websites AI browser automation via accessibility trees — agents that can navigate real websites CrewAI — Multi-agent orchestration — define roles, goals, and let agents collaborate on tasks Multi-agent orchestration — define roles, goals, and let agents collaborate on tasks LangGraph — Stateful, graph-based agent workflows — built for complex multi-step reasoning Stateful, graph-based agent workflows — built for complex multi-step reasoning AutoGen (Microsoft) — Multi-agent conversation framework — agents that debate, delegate, and self-correct Multi-agent conversation framework — agents that debate, delegate, and self-correct Semantic Kernel — Microsoft's SDK for integrating LLMs into enterprise .NET and Python applications Microsoft's SDK for integrating LLMs into enterprise .NET and Python applications Composio — Tool infrastructure for agents — 200+ integrations so your agents can actually do things Tool infrastructure for agents — 200+ integrations so your agents can actually do things

Developer Tools

Claude Code — Anthropic's terminal-native AI coding assistant — thinks before it edits Anthropic's terminal-native AI coding assistant — thinks before it edits Cursor — AI-first code editor built on VS Code — inline completions, chat, and multi-file edits AI-first code editor built on VS Code — inline completions, chat, and multi-file edits Windsurf (Codeium) — AI-powered IDE with deep codebase understanding and flow-state design AI-powered IDE with deep codebase understanding and flow-state design Astro — Content-first web framework — island architecture, zero JS by default, blazing builds Content-first web framework — island architecture, zero JS by default, blazing builds Tailwind CSS 4 — Utility-first CSS rebuilt from scratch — faster, simpler, CSS-native configuration Utility-first CSS rebuilt from scratch — faster, simpler, CSS-native configuration Supabase — Open-source Firebase alternative — Postgres, auth, storage, and edge functions in one Open-source Firebase alternative — Postgres, auth, storage, and edge functions in one

Cloud & Infrastructure

Microsoft Azure — Enterprise cloud with deep OpenAI integration — the default for regulated industries Enterprise cloud with deep OpenAI integration — the default for regulated industries AWS — Still the largest cloud — Bedrock makes it easy to deploy foundation models at scale Still the largest cloud — Bedrock makes it easy to deploy foundation models at scale Cloudflare Workers AI — Run AI inference at the edge — low latency, pay-per-request, globally distributed Run AI inference at the edge — low latency, pay-per-request, globally distributed Vercel — Deploy frontend and serverless functions with zero configuration — pairs perfectly with Astro and Next.js Deploy frontend and serverless functions with zero configuration — pairs perfectly with Astro and Next.js Railway — One-click deployment for backends, databases, and cron jobs — the simplest PaaS in 2026 One-click deployment for backends, databases, and cron jobs — the simplest PaaS in 2026

Cybersecurity

CrowdStrike Falcon — AI-driven endpoint protection — still the gold standard for threat detection and response AI-driven endpoint protection — still the gold standard for threat detection and response SentinelOne — Autonomous endpoint security with AI-powered investigation and remediation Autonomous endpoint security with AI-powered investigation and remediation Wiz — Cloud security posture management — finds risks across AWS, Azure, and GCP in minutes Cloud security posture management — finds risks across AWS, Azure, and GCP in minutes Huntress — Managed detection and response built for SMBs — 24/7 SOC without the enterprise price tag Managed detection and response built for SMBs — 24/7 SOC without the enterprise price tag

Reading List

The Coming Wave — Mustafa Suleyman on why AI and biotech are the most consequential technologies in human history Mustafa Suleyman on why AI and biotech are the most consequential technologies in human history Co-Intelligence — Ethan Mollick's practical guide to working alongside AI — essential for anyone using these tools daily Ethan Mollick's practical guide to working alongside AI — essential for anyone using these tools daily The Age of AI — Kissinger, Schmidt and Huttenlocher on how artificial intelligence will reshape geopolitics Kissinger, Schmidt and Huttenlocher on how artificial intelligence will reshape geopolitics Chip War — Chris Miller on the global semiconductor battle that determines who controls the future Chris Miller on the global semiconductor battle that determines who controls the future The Alignment Problem — Brian Christian on the challenge of building AI systems that actually do what we want Brian Christian on the challenge of building AI systems that actually do what we want Nexus — Yuval Noah Harari on information networks from Stone Age to AI — and what they mean for truth and power Yuval Noah Harari on information networks from Stone Age to AI — and what they mean for truth and power

Communities & Newsletters

AI Engineer — The community for builders shipping AI products — conferences, Discord, and job board The community for builders shipping AI products — conferences, Discord, and job board Latent Space — The podcast and newsletter tracking the AI engineering frontier The podcast and newsletter tracking the AI engineering frontier The Batch (DeepLearning.AI) — Andrew Ng's weekly newsletter — clear, concise AI news without the hype Andrew Ng's weekly newsletter — clear, concise AI news without the hype TLDR AI — Five-minute daily digest of the most important AI developments Five-minute daily digest of the most important AI developments Hacker News — Where the best technical discussions happen — still unmatched for signal-to-noise Where the best technical discussions happen — still unmatched for signal-to-noise

SUGGESTIONS? GET IN TOUCH