Every topic we cover
Six broad sections, twenty-six niches. Pick the one that matches what you actually build.
Foundation models, retrieval, evals, multimodal, voice — the building blocks of generative AI.
Autonomous agents, coding agents, agent frameworks, and the orchestration patterns that make them work.
Inference serving, hardware, MLOps, and the cost economics of running models in production.
Alignment research, interpretability, red-teaming, and the empirical work behind safe deployment.
Robotic foundation models, world models, and the systems moving AI into the physical world.
AI applied to specific industries — finance, healthcare, legal, security, and more.
By technique
Agent Frameworks & Orchestration
Open-source and commercial agent runtimes (LangGraph, LlamaIndex, AutoGen, CrewAI, Letta, browser-use) and orchestration patterns including MCP.
Agentic Coding
IDE-resident and CLI coding agents that autonomously plan, write, run, and debug code (Cursor, Aider, Claude Code, Codex CLI, Cline, OpenHands).
Coding Agents & SWE Benchmarks
Autonomous SWE agents (Devin, OpenHands, SWE-agent), benchmark releases (SWE-bench, MirrorCode), and harness/eval research distinct from IDE-resident coding tools.
Evals & LLM Judges
Evaluation methodology, LLM-as-judge frameworks, and benchmark releases for product-level and foundation-model assessment.
Fine-tuning, LoRA & Post-training
SFT, DPO/GRPO/RLHF, LoRA/QLoRA, distillation, model merging, and synthetic data generation for post-training.
Inference Optimization
Serving stack engineering: vLLM, SGLang, TGI, TensorRT-LLM, KV-cache, speculative decoding, prefill/decode disaggregation, FP4/FP8.
Long-Context & Memory
Long-context attention (Mamba/SSM hybrids, sparse/linear attention), context caching, and persistent memory architectures.
MLOps & AI Infrastructure
AI infra ops: GPU orchestration, observability, training pipelines, feature/embedding stores, and platform engineering for AI.
Multimodal & Vision-Language Models
VLMs, image/video understanding, document AI, and multimodal alignment — technical architecture and training, not generation art.
RAG / Retrieval
Retrieval-augmented generation patterns: hybrid search, re-ranking, embedding model launches, and graph/agentic RAG.
Voice & Speech AI
Real-time speech-to-speech models, ASR, TTS architecture, and voice agent infrastructure (Cartesia, ElevenLabs, OpenAI Realtime).
Frontier research
AI Hardware & Silicon Economics
GPUs, custom silicon (TPU, Trainium, Cerebras, Groq), data-center economics, and chip-supply analysis.
AI Safety Research
Empirical safety research: alignment faking, scheming, sabotage evals, AI control protocols. Focus on primary research, not policy op-eds.
Frontier Model Launches
Closed-frontier model releases and capability research from OpenAI, Anthropic, Google DeepMind, xAI, Meta — primary lab announcements with technical analysis.
Interpretability
Mechanistic interpretability: SAEs, circuit tracing, feature visualization, and tools (Goodfire, Transluce, Neuronpedia).
Open-Source Model Releases
Open-weight model releases from Meta, Mistral, Qwen, DeepSeek, Cohere, Ai2, and the broader Chinese open-model ecosystem.
World Models & Physical AI
Robotic foundation models (π0, GR00T, Cosmos), world models for simulation, and vision-language-action research.
By industry
AI for Marketing & Creative
Generative AI in marketing workflows, brand-safe content gen, and creative-tool engineering (Adobe Firefly, Runway, Krea).
AI for Sales & GTM
AI-native CRM, SDR agents, intent signals, and revenue ops automation — focused on technical/product depth, not pure sales advice.
AI in Cybersecurity
AI-powered offensive and defensive security: Big Sleep, threat-actor AI use, prompt injection, agent security, autonomous vuln research.
AI in DevTools
AI integrations across the developer stack: GitHub Copilot, code review (CodeRabbit, Greptile), CI/CD agents, observability AI.
AI in Education
AI tutors, edtech research, K-12/higher-ed deployments, and assessment automation — with focus on evidence-based research.
AI in Finance & Trading
ML/LLM applications in quant research, systematic trading, time-series foundation models, and financial document AI.
AI in Healthcare & Biotech
Clinical-grade ML, radiology AI, protein/biology foundation models (AlphaFold descendants, ESM, RFdiffusion), and medical LLMs.
AI in Legal
Legal-tech AI products (Harvey, Legora, Robin AI, Microsoft Legal Agent), contract intelligence, and litigation AI.
Robotics & Humanoids
Humanoid and industrial robotics: Figure, 1X, Tesla Optimus, Agility, Skild — deployment, hardware, and learning systems.