Tools & Resources I Actually Use

Every tool and book on this page is something I've used, tested, or read in the course of writing apick.net. Nothing is sponsored. If you purchase through an affiliate link, I earn a small commission at no extra cost to you.

How I make money from this site This site is funded entirely through affiliate commissions. I only recommend tools I have personally used in production or extensively evaluated. I do not accept sponsored posts, paid reviews, or brand partnerships of any kind. Full disclosure →

🔧 AI Cost & Observability

Datadog

Infrastructure and application monitoring that works for AI workloads. I use it to track API latency, token usage per model, and cost trends across providers. The LLM Observability dashboard gives you per-request breakdowns that most billing consoles hide.
Try Datadog →

Helicone

Open-source LLM observability platform. Logs every API request with cost, latency, and token counts. The caching analytics alone can save you 30-70% on inference costs. Lightweight enough to deploy in an afternoon.
Try Helicone →

Portkey AI Gateway

API gateway with built-in cost management, fallback routing, and caching. If you're running multiple models in production, Portkey gives you a single control plane for cost controls and usage alerts.
Try Portkey →

🚀 Cloud & Infrastructure

DigitalOcean

Straightforward cloud hosting for AI inference workloads. Simple pricing (no surprise bills) and good GPU options for self-hosting smaller models. Most of my deployment benchmarks in articles were run here.
Get started with DigitalOcean →

Pinecone

Vector database purpose-built for production RAG systems. Low-latency, high-throughput, and integrates with every major embedding model. If you're building AI agents that need long-term memory, this is the standard.
Try Pinecone →

Modal

Serverless GPU compute platform. I use Modal for running benchmarks and batch inference jobs without provisioning servers. Pay per second, scale to zero when idle. Perfect for the kind of cost-conscious deployments I write about.
Try Modal →

💻 Developer Tools

Cursor

AI-native code editor. I use it for all the Python analysis scripts and data processing behind the articles on this site. Tabs mode with Claude integration handles multi-file refactoring that Copilot can't touch.
Try Cursor →

Sentry

Error and performance monitoring for production systems. Essential for any team deploying AI agents in production — it catches the silent failures that don't show up in API logs.
Try Sentry →

📚 My Books

The AI Tax: 5 Decisions That Stop You From Overpaying for AI

Apick Lion · Kindle Edition · Now Available
The definitive guide to AI cost decisions. Most AI failures aren't technical — they're financial. This book gives you a practical framework for evaluating AI investments through the lens of cost structure, scaling economics, and organizational readiness. Benchmark traps, the replication tax, inference caching, and the cash-flow-first deployment checklist — all backed by real production data. This is the book for anyone making infrastructure cost decisions.
Buy on Amazon →

The $5.6 Trillion AI Bubble: A Survival Guide for When the AI Bubble Bursts

Apick Lion · Kindle Edition · Now Available
A survival guide for when the AI bubble bursts. $39 trillion in federal debt. A 16:1 investment-to-revenue ratio in AI infrastructure. Tech giants burning cash while the market prices in infinite growth. This book breaks down the three converging risks — the debt time bomb, the AI investment gap, and geopolitical fuel — and shows you how to navigate what comes next. Essential reading for investors, engineers, and executives.
Buy on Amazon →
All product links on this page are affiliate links. If you make a purchase through them, I earn a small commission. Prices are the same for you. I only recommend tools I've used and believe in.