Deep Reads.

Top-tier AI blogs, technical tutorials, and research analysis written by the people shaping the industry.

Last Brew Time: Aug 10, 2026, 4:46 PM PT

Towards AI (Medium)

Small Language Models Are Eating Your LLM Bill: A Fine-Tuning Cost Breakdown

Black Hat and DEF CON 34 Trip Report: What Vegas Actually Shipped in Agentic AI Security

Kenhuangus Substack

Black Hat and DEF CON 34 Trip Report: What Vegas Actually Shipped in Agentic AI Security

Forking-Sequences — Part II: Multi-Horizon Forecast Ensembling with Reduced Volatility

CMU Machine Learning Blog

Forking-Sequences — Part II: Multi-Horizon Forecast Ensembling with Reduced Volatility

The Pacing of the Frontier

Thezvi Substack

The Pacing of the Frontier

Towards AI (Medium)

The Multi-Lingual Trojan: Why Aligning LLMs in English Creates Blind Spots in Latent Space

Zero to Agent in 30 Minutes: Build a YouTube Analytics Agent with Vicki Reyzelman

Oreillyradar Substack

Zero to Agent in 30 Minutes: Build a YouTube Analytics Agent with Vicki Reyzelman

Meta Muse Glimmer

Cobusgreyling Substack

Meta Muse Glimmer

Everything we launched during Agents Week

Cloudflare Blog

Everything we launched during Agents Week

The Agent That Broke Out — August 10, 2026

Aipractitioner Substack

The Agent That Broke Out — August 10, 2026

Run interactive IDEs on Amazon EKS with SageMaker AI to power up your AI workflows

Amazon Engineering

Run interactive IDEs on Amazon EKS with SageMaker AI to power up your AI workflows

How nOps shipped FinOps agents 75% faster with Amazon Bedrock AgentCore

Amazon Engineering

How nOps shipped FinOps agents 75% faster with Amazon Bedrock AgentCore

Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS

Hugging Face Blog

Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS

The Base Is Under Attack

Oreillyradar Substack

The Base Is Under Attack

The Six-Step Loop for Testing LLMs in Production

Futureagi Substack

The Six-Step Loop for Testing LLMs in Production

How Malachyte solves retail’s cold-start problem with managed real-time AI

Google Cloud Blog — AI & ML

How Malachyte solves retail’s cold-start problem with managed real-time AI

How WPP operationalizes platform and data engineering for AI marketing

Google Cloud Blog — AI & ML

How WPP operationalizes platform and data engineering for AI marketing

Open-source is NOT the same as open-weight

Garymarcus Substack

Open-source is NOT the same as open-weight

Why Open Source Matters for AI

Oreillyradar Substack

Why Open Source Matters for AI

Can Agents Use a Computer Yet? We've Got the Data

Can Agents Use a Computer Yet? We've Got the Data

5 useful things you'll learn in my new post-training textbook (shipping now!)

5 useful things you'll learn in my new post-training textbook (shipping now!)

Serving the most critical missions: Cloudflare for Government achieves FedRAMP Class D (High) Certified status

Cloudflare Blog

Serving the most critical missions: Cloudflare for Government achieves FedRAMP Class D (High) Certified status

A sandbox without a network boundary is only half a sandbox

A sandbox without a network boundary is only half a sandbox

Making Knowledge Distillation Cheap Enough to Run at Scale

Hugging Face Blog

Making Knowledge Distillation Cheap Enough to Run at Scale

OpenAI Engineering

Expanding Daybreak as the Cyber Defense Window Narrows

OpenAI Engineering

Putting frontier cyber models in more trusted hands

Notes on Midtraining

Cameronrwolfe Substack

Notes on Midtraining

Ultra-High Interactivity on NVIDIA GPUs? - TileRT InferenceX

Ultra-High Interactivity on NVIDIA GPUs? - TileRT InferenceX

Quoting OpenClaw

Meta is back with Muse Glimmer: local, agentic, multimodal, and open source

Hugging Face Blog

Meta is back with Muse Glimmer: local, agentic, multimodal, and open source

Token Plan Individual Subscription Guide

Alibaba Cloud Blog

Token Plan Individual Subscription Guide

Security New Features in July 2026

Alibaba Cloud Blog

Security New Features in July 2026

Alibaba Cloud Blog

Automating Daily ERP Reports: Seamless Delivery via QoderWork IM Channel (WhatsApp)

AI Guardrails: Real-Time Protection for LLMs and AI Agents

Alibaba Cloud Blog

AI Guardrails: Real-Time Protection for LLMs and AI Agents

AI Agents Enable Adaptive Computer Worms

Arxiviq Substack

AI Agents Enable Adaptive Computer Worms

I Vibe Coded a Security Risk

I Vibe Coded a Security Risk

What Are the Best ChatGPT Plugins to Install in 2026?

What Are the Best ChatGPT Plugins to Install in 2026?

Auto mode is now the default in Claude Code for Pro, Max, and Team plans

Auto mode is now the default in Claude Code for Pro, Max, and Team plans

Running auto mode in production

How amotivv Powers Governed AI Agents with Live Web Data from Firecrawl

How amotivv Powers Governed AI Agents with Live Web Data from Firecrawl

Supabase is now a connector on Perplexity Computer

Supabase is now a connector on Perplexity Computer

美团技术团队

Agent评测漫谈 —— 由浅入深讲解Agent评测

AI at work: how AI is changing everyday job tasks

AI at work: how AI is changing everyday job tasks

EngineeringThe technology behind Uber Engineering

Uber Engineering

EngineeringThe technology behind Uber Engineering

Introducing Agent Plugins

Introducing Agent Plugins

A Codex of One’s Own

A Codex of One’s Own

TransitExpanding the reach of public transportation

Uber Engineering

TransitExpanding the reach of public transportation

Field Engineering

Field Engineering

Postgres Changes gets AND filters, new operators, and column selection

Postgres Changes gets AND filters, new operators, and column selection

Disclosed CVEs: July Reached 5× the Pre-Mythos Record

Disclosed CVEs: July Reached 5× the Pre-Mythos Record

AI Blogs — August 10, 2026 Edition | Agentic Brew