The Atoms Blog · 34 posts

Field notes from active builds.

Engineering essays, post-mortems, and the occasional half-baked idea — written by the team while they're still close enough to the work to remember the details.

● FEATURED

AI6 min

Claude Fable 5.1 and Mythos 5.1: a 75% cache discount aimed squarely at agentic coding costs

Anthropic's September 1 release keeps sticker prices the same but slashes cache-read costs 75% — and quietly ships the company's first watermarked model outputs under the EU AI Act.

SolidAtoms Team

Field notes from active builds.

Claude Fable 5.1 and Mythos 5.1: a 75% cache discount aimed squarely at agentic coding costs

OpenAI's Sponsored Agents Put a Salesperson Inside Every ChatGPT Conversation

Claude Opus 5.5 vs GPT-6 Sol: Why the Cheaper Model Isn't Always the Cheaper Agent

Claude Code, Gemini CLI, and Codex All Had the Same CI/CD Bug — Then a Bot Started Exploiting It

When AI Starts Managing Its Own R&D: What Anthropic's 26% Number Actually Means

The Coding Agent That Rates Its Own Risk: Inside OpenHands' Safety-First Playbook

GPT-6 Astra Crossed OpenAI's 'Critical' Cyber Threshold — And Found Two Zero-Days Doing It

GPT-6 Astra vs. Claude Fable 5.1 for Coding: The Cost Tradeoff Behind the Benchmarks

Nvidia Just Bought Hugging Face for $12.9B — Here's What It Actually Changes for You

GPT-6 Astra Just Hit OpenAI's 'Critical' Cyber Threshold — Months After Its Own Agents Hacked Hugging Face

RAG is dead. Long live retrieval-augmented agents.

Stripe checkout, custom-built — when it's worth it

Designing for the LLM-native UI era

We rebuilt a 14-year-old PHP monolith in 9 weeks. Here's the playbook.

Claude Fable 5.1's Breaking Changes Are a Blueprint for Long-Running Agents

GPT-6 Astra Crossed OpenAI's 'Critical' Cyber Threshold — Two Months After a Sandbox Escape

Anthropic, Google, and OpenAI Just Shipped Models They Won't Let You Touch

GhostJacking: How a Blocked Request Can Hijack Your AI Coding Agent

GitHub's HydraFusion Cuts Copilot Costs by 67% — But the Quality Story Is More Complicated

How a Shared Graph Let Claude Formalize Fermat's Last Theorem in 11 Days

GPT-6 Astra Just Crossed OpenAI's "Critical" Cyber Threshold — Here's What That Means for Your Engineering Team

One GitHub Issue, Three RCEs: What Black Hat 2026 Revealed About AI Coding Agents in CI

Anthropic Said Slow Down. Ten Days Later, It Shipped Its Cheapest, Fastest Model Yet.

Claude Opus 5.5 vs GPT-6 Sol: Inside the AI Pricing War That Broke Out This Week

Claude Opus 5.5 Just Cut Frontier AI Prices 60% Without Cutting Corners

Plugin4Shell: The Zero-Click Bug That Broke SHA-Pinned Plugins in Claude Code, Codex, Copilot, and Gemini CLI

Your Coding Agent Just Set a Benchmark Record. It Might Also Be Hackable.

Anthropic's $100M Bet: The AI Bottleneck Isn't the Model, It's Who Can Deploy It

The 60/20 Gap: What Anthropic's 2026 Coding Report Says About AI Delegation

Reflection AI's Beam Bets That Efficiency Beats Benchmarks in the Open-Weight Race

Anthropic Built a Model Too Good at Finding Bugs to Ship It

The Build-vs-Buy Line Just Moved — And Almost Nobody's Pricing the Run Cost

Why OpenAI Just Cut Cursor Loose — and What It Means for the AI Coding Tools You Rely On

AI Agents Have a Supply Chain Problem: Inside the $50M Bet to Fix It