🧠 Claude Levels Up, Your Tools Get Smarter

Opus 5 comes close to Fable 5's frontier intelligence at half the price. Same API cost as Opus 4.8 ($5 per million input tokens, $25 per million output tokens), which means you get a meaningful step up in capability without your bill moving. On coding and knowledge work evaluations like Frontier-Bench and GDPval-AA, Opus 5 is the new state-of-the-art.

The more interesting number might be the safety one. Anthropic's automated behavioral audit found Opus 5 to be its most aligned model to date, with the lowest rates of reckless or deceptive behavior of any Claude released so far. It's also gotten meaningfully better at cybersecurity work like spotting vulnerabilities, though Anthropic has intentionally kept it well behind their more restricted Mythos 5 model when it comes to actually building exploits.

SpaceXAI announced the release of its new Grok add-on for Google Workspace on Friday, embedding its conversational AI assistant directly into Google Docs, Sheets, and Slides. One install from the Google Workspace Marketplace covers all three apps.

Grok can answer from spreadsheets with cited cells, build formulas and charts, turn outlines into editable presentations, and draft or refine documents in place with connectors for Drive and email. The Grok add-on itself is free. If you're already in Google Workspace, this saves you the mental load of context-switching to a separate AI chat.

On July 24, 2026, Anthropic published that it removed over 80% of Claude Code's system prompt for Claude 5 generation models (Opus 5, Fable 5, Sonnet 5) with no measurable loss on coding evaluations.

This means newer Claude models have better judgment and need fewer guardrails. The old rules were written for older models that needed guardrails. Claude 5 models have better judgement. They do not need the same constraints. If you're still writing thick CLAUDE.md files and elaborate system prompts, you're actually getting in the way. Strip it down.

Heard is a macOS voice layer for your agentic workflows. It connects to Claude Code, Codex, and Cursor and turns their output into intelligent summaries you can hear. So instead of having five terminals running in parallel and losing track of what agent three is doing, Heard speaks to you—full commentary when your eyes are elsewhere, silence unless something errors or needs a decision. Open source, free for personal use. If you're running multiple agents at once, this solves the wall that hits around agent three.

Paradigm is an AI-driven personalized education platform built to level the playing field. The founder's mission is giving every learner a lifelong tutor: an AI that adapts, evolves, and understands them—one that keeps them curious and coming back for more. To that end, Paradigm's answer is a radical one—a 100% free tier that will remain for as long as the company exists. Students pilot their own learning and shape the curriculum in real time, which means engagement actually skyrockets. Not a generic tutoring app.

A roofing company owner floated an idea worth stealing: instead of a normal CRM, build one that plays like a tycoon game, where every real project is a level, AI agents handle the actual customer communication and paperwork behind the scenes, and leveling up in the game mirrors real growth in the business.

As projects complete, crews level up. Growth in the game mirrors growth in the business. Suddenly work feels like playing a tycoon game you actually care about winning. Now tell me that’s not fun!

What is worth noting for you is this: The specifics change by industry (a legal firm might turn case pipeline into a quest log, a sales team might turn prospect tracking into a progression system), but the principle holds: AI can help you build the boring infrastructure, leaving you free to make the boring parts fun.

Every Friday, the PAI team gets together to host a live AI build session where we build workflows that you bring to us.

Because we build in real-time, you get to see how we think through the problem, where we get stuck, and what the final outcome is. It's the difference between reading a recipe and watching someone cook it while talking through their decisions.

🗓️ When: Friday, July 31st @ 10AM PT (1PM ET)
 📹️ Where: Virtual