🪅 Heap Heap Hurray Ed. #548

Cheers to Thursday! The weekend is almost here!

#Google - Do you have an app in the Play Store? Chances are that Google may reach out to you - and pay you to ingest your source code to train their AI.

#Uber - After blowing off their budget for AI tools within 4 months (or was that 1 month, we are not certain), Uber has put a cap of $1500 per employee per month on AI tools.

#Meta - After forcing their employees to participate in a program to track employees’ mouse clicks and keystrokes in the name of training AI, and facing backlash for it, Meta has now decided to allow their employees a break of 30 mins in the event they need to “check something personal”.

Apple Agrees To Reveal India Revenue In Order To Avoid Massive $38B Fine
Apple agrees to reveal India revenue in order to avoid massive $38B fine. Company was found guilty of exploiting its dominant position in iPhone apps back in 2024, but the company initially refused to reveal financial data within the country. That potentially left it facing the world’s last largest antitrust fine of amassive $38 billion.

Infosys, TCS and Wipro Scale Microsoft Copilot To Over 300,000 Employees
Three global IT majors, Infosys, TCS, and Wipro, have each scaled their Microsoft 365 Copilot licenses to over 300,000 employees, marking one of the largest and fastest enterprise AI rollouts globally.

Microsoft Unveils Project Solara
Microsoft is building a platform for devices that run AI agents instead of apps. The platform, dubbed “Project Solara,’ is Microsoft’s bet that AI will open up entirely new scenarios for computing. Microsoft says it won’t ship these devices itself.

WhatsApp - AI-powered customer support bot, Meta Business Agent, now available globally within WhatsApp and Instagram DMs.

Microsoft AI - Microsoft Launches 7 AI Models Across Coding, Voice and Reasoning.

Google
- Generative AI Search features now include inline links and website previews to encourage users to visit websites.
- Added Preferred Sources in AI Overviews and AI Mode with new subscription labels.
- Introduced a new control to manage how website links and content appear in generative AI Search features.
- Rolled out new insights for website owners about the appearance of their pages in generative AI Search features.

In this panel from the AI Engineer conference, Google DeepMind’s Ian Ballantyne and KP Sawhney detail how they run AI agents at scale. They showcase "anti-gravity," an IDE-integrated agent manager with a planning harness, visual browser tracing, and human-in-the-loop validation. To combat massive token consumption and high costs, DeepMind mixes advanced commercial models with local Gemma models while using mock hardware environments for budget-friendly evaluation. For code quality, they deploy a Darwinian "survival of the fittest" selection process for agent skills and leverage automated, fine-tuned reviewing agents to maintain enterprise style guides.

This Conf42 Golang 2026 talk by Sonali Priya focuses on building scalable, reliable, and performant design systems for enterprise analytics platforms. She outlines five core pillars to prevent component drift and code duplication: Component Modularity (layering primitives, composites, and feature modules), Token-Driven Theming (using a 3-level token abstraction), Cross-Platform Consistency, Versioning Discipline (strict semantic releases), and a Federated Governance Model to balance developer autonomy with system alignment.

Mastering Claude Code: Beyond The Basics
This guide will teach you how to unlock the full potential of Claude Code by treating it as an autonomous agent, not just a chatbot. Learn how to give Claude a way to verify its own work, use plan mode to design and review code, delegate tasks, and properly understand the .claude directory.