This week, we’re shifting focus from the model layer to the data layer — where every great AI system begins. Learn how dlt, a fast-rising open-source Python library, is automating the toughest parts of data engineering for more than 3 million monthly users. It’s the kind of project quietly transforming developer workflows behind today’s AI products — letting teams move from “data chaos” to “data flow” in minutes.

Let’s get into it — from automated data pipelines to self-learning agents, this week’s releases all point toward one clear trend: AI development is becoming faster, more autonomous, and more deeply integrated into the tools we already use.

How dltHub's open-source Python library helps developers create data pipelines for AI in minutes

dlt is a free, open-source Python library that automates key data-engineering tasks, already being used by over 3 million monthly users across 5,000+ companies—including regulated industries like finance and healthcare.

For developers and teams building AI systems, data engineering is often the hidden bottleneck. dlt offers a way to speed up data workflows, reduce friction, and focus more on model development and less on plumbing.

Building AI isn’t just about LLMs or flashy models — it’s about reliable, scalable infrastructure. A tool like dlt underscores the shift: developers need efficient data pipelines, not just big models.

“The catalyst is dlt… [which] automates complex data-engineering tasks… the tool has reached 3 million monthly downloads and powers data workflows for over 5,000 companies.”

A library or tool is only as good as its integration and governance. Even with dlt, teams still need to handle data quality, compliance, and architecture choices.

Adobe showcased several new generative AI features, including Firefly Custom Models for personalized content, the Adobe AI Assistant to streamline workflows, and Natural Language Editing in Photoshop. Adobe Express also now integrates with ChatGPT, enhancing dynamic content creation. These updates highlight Adobe's commitment to AI-driven creativity.

Google redesigned AI Studio with conversational "vibe coding" that goes from prompt to working AI app in minutes. Developers no longer manually configure APIs—AI Studio interprets requirements and auto-wires Veo (video), Nano Banana (image editing), and Google Search. The platform includes Annotation Mode for visual editing, a model gallery, and contextual brainstorming suggestions. It's positioned as Google's answer to Vercel v0 and other low-code AI tools.

Perplexity released Patents, the world's first AI-powered patent research agent that eliminates complicated keyword searches. Users can ask natural language questions like "Are there any patents on AI for language learning?" The beta is free for all users, with Pro/Max subscribers getting additional quotas.

Microsoft released Agent Lightning, a reinforcement learning framework that lets AI agents learn from their own mistakes without requiring system rebuilds. The open-source framework watches agent performance, identifies what works and fails, and teaches improvement through automatic intermediate rewarding. Developers can now plug it into projects to continuously enhance AI agent capabilities.

The company released its “Creative Operating System”, integrated design-AI for editable layered objects (not just flat images), spreadsheets tied to app-widgets, and made its Affinity design suite free for all users.

Another week of separating AI signal from noise. If we saved you from a demo that would've crashed prod, we've done our job.

📧 Got a story? Reply with your AI tool wins, fails, or war crimes. Best stories get featured (with credit).

📤 Share the skepticism: Forward to an engineer who needs saved from the hype. They'll thank you.

✍️ Who's behind this? The Augment Code team—we build AI agents that ship real code. Started this newsletter because we're tired of the BS too.

🚀 Try Augment: Ready for AI that gets your whole codebase?