Heavybit

This week's #DevToolsDigest includes news and resources from Callum Akehurst-Ryan, Eric Lubow, Matt Overstreet, and more. https://lnkd.in/gNKDiJ8f

Heading to Heavybit's DevGuild: Write-Only Code Summit in San Francisco on May 7th. The thesis: a growing percentage of production code is about to be written by AI and never read by a human. Not skimmed. Not reviewed. It ships. The engineering practices we've built over the last 20 years weren't designed for this. One of the main reasons I signed up: Geoffrey Huntley is speaking. He invented the Ralph loop — a core pattern, used to build scalable, multi-feature, applications autonomously with a long-running AI agent harness. I've been applying his patterns to what we build at West Stack for months. Wasn't going to miss hearing him in person. The questions I want to explore: → When nobody reads the code, what replaces human review as the foundation of trust? For my clients in regulated financial services, "we didn't review it" isn't an option. → If one engineer with an AI agent can produce what a team of five used to, do you need fewer engineers or different engineers? → How do you measure quality when you can't read the code? In financial services, every output is potentially auditable. The frameworks for trusting unread code don't exist yet. These are the conversations I want to be in the room for. Link to the event in the comments — a few tickets left. DM me if you're going. #DevGuild

On episode 8 of High Leverage, Joseph Ruscio sits down with Scott Hanselman (Microsoft) to examine whether AI will accelerate software engineering or hollow out the path to mastery. From learning modes in coding agents to Star Trek economics, they explore how technology can support both productivity and craftsmanship. Tune in! https://hubs.ly/Q04djcNk0

Open-source software was declared dead many times. Yet today it powers most databases, operating systems, and web browsers in the world. LLMs will not be different. https://hubs.ly/Q04dg2vF0

On episode 50 of The Kubelist Podcast, Marc Campbell (Replicated) and Benjie De Groot (Shipyard) speak with Ivan Burazin (Daytona). The conversation dives into building high-performance sandbox infrastructure for AI agents, the technical tradeoffs behind ultra-fast compute, and the challenges of scaling to thousands of machines. Ivan also reflects on pivots, timing, and navigating rapid startup growth. Tune in! https://hubs.ly/Q04bwWVw0

The lines between researcher, engineer, and product are being redrawn faster than any org chart can capture. Google DeepMind's Paige Bailey joins Write-Only Code Summit on May 7 to decode what it means to build today, where the best work happens precisely because no one can draw a clean line around it. T-2 WEEKS: https://lnkd.in/g8bsXRFv

A little belated as I've been heads-down building: but very proud to share that 𝐌𝐮𝐬𝐢𝐜𝐚𝐥 𝐀𝐈 𝐡𝐚𝐬 𝐫𝐚𝐢𝐬𝐞𝐝 $𝟒.𝟓𝐌 𝐔𝐒𝐃, led by Heavybit with participation from BDC and Build Ventures. We're making attribution for AI-generated music part of core infrastructure. For publishers, labels, and artists, it means transparency and real participation in the AI economy. For AI companies, it means licensing paths, and the ability to scale without uncertainty. Very grateful to Jesse Robbins, James Lindenbaum and the Heavybit team, Daniel Nieto and Dinar Ahmed at BDC Capital, and Patrick Keefe at Build Ventures. And to our advisors vickie nauman and Elizabeth Moody. Extremely proud to be working on this with my co-founders Sean Power and Matthew Adell, whose conviction, drive and brilliance made this real, and the Musical AI team. Developing this core tech has been the culmination of a long journey: A very personal and special acknowledgment: to Philippe Pasquier at the Metacreation Lab for Creative Artificial Intelligence, who set me on this path of understanding the relationship between generative AI outputs and human input nearly two decades ago. And to James Maxwell, who built alongside me in generative music for a decade and whose unwavering commitment to artist rights helped shape the ethical foundation this work now rests on.

On episode 36 of Open Source Ready, Brian and John (Paper Compute Company) sit down with Jesse Vincent (Prime Radiant). They explore how Jesse’s Superpowers project turns AI coding tools into structured, reliable development systems. The conversation dives into agent orchestration, prompt engineering, and what it takes to make AI behave like a capable software engineer. Tune in! https://hubs.ly/Q04cD3cD0

At Notion, AI isn't just a feature. It's infrastructure. Sarah Sachs, AI Lead, is making sure it earns enterprise trust at scale. She joins Write-Only Code Summit on May 7 to share what responsible deployment looks like from inside one of the most widely adopted platforms in the world. Last chance for 20% off with code COMMUNITY: https://lnkd.in/gPk_Fgg3

On episode 35 of Open Source Ready, Brian Douglas and John McBride (of Paper Compute Company) sit down with Den Delimarsky of Anthropic. They discuss MCP’s rapid growth, its governance structure, and how it fits into the broader AI tooling ecosystem. Tune in! https://hubs.ly/Q04b1vzl0