DevGuild
Write-Only Code Summit
Write-Only Code Summit
AI code generation, multi-agent workflows, and shifting release practices have ushered in an era of write-only code, one where software outpaces human attention. But for those with real customers, velocity cuts both ways. Enterprises still expect a high bar for quality, performance, and trust. As engineering decision-makers, our work has never been more valuable.
DevGuild gathers technical founders and infrastructure leaders redefining the software development lifecycle. If you're building the culture, processes, and tools for a write-only future, join us.
Agenda
Session #1
Doors & Registration
Session #2
Opening Remarks
Session #3
Software Development Costs Less than Minimum Wage
Coding loop costs are outrageously cheap and the agents aren’t sullen in standup. Latent Patterns founder Geoffrey Huntley emerges from a goat farm on Kangaroo Island, creator of the enigmatic Ralph Loops, with a message equal parts warning and invitation: the developer's identity is being rewritten, and the curious ones are having the time of their lives. PMs ship product, tour guides run trading bots, and language tribalism is dead. The only tribes that matter now are small groups of wandering, observant flâneurs. Stay sharp and keep up.

Session #4
The Illusion of Competence: A Case for Governed Reasoning
AI is moving into messy, high-stakes domains where evidence is incomplete and often conflicting. It’s fine for code gen, but when it comes to managing engineering teams or systems diagnosis, agents collapse uncertainty into confident, untraceable answers. This creates an illusion of competence: outputs sound right, but the reasoning is opaque, inconsistent, and impossible to audit. Dr. Stephen Barrett argues for a control layer of “governed reasoning,” where AI decisions are structured, stateful, and constrained like software, making them inspectable, reliable, and safe to use at scale.

Session #5
Validation, Digital Clones & The Case for Rapid Fire Bets
The SDLC is no longer a pipeline for software, but a capital allocation engine. One where the only thing that matters is whether your bets pay off. Join StrongDM cofounder Justin McCarthy as he discusses his digital twin universe strategy, one where agents spin up clones, run thousands of simulations, and place rapid-fire wagers on what might work.

Session #6
Lunch & Roundtable Discussions
Roundtable Topics:
Multiplayer AI & Collaboration
Getting Return from the New AI SDLC
Software Factories Show & Tell
AI-Native Culture, Hiring & Org Design
Architecture Patterns & New Compute Primitives
Security and Trust in AI Code-Gen
The Future of Open Source
AI Ecosystem Dynamics: competition, partnerships, and platform wars
Scaling Dev Tools Interviews
VC Office Hours
Session #7
The Unbearable Lightness of Titles
Companies are redrawing the lines between researcher, engineer, and product faster than any org chart can capture. The sharpest people aren't climbing ladders or building fiefdoms, they're expanding into everything all at once. Join DeepMind's Paige Bailey as she decodes what it means to build today - where the best work happens precisely because no one can draw a clean line around it.

Session #8
Designing for Agents: Lessons from Sentry
Agents draft specs, revise plans and iterate continuously through design patterns and frameworks. The question we're circling in this fireside: What even is design when the product is as fluid as water?


Session #9
Code Review Has to Go
mAgentic development doesn't just change how we write software — it reveals how much of our traditional process is obsolete. Pair review, approval gates, bureaucratic guardrails: cozy rituals from a time when humans wrote every line. David Crawshaw, cofounder of Exe.dev, is ready to rant for change. He’s arguing for better version control that doesn't assume human authors, workflows that don't bottleneck on peer review, and the courage to ask which rules actually prevent disasters versus which ones are performative adulting. You don't have to be a cowboy. But you probably need to question your sacred cows.
Session #10
Control Plane Automation: Agent-Led Trust and Safety
Agents can write, review, and ship code — but how accountable can we be for systems we never fully scrutinize? Flags like --dangerously-skip-permissions capture a deeper tension: when speed and autonomy collide with the governance, security, and observability enterprises require. This panel uncovers what it takes to let agents loose on production systems without losing sleep: permissions, observability, and the guardrails keeping it all from imploding.




Session #11
Roundtable Discussions
Session #12
The Last 6 Mos of the Agentic Engineering Revolution
Coding agents got really good in November. Our digital coding interns can rise to real coding challenges provided we give them clear goals, the right tools, and context. But steering them well is its own craft. After a full day of agentic engineering talks, we've seen what's working: some of it echoes how great teams have always collaborated, some of it is deeply weird, and very little of it involves a pelican riding a bicycle. Simon Willison shares what’s working and brings it home.
Finale + Closing Statements - Joe Ruscio






