Dustin Warner

AI and Technology Leader!
Build It, Break It, Fix It, and Repeat.
"memento mori"/"memento deployment"
T1D Cyborg.

[Erik Meijer @headinthebox]

Probably this book should be forbidden https://t.co/dFZiSrCVHh. Read it before it disappears just like your favorite LLM ;-)

(via https://t.co/tva8fgPLRk, not on X)

[Gergely Orosz @GergelyOrosz]

Things I really dislike about Fable:

1. Anthropic collects my prompt history, stores it, and does whatever they want with it for 30 days. No opt-out

2. They can nerf their most expensive model without telling me, billing me the same amount, wasting my time. Whenever they want

[Dustin Warner @DWarner90]

It’s not done if it’s on localhost

[dax @thdxr]

it's not done if it's not implemented
it's not done if the implementation is ugly
it's not done if it's not documented
it's not done if users can't discover it
it's not done if you can't market it

[Dustin Warner @DWarner90]

Supply chain attack every other day.

[Rhys @RhysSullivan]

2026

[Dustin Warner @DWarner90]

Love this part:

“What companies need is to empower their own employees, the ones who actually work together (even across departments), the ones who know how the business works, to shift the company in new directions together.”

I recently heard that I saved two separate departments ~20 hours a month just by taking time to do a 1-hour session on Codex.

We did not build a production system for everyone. We automated processes that never would have met the threshold for a traditional tech team to prioritize.

AI innovation is being killed by premature ROI calculations. We do not yet have enough historical data to confidently model future cash flows, determine appropriate discount rates, or estimate the long-term impact AI will have on organizations.

[Dustin Warner @DWarner90]

Check out my latest article: Why is this taking so long? https://t.co/TVM3O9JtAs via @LinkedIn

[Dustin Warner @DWarner90]

Very interesting article. I had this same discussion with some friends at MS Build last year, following a firm's model of partnership and how we could apply that to building out delivery teams. For us, it really came down to incentives. If delivery cost is equal, the more effective a pod is, the more profit each could earn. Pods would quickly lean into efficiencies that don’t show or have the potential to deliver value.

I am trying to think of the impact this has on internal development teams. Do those shrink just like in-house counsel over the years? What is the cost for companies that are technology enabled but not technology first?

[Gene Kim @RealGeneKim]

Here are more the amazing lineup of amazing speakers!

- Kent Beck, Software Engineer, Programmer & Author
- Joe Beutler, Head of Solutions Engineering, Strategics, OpenAI
- Tim Cochran, Principal, Amazon Software Builder Experience, Amazon
- Jason Cox, Executive Director, Global SRE, The Walt Disney Company
- CDR Collin Fox, Chief Strategy Officer, U.S. Naval Surface Forces / TF Hopper (AI/ML)
- Ben Grinnell FBCS, Board Member, Newton
- Devlin McConnell, Senior Manager - Emerging Technology | Generative AI at Vanguard
- Charity Majors, Co-Founder/CTO, Honeycomb
- Ryan Martens, Cofounder, Manifest AI
- Tisson Mathew, Founding CEO, Skypoint (famous for shipping Amazon Prime Now as an engineering director)
- Dr. Erik Meijer, Research Scholar, Leibniz Labs
- Daniel Neff, Senior Principal Cloud Architect, Adobe
- Dr. Tapabrata "Topo" Pal, VP of Architecture, Fidelity Investments
- Melissa M. Reeve, Founder, HyperadaptiveSolutions
- Ezra Savard, Engineer, Netflix
- Brian Scott, Cloud Engineering Leadership, Adobe
- Annie Vella, Distinguished Engineer, Westpac NZ
- Steve Yegge, Co-author of "Vibe Coding," Author of Gas Town agent swarm
- Dustin Warner, Director of Software Engineering, NRC Health
- John Rauser, Sr. Director of Software Engineering, Cisco Cloud Security

And my heartiest thanks to the programming committee who have been helping evaluate all the amazing proposed talks that you've sent me! Amy Willard, Christina Yakomin, Cornelia Davis, Tapabrata "Topo" Pal, Jeff Gallimore, Ben Grinnell FBCS, Scott Prugh

And our sponsors,
@gitlab, @LucidSoftware, @Sourcegraph, @xaltbc!

[Dustin Warner @DWarner90]

@AnthropicAI I have been trying to sign up for an Enterprise account for 2 weeks now. No email no follow up after the initial sign up link.

I think your sales agent may be broken 🤖

[Dustin Warner @DWarner90]

Two nerds. A lunch table. A napkin.

Dr. Bill Cross and I were talking about AI adoption, and he did what mathematicians do — turned it into an equation.

The insight: every AI tool decision has an efficiency curve and a cost curve, and they pull in opposite directions. Most teams optimize for one and ignore the other. The sweet spot exists, but only if you do all the math.

Article three in the series is live. Next up: how to use technology scouts to establish the efficiency value before your org commits.

#AIAdoption #TechStrategy #EngineeringLeadership #DecisionFrameworks

https://t.co/8FlLaJjHpa

[Dustin Warner @DWarner90]

Second article in my series on AI adoption in engineering teams is live. This one covers why the race to eliminate roles is the wrong move, and what's actually working when product, design, and engineering use AI together.

#SoftwareDevelopment #AI #EngineeringLeadership #ProductDevelopment #CrossFunctionalTeams

https://t.co/0rQ3G0NfmM via @LinkedIn

[Scott Hanselman 🌮 @shanselman]

From my TikTok - non-technical family friends have asked me to explain Clawdbot/OpenClaw/Moltbot and agents in the context of what they use, which is ChatGPT

[Dustin Warner @DWarner90]

Your Security, Compliance, and UX teams already have standards.

The problem?
Your engineers can’t easily find or use them.

So we built a framework to fix that.

The Expert Organization framework lets non-engineering teams define and own their standards in simple markdown repos. From there, AI skills automatically evaluate projects against those standards—before issues ever reach production.

How it works:
1️⃣ Stakeholders own their spec github repositories (Security, Cloud, UX, QA)
2️⃣ Spice AI indexes those specs using hybrid semantic (embeddings) + keyword search
3️⃣ Skills integrate directly with Claude Code, GitHub Copilot, and other AI tools
4️⃣ Engineers ask questions in natural language - right from their IDE or CLI
5️⃣ Issues are caught before PRs, not after

The key insight: stakeholders own their own context.

This isn’t generic advice scraped from the internet. These are your organization’s standards - complete with real-world constraints, exceptions, and edge cases.

Your Security team knows your exact compliance posture.
Your Cloud team knows which AWS patterns are approved.
Your UX team knows your accessibility requirements.

They document that knowledge once. The AI delivers it instantly, on demand.

Need to know the authentication requirements for HIPAA?
Ask the security-expert skill.

Wondering about AWS tagging policies?
Ask the cloud-expert skill.

No more scheduling meetings for quick questions.
No more waiting days for Compliance to respond.
The knowledge is always available.

Best of all, it’s implemented once and works for all engineers across the most popular AI coding tools.

We ran this against a personal project and saw:

• A critical HIPAA risk flagged in seconds
• 12 cloud infrastructure gaps identified
• 9 accessibility issues caught
• A prioritized 28-item work plan generated automatically

Stakeholders own the context.
Engineers get instant answers.
Everyone skips the meeting.

Demo and repositories coming soon. 🔥

#AI #Development #vibecoding

[Dustin Warner @DWarner90]

Dumbbell Development: A Director of Engineering's Guide to Progressive AI Adoption https://t.co/wOsHYQS2mJ

#AI #developer #SoftwareDevelopment

[Dustin Warner @DWarner90]

After 20 minutes your agent hung up. No flight to the destination until Tuesday.

[Dustin Warner @DWarner90]

@united come on do better. Every flight out of Chicago has been delayed or canceled. Gate agents are rude and your updates are worthless. How about stop putting “(Powered by GenAI)” it just makes a bad experience even worse. I say this knowing fully well you will use AI to respond to this post.

[Dustin Warner @DWarner90]

@bryan_johnson @Benioff are 🍄 coming to the Ohana floor next to the snacks and drinks?

[Dustin Warner @DWarner90]

@elonmusk #dietcoke created ai

[Julio Casal @julioc]

Microsoft Build 2025 just dropped tons of new stuff for devs.

Here’s what’s new:

𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝟯𝟲𝟱 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗧𝘂𝗻𝗶𝗻𝗴
Use your own company data to fine-tune Copilot responses.

𝗔𝗴𝗲𝗻𝘁𝟮𝗔𝗴𝗲𝗻𝘁 𝗶𝗻 𝗧𝗲𝗮𝗺𝘀
Let agents securely talk and work together inside Teams.

𝗔𝗴𝗲𝗻𝘁𝘀 𝗧𝗼𝗼𝗹𝗸𝗶𝘁 𝗳𝗼𝗿 𝗩𝗦
Build enterprise-grade agents in Visual Studio.

𝗔𝗴𝗲𝗻𝘁𝘀 𝗦𝗗𝗞
Programmatic access to M365 data for custom agents.

𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗔𝗣𝗜𝘀
Connect your app’s data to Copilot with structured APIs.

𝗔𝗴𝗲𝗻𝘁 𝗦𝘁𝗼𝗿𝗲
A marketplace to discover and deploy Microsoft 365 agents.

𝗠𝗖𝗣 𝗶𝗻 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗦𝘁𝘂𝗱𝗶𝗼
Use Model Context Protocol to wire tools and actions into flows.

𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗣𝗮𝗴𝗲𝘀
Turn Copilot answers into editable, shareable content.

𝗔𝘇𝘂𝗿𝗲 𝗔𝗜 𝗙𝗼𝘂𝗻𝗱𝗿𝘆 𝗦𝗲𝗿𝘃𝗶𝗰𝗲
Now GA for building and chaining multi-agent systems.

𝗣𝗼𝘄𝗲𝗿 𝗕𝗜 𝗖𝗵𝗮𝘁
Ask questions in natural language, get answers with visuals.

𝗙𝗮𝗯𝗿𝗶𝗰 𝗗𝗶𝗴𝗶𝘁𝗮𝗹 𝗧𝘄𝗶𝗻 𝗕𝘂𝗶𝗹𝗱𝗲𝗿
Model systems with a no-code editor—now in public preview.

𝗣𝗼𝘀𝘁𝗴𝗿𝗲𝗦𝗤𝗟 𝗶𝗻 𝗩𝗦 𝗖𝗼𝗱𝗲
Explore Postgres with the new Azure extension for VS Code.

𝗦𝗤𝗟 𝗦𝗲𝗿𝘃𝗲𝗿 𝟮𝟬𝟮𝟱
Private preview adds vector support and AI search.

𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗶𝗻 𝗦𝗦𝗠𝗦
Get help writing and debugging SQL in SSMS 21.

𝗣𝗹𝗮𝗻𝗲𝘁𝗮𝗿𝘆 𝗖𝗼𝗺𝗽𝘂𝘁𝗲𝗿 𝗣𝗿𝗼
Premium access to geospatial and environmental data APIs.

𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆
Share and reuse scientific datasets and models.

𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗰𝗼𝗱𝗶𝗻𝗴 𝗮𝗴𝗲𝗻𝘁
Copilot now reasons and edits code based on goals.

𝗦𝗥𝗘 𝗮𝗴𝗲𝗻𝘁
Get help resolving incidents with a new reliability-focused agent.

𝗚𝗶𝘁𝗛𝘂𝗯 𝗠𝗼𝗱𝗲𝗹𝘀
Self-host and fine-tune open-source models.

𝗢𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗖𝗵𝗮𝘁
The VS Code Copilot Chat extension is now open source.

𝗠𝗖𝗣 𝗜𝗱𝗲𝗻𝘁𝗶𝘁𝘆 & 𝗔𝘂𝘁𝗵
Add secure access control to MCP-based agents.

𝗠𝗖𝗣 𝗥𝗲𝗴𝗶𝘀𝘁𝗿𝘆
Public registry for discovering trusted MCP servers.

𝗡𝗟𝗪𝗘𝗕
Spec that lets websites expose agent-ready conversational UIs.

𝗪𝗶𝗻𝗱𝗼𝘄𝘀 𝗔𝗜 𝗙𝗼𝘂𝗻𝗱𝗿𝘆
Run small LLMs locally with new APIs and tools.

𝗠𝗖𝗣 𝗼𝗻 𝗪𝗶𝗻𝗱𝗼𝘄𝘀
Apps and agents can now call each other securely.

𝗪𝗶𝗻𝗚𝗲𝘁 𝗖𝗼𝗻𝗳𝗶𝗴
Automate dev environments using declarative config files.

𝗪𝗦𝗟 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲
Windows Subsystem for Linux is now fully open source.

𝗡𝗲𝘄 𝗖𝗟𝗜 𝗘𝗱𝗶𝘁𝗼𝗿
A new lightweight edit tool now ships with Windows.

𝗣𝗼𝘄𝗲𝗿 𝗔𝗽𝗽𝘀 𝗪𝗼𝗿𝗸𝘀𝗽𝗮𝗰𝗲
Copilot now helps structure and build solutions in Power Apps.

𝗔𝗴𝗲𝗻𝘁𝘀 𝗶𝗻 𝗣𝗼𝘄𝗲𝗿 𝗣𝗮𝗴𝗲𝘀
Embed agents into Power Pages for interactive customer flows.

𝗗𝘆𝗻𝗮𝗺𝗶𝗰𝘀 𝟯𝟲𝟱 𝗠𝗖𝗣 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻
Dynamics 365 apps now support agent-driven workflows.

Which announcement are you most excited—or skeptical—about?

--
Get my Free .NET Developer Roadmap 👇
https://t.co/Ed7r4wgPJK

[Ryan Holiday @RyanHoliday]

“When a man can’t find a deep sense of meaning, they distract themselves with pleasure.” — Viktor Frankl