Product
A product gives someone a capability.
Good products lose to good ecosystems.
I engineer ecosystems. Product is one piece of that. So is community, knowledge, self-service, support, and how someone learns to use the thing in the first place. Most companies build these separately. I build them as one system.
I write about it every week in Hive Five, along with whatever else caught my attention in security, AI, and work. Drop your email if you want it.
Join more than 4,000 security pros, hackers, builders, and curious minds.
I think of this work as world-building.
Every product creates a world around itself.
That world has rules people must learn, paths they can follow, knowledge they can trust, places to ask for help, and ways to contribute what they know.
Most teams design the product and let the rest of that world emerge by accident. I engineer the whole thing.
System
A system makes that capability dependable and repeatable.
Ecosystem
An ecosystem lets people participate, combine what others have built, and make the whole thing more valuable for the next person.
A pile of disconnected systems is still a pile. An ecosystem carries knowledge, context, and trust between them.
That is where products stop being something people merely use and become something people can build upon.
Selected work
A few worlds I've built.
What I'm building now
CyberAgents Exchange
At Tenable, I help build CyberAgents Exchange, an open platform where security practitioners can discover, share, and combine AI agents, skills, MCP servers, and playbooks.
The website is part of the product. So are the code, contribution paths, shared vocabulary, review process, creator attribution, and the rules that help people decide what they can trust.
That is what I mean by ecosystem engineering. I build the product, then build the conditions for other people to extend what it can become.
Before that
ExtraHop
At ExtraHop, I led digital support for engineers troubleshooting their own networks: community, knowledge, self-service, and AI-assisted help.
The community is part of the product. So are the knowledge articles behind it, the governance that keeps those articles true, the search that has to surface both, the routing that decides who sees a question first, and the Salesforce work I shipped myself so fixes never waited in an engineering queue.
Built separately, those pieces send people in circles. Built as one system, they took $192,000 a year out of cost-to-serve and grew community-assisted resolution from 4.96% to 8.8% of incoming volume.
2021 to 2022
Bugcrowd
At Bugcrowd, I owned researcher experience for hackers working through programs, policy, payouts, and the occasional rejected submission.
The help center is part of the product. So are the Discord, the forum, the escalation path, and the moment a researcher decides whether the answer they got was honest. Each of those belonged to a different team, and no one owned the route between them.
I rebuilt the whole route around what researchers were trying to do instead of who internally owned it. Tickets dropped 10%. Repeat questions dropped 25%.
Earlier
Federacy
At Federacy, a bug bounty platform for startups, I built the entire front end for both sides of the market: the researchers hunting and the companies paying them.
The interface is part of the product. So are registration, onboarding, the dashboards, the submission workflow, the commenting system, and the payout flow. Design, UX, and engineering, all of it mine.
A researcher who signs up on Monday and never submits anything is not a product problem. It is a path problem. I owned the whole path.
On my own
Hive Five
Every week I send Hive Five, a newsletter where security meets humanity. Research, AI, tools, work, and the human ideas underneath the technology.
The email is part of the product. So is the site, the archive, the people who reply and tell me what I missed, and the taste that decides what earns a spot in five slots.
It is the one ecosystem I own outright, which makes it the place I test everything I believe about the rest of them. More than 4,000 people read it.
I study these worlds in public.
I walk through products the way a customer or user does and look for the moment the system stops helping.
Sometimes I find friction. Sometimes I find a decision worth stealing.
Flow
Examples of products that anticipate the next question, carry context forward, and quietly keep people moving.
Community sits on the path, not beside it
A peer thread gets someone to the answer without a ticket, because community sits on the path instead of beside it.
The article closed the job, not the feature
An article closes the job the customer came to do, instead of describing the feature.
The handoff carried what was already tried
When self-service fails, the handoff carries what was already tried, so the customer does not start over.
Friction
Short records of the moments where the system drops context and makes the person do its work.
Spotify: the help path hides behind account chrome
Finding support is a thirty-minute maze; the help path hides behind account chrome.
1Password: the journey breaks when platforms do not share context
Help-center search is excellent; the journey breaks when platforms do not share context.
New York City: capability without access
Capability exists (inventory, schedules, programs); access does not. The product behind the product is tribal knowledge.
Reviews
Full, recorded walks through a real product and the world built around it.
Longer arguments, weekly signal.
The longer pieces are where I work through what those moments teach us about the whole ecosystem.
And every week, Hive Five collects the security research, AI developments, tools, and human ideas that made me stop and think.
Writing
Longer arguments about products, systems, ecosystems, and the people living inside them.
Community Is the Routing Layer
Pauline Narvas asked what job community is being hired to do. In CX, the answer is bigger than engagement: community turns customer intent into answers, access, and action.
The Product Behind the Product
New York didn't lack inventory. Every feature existed: Switch consoles, projectors, court schedules. But capability isn't access. A feature nobody can find isn't a feature. It's Shadow CX.
The CX Gate
Evan Spiegel said design is a bottleneck at Snap, and that's intentional. CX needs the same kind of gate: a formal moment before launch that asks not 'does it look right?' but 'does it feel right to be a customer of this?'
Hive Five
The weekly newsletter: security research, AI developments, tools, and human ideas worth passing along.
Hive Five 286 - How Will You Measure Your Life
Stealing Reasoning Traces from Proprietary LLM APIs, Bing Images RCEs, Cure all diseases, unslop, Claude Skills to Run a Newsletter, The Task Isn't The Job, Why Is Everyone In Tech So Sad
Hive Five 285 - The End of an Era
Meta Engineer's Agentic Engineering Workflow, GPT 5.6 Sol Found a Pre-Authentication RCE in WordPress With Minimal Human Input, Anthropic Removed 80% of Claude Code's System Prompt for Claude 5 Models, How To Use /goal to Find Bugs in Patch the Planet, How to Make a Nintendo 64 Game in 2026
Hive Five 284 - Control the Ideas, Not the Code
Anthropic Finds a Global Workspace Inside Claude's Mind, The Memory Heist: Tricking Claude Into Leaking Your Secrets, The AI Colander: Why Customers Pour In and Out, Labor is Now Free. Make Something, How to See the Most Popular US National Parks, Without the Crowds
A product has to do more than work.
Discover → Understand → Trust → Use → Share → Improve
This is the path a product travels when it becomes part of an ecosystem.
People have to find it. They have to understand what it lets them do and trust it enough to try. They need a path forward when something breaks. Then they need a way to share what they learn so the next person starts further ahead.
When those steps reinforce each other, the product gets easier to adopt, extend, and improve. When they do not, every new user pays the same friction tax.
Systems should change what people can do.
Work I have helped design and operate has produced results like these:
$192K
In estimated annual savings after customers could resolve more problems on their own.
226%
Increase in problems resolved before someone needed to file a support ticket.
Nearly doubled
Peer self-sufficiency after giving community a real role in how people found answers.
Lower cost-to-serve was the consequence. More people getting useful outcomes from the technology was the point.
Start with the gap
Is your product stronger than the world around it?
A capable product can still lose people in the gaps between product, knowledge, support, community, communication, and AI.
The Intent Gap Diagnostic shows you where that happens.
I walk your customer experience as an outsider, document where people lose momentum, trace the systems behind those failures, and show you what to fix first.
$1,500 · 5 business days · no internal access required
You receive:
- A recorded walkthrough
- An intent gap map
- A routing diagnosis
- A prioritized fix list
For deeper work, I take on a small number of engagements where I help teams build the product, systems, and participation model together.
Before you go, want the next Hive Five?
Every week, I collect the security research, AI developments, tools, and human ideas that made me stop and think.
I send the pieces worth passing along, plus a note about why each one earned your attention.
Free. Weekly. Written by me.
