Author Archives → David Hoang

Very few people are talking about the power of using Cursor and Obsidian together. I write and enjoy Obsidian in my day-to-day use, but can also use an IDE to batch edit Markdown files. There are likely loads of plugins in each ecosystem to do this, but the workflow is so simple and intuitive for me.

This is the power of file over app. More software interoperability, please.

Skills are the new currency

Originally posted on Proof of Concept

One of the most anticipated sequels during my younger years was The Matrix: Reloaded, the follow-up to the 1999 Keanu Reeves film that had already cemented itself as a sci-fi classic. Though visually stunning with high-octane action scenes, I felt underwhelmed by it, but this is not a movie review. One of the characters they introduced was Seraph, a pivotal protector in the world of The Matrix. I learned years later that the role was designed for legendary martial artist Jet Li, who rejected the role, which eventually went to Collin Chou.

Li turned down the role because the filmmakers wanted to digitally capture and own his martial arts choreography. While the opportunity was massive, Li balked at the idea that a studio could archive his movements, effectively owning his life’s work and repurposing it indefinitely. “We martial artists can only grow older,” he said, “yet they could own (my moves) as intellectual property forever.” In hindsight, Li’s refusal feels remarkably prescient. What he resisted then is now at the heart of conversations around generative AI and digital likeness.

NCAA Football 1999 and games before the NIL case only showed numbers, not the names of the players.

Attribution and credit to any craft is not a new phenomenon. In collegiate sports, there was a decade-plus-long lawsuit known as the Ed O’Bannon case, which challenged the NCAA’s ban on compensating college athletes for the use of their name, image, and likeness (NIL), after O’Bannon saw himself featured in an EA Sports video game without pay. The 2014 ruling found that the NCAA’s rules violated antitrust laws, helping pave the way for future NIL reforms. It marked a turning point in how college athletes are viewed, not just as students, but as individuals with marketable rights. Despite the “win” of the case, the players eligible for the lawsuit outcome were only compensated with less than $10,000 as a one-time payment.

In 2016, Star Wars: Rogue One was released with a scene of the iconic Grand Moff Tarkin in the prequel to A New Hope. There was one issue, however: Peter Cushing, the legendary actor who portrayed the Imperial commander, passed away 22 years prior. What was released was a CGI-generated portrayal of the actor with quite an uncanny valley.

As models increasingly train on the labor of artists, writers, and performers, new questions emerge around consent, ownership, and compensation. Li’s stance wasn’t just about martial arts; it was about control over the soul of your craft in a world where replication is effortless.

That tension is now at scale. When everyone can build everything with AI, the scarcity shifts. Execution is no longer the bottleneck—taste, judgment, and originality are. And that’s why I believe foundational human skills are becoming the new currency.

Take H&M’s announcement that they’ll create digital twins of 30 models. While the models retain rights to their likeness, key jobs in greater iconic photoshoots, such as photographers, stylists, and the rest of the crew, could be rendered obsolete. What happens when the commercial value of a person becomes detached from their physical presence?

This may sound like science fiction, but it’s already unfolding. When technology commoditizes an industry, one of three paths tends to occur. The first is that the entire industry may collapse entirely, replaced by more efficient systems. Computer word processing did this to the typewriter industry. The second path may be the remaining artifacts of the disrupted industry that increase in value due to scarcity, such as mechanical watches in the digital watch age. Third, entirely new economies and creative surfaces emerge from the disruption. I’m not arguing for one outcome over another, but these are the common patterns we see.

As generative AI continues to expand what’s possible for anyone to create, the most valuable asset may no longer be how much you can produce, but the quality, meaning, and how uniquely you think. Foundational skills like discernment, storytelling, and design intuition are no longer just ways to shape output. They are the product.

Commoditizing your skills

Consider this: we are doing this already in a mundane way without AI. Any person selling a course, doing consulting, or selling a digital product in any way is commoditizing their skills they believe are valuable.

Licensing human intelligence

We’ve long accepted that musicians can license beats, melodies, and samples. But what happens when we start applying that model to cognition? Imagine a future where a UX designer licenses their way of solving interface problems—an aesthetic judgment API. Or a therapist licenses their method of guiding a conversation, embodied in a chatbot that mimics their voice and approach.

This isn’t about monetizing outcomes. It’s about a valuing approach. The IP isn’t just what you create—it’s how you create it. Your worldview, methodology, and decision-making frameworks could become the licensed layer, not just the artifact you produced.

Your skills as a module

In this future, foundational skills might become modular—think of them as plugins for AI systems. You might not hire a strategist in the traditional sense, but you might integrate their framework into your product stack. Need help prioritizing your roadmap? Load the “first-principles PM” module. Want to generate a story world with internal coherence? Snap in the “novelist worldview” engine.

These aren’t just prompt templates—they’re encoded forms of human judgment. Your skill doesn’t get replaced; it gets abstracted, amplified, and distributed.

Outsourcing your digital twin

Let’s return to H&M’s digital models. While the individuals retain their rights, this raises a deeper question: What happens when the most valuable version of you is a digital one? When is your presence decoupled from your labor?

We’re entering an era where people will train models on how they speak, decide, and create, effectively generating personal APIs. In that world, owning your “digital twin” becomes owning your intellectual property. Just as actors fought for rights to their likeness under SAG-AFTRA, we may soon see similar protections demanded by designers, strategists, and domain experts.

Unique skills are the differentiator

I wonder if Jet Li would’ve made a different decision if his choreography had been treated not as a one-time capture, but as licensed IP, valued, protected, and compensated accordingly. Who knows? He might still have turned it down. But he saw early what we’re now being forced to reckon with: that our most irreplaceable asset is our skills.

As AI saturates creative and operational domains, the rarest trait becomes being deeply, idiosyncratically human. Whether you use AI or not, what you choose to create and bring into the world represents what you stand for. Digital twins, voice clones, and even the humble “pick your brain” email all point to one thing: the value of your ability is rising. The difference now is you may have a say in how it’s used and a share in what it creates.

Am I thrilled about this future? No. Working in AI doesn’t mean you have to agree with every development, but I believe it’s the evolving future, and I choose to help shape the implications of it.

Generative AI has flattened the effort curve. You no longer need to be a seasoned video editor to create a cinematic sequence or a programmer to ship a full-stack app. There lies the challenge of new originality. For skills to be perceived as valuable in the AI-native world, you need to be better than what AI can allow everyone to create. It’s sometimes sloppy and generic, but don’t underestimate the general population compromising for quick and cheap.

But if everyone can do everything, how do we distinguish good from great? That’s where we come in. Taste, context, ethics, and lived experiences are the new moat.

Grocery store was playing “Sweet Child o’ Mine” by Guns N’ Roses; had to do another loop in the aisles to hear Slashes’s guitar solo before leaving.

AI Design at Atlassian: we’re hiring

July 2025 marked my first anniversary joining Atlassian. When I first decided to join the company, people were surprised. Why would someone who has primarily worked at hyper-growth startups join a larger-scale company that had existed for two decades? The decision wasn’t about age, scale, or company stage. There were two areas I was looking for in my next role: leading through the craft of design and re-imagining the application layer experience in the era of AI.

I had a lot of trust that design craft matters with Charlie Sutton at the helm of Design. I was impressed with the continuous improvement Atlassian rolled out with the navigation refresh work and Atlassian Design System. Charlie and Josh Higgins gave a talk at Team ’25 in Anaheim about how Atlassian redesigned its products around a unified navigation and refreshed brand identity, aiming to create a more cohesive, modern, and collaborative user experience. They shared behind-the-scenes insights into the design and technical decisions that brought these improvements to life, inspiring large-scale design transformation.

Atlassian is a company that’s been around since the beginning of my career. It found business success early on and built strong foundations. I wanted to make sure they didn’t hit the innovator’s dilemma. I’ve primarily worked at founder-led companies throughout my career, and it’s my favorite type of leader to work for. In addition to Scott Belsky joining the board, I was very excited about the AI conviction the company has.

Rovo, Atlassian’s flagship AI experience

I joined in the midst of prepping for General Availability (GA). Jamil Valiani, my product partner, knew this was a must-ship moment and really fosters a startup environment within Atlassian. With the speed of AI changing, we have to. In October of 2024, at Team ’24 Europe in Barcelona, Rovo went GA! This was very much the beginning of our AI journey. At Team ’25 Anaheim, a few months later, we announced that Rovo will be available for most subscriptions with the aspiration to allow every customer to harness the power of AI.

Here’s a list of a few things we’ve designed and shipped in the short time Rovo’s existed:

  • AI-Powered Search – Instantly find information across Jira, Confluence, and connected third-party apps using natural language. Rovo Search delivers personalized, context-aware results, including Smart Answers and Knowledge Cards that summarize and explain key information.
  • Rovo Chat – Ask questions in plain language and get real-time, context-rich answers. Rovo Chat can summarize documents, provide insights, and even take actions like creating Jira tickets or sending Slack messages. It’s available in the browser, Confluence mobile, and as a browser extension
  • Deep Research – Use AI to search, analyze, and generate comprehensive, fully cited reports on any topic in minutes.
  • Studio – A unified space to build, manage, and deploy AI agents and automations—no coding required. Studio lets teams create custom workflows and automations powered by AI. ​
  • 3P Connectors – Rovo connects with tools like Google Drive, Slack, and GitHub, bringing AI-powered insights and automation to your entire workflow. 
  • ​Rovo Agents – Configurable AI teammates that automate tasks, analyze data, generate content, and help with decision-making. Out-of-the-box agents are available for common workflows, and you can build custom agents in Rovo Studio.

What’s exciting in the next horizon

Despite all the experiences designed and shipped, our team remains ambitious and hungry to push AI Design even further. There are three key investment areas I’m personally excited about.

The first is bringing consumer-grade AI experiences to the workplace. Most people’s experience with AI comes from their personal use of consumer experiences. For enterprises, matching the quality and craft of consumer AI isn’t just about keeping up with trends; it’s about meeting rising employee expectations, boosting productivity, and fostering engagement. When workplace AI feels as polished and empowering as the tools people choose for themselves, organizations unlock not just efficiency, but genuine enthusiasm and innovation from their teams. Well-crafted experiences are no longer a nice-to-have in Enterprise; they’re a differentiator.

The second area is empowering makers and builders with Studio. I have a close place in my heart for builders and makers throughout my tenure at Webflow, and have seen firsthand what impact people can make when the tools are empowering to people. With Studio, we’re continuing to invest in the Atlassian ecosystem, often without writing a line of code. This means teams can innovate faster, automate the “work of work,” and create tailored solutions that bring the magic and agility of consumer-grade AI directly into the enterprise.

I’m excited to expand Rovo’s surfaces to meet customers where they are. Work happens across various surfaces, devices, and apps; not only in the contained app we’re using. With AI capabilities such as MCP and other technologies, we’re moving into a world where a new MVC is emerging, and expansive surfaces allow different ways to interact with AI.

AI Design at Atlassian

The team I look after is Central AI Design (CAID). We are the center of excellence for AI and partner closely with teams across the Atlassian portfolio, such as Jira, Confluence, Trello, Loom, and many more. I break down our core work in three areas: Product, Platform, and Practice.

Product is the continued investment in making Rovo the best AI experience in the enterprise. By elevating the experience for AI Search, Chat, and 3P Connectors, we increase the capabilities for other teams to incorporate Rovo into their work.

AI is at the center of Atlassian’s System of Work, a new era of teamwork that spans beyond one persona or department. CAID’s Platform practices include our AI design guidelines, patterns, and working with other teams on how they apply AI to their experiences.

Amongst our design organization, we are all at different points in the journey of Designing AI with AI. Practice is AI enablement in our own org. Our goal is to ensure each human on our team can develop a point of view and understand the material of AI throughout their career at Atlassian. One highlight of this is Joel Unger, Principal Product Designer and long-time designer on Trello, who built a vibe-coded prototype in Cursor, testing breakpoints and generating images with Midjourney.

Why Atlassian Design?

What has always drawn me to Atlassian is the strength of its design community—a network of people like Jennie Yip and the collaborative spirit behind our design systems. Community is essential for design teams because it fosters peer mentorship, collective learning, and the open exchange of ideas. It’s where designers grow together, support each other through challenges, and build a shared sense of purpose and belonging. Especially after years of pandemic-driven disconnection, rebuilding these bonds isn’t just a nice-to-have; it’s the foundation for creative resilience and long-term impact.

  • Shipping impact at scale – Shipping to over 300,000 customers means every update has a real impact. Our job isn’t just to launch features—we help a huge, diverse community navigate the shift to AI. The scale is massive, but our team moves fast and keeps a startup mindset, making sure our customers can adopt new tools and stay ahead.
  • Community – This year, Atlassian Design hosted a series of Intentional Togetherness Gatherings (ITG) globally. We recently wrapped up the tour in our Sydney headquarters building with LEGOs and vibe coding.
  • Team Anywhere – Though I love spending my time in the San Francisco office, we’re a remote-first company and proudly Team Anywhere.

Join our team

We have an opportunity to build something impactful for our customers in this next era. Between developer tools, agile, AI, low-code, and no-code, the work at Atlassian is the amalgam of my entire career; been enjoying every moment of it. When it comes to the ingredients of a winning AI Design team, I’m looking for three key areas:

  1. AI Literacy – Applied AI experience and the desire to experiment and tinker with the new material of AI
  2. Craft – Strategy to pixels leadership—show, don’t tell
  3. Talent Density – Central AI Design has a high concentration of talent—excellence at all levels. To work on our team, one must have short toes and collaborate across the product, platform, and practices

We have various roles open across our team based primarily on the West Coast of North America West Coast (British Columbia, Seattle, and San Francisco) and Australia (Sydney and Melbourne).

Atlassian’s other teams are hiring key roles working on AI too:

I’ll close with notes I jotted when considering joining Atlassian. In any role I consider taking, I write down the end state of impact I’d like to have. I only join if I believe there is a possibility to pull it off. Here’s what I wrote:

  1. Atlassian is known as The Rovo Company.
  2. Define the winning AI interaction models that propel the industry moving forward
  3. Transform the Design Org to Design AI with AI
  4. Work with Charlie Sutton to make Atlassian a world-class design team

I’m so excited about the road ahead and hope you are too.

All the designers I know from the early mobile design era in the 2000s are now investors, starting companies, or forming studios.

The future of Lickable UI and high craft is essential in this next phase.