Last week, I had my first-ever viral post on social media. With over 100,000 views and nearly 1000 comments on LinkedIn, it is quickly becoming known as “the iceberg UX post” – you can visit the thread here: https://www.linkedin.com/feed/update/urn:li:activity:7310349394065666050/ 

(And if you are already a subscriber to the UXforAI.com, you will get my SPECIAL REPORT:  How to Become Indispensable: AI-First User-Centered Process Guide for the New Normal shortly in your email inbox.)

I wanted to share 7 lessons I learned in the process with all of you who have made this amazing journey possible. Thank you! Please keep your comments and questions coming – and let’s figure this new normal out together.

The Challenge

Having a viral post has been a wild experience for me -- I got way more responses than I expected. In the post, I promised to send my new special report to everyone who commented on the post. LinkedIn didn't offer any tools or support to help me scale, and when I tried to work with humans, they were even less trustworthy and reliable than AI. So, scaling wasn't really an option. I wanted to give everyone who responded the personal touch I committed to, but with traffic to the post being insane, I had to prioritize. So, I had to stop and figure out what was important to me – why was I doing this? What was I trying to get out of this viral post? This led to 7 lessons, which include profound insights about myself, what is happening with the UX industry, and how we can together build a future where UX can thrive and provide incredible value:

1. AI is extraordinarily capable at marketing

2. AI is easier than VA

3. Vibe Coding is super-capable, but deploying the code is another matter entirely

4. You need to know your why

5. There's simply no substitute for authenticity

6. The UX stress is real

7. Find your “UX Moat”

I hope these lessons and insights help you in this singular crazy moment in UX history.

Lesson 1: AI is extraordinarily capable at marketing

While there's no substitute for judgment, taste, and meaningful human connection, today’s AI is quite capable of an extraordinary level of content creation. AI will not give you a single original thought, but it will absolutely help you package your creative thought in a way that will resonate with your audience even better than you know how to do yourself. AI is not just (in the words of Christopher Noessel) a “bullshit generator” – it is a Zeitgeist Hype Machine. (Which is actually terrifying… Because of, you know, politics. Having had this viral post happen gave me tremendous insight into the results of the 2024 US election that I have not had before. More to come on that front.) 

For now, just imagine a miracle Zeitgeist Hype Machine that can, in near-real-time, analyze what your target audience cares about, then skillfully tie what you are writing about to that immediate need in that moment. And create 10-15 awesome versions of your ad in seconds.

That’s AI today. 

And it is already extraordinary. 

However, if you consider that all that content can be tied directly to individual needs and preferences at virtually no additional expense, it’s unprecedented. We’re not just talking about individualized marketing – we are talking about complete individualized product experience: information architecture, product selection, imagery, content, font, layout, etc. – essentially, the entire end-to-end UX can be 100% individualized at almost no cost to the company. I wrote about this AI-first trend and how to let go of control and successfully embrace AI-first design techniques for complete individuality in Chapter 12 of my book, UX for AI: A Framework for Designing AI-Driven Products (upcoming May 13th from Wiley, pre-order on Amazon: https://amzn.to/4l2ShyL). 

Lesson 2: AI is easier than VA

When I saw the post take off, I tried to do what I’ve done in the past – hire a Virtual Assistant (VA) to help me handle the extraordinary levels of traffic, respond to queries about the new book, requests for conference talks, and the like. 

Having done business with Freelancer dot com in the past, I decided to try there. Unfortunately, things have really, ahem, “evolved” a great deal since my last engagement with them. This time, I tried three different VAs, each of which profoundly and completely disappointed me in various ways:

1. The first VA wanted to immediately charge me a fixed price at triple my original agreed-to rate for answering queries using a template, which I thought was excessive. 

2. The second VA I hired was utterly incompetent, claiming 13 years of experience but not having the foggiest idea how LinkedIn comments worked or that LinkedIn had 2nd- and 3rd-level connections. Failing a simple task after 1 hour of instruction, that person, too, had to go. 

3. The third VA was the charm: literally. The charming person without a last name in another country wanted my LinkedIn password credentials before she even accepted the contract. When I objected that we don’t even have a contractual relationship yet, she said, “I am a verified Freelancer, and you can trust me.” Whew. Talk about A HUGE RED FLAG. 

Now, I’m sure there are many trusted, awesome people doing great VA work out there (I’m still looking, so ping me if you can recommend someone). However, the space has become so saturated with scammers and nincompoops that it’s now become quite difficult to sift out honest, skilled folks who will help you with basic routine tasks at a reasonable price. 

This brings me to AI: 

Lesson 3: Vibe Coding is super-capable, but deploying the code is another matter entirely

Much has been made of AI not completely replacing human beings, which is true. However, for simple tasks like digital virtual assistance, a somewhat mediocre mainstream AI agent would be preferred by most people over the stressful, negative experience I had with my VAs. With AI, there is no negotiation, no surprising price increases, no weird requests, etc. AI takes instruction extremely well – no hours wasted on nincompoops! As to the question of trust, I’d much rather give my credentials to a mainstream AI agent than to a potential scammer pretending to be a legitimate VA.

“Vibe Coding” (or a fancy way of saying “simply tell the agent what you need and have it write the code for you”) is all the rage. AI is actually pretty awesome at this. However, even with me knowing how Selenium plugins work really well (my team and I designed a highly-rated Selenium-based Synthetics monitoring UI for LogicMonitor), it was just way too much hassle to completely deploy a working Selenium code for my LinkedIn tasks. Interestingly, ChatGPT was supremely confident it could “totally make this work,” but in reality, it came up short even in the “deep research” mode. Maybe Manus.ai would have done better. However, I discovered that doing even basic LinkedIn tasks with Selenium would violate LinkedIn policies designed to prevent exactly this kind of thing from happening, so that was another failed avenue of streamlining my communications.

BTW, I think there is an exceptional market opportunity to be had for creating an AI hosting platform capable on interacting with other agents that want to deploy code – some kind of a trusted one-time key can be established to push code to this platform and then the code agent can set you up with a low-volume production software stack: maybe a Dynamo DB, EC2 for compute with K8s clusters, S3 Buckets for images, something to host a React front-end, etc. So all you’ll need is to pass those 1-time deployment keys to your local agent and your code will be deployed on the other end. Then the keys would expire, and the whole agent-to-agent code deployment handover would be handled securely. Then, the human operator would be able to log in to the hosting platform with a different private password and do the deployment magic there with a deployment agent. Perhaps we’ll see this soon.

At this point, I started to really freak out. How would I process 100s of LinkedIn communications when I am working full time and have a key deadline at work to boot? 

Which brings me to my next lesson:

Lesson 4: You need to know your why

Sure, a viral post was amazing! Who would not want that? But I was really, really starting to lose my sh*t. I had to stop and take stock. I went for a walk out in the California hills and let my mind relax and let go. I asked a simple question: Why? Why am I doing this? Do I want to sell more of my upcoming book? Sure! But I know it’s not going to be another 50 Shades of Grey, so I’m not going to become rich off of it. (Let’s not kid ourselves here: it’s a bloody design techniques book. Some nerds like me find that kind of stuff titillating, but we are a rare breed.) Besides, I never really gave two sh*ts about being Scrooge McDuck or Muskrat – I just wanted to build cool things, travel the world, and be free from want.

So what then? Did I want to scrape all of your emails for something? So I can claim my list now has 50,000 followers? Who the f*ck cares! My list is already growing faster than I can currently keep up with, so that clearly was not the goal. 

(That’s what happens when the universe gives you what you want – it causes all sorts of uncomfortable questions in your mind.) 

So…Why exactly – why am I doing this? 

Because I want to learn together. I believe in the power of UX design to bring out the best in technology. And I want to continue to be part of our special tribe that empowers people with cool technology. 

However, our industry is going through a major transition. The cost of 90% “good enough” design for routine projects is fast approaching zero. And many companies are throwing their awesome UX people out with the bath water. I feel that I have both the experience (35 AI-driven projects) and the motivation (6 books, duh) to contribute to this conversation in a tangible way, and I want to help people who are struggling to understand how to deliver value in this new AI-first world. 

So, what I really wanted from my viral post is to connect with all of you who want to continue the conversation to understand what is truly happening with our industry and how we can together reinvent the next incarnation of our craft. (And yes, to promote my new upcoming book, because I wrote it to help my tribe to do precisely that.) 

That’s my “why.” 

And you are my community.

But here’s the hard truth: there's simply no substitute for authenticity, no substitute for real human connection. We can’t outsource that because that is the very core of our UX being. 

Lesson 5: There's simply no substitute for authenticity

The real revelation came when I read this supremely self-aware, kind, and caring response from Theresa Carranza-Fulmer:

"No worries, was not expecting something immediately... you can take a step back at any point or ask for support from others you trust." — Theresa Carranza-Fulmer, Senior Engineer Technology and Innovation, Hyundai America (BTW, Theresa is awesome, and you should totally follow her on LinkedIn: https://www.linkedin.com/in/tlcarranzafulmer/)

Once again, as many times in the past 20+ years, my UX tribe was much wiser than I was – you understood my “why” at that moment even better than I did myself.

So, authenticity and connection being the goal, I answered every one of those communications myself, bringing my best presence and full attention to each and every message. I stayed up till 2 AM multiple days responding to messages and sending out the guide. All in all, I've sent many hundreds of LinkedIn communications. It's been a grueling week, but I've been amply rewarded with incredible connections to some of the most amazing UX practitioners all over the globe. Some of you have already written detailed, article-length messages to me about my guide, and I am supremely grateful to all of you who responded. 

My cup overflows. 

There are so many gems in your responses, and I will do my best to explore your questions in the coming weeks and months. (And if I haven’t yet responded to your message, please be assured that I am doing my utmost best to get to it, but it might take a bit of time, given the volume.) 

However, one thing is clear:

Lesson 6: The UX stress is real

Overall, my connection with many of you has been extremely informative about the stress our industry is currently under and why we have an urgent need to reinvent ourselves. Many people really are on the Figma Titanic right now, heading toward an iceberg they can already see right in front of them, unavoidable, rushing toward them fast. 

While just a few short months ago, most UX people were just excited about the cool mountain of snow in the middle of the ocean and thinking about throwing a few snowballs, now the sh*t is getting real with the layoffs and demands for seeing value, and increasingly aggressive handover of basic design functions of simple pages to AI. There are many reasons for this, as we’ve covered in the original article (https://www.uxforai.com/p/ai-is-flipping-ux-upside-down-how-to-keep-your-ux-job-and-why-figma-is-a-titanic-it-s-not-for-the-re), so I'm not going to repeat them here. 

One thing is clear: the UX industry is clearly going through some “interesting times,” and the pain and stress are very real. 

Which brings me to my last point: 

Lesson 7: Find your “UX Moat”

At this point in the evolution of the UX industry, it is more important than ever to “Find your moat.” Your moat cannot be “Figma auto-layout skills,” or “I make mockups based on PM’s Jira tickets,” or “I conduct usability surveys” – it has to be something more tangible, more valuable, more human-centered, preferably something AI has trouble replicating. 

Mine is being a Practical Futurist. I combine my understanding of AI technology, business ROI, market fit, and customer needs to help teams invent and ship novel AI-first experiences.

You need to find your own moat. 

It will probably be different than mine, simply because you are not me – your skills and motivation are different.

This is why I wrote the entire book on the subject:

(UX for AI: A Framework for Designing AI-Driven Products is shipping May 13th from Wiley. You can pre-order your copy now: https://amzn.to/4hXMYyb) 

In closing

As Bradley Hebdon so eloquently said:

"To be fair, I feel like the industry needed a shakeup anyway... I've been doing UX before it was UX, and I've witnessed how designers have slowly turned into "Figma Operators." No real UC problem solving, with focus rather on high-fidelity prototypes, design systems and hand-off to devs. While Figma has done some things really well, like collaboration - I think it's done some damage to a true UCD process. I know it's just a tool, but it has played a part in devolving and devaluing UX/UI roles." – Bradley Hebdon, UX Design Lead, SCE (Bradley is an active contributor to the topic of #UXforAI, and you should follow him on LinkedIn: https://www.linkedin.com/in/bhebdon/)

Bradley’s description of Figma’s role in bringing about our current situation reminds me of one of my favorite stories by Neil Gaiman, “The Sandman”: 

Morpheus (Dream) creates a ruby called the Dreamstone to house a portion of his essence to help him focus his power and control his realm. (Hey, it makes his job easy, and who would not want that?) However, in the process, Morpheus puts too much of himself into the ruby, causing him to become weak. This weakness allows a mere mortal to trap and capture Morpheus, using the ruby as leverage. 

This angry man then smashes the ruby, thinking he is killing Dream. 

But that does not kill Morpheus. 

Instead, Morpheus is able to reclaim the essence that had been trapped within the ruby, allowing him to become more powerful than he had been in centuries.

Figma is our ruby. AI is about to smash it. 

And for those who are ready to reinvent themselves and embrace their true power, this is an opportunity of a lifetime.

Godspeed,

Greg Nudelman and Daria Kempka (Contributing Editor)

P.S. The STRAT AI bootcamp is now online! And as my subscriber you can enjoy 20% off discount for a limited time. (And while it’s slightly more expensive than an average carton of eggs, it’s still cheaper than letting AI take your job.)

STRAT’s AI Bootcamp for UX Teams is the only workshop of its kind​ – a rigorous 3-day online training (May 13–15, 2025) that will reforge your skills for AI-driven design. This isn’t some fluffy conference or sleepy lecture series. It’s roll-up-your-sleeves learning with immediate impact:

• No panels, no fluff – just action. You won’t be stuck listening all day; you’ll be designing real AI solutions in team-based exercises from the get-go​.

• Learn from the best. Work side-by-side with top instructors from Amazon, IBM, Sumo Logic and more​. These are UX leaders who live and breathe AI innovation – and they’re ready to share what they know.

• Design beyond the screen. Practice shaping actual AI use cases, models, and data – focusing on UX strategy, not just pretty UIs​. You’ll learn to collaborate with data scientists and engineers, and emerge as the UX leader your AI team needs​.

• Lead the charge. This bootcamp isn’t about turning you into an AI researcher – it’s about equipping you to lead product design in an AI world with confidence and credibility.

Spots are limited and going fast – this is a live, interactive experience and we’re keeping it intimate. Now’s the time to act: use code SPKBC20X for 20% off your ticket (limited time)​. Don’t wait and miss out. When AI is driving the industry, there’s no “later” for those who fall behind. Join us at STRAT AI Bootcamp and level up in real time.

Get your ticket here: https://www.strat.events/