Summary: Jakob gains 100,000 followers on LinkedIn: How? | Video from Jakob’s live meeting with ADPList | AI model unifies document types for unified | Write like a machine, think like a human | New documentary about UX halfway done: watch the trailer | Meta’s new generative AI for images
UX Roundup for December 11, 2023. Hat tip to Bob Goodman for the Midjourney prompt used to generate the feature image for this newsletter issue. Data-driven design is a good UX lifestyle.
Celebrating 100,000 LinkedIn Followers 🎉
Thank you to each and every UX friend who is following me on LinkedIn. I recently passed 100,000 LinkedIn followers, which brings me into the medium league on that site. The big influencers have followers in the millions.
What’s more impressive than the absolute number is that I only started posting to LinkedIn in May 2023. I am amazed that anybody followed me before then when all I posted was an intermittent job opening every two years. However, after abandoning administrative duties, I got my old creativity back and have started posting original content.
My follower count is currently growing at an annualized rate of 155%. This growth rate may be hard to sustain as the absolute number of followers grows, but it might be possible to double in a year.
Celebrating 100,000 LinkedIn followers. (Illustration by Dall-E.)
There isn’t a “secret formula” to my growth. My mantra has always been simple: share content that’s original, insightful, and consistent. Note those three keywords: original, insightful, and regular posting. I see too much content on LinkedIn that’s regurgitated or lacking insight. And, of course, if you only post intermittently, don’t expect sustained follower growth. Regular postings are why I gained more followers in the last half year than throughout the previous 20 years.
In fact, I have violated most of the advice for LinkedIn growth hacking. I post long and detailed articles, which don’t do well on social media. But going deep is what I like. I prefer linear content over carousels with poor usability, even though the latter forces users to “engage” more to see content, which the algorithm likes. I also refuse to play tricks like “link in comment,” which supposedly makes you a darling of the LinkedIn algorithm. I am a usability guy, and the only usable place to link to the full article is within the post that announces that article. If you post the link in the comments, it risks being flooded by real comments. And in any case, it’s positively user-hostile to make users go hunt for where to click when they want more info.
I refuse to impose the extra interaction cost on my readers that LinkedIn’s algorithm supposedly prefers, with tricks like “link in comment” and carousels. This costs views but preserves my pride as a usability guy. (Maze by Dall-E.)
Once again, thank you for being part of my journey! 🚀
Video: Jakob Live With ADPList, Discussing AI, Heuristic Evaluation, & More 🎬
I am a huge fan of ADPList, which is likely the best way for UX newbies to learn about the profession and a great way for seasoned UX professionals to get inspiration by mentoring new folks and hearing what they’re up to.
They had invited me for a “fireside chat” a few days ago, and the video from that session is already available for your viewing pleasure on YouTube, in case you were not among the 2,400 live participants:
🎬 Fireside Chat with Jakob Nielsen | ADPList Growth Series 🎬 (79 min. video)
Screengrab from Jakob’s discussion with ADPList founder, Felix Lee. Live from Singapore and California, meeting with 2,400 UX fans from around the world.
Here are the main topics that we covered:
🌱 Jakob’s rebirth as a creative scientist and writer after a decade as a manager
📈 AI becoming a significant reality after decades of anticipation
🚜 AI's transformative potential in cognitive tasks as a forklift for the mind
🖥️ Drawing parallels from the PC and Web revolutions to predict AI's trajectory
✍️ Advocating for improved usability in AI products
📅 Jakob’s career journey through several tech revolutions
🔄 The enduring relevance of the 10 usability heuristics
🤔 Contemplating minor updates next year to reflect contemporary UX practices
🧠 AI as a tool to enhance, not replace, human creativity and intuition
⚖️ Balancing AI capabilities with human oversight and ethical considerations
🌍 Global expansion and growing demand for UX design
📉 Addressing recent setbacks like layoffs and salary adjustments
🔮 Predictions of continued growth and evolution in UX design and AI integration
Paul Smith used his superior Midjourney skills to generate this image of me having a fireside chat. It looks more like me than the image I made and included with my announcement of the session in my December 4 newsletter.
AI UI Beyond Chat
Luke Wroblewski is one of the world’s best user interface designers, so leave it to him to be one of the first to fulfill my vision of an AI UI beyond the chatbot: what I called the hybrid UI, combining prompting and GUI interactions. (Chatbots have horrible usability problems, such as the articulation barrier and the need to engage in laborious accordion-editing and apple-picking behaviors to cope with the infinitely scrolling chat).
LukeW has created an AI system based on his extensive content to provide design advice. So far, so useful, but not revolutionary. In fact, I wrote recently about how AI systems like Perplexity outperform traditional search as a way to answer users’ questions. I like his AI content, but I am particularly impressed by LukeW’s use of simple design ideas to vastly improve usability, relative to what we get from billion-dollar AI companies like OpenAI. (It’s disgraceful how big-AI ignores usability and don’t even fix design flaws that can be found in a day of user testing.)
LukeW has published a series of short videos discussing “how AI ate my website” — meaning how he fed all his content to an AI model and created a highly targeted way for his users to get their questions answered. It’s quite possible that this new AI question-answering model makes the “Ask Luke Wroblewski” bot the most useful current way of addressing your UX questions.
These videos are short and worth watching if you’re interested in improved ways of bringing knowledge to life with AI, without the constraints of the ChatGPT UI:
AI Models Enable New Capabilities (3 min.)
Expanding Conversational Interfaces (4 min. — the best of the videos IMHO)
Structuring Website Content with AI (3 min.)
AI is now standing by to answer your UX questions after consuming all Luke Wroblewski’s articles, PDFs, videos, audio recordings, and lectures. I like the unified interface to all these document types, focused on the parts of each that answer the user’s current question. (Dall-E)
Write Like a Machine, Think Like a Human: AI’s Writing Revolution
I was on the Writers in Tech podcast for a great discussion about the interplay of AI, UX, and content. I see enormous opportunities for writers and writing skills in our increasingly AI-driven world. I believe AI will vastly expand the demand and need for quality writing and human creativity.
The full 52-minute interview is available here:
In the past, AI tools that can generate text content automatically were feared as potential job killers that could make human writers obsolete. However, I argued that this is not the case at all.
Modernizing writing will create more jobs for writers. Not fewer. (Dall-E)
While AI tools can now produce raw text and drafts with unprecedented speed, the judgment, goals, branding expertise, and communication skills of human writers are still essential. AI-generated text on its own is rarely publishable quality without human guidance.
I believe AI will shift the role of writers from just generating text to being editors and evaluators of text. Future writers will spend more time providing creative direction to AI tools, reviewing outputs, revising based on brand needs, and selecting the best options.
The massive productivity gains of AI writing tools mean generating multiple text samples becomes almost free. Writers can get AI to produce many variations or alternatives on demand and then apply their skills to pick the best ones. This also facilitates experimentation and creativity.
By making high-quality writing scalable, AI enables new applications of writing for human experiences. More touchpoints, both physical and digital, will incorporate tailored writing. Demand for writing will likely grow faster than writer productivity.
I advised writers to start experimenting now with AI text tools like ChatGPT. Building skill directing and iterating with AI over time is essential to remain competitive. Writers adept at applying human creativity to AI capabilities will thrive.
When AI creates massive volumes of draft text, the writer’s role becomes that of the editor in charge of the final product. (Dall-E.)
The future role of writers is to provide the judgment, goals, framing, and editing to channel AI's raw productivity into final texts that meet audience needs. AI enhances rather than replaces the uniquely human parts of writing.
Writers who embrace AI as a productivity multiplier will have an advantage over those who fail to add this skill. I see writers as one of the top two professional roles that stand to benefit most from properly utilized AI collaboration. (The other being programmers.)
To summarize the podcast, I strongly believe integrating AI makes individual writers far more capable, effective, and valuable. By mastering AI as a tool, writers can reach new heights of creativity, productivity, and impact.
As an example of using AI to help writing, here are some of the headlines I got from pasting the above writeup into ChatGPT. Headlines 1-5 are “standard” offerings from the AI, and headlines 6-10 are what it produced when I asked it to be more daring and creative.
Revolutionizing Writing: How AI and UX Are Creating New Opportunities for Writers
Embracing AI in Writing: The Future of Content Creation and Writer's Role
Beyond Fear: How AI is Expanding, Not Replacing, the World of Writing
The New Age of Writing: AI's Role in Enhancing Human Creativity and Productivity
From Writers to AI Collaborators: The Evolving Landscape of Content Creation
Text 2.0: When AI Becomes a Writer's Co-Conspirator
Words at Warp Speed: AI's Turbocharge Effect on Writing
Ink Meets Algorithm: The Dazzling Dance of AI and Words
Prose in the Machine Age: AI Ushering in the Era of Super-Writers
From Quill to Quantum: How AI is Scripting the Future of Writing
Write at warp speed by using AI in your content-creation process. Image by Playground AI.
UX Documentary in Progress
Mitja Mavsar and Aleks Bezjak are halfway through a project to film 50 UX experts to create a documentary and a book. I was in the first interview round. They have released a 4-min. trailer, which is quite engagingly cut, so I think the project will turn out well.
They are crowdsourcing suggestions for whom to film in the second half of the project: the email address is in the description for the trailer video.
Mitja Mavsar and Aleks Bezjak are halfway through filming a documentary with UX experts. Who should they interview next? (Dall-E)
Meta Imagine: Not Impressed
Meta (the company behind Facebook and Instagram) has released a new generative AI tool for image generation. It’s free so far, so comparing it to the paid services is a little unfair. On the other hand, since there are plenty of image-generating tools already, we need something better, not just an also-ran, which this tool seems to be.
(To be fair, this is Release 1.0, and AI improves fast as it learns from user data. So, hopefully the next version will be better.)
As a simple example, I tried to generate an antique Greek marble bust of Socrates. Here’s the Meta Imagine philosopher:
Socrates as generated by Meta Imagine.
For comparison, here’s what I got from Leonardo:
Socrates as generated by Leonardo.
Neither image looks like Socrates: In Xenophon’s Symposium, Socrates himself mentioned that he had protruding eyes, a snub nose, thick lips, and a paunch. The images also don’t look like surviving busts of Socrates in leading museums, which follow the written description of his likeness.
These images are idealized Hellenistic depictions of a generic philosopher, not individualized portraits of history’s most famous philosopher. That nose in Leonardo’s version: more Roman than Greek! And a full head of hair, whereas Socrates is depicted with a hairline like mine (semi-bald) in all ancient busts.
To me, accuracy is the ultimate in prompt adherence. Image generation fails if the image doesn’t accurately depict the user’s request. Both tools thus failed my test.
Leonardo wins in several other ways:
Leonardo allows for multiple aspect ratios, whereas Meta Image is limited to square images.
Meta Image includes an annoying watermark.
Leonardo creates a vastly higher resolution, supporting many use cases like print posters. (These images were 3776x2122 pixels from Leonardo and only 1280x1280 from Meta Imagine.)
While it’s a matter of taste which image you prefer, Leonardo’s is superior to me. Meta Imagine’s bust seems made of some cheap 3D-printer resin instead of marble.