UX & AI Bites / No. 10

Hey! 👋

This time, we explore the following topics:
❌ Why did Duolingo abandon the UX Term?
🖼️ What exciting features does ChatGPT’s new image generation model offer?
🚀 What surprises does Google Gemini 2.5 PRO have?
✨ Can AI replace real user research?
🔍 What delightful update did the Postimees news portal implement?

1. Why Did Duolingo Abandon the UX Term?

Mig Reyes, Duolingo’s Vice President of Product Experience, announced on social media that they are replacing the term UX (User Experience) with PX (Product Experience).

He explained that their product drives their business, culture, and priorities. As a Duolingo user, I was somewhat surprised that the user perspective was omitted from this explanation.

In communication, I would have expected that Duolingo, as a product, brings all stakeholders—users and company employees—together for a common goal.

Does this mean that Duolingo no longer engages with UX and UI? They still do. It’s a case of “a beloved child has many names.” For several years, companies developing specific software, such as apps, have used the term “product,” and the designer’s title in these companies is “Product Designer,” with responsibilities similar to those of a UX/UI designer.

2. What Exciting Features Does ChatGPT’s New Image Generation Model Offer?

OpenAI announced several updates to their image generation model.

Previously, generating images with correctly displayed text was problematic, if not impossible. Now, the results are quite good, at least in English.

In Estonian, however, there are still issues with the letter “õ.” For example, AI tends to write “Palju Önne!” instead of “Palju Õnne!” on images. This persists even when repeatedly pointed out in chat or when attempting to clarify the issue to the AI using Unicode encoding. I hope this gets resolved eventually.

Regarding other updates, I appreciate that ChatGPT now better considers user-provided images as input and allows for the reuse of characters/elements.

3. What Surprises Does Google Gemini 2.5 PRO Have?

Google Gemini 2.5 Pro can create exciting simulations and prototypes, such as Rubik’s Cube, Lego builder, snake, flight simulator, and other games. Reportedly, it performs better than other AI models.

We live in exciting times where almost every day brings news of one AI model surpassing another. 🏁

4. Can AI Replace Real User Research?

Vitaly Friedman wrote in his blog that some companies have started using AI-based “synthetic” user research, where real people are replaced with AI-generated “users.” One service provider in this field is Synthetic Users.

Although this method is quick and convenient, it doesn’t replace observing real user behavior, emotions, or unexpected discoveries. AI creates only a plausible illusion but doesn’t provide actual understanding. Truly valuable insights still come from conversations and tests with real people. AI can be a useful tool but not a complete substitute for UX research.

I agree with Vitaly that superficial solutions usually don’t lead to the right path. I see AI’s benefit in making an initial assessment of a website’s hygiene factors—text size, contrast, etc.—before conducting user research.

5. What Delightful Update Did the Postimees News Portal Implement?

In Postimees’s mobile version, a new button is visible, offering the option to change text size.

I appreciate how Postimees has temporarily highlighted the new feature. The “Aa” button is surrounded by a conspicuous purple frame, accompanied by a small animation that makes the button slightly wobble. This invites users to notice and try the new feature.

Pressing the button opens a slider that allows users to change the text size. It’s convenient because you immediately see a preview of the text’s appearance, helping you decide whether the chosen text size is suitable.

Thanks for reading the newsletter to the end! 🙏