Catherine Zhang

👩🏻‍💻 Head of Product @ AI health startup · 📸 Lifelong photographer

[Catherine Zhang @CatherineYZ]

Can't wait to try! But if the minor HTML issues in the desktop UI were fixed, we would have more confidence in the details.

[Claude @claudeai]

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.

Its capabilities exceed those of any model we’ve ever made generally available.

[Catherine Zhang @CatherineYZ]

Someone who sees further than I do says we have about 1,000 days before everything changes.

So I’m writing until then. Day 1 starts with a question:

If a button could erase you completely, and no one would even remember you existed, would you press it?

https://t.co/SjpudclbJF

[Catherine Zhang @CatherineYZ]

Recently, I’ve been wondering if AI chatbots aren’t actually creating their own future clients, rather than stealing therapists’ jobs. It’s a thought-provoking idea.

With 62% of Americans now having seen a therapist (up from 48%) + 21% using AI mental health bots, we might be witnessing the ultimate gateway effect.

Think about it: AI chatbots could be the training wheels for therapy-hesitant people. They reduce stigma, teach mental health vocabulary, and let users practice vulnerability in a "safe" space before facing a human professional.

But here's what we don't know: Are people actually transitioning from AI → human therapy? Or are they getting stuck in the digital loop?

The data shows record therapy adoption alongside AI usage, but we need longitudinal studies to understand the real relationship. Is AI preparing people for deeper therapeutic work, or creating unrealistic expectations about quick fixes?

This isn’t just about technology disrupting our lives; it’s about whether our mental health “progress” is fostering genuine human connections or creating barriers.

What's your take? Have you seen AI tools lead to real therapy, or replace it entirely?

#MentalHealth #AITherapy #DigitalWellness

[Catherine Zhang @CatherineYZ]

Everyone writing about vibe coding addiction is writing from the engineer's seat.
But there's a reason it hits designers and PMs harder: our work has no ground truth. The bet has no settlement. The slot machine never says, "You won."
Three weeks in. Here's what I noticed. https://t.co/ZEM7ZeQT34

[Catherine Zhang @CatherineYZ]

Sources:
Schyvens et al. (2025) https://t.co/4JlPoBd5xY

Robbins et al. (2024) https://t.co/K53NKYSkgS

Doherty et al. (2025) https://t.co/y3fQI8aFm4

[Catherine Zhang @CatherineYZ]

I wore both an Oura Ring and Apple Watch to bed.
Oura: 1h 11min deep sleep.
Apple Watch: 29min.
Same night. 42 minutes apart. I work in AI health, so I went and read the actual research. Here’s what I found.

Both devices have been tested against polysomnography, basically a full sleep lab test where they wire you up with electrodes to monitor brain waves, eye movement, heart rate, everything. It’s the gold standard for measuring sleep accuracy.

A 2025 independent study (Schyvens et al., SLEEP Advances, 62 adults) found that Apple Watch overestimated light sleep by about 59 minutes and underestimated deep sleep by about 25 minutes compared to the lab. A 2024 study (Robbins et al., Sensors) found Oura showed no significant difference from the lab on any sleep stage.

The researchers themselves pointed out that most wearable companies “neither provide access to raw data nor insight into the algorithms.”

Credit where it’s due, Oura published their sleep algorithm in a peer-reviewed journal. Apple’s algorithm is still proprietary. WHOOP Recovery and Garmin Body Battery are the same story, nobody really knows how those numbers are calculated.

I happen to have the background to find these papers and read them. But millions of people check these numbers every morning to make health decisions, and most don’t have the time or access to audit the science behind a sleep score. That shouldn’t be their job. If you’re charging $300 for a health device, explaining what the data means and where it falls short is on you.

76% of major wearable manufacturers scored “high risk” on transparency reporting (Doherty et al., 2025, npj Digital Medicine). No margin of error in the app. No note saying “we tend to underestimate deep sleep.” Just a confident number.

When two devices disagree by 42 minutes on deep sleep, what is a regular user supposed to believe?

Should wearable companies be required to show margins of error? Or would that just confuse people more?

#WearableTech #SleepTracking #DigitalHealth #HealthTech #OuraRing

[Nav Toor @heynavtoor]

https://t.co/z4kQwYt7EA

[Catherine Zhang @CatherineYZ]

@claudeai My Claude won’t respond to anything now. Does anyone else have this issue as well?

[Deedy @deedydas]

If you're not using AI apps in 2025, you're at a huge disadvantage.

I've tried 100s of new apps this year, and here are the best 14 AI tools that actually stuck with me!

[Catherine Zhang @CatherineYZ]

@iconicselpink Did you went to Coachella last year?

[Catherine Zhang @CatherineYZ]

@duolingo My username is CatherineYZ. Please remove the number. Thanks!

[Catherine Zhang @CatherineYZ]

@duolingo Please remove the wrong number from my account. I can update my Candian phone number if needed. Thank you so much. So I don't need to merge my new account into my inactive account.

[Catherine Zhang @CatherineYZ]

@duolingo @duolingo Also, I searched online it says do not add any Chinese phone number. Once I did, it will mark me as a Chinese user and it cannot be undone. Is this true?
Now I am back to Canada and I cannot find the input box on UI to add my Canadian number.

[Catherine Zhang @CatherineYZ]

@duolingo Hey @duolingo I submit issues three times but nobody reply me. I went to China for travel and now all my friends related features gone. I cannot add friends, my friends cannot see me, and I cannot turn on friend quest. Can anyone help me with this issue?

[UX Links @uxlinks]

Learn how to create a Design System with this useful step by step guide 🧨🔖

Keep leaning and keep moving forward 🤜

Download the guide:
https://t.co/Gs14O2hAg7

by @kamrify

#ux #ui #uxdesign #uidesign #userexperience #research #userresearch #usertesting #uxresearch #usabilitytesting #productdesign #designprocess #design #designsystem #visualdesign #webdesign #appdesign #uxstrategy #uxprocess #prototyping #business #startup #entrepreneur

[Catherine Zhang @CatherineYZ]

@10xDayDreamer I would like to see some portfolio created by designers who have front end background and specialized in 2B product. Portfolios have more storytelling , solutions for complex problems and less visual design.