Not Bear Liu

After Cube type launched, I was busy with other things for a while and didn’t have time to pay attention to it. But yesterday, it suddenly got its biggest traffic spike since launch. I checked and found it had been featured on Hacker News—one or two thousand people showed up out of nowhere. The comments were mostly positive. I’m happy; it’s great to get such positive feedback :)

Day two of water rescue instructor training. I feel a little more confident today than I did yesterday, mainly because I wore a wetsuit and avoided shivering from the cold like I did yesterday.

Apple asks you the fewest questions; Google never stops asking. This isn’t about one being better than the other—it’s about two completely different product philosophies.

I’ve been wearing a Fitbit and using Google Health for the past three weeks, while comparing it with Apple Health. Here’s my breakdown of the two products.

Let’s start with Apple. Apple Health barely says a word. It quietly records your data. There’s no AI coach commenting on how you slept last night, and it doesn’t follow up to ask whether you took its advice. It’s just there: you check in, and it shows you what you need.

Google is the complete opposite. Open Google Health, and the AI coach might jump right in: “You slept a little less last night, and your heart rate has changed recently. How are you feeling?” I asked what my resting heart rate meant. It pulled up 13 days of data, made a chart, laid out some ranges, and even cited external health sources. Then it asked me, “Does this overview give you the reassurance you were looking for?” Later, I mentioned that I was taking heart medication, and it brought that information into our next conversation. When I switched to Chinese, it switched too.

My first instinct was to score the two products and decide which one was better designed. But that instinct was wrong.

Apple’s health strategy is built on the iPhone and Apple Watch ecosystem. It needs you to trust that years of sensitive data can stay safe in its hands. That trust is what gives you a reason to keep buying its next device. So it doesn’t need to say much.

Google doesn’t carry that same baggage, nor does it have that same moat. Conversation, initiative, and back-and-forth—that’s how it builds relationships.

These two strategies reflect two companies in completely different positions, with completely different ambitions. Measuring them by the same yardstick was the mistake all along.

I unpacked this whole process in this episode of my podcast. Give it a listen if you’re interested: https://redcircle.com/shows/936a3212-92e5-41c7-b6f9-39e3596adeb8/episodes/7953b746-2644-4f9c-b03e-b781323a542c

New video: The feature you charge for—is it a strength or a weakness? Let’s take Raycast apart and find out.

I open Raycast thousands of times a week, yet I’ve only ever used its most basic features. So I took it apart, with one question in mind: Is the feature you charge for a strength or a weakness?

https://youtu.be/T1EmD3Sp7Fs

The product window in the AI era is incredibly short. Google just introduced the concept of a “smart pointer.” A few months ago, I came up with exactly the same interaction and even sketched it out, but I didn’t pursue it further at the time—and now a major tech company has already built it.

This period of transition is a bit like the stage decades ago when the command line had matured but the GUI hadn’t yet appeared: there are so many possibilities, and time is tight. I’m sure anyone actually building products can feel it.

In this episode, we also talk about how the habit of recording voice notes gets easier the more you use it; Paulo Coelho’s pilgrimage and the creative path behind *The Alchemist*; why earning 33% less in 2025 has actually made me happier; and the use of Progressive Disclosure in the game *Control*.

https://bearliu.substack.com/p/79

Starting in 2026, making a serious effort to be an active, engaging person online may be one of the highest-ROI things you can do.

Dr. Andrew Huberman says that after you consume caffeine, whatever you’re doing feels more positive psychologically. In plain English, after drinking coffee, you’re more motivated to do things.

Experiments with bees found that when given a certain dose of caffeine, they worked harder and produced more honey.

I’ve always had coffee before work. I was wrong!

I should have it before doing something fun.

After working on a project for nine months, the contract listed four metrics: launch completion rate, conversion rate, daily and monthly active users, and 7-day and 30-day retention. But in the end, we couldn’t get a single number, because the product was never publicly launched.

Have you ever run into this situation?

A lot of teams assume, “No launch = no users = no data,” and then design decisions turn into: “I think this color looks nice” or “This flow feels smoother.”

But actually, you can plan ahead and get real user feedback before your product officially launches.

In this video, I share three things I would do:

First, decide what you’re going to measure before you even know whether the product will launch. Metrics need to exist before the data does; otherwise, you’re just making excuses after the fact.

Second, don’t wait for the product to launch—find real users to test your prototype ahead of time. Even a clickable wireframe can give you genuine signals about the flow, logic, and priorities.

Third, think of “whether you can launch” and “whether you can learn” as two separate things. Launching is a milestone, but learning can start today.

If you’re currently working on a project with an “uncertain launch date,” this video might be just what you need.

https://www.youtube.com/watch?v=PqC0aw-WkWM

Clearing your calendar isn’t actually hard. The hard part is deciding what gets in when two things both want the same time slot.

I’ve been reflecting on how I use Apple Health lately, and noticed something interesting: I thought I was “managing my time,” but most of the time, I was really “managing who gets to show up in my time.”

Health data, work rhythms, focus blocks… These things aren’t competing over what matters most. They’re competing for ownership of that hour.

Who you let into the room determines the kind of person you become.

I’ve written about my thought process in the latest newsletter. Give it a read if you’re interested:
https://newsletter.bearliu.com/p/apple-health-review

If you create content on YouTube, remember to connect Claude or Codex to the YouTube API. For example, you can compile the thumbnails for your most popular videos, or research which thumbnails perform best on other channels.

Just tell your agent what you want, and it can put everything together. I’ve connected mine to Figma via MCP, so it can save everything directly to a Figma file. It’s really convenient.

Once you have enough data, you can probably spot some patterns. But with the modest number of videos I have, I haven’t really found any patterns yet, haha!

Recently, while taking a closer look at Linear, what made me pause wasn’t a particular feature—it was the first impression it gives you before you even use it.

Dark backgrounds, gray-white text, and a high information density, with almost no illustrations or decoration. It feels more like a code editor or terminal than the note-taking apps we’re used to.

There’s more going on here than meets the eye. People rarely read through a tool’s feature list before deciding whether it’s right for them. Many decisions happen in the first few seconds. The colors, density, and typography are already answering a question on the product’s behalf: Do the people here seem like people like me?

Linear’s answer is very specific. It serves engineers and product managers who are used to thinking like engineers. Technical teams at OpenAI, Cursor, Perplexity, and Scale AI all use it. The data tells the same story: adoption is higher among small, fast-moving, tech-driven teams, and lower at large enterprises—exactly what the interface promises.

When we build products, we often treat visual design as the packaging we add after the features are done. But users encounter things in the reverse order.

**Visual language is the first sentence of a product strategy.**

It’s worth asking yourself this about your own product: Are the people drawn in at first glance the same people who are ultimately willing to pay for it?

I’ve been using Plaud for almost three days now, and it’s been incredibly satisfying. The best part is that it fits perfectly into my existing workflow. I can transcribe my voice memos and recordings directly using a command I created with Claude Code. I use it to record and neatly organize conversations over coffee and in-person meetings. The other day at a coworking space, I ran into a Kiwi woman who quickly noticed she was using one too.

So we agreed to meet for coffee next time and exchange tips on using Plaud.

Turns out digital devices can help you score a coffee date with a beautiful woman.

Something is happening again: AI slop.

In 2023, once ChatGPT became more or less reliable to use, a flood of AI-generated text hit the internet. More than 90% of it was junk. Many people explored traditional ways to distribute and monetize it, such as Kindle and Amazon self-publishing, and there really was a boom during that period. But very few people actually sold books or made money.

Now, as new AI models’ video-generation capabilities rapidly expand, the same thing is happening again. A flood of AI videos is going online, and just like back then, 90% of them are junk.

People have a limited amount of time to consume content. Viewers can tell whether you made something with sincerity and genuine emotion you want to express, or whether you’re just trying to treat them as traffic and cash in.

Don’t underestimate your audience. If you’re making content, you may still need to work on the fundamentals.

It’s common for people in the entertainment industry to keep upside-down schedules. For one thing, they don’t really have anything they need to get up for in the morning. Another reason is that it lets them avoid the hours when the public is out and about.

A couple of years ago, I was back in China and went to a get-together at a teacher’s studio. He said that Old Cui—Cui Jian—might come by that night. A whole bunch of people excitedly said, “Let’s wait for him.” I’m the sort who can’t stay up late, so I left before eleven and missed my chance to meet Mr. Cui Jian.

For Cui Jian and the others, that was when the day was just getting started. I don’t know whether keeping such upside-down hours has any long-term effects on your health.

Let’s chat about two books that are a little woo-woo: *Why Buddhism Is True* and *The Artist’s Way*.

People say a book only counts if you finish it. But the two books I rate most highly are both ones I stopped halfway through the first time: one at 40%, the other at 15%.

The first was *Why Buddhism Is True*. I started listening to the audiobook in October 2019. I was preoccupied with work and couldn’t quiet my mind, and I was still pretty inexperienced at listening to audiobooks. By the time I’d reached 40%, I hadn’t gotten much out of it, so I decisively put it on pause. I even wrote myself a little justification in my notes: This is kind of in keeping with what Buddhism teaches about letting go.

In the spring of 2021, I started meditating again, sitting for thirty minutes each evening. Some nights it came easily; other nights, not so much. Two months later, I picked the book back up, and this time I couldn’t stop listening. The day I finished it, I wrote: There are no bad books, only bad timing.

What it left me with was very tangible. One day, I spilled coffee on my slippers and the floor. In the past, my first reaction would definitely have been to swear. This time, I first noticed the urge to swear, and then it passed on its own. At the time, I was meditating for only about ten minutes a day.

The second was *The Artist’s Way*, and I put it down twice. In January 2021, I got 15% of the way through and realized it was a 12-week program that involved writing three pages every morning. My focus was on work at the time, and I couldn’t spare the energy, so I stopped. In March 2022, I felt the timing was about right and started reading again, but after that, it never showed up in my journal.

I finally got into it at the end of 2024. Around then, I realized that putting my thoughts down on paper worked better than anything else, which reminded me of the book. On December 27, I wrote my first Morning Pages, and I finished the book the following April. It was the book that had the greatest impact on my behavior in 2025.

There was also *Principles*. In 2019, I spent 17 hours listening to the audiobook, and I’d guess at least 70% of it went over my head. When I read it for the third time in 2022, I finally felt I understood it.

Looking back at these experiences, I think there are at least three reasons you might not be able to keep reading. The book itself might be bad—in that case, put it down; there’s no need to feel guilty. Or you might not be ready for it yet. For example, my English listening skills weren’t up to the density of *Principles* back then. You can skim through it over a few hours and come back to it in a few years. The last possibility is that the timing isn’t right: you haven’t yet run into the problem the book is about.

I used to think the right time meant having free time and a quiet mind. Now I think it’s more like this: the book happens to answer a question you’re dealing with. I needed to quiet my mind before *Why Buddhism Is True* could speak to me, while *The Artist’s Way* clicked during the messiest period of my life.

The next time you can’t get through a book, you could ask yourself: Is it the book, my ability, or the timing? The first two are easy enough to deal with. For the third, just remember the book, and wait until you encounter the question it has an answer to.

I only open Taobao once in a blue moon, and when I did, it recommended this hat to me.

Looks like Alibaba’s recommendation algorithm is nothing to sneeze at.

Jing Yidan passed away last month, and two weeks ago, Liu Huan passed away.

There was a wave of lamenting online: they were gone too soon.

But if you think about it, at most we’re surprised for a few minutes, and then everything goes on as usual and we return to our own lives. At least that’s how it is for me. I suddenly think, “Wow, they were still so young,” and then I go right back to my life as usual.

And my surprise has more to do with a basic fact: at this age, life becomes more precious—or, put another way, it can end at any moment.

I’m thinking about myself, not Jing Yidan as a person. Our lives never crossed; she was just someone whose name I knew.

So, online and in real life, there are plenty of people we pay far too much attention to, and we only have so much energy. Turning our attention back to ourselves, exploring who we are, caring for our families and for the people who are truly worth connecting with—these are all good ways to make our own lives feel longer. If you don’t spend your time on yourself, or on the people who deserve your attention, you’re wasting your life.

At first, I thought writing was to help the world understand us.

Now, I think writing is to help us understand the world.