Slices of Our Brain

The most recent development in AI makes me feel like I’m living in a condensed timeline where progress flies at a 10x speed. It wasn’t that long ago that people were shocked by the image generation capabilities of DALL-E 2, a GPT-based multi-modal model. Suddenly, we have tools that are able to make functional games, websites, and demonstrate complex reasoning skills with a surprising range of flexibility over images, text, and even videos. Really, once you grasp the power of the building blocks of GPT models, there’s virtually nothing you can’t do by training and combining them in ever more creative ways. Here are some examples that blew my mind:

These use cases technically are achievable with purpose-built tools (e.g., “an app that gives you recipes based on pictures of foods.) The remarkable thing here is that GPT-4 models can do these highly complex tasks out of the box without any specialized training.

It’s as if…a human were behind it.

Well, maybe small, tiny parts of humans.

Read More

Time Capsules

The Bay Area has seen more rainy days than I can remember. On days like these, I usually spend time by making a cup of tea or coffee, putting on some music, and working away at some project at my morning table.

This other day, I stumbled upon a background music video that uses a few piano instrumentals from classics like 5 Centimeters per Second. One of the top comments read:


“(This made me think:) ’really, music is a medium for storing memories.'”

Music really is a medium for memories. Last time, I wrote about the physicality of music. But the temporal dimension of music is also equally important.

Read More

“The last best physical music format”

The weather was beautiful outside, but I was exhausted and in pain. Coughing, blowing my nose, I couldn’t do anything except press the TV remote while in bed. It may have seemed like a misfortune in the moment, but being quarantined in Singapore for a week had at least one silver lining. After watching hours of videos about obsolete media formats, I discovered one of the coolest things I’ve seen: MiniDiscs (MD’s).

Read More

Repair before replacing

The ice maker at our new townhome didn’t work, but it didn’t bother us as much as we rarely used ice. Earlier this month, our landlords kindly replaced the fridge. I suppose they replaced it because there’s no one offering repairs for a fridge as old as the house.

Still, it seemed like a big waste. Other than the ice maker, it was a perfectly functional fridge. Perhaps our landlords wanted a new fridge anyway, I guess.

The old Frigidaire unit was taken apart and hauled away like trash. It hurt a little, even though I didn’t pay anything for it. Like killing a life still going strong, it felt mottainai.

Read More

The morning table

When it comes to personal work, if I’m honest with myself, I’d say that I probably spend more time working away from my work desk than at it, no matter how decked out and ergonomic it is. So, I started wondering why.

In sociology, there’s this concept called “third place,” a place other than your home or work that you’d visit to socialize. I think the proper definition is more about a place to gather with the community, but I started thinking of it as also a place to focus, away from the inertia of familiar environments. And I started wondering if this place could exist at home as well.

For one, during most of the pandemic, my home is my world. Perhaps our regular desk has become too associated with work that I seek another spot to escape it. But even before the WFH era, I’ve always started my morning with hours of sedentary activities at such a place. So perhaps it’s the suburban environment, where things are far apart so I can’t exactly take a short and casual walk to a nearby coffee shop. Not knowing whether this qualifies as a proper “third place,” I started calling it “the morning table,” since morning is when I usually do my personal work.

I found that, for myself at least, my morning table wherever I lived has always had the following 4 traits:

  1. By or facing the window, with abundant natural light
  2. Not my “serious work” desk
  3. Somewhere I can make myself a nice drink
  4. Quiet and not distracting

Do you have a spot like this at home? Where do you do most of your nonwork work?

A Love Letter to Pho New Saigon

Unassuming and nestled between a boring UPS store and a Roundtable Pizza, you have a special place in my heart.

From the day I discovered you and your #56, to the 56th time I’ve devoured this one-plated feast with an extra egg cake entirely by myself, you’ve continued to deliver the same satisfying taste and generous portions through the last three years. Pandemic or not, dine-in or take-out, you were quick to bring out the food and accommodating to my indulgent add-on’s.

Pho New Saigon is the kind of mom ‘n’ pop places that feeds the neighborhood. On any given day, I see families, construction workers, policeman and women, tourists, office workers, all enjoying their conversations, solitude, or meetings over a meal. It’s a hub, In an otherwise isolating suburban city forced apart by concrete and asphalt, people from all walks of life are united by beef broth and fish sauce. It’s a comfort zone of serendipity.

As a fan of Vietnamese com dia or rice platters, I love it above even the more popular pho. The variety of juicy, savory, salty, fatty, and sweet meats overwhelm my mouth in a nuclear explosion of umami. With a simple spoon, I can combine any given number of these ingredients and create variations from bite to bite. It’s a buffet of flavors on a plate.

But com dia is, for some reason, hard to find. Perhaps it’s just underappreciated, or perhaps it hasn’t been “discovered” by “the mainstream” yet. I hunt for it and even then, many of them are unreasonably expensive or disappointing in quality.

But Pho New Saigon, oh you and you. Your #56 puts everybody else to shame. Have you not realized how large and juicy your thicc pork chop is? Did nobody tell you that no one else serves fried shrimp cake in 4 plump chunks of tenderness? Oh, don’t even get started with the generous portions of fried pork skins. The fattiness, beautifully infused with a thick layer of rice powder coating, is a whole experience of naughty sensation as it wraps around each bite of rice. And the Chinese sausage is just a pure indulgence of delicious salty cured pork, pushing you to a dangerous new height that no pho-slurping neighbors in the restaurant understands.

And I always, always want your egg cake added on. Always warm, and always big. The chonky slice is a soft landing after every fish-sauced filled climax. It’s light, fluffy, and full of crab and egg flavors to soothe your assaulted taste buds and ready them for the next attack.

Your restaurant is roomy, clean, and unassuming. Not intimidated by pretentious high-end interpretations that often make me feel like an outsider in my own culture. Not tarnished by the unsavory attitudes and interiors like some neglected establishments. It feels like home, even 30min away from where I live now.

All the utensils, napkins are self-serve, allowing myself to really make myself comfortable. And for me, that means a fork and knife to cut the pork chop, and a spoon and a pair of chopsticks for plate-to-mouth delivery. Two of them to the left of the plate, on a clean sheet of napkin. And two of them to the right of the plate, also on a clean sheet of napkin.

It’s a whole thing. A ritual. An experience that comes with a process, a process to my liking, without having to badger the staff repeatedly for extra utensils.

None of these things, ingredients or service alone, are probably “the absolute best” if you scrutinize hard enough, but nobody else serves them in this perfect balance of portion, taste, and price. As we always say in the user experience industry, it’s not about the individual features, it’s about the holistic experience.

Please realize how beautiful you are. And yes, a #56 with an added egg cake please.

What is an artist in the DALL-E 2 Era?

Recently, DALL-E 2 made headlines everywhere: an AI illustration program that turns descriptions of images into actual ones with believably good results.

For example, if you type “an astronaut riding a horse,” this is one of the options DALL-E generates for you:

With DALL-E, it is suddenly possible to ”express one’s ideas” without mastery of any artistic technique. So, that got me thinking: in a world where hyperrealistic and artistic content can be “commissioned” simply by uttering words of imagination, what will happen to artists in the future?

Or more precisely, what will be an artist?

I consider an artist to be someone who combines creativity and technique to create art, which is something that evokes a visceral response and consequently impart a message. At least this is how I interpret these two words.

One could argue that DALL-E is not unlike photography. Photography first appeared to have replaced painting, but it soon became clear that photography, by offloading the job of documenting people and things from painting, actually elevated its essence: artistic expression.

It feels to me that tools like DALL-E 2 can evolve in two different ways, which ultimately is a matter of perspective.

Scenario 1: artists will continue to evolve their skills as tools become more powerful, they will master new techniques that allow them to precisely control their artistic intent. In this scenario, “artists” are still an influential, “elite” group of people who have skills that most don’t. This group might even shrink due to its higher skill barrier.

For example, with DALL-E 2, an artist equipped with ML understanding can leverage it to generate and test out ideas, composition, palette, etc. quickly. They can then fine-tune its input parameters to create exactly what they envisioned before hand-painting over the nonsensical details that ML doesn’t understand. AI does 80% of the grunt work, and the artist elevates the piece to their unique vision with the last 20% of manual work. These future artists can output a larger quantity of content, supply to a larger base of audience, resulting in an even smaller demand for artists per unit of population, adding to the phenomenon of pop culture oligopoly.

Scenario 2: more and more people can express their imaginations with access to more powerful and more precise tools. “Thought → Result” becomes a reality, being an artist is simply an “activity” that anyone can partake with minimal experience needed. This is analogous to blogging: it used to take a programmer to publish a website, now any budding writer can.

In both scenarios, there will be an explosion of content. The net might be: both scenario 1 or 2 could be true, it’s just a matter of how you define what an artist is vs. what a remixer is.

You could argue that great meme gif makers are in an artist category of their own, even though their raw materials are creations from others. But then again, is any art truly independent of other people’s works? The best artists and musicians reference other sources for inspiration. Is DALL-E 2 simply a more efficient reference and brainstorming tool? Will DALL-E remove the burden of documenting ideas and let artists focus on expressions, like photography did for painting?

Bruce Barnbaum advocated the idea that what makes one an artist is their way of seeing, and an artist’s unique way of seeing is consistent throughout their entire body of work. Even if DALL-E takes away the laborious work of painting stroke by stroke, it would still take an artist’s eye to select the exact composition that matches the artist’s vision.

If we consider an artist to be not defined by their kind of technique or level of technique, then an artist is really about what happens in their head. Do they connect the dots? Do they know how to translate ideas into an artistic medium? If this is the definition, then the explosion of low-barrier tools paints an optimistic color for the future of art, where technology indeed unlocks the full potential of human creativity.

Incomplete Install

Have you used an older Windows PC running, say, Windows XP or 2000 before? You might remember that you had to use an “installer” .exe program to install a piece of software. Larger programs sometimes took hours to install. You absolutely must not disturb the process. One time I did, and the program left zombie footprints all over my disk drive that was almost impossible to clean out. Files would be littered everywhere but the program would only throw errors if you tried to start it. Of course, because the installation wasn’t done. There was no way to “uninstall” it either because, well, the “uninstaller” wasn’t compiled yet.

I have two incomplete installations in my brain. The Chinese language. And the English language.

It first occurred to me that my confidence in my mother tongue was no longer justified when I visited China after spending my first four years in California. There were expressions I knew perfectly how to articulate, and as I fluently layered one sentence upon another, I’d suddenly hit a nasty speed bump, a hit of silence that gives me a small identity crisis every time. The word lacked a Chinese equivalent. Not because it was untranslatable, but I just never learned that concept before I moved to the states.

The fourteenth year is a formative one, and so are the years that come after. My Chinese language installation was abruptly stopped. And in its place, a completely different language resumed its role, but did not supplant it completely.

People looked at me aghast when I didn’t know something that a six-year-old American kid does. “What kind of childhood did you have?” Their eyes were more honest. “One where I can’t even write with Chinese characters, because some of those names only existed in a dwindling dialect.”

It did not get better either. Year after year, I accumulated more knowledge and expressions in the English language, without ever knowing the equivalents in Chinese. To date, I still can’t pair up the names of fish I eat even though I’m familiar with all of the individual words. Is “catfish” 鳕鱼 or 鲈鱼? I don’t know, but I’ve eaten all of them and am confident that it’s not “猫鱼”. Last week, I sent out calendar invitations as a 28-year-old adult titled “greet and meet.” If Kelsey didn’t point it out, I’d probably gone for another ten years using it while people politely ignore my odd expressions.

I started learning English with English a few years after settling down here. It was a big linguistic milestone for me. Instead of relying on translations, it freed me from Chinese constructs and lit up a new circuit of thinking. But as the two languages grew apart, I got farther and farther away from the culture that I used to breathe like air. If I were a piece of software, I’d no longer be backward compatible.

How do I say “density” or “gender-neutral bathroom” or “cross-functional collaboration”? I don’t know. Is “坦率” the same thing as “forthcoming”? Sounds like it. While I know that “底片” is “film negative,” I’m not sure how to say “film advance lever.”

I continue to struggle to explain my life in front of my relatives. But hopefully, word by word, I’ll painstakingly stitch the two worlds back together one day.