Taiwan's Software's Terrible?

So I happened to catch an augmented Youtube video about a Stanford interview with Google's ex-CEO Eric Schmidt. As I am from Taiwan, naturally whenever the keyword "Taiwan" is uttered I'd pay attention.

The remarks in the video are along the lines like

• TSMC is impressive; they have rules that let starting Ph.D physicists work in the factory in the basement floor,
• "Taiwan is a fantastic country, whose software is terrible," followed by "that's not going to work," and then reiterated "amazing hardware."

The context for the first point is about competing in the heated AI space, where people in startups, in the US or from other countries, work like hell to stay ahead or to catch up, and prioritizing on work-life balance may be a mistake. Hence TSMC is brought in for comparison, and the importance of "time management" in the AI space that follows. Of course I don't know the main rationale behind TSMC's practice, the nature of the field work in the basement, or the qualification of good physicists compared to those of America. I do imagine engineers, physicists, etc. work like hell to help TSMC stay ahead.

The second point is related to US national security, and the potential allies that can help. Schmidt mentioned about India's importance and attracting most, if not all top AI people to work in/for the US. In this context, Japan and Korea, but not Taiwan, can help in terms of software development. As to hardware, I guess Taiwan is very welcome.

You'll have to look up the video or transcript yourself, as any out-of-context citations will do more harm than help. As usual, I am just murmuring without in-depth research to back myself up. You are encouraged to have your own opinions.

Are both Schmidt's points regarding Taiwan mind triggering? Although it's not exactly like "what's new about it?" to me, it is also close enough. Actually my first reaction was: Schmidt's remarks are just straight from his heart, what he regards as facts or common sense, due to the atmosphere of the talk (he believed it was a private session) as well as all the questionable wording he gave when discussing other topics. It seems people in the US, or in the AI field in particular have opinions about Schmidt himself, but I think at least he has broader view than a lot of people. So is his perception representative – that people in general don't associate software with Taiwan, that Taiwan's software is indeed terrible?

I purposely leave the word software ambiguous. Interpret it in any way you want, be it software being built or delivered, competence of the developers behind it, capability of the organization that produces it, and so on.

Initial Thoughts

Schmidt has his own opinions, so do others, whether they have equal status or not, and me. I haven't made up my mind to organize and theorize. I normally won't dig into the web to figure out what others have to say about trendy things, but since I'm writing this post, let me just do it. But before searching, I'd like to lay down my basis up front, just to refrain myself from echoing what other people are saying.

• First, it is not just software, but also genuine fundamental scientific research, and other professions of creativity. The reasons are more fundamental, rooted deeply inside people.
• It is not just the government, but also the people. We are all accomplices driving ourselves to whatever it is today, and this has nothing to do with intricacies from the other side of the straight or overseas, since these kinds of influences will always be there.

Again I only state the abstract without further elaborations. Therefore it can not be proven right or wrong, just some opinions.

What Reflections out There

Searching the web for articles from news agencies, blog posts, social media, and so on, and I couldn't find much that I don't anticipate. Social media and discussion boards may have more "spontaneous" feedback compared to more serious articles, as usual. I was hoping to find threads from internationally discussing Taiwan's software, but I didn't go the extra to find them since understandably it's not something interesting. I was a bit shocked there are quite some articles sitting behind pay walls or requiring registration, as if their opinions are so valuable. You have to look up on your own, and come up with your own conclusions if you care.

Anyways, I am just going to write down some notes. Remember these are all opinions from Taiwanese people (with a bit rephrasing from me) and remain subjective at best.

• Few people object the idea. Many kind of accept the fact, however reluctantly.
• Some distinguish software from software engineers, often implying that the former is genuinely not good, but the latter is not bad at all.
• Taiwan's strength is hardware, which also attract much more resources and talent, leaving little for software.
• Taiwan's software is very good, and doesn't have to be as great as hardware, which is top notch already and hard to surpass. We need to find some niches to our own advantage.
• We will create AI science parks or provide big funding to help improve it.
• What we need is some open and free ecosystem? And know how to work collaboratively among people with different specialities?
• Non-functioning government policies regarding science and technology.
• Software demands very large market. Taiwan's market is too small. People keep competing each other harshly, being shorted-sighted and not daring to take risk to concur the global market out there.
• People blind themselves like ostriches, or frogs, ignoring the fact that people in our neighborhood have move forward long time ago

Again I purposely leave the word hardware ambiguous. Interpret it in any way you like, be it semiconductor technology, engineers, IPs, innovations, process improvements, etc.

To Ponder

So far I think people mostly touch on the surface only, i.e. the situation we have and the background and some reasons leading to it. I sure can find more in-depth documentation or discussion boards that touch on the deeper "whys." Maybe if you ask some foreign people who know a bit about Taiwan, they will also be puzzled. Indeed, with so many people with high education and skills, so many high-tech (hardware) companies, and supposedly alerting environment we are in, we should establish and flourish our intellectual capacities in important, if not all, front tiers, and even try to export them to and influence the world, instead of mostly the other way around.

It will be an endless debate what's going on, or why I am wrong and you are right. So it is kind of meaningless to go on here. I'll just give some imaginary scenario to ponder on:

Suppose we can still maintain hardware leadership (in some aspects) and make related industries and people happy for a while, in a relatively peaceful spacetime, and suppose we have some generous leeway to grow for ourselves in other fields (to be determined), hire outside experts to train use whenever necessary, but we can only plan and execute the steps towards the goal by ourselves. Can we do it? What will be the likely outcome(s)?

Well call it rhetoric, a straw man, or whatever. This is the best I can come up with, so just think of it as a brain exercise. I have bets on good outcomes and the bad ones with different probabilities, simply because there are still more unknown factors than what the over-simplified condition given above can control.

Schmidt's Another Observations about Taiwan

As I tried to search for more meaningful articles, I happened to run into this "trip" report, also from Schmidt – Hardware Island: Taiwan's Indispensable Role in the Global Supply Chain.

Here are the usuals, about TSMC and Taiwan:

Taiwan has mastered a level of complexity in manufacturing processes that is beyond human comprehension. This level of excellence will serve it well in emerging technologies.

Taiwan’s hardware-centric technology ecosystem has some drawbacks however. Amid the global tech talent shortage, the gravitational pull of Taiwan’s successful hardware companies is resulting in a lagging domestic software ecosystem. As a result of this, there are only a few Taiwanese firms working on large language models (LLMs), and they are small and not well-resourced.

Here is the reason I cite the short report:

Taiwan’s lackluster software industry may also be a byproduct of the East Asian educational systems that incentivize conformity and structure over a looser and more creative approach. This can be addressed, but will take some time.

I can only add that, as hinted previously, it is not just because of the "department of education", but also the teachers, parents, adults, and the society as a whole (but let's leave youngsters out of this), who cherish less on creativity, curiosity, intellect, and other virtues than success, wealth, and fame. My hope will be people can get richer doing their own diligence, and then pay more for quality goods or service(and reject otherwise) to encourage more people to perform even better in different fields. But it takes not only time, but also strong, collective determination.

Other Important Topics

There are topics related to computer programming and education, thus less relevant for this post but important for this blog. Therefore I keep it here for future exploration.

• The text to action model will be very powerful. Instead of text to text generation, people can feed a large amount of data, give simple commands in plain English, and ask AI go create software in seconds.
• CS undergraduates, when learning programming, all have a programmer buddy to help, and professors/lecturers can just talk about concepts.

To me the first point is obvious (in the sense that everybody has thought about it), though we don't know the time line. Hence in the second point, people and students are genuinely concerned about whether it is worthwhile to learn programming if AI's so powerful in near future. I don't know what the "buddy" would be thinking when mentoring students, based on their "capabilities," but it is clear that learning advanced and hard CS topics is still needed, not the simple things.

To be continued...