Greetings from a world where…
for the rest of the college football season, this status update will be devoted to tracking the Iowa Hawkeye offense’s march to mediocrity
…As always, the searchable archive of all past issues is here. Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).
Feature Translation: Tiancheng Lou — A Profile
Context: Thanks for engaging with the previous Around the Horn issue. The winning pick was this longform profile of Tiancheng Lou, known as China’s No. 1 Programmer. After gaining fame by winning basically every global programming competition, in 2016 Lou founded Pony.ai, an autonomous vehicle technology company co-located in Silicon Valley, Beijing, and Guangzhou. This Leiphone article (link to original Chinese), by Caixian Chen, profiles Lou’s journey.
Key Passages:
On Pony’s early beginnings and the connections to global VC firms. Here’s the scene in October 2016 in Hangzhou:
“The Shangri-La hotel was full of people drinking and gambling, full of uproarious hubbub. Tiancheng Lou sat opposite Nanpeng Shen (Neil Shen), then the top leader of Sequoia China. The atmosphere must have been delicate and the conversation tense…At that time, autonomous driving was booming in China, and there were many people but little money. There were rumors in the industry that Lou and Peng Jun intended to leave Baidu to start a business. IDG Capital and Sequoia were the first to find them and expressed their investment intentions…The dinner was short. They didn't chat in depth that day, and Tiancheng Lou went back slightly confused. But after Pony.ai’s first round of financing, IDG and Sequoia invested a total of US$15 million in the same round, and Sequoia proactively added a little more.
On Lou’s technical skills:
After working at Google X (the moonshot lab), Lou joined Baidu in the spring of 2016: “Within half a year he unified the communication interfaces of the chaotic modules of Baidu's ADU (autonomous driving unit) at the time. This was equivalent to reconstructing each module while ensuring the performance of the original system, which was very laborious, but Lou completed it in a short time. When he was at Baidu, Li Hengyu also collaborated with Lou on a project, which was to reconstruct the bottom layer of the perception system and turn the serial (processing) system into a parallel (processing) system. This was the first attempt in the field of autonomous driving. Li Hengyu served as the project leader, and Tiancheng Lou, then head of the technical committee of Baidu’s autonomous driving division, was responsible for checking the technical plan and code review. The project was very successful. It is said that this project later became part of Baidu Apollo.
What makes Pony unique?
The concentration of talent is pretty impressive. Many of Pony’s people secured “guaranteed admission” [保送] to top Chinese universities for winning certain competitions, meaning they didn’t have to go through the gaokao route. Some core staff graduated from Tsinghua’s “Yao Class,” run by former Turing Award winner Andrew Yao. These Yao class graduates also include academic researchers at Stanford, Princeton, etc., as well as all three co-founders of Megvii, one of China’s leading facial recognition companies.
Mo Luyi, Pony’s VP and head of its Robotaxi business, is the only female champion of the ACM International Collegiate Programming Contest (ICPC) in the past 30 years.
From the article: “Pony.ai’s founding team and core members are very stable. Li Hengyu revealed that about 90% of the first group of founding members who joined Pony.ai in 2016 and 2017 have still stayed with Pony. This level of stability even exceeds that of some mature companies.”
Some cool notes on trans-pacific differences in autonomous driving:
Lou’s reflections on why he left Google: “When Google was researching autonomous driving, he discovered that there was a large gap in Google’s cultural understanding of China’s transportation, and it did not understand Chinese driving habits at all. Americans and Chinese have different ideas and ways of driving. Americans’ minds are full of rules, but the most important thing when driving in China is communication. (Cars) need to understand people’s intentions, analyze and communicate, and form compromises and guidelines when conflicts arise.”
Pony’s first road test in Beijing, in 2017, was actually a “black run” because they were violating government policy. Recently, Pony secured approval from the Beijing government to run its unmanned Robotaxi program in Yizhuang, a district in the southeast suburbs of Beijing, where many high-tech research centers are located.
FULL TRANSLATION: Tiancheng Lou, ideals will never die
ChinAI Links (Four to Forward)
Must-read: The New AI Panic
Karen Hao, in her first piece for The Atlantic, analyzes a possible new front of economic warfare between the U.S. and China in AI:
The battle lines may soon extend beyond chips. Commerce is considering a new blockade on a broad category of general-purpose AI programs, not just physical parts, according to people familiar with the matter. (I am granting them anonymity because they are not authorized to speak to the press.) Although much remains to be seen about how the controls would roll out—and, indeed, whether they will ultimately roll out at all—experts described alarming stakes. If enacted, the limits could generate more friction with China while weakening the foundations of AI innovation in the U.S.
Should-read: The Authoritarian Data Problem
Eddie Yang and Margaret Roberts in Journal of Democracy:
As the race to develop artificial intelligence (AI) accelerates, access to more and higher quality data is becoming increasingly crucial for AI systems. Yet the search for more data for AI facilitates information flow between authoritarian and democratic states in a way that has important implications for the behavior and output of AI. In particular, the homogenization of data, through institutions such as censorship and propaganda in authoritarian regimes can influence the output of AI developed in democracies. On the other hand, data from democracies provide valuable information for AI that is used for repressive purposes in authoritarian regimes. The authors call for greater scholarly and policy attention on the dual effect of the two-way AI-mediated data flow between democratic and authoritarian states and lay out a research agenda that would enable us to better understand the political influences on AI.
Should-read: A short drive across the Pacific
I may have stacked the deck in favor of that Lou profile, in part because I’ve been following Pony for a while now. Way back in the day — December 2018 — I wrote my first and (probably only) news article for MIT Technology Review. My reporting covered autonomous vehicle companies with home bases in both China and Silicon Valley, such as Pony.ai, Roadstar.ai, and WeRide.
Should-read: China proposes blacklist of training data for generative AI models
For Reuters, Eduardo Baptista reports on draft security requirements for public-facing generative AI models, which includes “a blacklist of sources that cannot be used to train AI models.”
Thank you for reading and engaging.
These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is an Assistant Professor of Political Science at George Washington University.
Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).
Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99
Tiancheng Lou sounds like the Chinese George Hotz
Your best issue yet.