Jing Conan Wang is the co-founder and CEO of Hachi AI. At Hachi AI, we believe that AI can do much more than serve as assistants; they can be our friends. Our goal is to build Large Language Models (LLMs) that interact with you just like real friends, providing joy and companionship.
Jing is a domain expert and pioneer in the LLM space.
From early 2022 to 2024, Jing co-founded and served as the founding CTO of Storytell.ai, a startup focused on using LLMs to turn unstructured data into actionable insights. He founded Storytell.ai based on his unique insight that LLMs have the innate ability to help humans distill signals from noise through conversations, a concept later popularized by ChatGPT.
From 2019 to 2022, he was a founding member and Head of Science & ML at Intellimize, a B2B SaaS startup that dynamically personalized websites for each unique visitor in real time (acquired by Webflow). At Intellimize, he also led efforts to use generative AI to create marketing campaigns.
From early 2018 to 2019, he was a Senior Research Engineer at Google Brain, where he applied reinforcement learning-based approaches to align humans and LLMs, as well as other deep learning models. His inventions have been integrated into various Google products, significantly improving core systems like YouTube and Google Ads.
Prior to Google Brain, he worked at Google Ads between 2014 and 2017, where he drove 20+ modeling launches that bring Google more than $200M in revenue annually.
He obtained a Ph.D. degree from Boston University (advisor: Yannis Paschalidis) and B.E. degree from Huazhong Univ. of Sci. and Tech.
He has good knowledge of mathematical modeling. He received a prize of Honorable Mention in Mathematical Contest in Modeling (MCM) 2009, and the second prize in the Chinese Undergraduate Mathematical Contest in Modeling (CUMCM) 2009. He has extensive experience in research and open-source development. He was a participant of Google Summer of Code 2012 and a mentor for Google Summer of Code 2013, both for the Honeynet Project. He was a finalist for Microsoft Young Fellow in 2009.
I am thrilled that my project of “network malware simulation” has been accepted by Google Summer of Code 2012! I will work on this project from May to August, and a new interesting open source software will be born in this summer 🙂
In the 19th century, engineering was almost a synonym of mechanical engineering. A typical way to solve problem at that time was analyzing the mechanical structure, building the machine with gears and wrench and letting steam engine to drive the machine.
Steam engine is a revolutionary innovation as it introduced a “paradigm” for problem solving. If you have a mechanic description of the solution for your problem, steam engine can take care of the rest. There were tons of derivative innovation following this “paradigm”, such as plane, steamship, submarine, automobile and so on. (some of them used internal combustion engine instead, an improvement of steam engine). Theoretically steam engine and its derivatives can solve any mechanical problem, what people need to do is to describe the problem in mechanical language.
Nowadays people have been accustomed to resort to computers when they have problems. Computer is another revolutionary innovation because it provides a similar paradigm. The paradigm for computer is to formulate a arithmetic model, propose an algorithm correspondingly and run the algorithm with computers. A computer is basically a calculator that can do addition very fast. Since all arithmetic operations can reduce to to additions, theoretically computer provides us the ability to solve any arithmetic problems, what people need to is just to describe the problem in arithmetic language.
Despite their different appearance, steam engine and computer share some common points. Both of them provide a good solution for a fundamental problem, steam engine deals with the problem of “generate rotary movement with strong force” and computer deals with the problem of “do addition with fast speed”.
When a problem can be divided into a large sequence of fundamental problems, we can get the solution of this problem by solving those fundamental problems. With this in mind we can easily conclude there are a lot of other paradigms except for those described above. The figure shows the paradigm for steam engine and computer.
Why paradigm?
The reason why we need paradigm is that paradigm can save us time. The real world is too complex and we cannot do everything well. A reasonable way is to solve a small set of problems perfectly and transform the problem we want to solve into these problems.
Limitations of Paradigm
Every paradigm has its own limitation, which is determined by the two factors. The first one is the fundamental problem itself, you cannot use steam engine to read this blog since the transmission of bits cannot be modeled as a sequence of physical movement.
The second factor is our ability to solve fundamental problems. Doubtlessly computer is the most awesome invention we human beings have ever had. However, there are still many things that computers cannot do in spite of the efforts in the past 50 years. They cannot understand imprecise information, they have very limited communication skills, and so on. All these problems belong to the field of artificial intelligence. The reason is that although theoretically there exists an arithmetic model for all these problems, the number of arithmetic operations is beyond the ability of any existing computer even for a small problem.
In the dawn era of computers, scientists estimated that it should take only 10-20 years for machines to catch up with humans in terms of intelligence. However, the expected date was missed again and again. There was no big improvement in the recent 30 years and it seems that all researchers are just waiting computers to be stronger and stronger under moore’s law.
However, I strongly believe most of the scientists are in the wrong track. The paradigm of computers cannot provide us a perfect solution for artificial intelligence. Human intelligence cannot be described in arithmetic model with appropriate complexity. We need other types of paradigm to get perfect solution of artificial intelligence.
Steam engine makes it possible for machines to move by themselve. Later the internal-combustion engine took the place of steam engine, but steamer was the beginning of this revolution.
be attention that, if you want to have the right resolution, download the driver from the nvidia offical site and don’t use the nvidia-current in apt-get, it will crash the GUI.
if you have installed nvidia-current and cannnot enter the GUI, just delete /etc/X11/xorg.conf and then type startx in the command window. then you will enter a GUI with low resolution interface. Then download the nvidia driver from official site and reinstall again.
I have no class this term so have more time for my hobby– entrepreneurship and music. I attended a technology lecture in last Wednesday and a one-day technology workshop in last Saturday.
The lecture was given by Brendon Howe, a BU alumni and the vice president of NetApp. Inc. a NASDAQ-listed storage company. To be honest, it was a little bit boring.
Mr Howe shared us his experience in both BU and NetApp. Since Howe grew up in the eastern coast, while NetApp Inc is a typical silicon valley company, I asked him about his opinion about the difference of east coast and west coast. I also ask whether he felt there is disadvantage for east coast graduates to find a job in high tech industry compared with west coast graduates as most of tech companies are silicon valley based.
His answer was that there are some difference, but the difference is not as large as people imagine. A east coast graduate student can easily adapt to the life in silicon valley. For the second question, he admitted that many companies are more willing to recruit local students, however, since globalization is the future trend, more and more companies set up their offices in the east coast, so this gap should be more or less bridged.
In terms of Saturday technology workshop, it was pretty much a MBA lecture. The workshop was organized by ITEC, acronym of Institute of Technology Entrepreneurship and Commercialization which is a very active technology organization under the school of management . About 30 students, mostly engineering and management students, joined the workshop. The students were divided into 5 groups to brainstorm biz ideas. Eventually each group will focus on one biz concept and try to draft a biz plan and give eight elevator pitch or biz presentation.
Our group focused on a gaming device in which you can control the avatars using your brain directly and watch game scenes through displaying glasses. It is actually a very interesting idea which can revolutionize the way people interact with gaming device. I eventually gave a elevator pitch. It was actually challenging for me since elevator pitch request very fast response which is harder for even native speakers. I tried to speak more passionate to overcomes the shortcoming of language and it seemed to work well.
For the music, I watched a very funny opera named secret marriage in Friday night. It was performed by BU students and free to all BU students, actually there is another good Boston Symphony Orchestra at the same time so I can only choose one. Choosing the opera turned to be not a bad idea. It says a young couple hide their marriage to the father of the wife. However, the betrothed husband of the wife’s sister falls into love with the wife, which make things complicated. The couple conspired to escape but was found. Eventually the young couple got permission of the wife’s father and show ends happily.
The only unhappy thing in the last week was research. The simulation results turned out to be crappy, so I switched to a new method, unfortunately the new method didn’t work, too. I stayed up last night but still didn’t get the result for meeting with advisor today, so the meeting is a little bit embarrassing as I made no progress in the past week. Anyway, research is only a small part of life.
Last Saturday, I attended the Sloan Asia Business Conference. Here is a brief summary of the conference.
The opening remark was given by Yasheng Huang, who is a faculty at MIT Sloan school of management. He just had a controversial talk about the development of India and China recently in TED2011 held in Edinburgh. In the talk, he attributed the success of China compared with India to the Human Capital—Better Literacy Ratio, Higher Life Expectation and less systematic discrimination of women. He also points that it is not fair to compare with India with China, which is a superstar in terms of economic development. Actually the performance of India actually is quite impressive among the emerging countries, while the slow movement of political reform in China may potentially hinder the economic growth in the future. The link of the talk is: http://www.ted.com/talks/lang/en/yasheng_huang.html
Lenovo Chief Marketing Officer. David Roman
As the chief marketing officer of Lenovo Corporation, David Roman introduced the global campaign named “For Those Who Do” that aims to promote the visibility of of the “Lenovo” brand outside China.
Despite the sluggish PC-market worldwide, Lenovo maintains 2-digits growth in the past several quarters and jumped to 2nd largest PC Maker, passing Acer and Dellconsecutively. But an embarrassing fact is the although the compony is a household brand in China, few people in western world know it in spite of its ThinkPad Series.
It is actually a common problem for many Chinese companies–so called the invisible Giants. Few Westerners know Baidu, SinoPec, ICPC, yet they are actually real mammoths even in western standard.
Now Lenovo’s focus is its ThinkPad Series, which formerly belonged the IBM. The deal was unbelievable then, since IBM PC department was triple the size of Lenovo at that time in terms of revenue.
It took Lenovo several years to learn how to be an international company. Now it seems that it has learnt the trick. In terms of my personal experience in United States, Lenovo is the most well-knowned Chinese Brand in United States. There is still a long way for other Chinese brands to go in United States.
SVP Finance & CFO of Global e-Commerce and Global Sourcing. Andreas Schulmeyer
This is the keynote I like most. The reason I favor this keynotes towards others is that it contains a lot of data rather than objective opinions. One figure is the comparison of the total population and GDP of China and United States. The left figure is for population and the right figure is for GDP. I guess the y axis is the GDP per capica. An interesting phenomenon is that the US’s value is also low in 1960s. The data actually needs to be validated.
Customer relationship Online and offline. There are five most components in culstomer Relatinship..
1. Social Network 2. Content 3. User Experience of Online Shopping 4. Brand Name 5 Search Engine.
Another interesting figure is the growth of US and China’s retail section.
Yuan Wang
Yuan wang, a high-level official in Chinese Development Bank and a consutant of world bank, gave a keynote about Chinese Economic situation. Her talk was a little bit boring. As a representation of new-generation Chinese policy maker, Wang reflected the impression of Chinese officials on westerners–well-educated, with international view yet still have bureaucratic ways of doing things. Most of the new Chinese policy makers,especially in economic field, receive well education. Some of them have western background and can speak pretty good English. However, different with their counterparts in Western Countries, Chinese policy makers always give people a feeling of commanding. Maybe it is caused by the booming Chinese Economy. All the investors worldwide want to go to China for new opportunities, and as the door-keeper of this huge market, the Chinese economic policy makers take it granted for the decisive power they have to let foreigner go or not.
Allan Kwan
Allan Kwan is a partner of Oak Investment. Now he focus the Investment on Northern Asia region, mostly in in Internet and food & beverage Industry. Before joining the Oak Investment, he worked in Motorola and Yahoo! consecutively. In Yahoo,
he took charge of the acquisition of 3721 founded Hongyi Zhou, which turned out to be a disaster, and the investment of Alibaba. What interests me is that his original major was mechanical engineering. His trace is quite typical and a lot of Chinese Students follow this way. Go to U.S for a Engineering Degree and then switch to financial field. He is a quite nice guy and gave two keynotes. In the second keynote, he noted the case study of LianLian Pay, one company he has invested in China. He used this case to illustrate the difference between business in China and that in America.The biz model of Lianlian pay is to construct the selling network of pre-paied card in China. Different with U.S, all the phone bills are charged using pre-paied card–user use to buy a pre-paid charging card and use the password in the card to charge their account. Lianlian pay construct the selling networks by setting partnership with thousands of grceries and convience store around China.
九点到十一点大家会开始选择自己的团队。每个团队需要在两天内完成一份商业计划,一个产品的原型,并在周日的demo大会上面做demo。来的人被分为三类,Tech People, Biz People and Design People, 注册的时候你需要选择自己的类别,然后去寻找合适的团队。
我开始想加入一个做吉他教学器的团队,他们的产品是一组夹在吉他上面的LED灯,通过LED灯的闪烁来显示你应该談那一首曲子。无奈他们只需要硬件方向的人,我对此又一窍不通。最后我加入了一个杂志出售的团队。团队的主题是Itune for Articles. Itune将现有的专辑打散,使得用户可以单独购买一首歌曲,从而开发了音乐市场的长尾。这个团队要做的就是将杂志打散成为一篇篇文章,然后放在网站或者移动应用商店上面出售,从而发掘杂志市场的长尾。
我们主要都是Biz Team在讲,但是因为时间有限,Biz Team将我的Simulation Result那一部分略过呢。等到问答环节的时候正好一个judge问我们的产品相对与Competitior来说有什么优势,我就接过来说了一下我的recommedation System和仿真的结果。demo完毕了之后还有个social的晚餐会。因为我得回去看第二天的Homework的due,所以我就没参加直接回去呢。最后我们团队获得了The most comerical potential 和 the 2nd place at the technology use.
波士顿地区的科技创业公司大多比较传统,大多是生物制药和机械物流。不过这次看到了绝大多数都还是Web或者App相关的,可见互联网,和移动互联网之火热。美国风险投资和技术创业最多的两个地区就是加州和波士顿。从Tec Transfer Center这样的学校组织到像Boston Startup Weekend的社会活动,这边确实有很多接触到这些信息的机会,也使得很多的人被鼓动到从事这个行业。为什么Microsoft,Akarmai和Facebook能够发轫与波士顿,很大程度和这边的环境是有关系。