🌇 China Community Day Deep Talk: Summary of key points from an in-depth conversation with nine Chinese IT practitioners

#communityday #usergroup #ai #startup

Danny Chan speaking AI topic in AWS User Group (Shen Zhen, China) on 2024-07-06

More Photos of AWS COMMUNITY DAY in Shen Zhen

[Chinese Version] - AWS China UG Deep Talk：与八位IT从业人的深度对话要点总结

1. Traditional IT embraces the GenAI wave - Boss Mu

"Founder's Leadership Sharing" Business Strategy of an IT Company

"Asset-light, fast-cycle, stable cash flow" IT company leverages GenAI to develop larger application scenarios.
"Asset-light + fast cycle + stable cash flow" has two business advantages: replicability and ease of promotion. After completing a project, GenAI can be sold to other buyers in the same type of industry, attracting new buyers through word-of-mouth marketing among industries, and thus accumulating into an "industry solution.
This kind of industry solution has the characteristics of "short business and shallow domain knowledge", such as debt companies and dentist companies.
AI scenarios can be developed in one of the pain points of a long business. For example, since many Chinese SME factories in China are not fully digitized, there is a disconnect between "manufacturing" and "sales". Therefore, AI can only cut into one of the pain points of the long factory production chain, for example, is only to assist the sales staff to predict the shipment volume of the next quarter, and so on.
Promote knowledge sharing in the team. Since only by improving the overall knowledge level of the team can we embrace the AI wave to develop new business, employees are encouraged to participate in industry sharing activities, such as AWS User Groups.
Employees' ability to master time is important. Since each task is different in importance and time-consuming, it is important to choose the most important tasks first and evenly allocate time for each session to deliver a complete result.

2. Empowering AI in Childhood Education - Miss Xi

"AI Scenarios for childhood education" zero code AI products are still tools for domain experts.

Currently, AI product tools are still not easy to use. Even though Baidu and Tencent have launched zero code AI products for the general public, it is still difficult to create AI products with specialized domain knowledge.
The user interface is not easy to use. Because family early childhood education is a multi-clerical scenario, although it is very suitable to use AI to help, but the process of entering information in the UI is still not friendly.
Model performance is hard to understand. Because zero code AI products do not have a visual model performance page, plus can not read the model data, so the public can not understand how to use.

3. Financial Data Analytics BI to AI Products - Miss Hai

"When Financial BI Meets Traditional Finance in Hong Kong" - Miss Hai (Q)
Miss Hai (Q): Is traditional finance passively accepting AI transformation because of catching up with the AI wave?
Mr Kuang (A): As traditional financiers, we embraced AI more than 10 years ago, but financial scenarios where AI can be applied are rare, so only in the last two years when AI technology has been perfected can we get through to the financial scenarios on the ground.

Miss Hai (Q): So what financial scenarios can AI be applied to.
Mr Kuang (A): Mainly in text-based financial scenarios, such as news sentiment analysis, research report attitude analysis. But these financial AI scenarios have existed for many years, only in the past two years AI technology has become more perfect, such as a large model with 650,000 parameters, in order to realize the AI fantasy scenarios of ten years ago.

4. Member of PhD research team, The University of Hong Kong - Mr Xu

"Business coordinator of the cross-sector "Lab Project Sharing on the Ground

You need to know the "business logic" of your field, but you should leave the research and development process to the field experts. It is a waste of your time to learn because you can't do it by yourself, but it is a waste of your time.
It is more of coordinating resources, but first of all, you should understand the "business logic" of the field, so that you can better coordinate the resources.
I have worked on AI, financial and medical projects. Adopted group thinking, enjoying the 'brand reputation' advantage even though the three are completely independent projects.

5. Side hustle as a serial entrepreneur - Mr Li

"Work Life Balance Sharing": The balance between a full-time worker and a side hustle entrepreneur, the balance between two sides of life.

China's VC competition emphasizes more on business logic, even in the public interest track, it still focuses on business logic and application scenarios.
Hong Kong's VC competition allows for more innovation, but China's focus on business logic undermines innovation.
I work for a foreign-owned global logistics company, which allows me to work remotely, so I have more time to be an entrepreneur.
Have done AI novel generation video project, AI chat robot project. For the financial chat robot, only focus on the application scenario development, and leave the professional financial domain knowledge to the professional financial people.

6. AI Entrepreneur - Mr Jack

"AI product commercialization sharing" VC funds and investors are two different ways of playing the game.

VC funds just help the first stage of product landing, but the support for commercial realization is not enough, so it's more of a learning stage.
Prioritize building a good relationship with investors. Since investors are the biggest VIP users, we should prioritize the investors' point of view and build a good relationship with them.

7. Senior Solution Engineer - Mr. Zhang

"GenAI Concept Project Technology Sharing" Multi-Round Dialogue Project with Multiple AI agents

We need a moderator AI agent because multiple AI agents have the following problems: (1) No common goal. (2) speak in random order. (3) they don't know what they are saying to each other.
Therefore, the role of the moderator agent is hidden behind the scenes: (1) Define each agent's identity. (2) pre-input the knowledge base to the group. (3) define the common goal. (4) control and coordinate the flow of the presentation. (5) controlling and coordinating the flow of presentations. (6) broadcasting the conversation.
But there are still some limitations: (1) If the AI agent proposes wrong content, the moderator can't correct it, which leads to the failure of the whole process.

8. Marketing Planner - Miss Nan

"Industry Experience Sharing" Game Publishing, TikTok E-commerce

"Game Publishing" means helping to match studios and marketing resources.
Small, medium, and large game studios all need game publishers to help them connect with marketing resources. The value point of "Game Publishing" is human resources, and the profit point is X% commission.
The essence of "TikTok E-commerce" is to sell traffic conversion rate. For example, XX Games spends 70 million dollars on TikTok advertisements, which is to buy the traffic conversion rate, and the goal is to guide users to take actions through TikTok advertisements, such as downloading games and buying goods.
"TikTok E-commerce" has a lot of user data and operational data, so it is necessary to analyze which traffic is the most efficient and valuable, and how to improve the profitability of the business forecast.

9. Front-end Developers Transition to Marketing Planning - Miss Ya

"Female Developers' Dilemma in "Sharing the Female Workplace Environment

Female developers are still being treated unfairly.
First, the culture of front-end involution. Due to the large number of front-end developers and the complexity of tasks, important tasks are given to male developers, resulting in no promotion opportunities for female developers.
Secondly, women's life in the workplace is short. Because in China, married women are mainly housewives, and even if they return to the workplace, they are regarded as lacking work experience.
Third, the instability of women in the workplace. Because women have maternity leave, maternity leave, etc., executives do not want to hire female developers who may leave their positions vacant.
The reason for the transition from front-end developer to marketing is not only the culture of the industry, but also the fact that the Chinese workplace is not friendly to women.

Developer Ecosystem Viewpoint:

1. How to see 996 culture?
Boss Mu: Never asked employees to work overtime, more like hoping employees have excellent time control ability to finish their work more efficiently. For example, some of the work that is internal to the company can be reordered in terms of importance rather than solved by working overtime.

Mr Zhang: If there is a project delivery, occasionally it is limited overtime. The most important thing is the completion of the project. For example, for exhibitions, you need to do the site layout, hardware debugging, docking with the team, etc. early.

2. How to see the eight-legged culture?
Boss Mu: will not assess the eight-legged text, more assessment of the actual work ability, such as is how many development languages and tools, how much development experience, etc..

Mr Li: Playing workers have the need to memorize the eight-legged text, due to the peer competition, so can only rely on the eight-legged text to compete for the fight.