The collaboration between Source Code Technology and NeusAI has already begun in earnest.
Unlike previous projects, the Source Code Technology employees involved in this project were both nervous and excited because they were collaborating with a team in Silicon Valley.
Li Ting's prediction proved correct; Jiang Yixin became the actual person in charge of the project.
Although the job title and salary haven't changed, the title of project manager sounds incredibly prestigious!
Moreover, because Jiang Yixin a good personality and gets along well with everyone in the development team, and the team leader Liang Wei gives her his full support, there is basically no resistance when she assigns tasks.
In addition, due to the special circumstances, the development team members generally lack confidence in their spoken English, and they would bring Jiang Yixin along whenever they needed to hold meetings with Silicon Valley, which has resulted in her having a strong presence in the company.
Jiang Yixin is now full of energy every day and arrives at the office early.
It wasn't entirely for the sake of showing off; the workload this week was simply beyond her capacity, and she needed to come here early to be quiet, undisturbed, and to process everything slowly on her own.
With rising status comes responsibility. Although Su Niannian hadn't explicitly told her, based on her experience as a workhorse for the past six months, she had already sensed it: if the project didn't go well, she would bear the main responsibility. Putting aside punishment, she didn't want to betray Su Niannian's trust. So at eight o'clock in the morning, Jiang Yixin was already sitting at her workstation, her desk piled with things: a printed copy of the Mercury Model's API documentation, three product comparison tables, and a flowchart she had drawn herself—all marked with four different colors of highlighter, making it look like an abstract painting.
Jiang Yixin first sat down to organize her to-do list. This week's main tasks were three things: figure out the capability boundaries of the Mercury Model, design a preliminary product plan for the overseas version of Open Things —where to keep the current features and where to do localization, and conduct the first round of communication with the technical window on Neus' side.
Many people think that localization is simply translating the interface, but that's far from enough.
Jiang Yixin conducted extensive research and found numerous examples of companies achieving impressive results in their home market but failing to localize successfully.
Take Airbnb's entry into China as an example.
Back in August 2015, Airbnb, with its top-notch product and technology team in Silicon Valley, made a grand entrance with the slogan "The sharing economy will change the world." At that time, everyone was optimistic that this unicorn, valued at hundreds of billions of dollars, would disrupt China's homestay market and even cripple the hotel industry. But what was the result? In 2022, Airbnb officially announced its withdrawal from the China market. From its entry to its exit, they burned through at least one billion dollars in China, but their revenue from business in China never accounted for more than 1% of Airbnb's total revenue.
So how exactly did they fail?
As a classic case of business failure, it has been analyzed by far too many people.
But after Jiang Yixin felt there was only one fundamental reason.
Arrogant.
When Airbnb first entered China in 2015, what were the login methods?
Log in with your Facebook or Google account.
Payment methods?
Visa and Mastercard credit cards only.
What is this restriction if not a joke in the China market?
At that time, it was mainly used by foreigners traveling to China.
It wasn't until 2017, two full years after it started operating, that Airbnb received its official Chinese name.
It's called Airbnb.
It was immediately ridiculed and became a trending topic as soon as it was released.
Ultimately, having tasted success in the United States market, Airbnb believed that beautiful design, compelling stories, and a vibrant community were all it took to achieve success. This is the reason for their success in the United States, and also the reason for their failure in China.
Because they have never really thought about what China users want.
President Su taught me, the first step in defining a product is creating user personas. If you don't even know who your target users are, how can you understand what they want?
"Make a tool for United States, not an English version of Open Things." Jiang Yixin wrote this sentence on a sticky note and stuck it in the center of the monitor. The United States is the first target market for the overseas version of Open Things. Since it has been determined that studying the usage habits of United States is the next focus of work, Jiang Yixin goes to every meeting with Neus' product manager with a learning mindset.
The product manager she was dealing with was a white woman named Sarah. She couldn't tell her age; she was like a typical United States sweetheart from a TV show. But despite her polite demeanor, Jiang Yixin always felt that the woman opposite her had a slightly condescending attitude.
Typing is fine.
Jiang Yixin's English is very good, and her writing is indeed excellent. She had a text exchange with Sarah on Sck, and although she occasionally had to look up a couple of terms, she could generally keep up.
Video conferencing is a different story.
During yesterday morning's video conference, Sarah spoke at an astonishingly fast pace, like she was on 1.5x speed, rattling off a string of gibberish mixed with various abbreviations and slang. After three minutes, Jiang Yixin began to doubt whether her CET-6 listening comprehension test was too easy.
Fortunately, she also enabled AI voice transcription, which automatically converts meetings into Chinese and English subtitles and sends out meeting summaries afterward. It just goes to show how technology changes lives.
After using AI, she wondered if she had wasted her time learning English.
Not all the technical support colleagues in the team are great—in fact, most of them aren't, but it doesn't affect the work. They use AI translation plugins to translate all kinds of API documents instantly without affecting understanding at all.
The only problem is with Chinese-to-English translation. It's not that the AI translation is bad, but when you can't understand the text, you can't tell if the AI translation is good or bad. In this case, you need to bring in Jiang Yixin or Li Ting to oversee it.
AI translation is flourishing everywhere: translation plugins, speech-to-text, automatic document localization... Jiang Yixin sometimes thinks that no matter what other applications of AI are like, at least in the field of translation, things can never go back to the way they were.
Bringing her thoughts back, Jiang Yixin added another note to the product proposal document, then checked the time; it was 9:40. Her colleagues were arriving one after another.
Qian Xiaole burst in right on time, at 9:59, with a sandwich in her mouth and her backpack slung over one shoulder, looking like she had just popped out of a portal.
"Good morning!" she greeted, taking the sandwich out of her mouth and gesturing to Jiang Yixin.
You said the word "early" with great confidence.
Qian Xiaole didn't care. She plopped down at her workstation, put her sandwich next to the keyboard, turned on her computer, and then opened her Little Red Book —this was her startup ritual.
Jiang Yixin was already used to it. Qian Xiaole's work rhythm was strange: for the first half hour, he would look at his phone, eat, and daydream, seemingly completely out of it. But once he got into the zone, his code output speed was the fastest in the group.
Is this the academic superstar who came from Shuimu Forum?
Qian Xiaole's work style is a little different from others.
She is a heavy user of AI.
Very severe.
While others write the logic and AI fills in the details, she writes the code and she reviews the logic. Her computer always has three windows open: a Heavenly Craft editor, a web browser, and a browser. She Open Things the code from the web browser directly into the editor, makes a couple of changes, and submits. After submission, the automated CI (Continuous Delivery) system automatically uses BugKiller to detect bugs, and she only fixes them if they are found. Her method is simple: she just passes the bugs identified by BugKiller back to the Heavenly Craft.
Some people also said privately that she was lazy.
But the data speaks for itself: her demand fulfillment volume was 30 percent higher than the second-place competitor.
It's not that she's smarter than others, it's that she believes in AI more than others.
"Xiaole," Jiang Yixin called out, "Are you going out for lunch? And Sister Ting too?"
"Great, I'll eat my sandwich now." Qian Xiaole turned around from in front of the screen.
So you were eating lunch just now?
What is your usual breakfast?
Two questions popped into Jiang Yixin's mind.
At lunchtime, the three of them sat in the dining area of the nearby shopping mall, each carrying their own lunch.
Li Ting was eating a light salad—she's still on a diet.
Jiang Yixin ordered a roujiamo (Chinese hamburger) and a takeout liangpi (cold skin noodles) from a newly opened shop in the mall—a place she wanted to explore.
Qian Xiaole bought another sandwich.
No, have you joined the Sandwich Cult?
Noticing Jiang Yixin's gaze, Qian Xiaole seemed to understand what she wanted to ask, "I had a tuna sandwich for breakfast, and this is a cheese and egg sandwich."
Jiang Yixin narrowed her eyes to slits. What's the difference? Aren't they both sandwiches?
Qian Xiaole, still reading her expression, began to explain, "The essence of a cheese and egg sandwich lies in the ratio of eggs to cheese. The eggs must be hard-boiled, chopped, and mixed with mayonnaise. The egg whites must be finely chopped but not mushy; the texture determines the bite. A gram too much or too little mayonnaise will make it tough; this place has the perfect ratio. The cheese must be cheddar slices; the creamy flavor only comes out when it's cold. If you use cheap, plastic-looking American cheese, it's a complete waste." Jiang Yixin sighed and looked at the roujiamo (Chinese hamburger) in front of her. "I'm a seasoned foodie, but... never mind, I still prefer my China palate."
Li Ting, watching from the side, chuckled, "What kind of encrypted call is this? Why doesn't Xiao Jiang even have to open his mouth?"
"Did you try out Mercury today?" Jiang Yixin asked while eating.
"I tried it," Qian Xiaole mumbled, chewing on his sandwich. "The code turned out alright, about on par with Heavenly Craft."
"Have you tried that automated workflow mode? It sounds like a big gimmick," Jiang Yixin asked.
"Oh, I tried that too." Qian Xiaole swallowed the sandwich in her mouth, then said a long speech, "You give it a task, like 'Add One user feedback collection function to this project,' and it will scan the project structure, plan the modification scheme, write code, and run tests. But the annoying thing is that it stops at each step and asks you: 'These files need to be modified, should we continue?' 'Is this implementation correct?' 'Should we run the tests now?'"
"This is so troublesome! Why don't you just write the final test and pass it? I just kept clicking 'yes' without even looking at the screen, clicking and...
As Qian Xiaole spoke, he pointed a finger wildly in the air.
Jiang Yixin, who was eating with her head down, suddenly remembered something and looked up abruptly when she heard this.