Here is a concise summary of the chapter:

Key Plot Points & Events

• Strong Financial Performance: Lu Mingzhou (CFO) reports November financials. BugKiller earned $1.12M (boosted by enterprise clients from the code-stealing scandal); Open Things earned $3.46M (4x projections, driven by a competitor's collapse and GAIS traffic). Total revenue is $4.58M, with a net profit of $1.2M.
• December Projections: The Thousand Sails white-label integration launches, projected to add $3.87M in December revenue, pushing total monthly revenue past $8M.
• Model Breakthrough: Zhao Wenyuan's team improves post-training accuracy from ~60% to ~85%, reducing inference costs by half. This saves $1.2M in December alone and makes the earlier $20M model investment highly worthwhile.
• Strategic Dilemma: Han Luyi reveals the model team's next goal: training a proprietary foundation model from scratch. The cost is $30M per attempt, with no guarantee of success.
• Proposed Solution (Internal): Han Luyi quietly considers spinning off the model business into a separate entity to avoid dragging down Source Code's application-focused valuation, but decides not to voice this idea yet.

Character Developments

• Han Luyi: Shows ambition to transform the company from a "house builder" on someone else's foundation into a company that owns the foundation itself. He is willing to take big financial risks for long-term control of the ecosystem, but is cautious about sharing radical restructuring ideas.
• Lu Mingzhou: Serves as the pragmatic financial counterweight. He is thrilled by the current success but deeply concerned about over-extending into the capital-intensive large-model race. He acts as the voice of reason regarding investor expectations and dilution.

Significant Themes

• Application Company vs. Foundation/Model Company: The core conflict is whether to remain a profitable application business (cash-flow positive) or enter the "burn-first" infrastructure race dominated by massive capital injections and heavy dilution.
• Financial Sustainability vs. Strategic Independence: The tension between relying on another company's underlying model (vulnerable to price hikes or control) versus spending hundreds of millions to build one's own, risking bankruptcy.
• Investor Expectations and Dilution: Lu Mingzhou highlights that investors funded a post-training strategy (fine-tuning), not a from-scratch training strategy. Entering the foundation race would require either massive dilution or an unprecedented "win two wars" pitch.

Character Interactions

• Information Sharing: Lu Mingzhou delivers the stellar financial reports; Han Luyi, happily surprised, reciprocates by sharing the model's accuracy breakthrough, causing Lu's excitement to spike over the cost savings.
• Strategic Debate: Han Luyi proposes training a foundation model. Lu Mingzhou pushes back, explaining that the economics of model companies (lose now, monetize later) don't mix well with application companies (profit now), and warns that current profits can only fund one 30M attempt.
• Unspoken Tension: The chapter ends with Han Luyi internally formulating the spin-off idea but withholding it from Lu Mingzhou, implying a future strategic conflict or major announcement.

Chapter Summary: "Beyond the Coordinates"

1. Key Plot Points

• Strategic crossroads: CFO Lu Mingzhou tells Han Luyi that building foundation models can't be done with incremental savings—the company must find a big investor and go all-in. He warns the application route is safer since Source Code already has a profitable model. Worst case scenario: "go home and sell sweet potatoes." Han Luyi says he needs time to think.
• Han Luyi's existential reflection: Heavenly Craft's code model hits 84.7% accuracy (vs. Qianyuan's 85%). Han Luyi realizes once they surpass Qianyuan, there's no benchmark left—no finish line for AI models. He recalls his GAIS speech about how the current AI interaction loop (prompt → generate → test → revise) isn't fundamentally different from emailing requirements twenty years ago, just faster. He senses Open Things is far from the real future but can't yet articulate what that should be.
• International virality discovered: Li Ting and Jiang Yixin find that the GAIS speech has exploded overseas—X posts calling it "mysterious oriental power," a former Google engineer (120k followers) posting a detailed architecture analysis, and the subtitled video passing 200k views on TickTick.
• Proposed overseas PR push: Jiang Yixin suggests the company actively manage its overseas narrative instead of letting spontaneous posts distort it. Li Ting agrees; they start compiling a report.
• Team dinner + mysterious email: The team heads out for dinner; no one dares ride with Han Luyi except Su Niannian. Before leaving, Han Luyi receives an email from Ryan Miller, CEO of NeusAI, titled "An Open Proposal for Source Code Technology."

2. Character Developments

• Han Luyi: Shows genuine insecurity about the company's strategic direction. He's less concerned with beating competitors than with finding a fundamentally better human-AI interaction paradigm. Also confirms his lingering discomfort with being "the boss" everyone keeps distance from.
• Lu Mingzhou: Established as a pragmatic, honest CFO who delivers tough truths without sugarcoating. His "sell sweet potatoes" joke reveals mutual familiarity with Han Luyi.
• Li Ting: Proactive and strategic—she's monitoring overseas sentiment before being asked and already thinking about media control and PR planning.
• Jiang Yixin: Bubbly but sharp; she spots the overseas opportunity first, understands the risk of uncontrolled narrative, and volunteers with a proud CET-6 score of 600. Shows growth from junior staff to more engaged contributor.
• Su Niannian: Her small gesture (volunteering to ride with Han Luyi) highlights her unique, comfortable relationship with him—she acknowledges other employees' discomfort without judgment.

3. Significant Themes

• The benchmark problem: Surpassing competitors leaves you with no reference point. AI model development has "no finish line, no referee—only a track stretching beyond sight." Han Luyi questions what 100% code accuracy would even mean to a real user.
• Safe profit vs. ambitious gamble: The classic startup dilemma—Lu Mingzhou explicitly frames it as a choice between the known profitable path and a risky leap that would transform Source Code's identity entirely ("If we win, Source Code won't just be an application company anymore").
• The interaction paradigm: Han Luyi's core dissatisfaction—that current AI-assisted coding workflows are just accelerated versions of old communication patterns. Quantity of capability isn't the same as a fundamentally better way of working.
• Going global / public perception: The chapter shows both the thrill (overseas recognition) and danger (uncontrolled narratives, "demo scam" accusations, cultural exotification like "mysterious oriental power") of organic international attention.
• Leadership isolation: The empty elevator moment—employees prefer cab rides over sitting with their boss.

4. Character Interactions

• Han Luyi & Lu Mingzhou: Deeply familiar, direct. Lu finishes Han's sentence ("sweet potatoes"), then laughs. Their conversation is a genuine advisory exchange, not a hierarchical command.
• Li Ting & Jiang Yixin: Playful workplace banter that turns into professional collaboration. Jiang teases Li Ting about browsing Twitter during work; Li Ting redirects her to the viral posts. Their dynamic is mentor-mentee with mutual respect.
• Jiang Yixin & Xiao Yang: Light comedy with the "CET-6" exchange and her teasing him about "that kind of website"—shows the team's casual atmosphere.
• Han Luyi & the team: His invitation to share his car is met with silence; no one accepts. Su Niannian rescues the moment, naming the awkwardness openly ("They'd feel awkward sitting with you"). A small but telling moment of the distance his position creates.
• Han Luyi & Su Niannian: Comfortable, unhurried presence with each other. She's the one person who navigates his status without friction. He immediately shares the NeusAI email with her ("I'll tell you in the car").

Cliffhanger: Ryan Miller's "open proposal" arrives at the exact moment Han Luyi is weighing whether to gamble on foundation models—potentially offering external validation (or a competitor's interest) in the very direction he's hesitating about.

Chapter Summary: “Letter from Silicon Valley”

Key Plot Points and Events

• While driving along Pudong Avenue, Han Luyi reads an email from Ryan Miller, CEO of Neus AI, a leading Silicon Valley AI startup.
• Ryan introduces Neus AI’s strong position: $580 million in Series B funding, a 40-person research team, and researchers from OpenAI, DeepMind, and Meta.
• He outlines Neus AI’s new focus on “autonomous workflow”—AI that can independently break down tasks, use tools, execute multiple steps, and verify its own results, with humans providing only the initial instructions.
• Ryan praises Source Code Technology’s product design, particularly its “remarkably complete loop” for helping non-developers use AI, but points out that Source Code lacks a general-purpose foundation model.
• Han realizes that Ryan’s proposal is not merely an invitation to cooperate. It is also a strategic probe and a warning that Neus AI may become a direct competitor.
• Han discusses the email with Su Niannian. They conclude that the future of AI may move beyond chat-based interaction toward AI that performs actual work.
• They speculate that once autonomous workflows become viable, companies such as OpenAI, Google, and major domestic AI firms will quickly pursue the same direction. Companies without general models may be left behind.
• Han and Su decide to schedule a private video meeting with Ryan, limiting attendance to themselves for the time being.
• The chapter ends as they arrive at a team dinner. Jiang Yixin jokingly calls Han and Su her “left and right arms,” while Han reflects that Source Code has already built the product-side closed loop Ryan praised.

Character Developments

• Han Luyi recognizes that his privately developing ideas have independently emerged as Neus AI’s strategic direction. He is unsettled by Neus’s superior technical foundation but also energized by the challenge.
• Han’s understanding of Source Code’s limitations becomes sharper: Heavenly Craft is a powerful code model, but it cannot yet support broad natural-language reasoning, tool coordination, and autonomous task execution.
• Su Niannian demonstrates strategic insight by identifying the implications of Neus’s model-product distinction and predicting that autonomous workflows will become the next industry-wide paradigm.
• Han and Su’s partnership is strengthened as they analyze the threat together and agree to manage the issue discreetly.
• Jiang Yixin remains confident and playful in the team setting, showing her sense of belonging and trust in Han and Su’s leadership.

Significant Themes

• Technological competition: The email represents both potential cooperation and an emerging confrontation between a company with a strong model and one with superior product design.
• The next AI paradigm: The chapter contrasts AI as a conversational assistant with AI as an autonomous agent capable of completing real-world tasks.
• Model versus product capability: Neus has the stronger general model, while Source Code has created a more complete user-facing workflow. The future may depend on combining both strengths.
• Strategic anxiety and ambition: Han feels threatened by Neus’s lead but responds with curiosity and determination rather than fear.
• Industry convergence: Once autonomous workflows prove effective, major AI companies are expected to rapidly follow, making the race much broader than a single rivalry.
• Recognition and vulnerability: Ryan’s praise validates Source Code’s product achievement, but it also exposes the company’s greatest weakness.

Character Interactions

• Han shares the email with Su, and they carefully dissect Ryan’s language and intentions.
• Su challenges Han to confront the implications of Neus’s technological advantage, while Han explains the limitations of Source Code’s code-focused model.
• Their discussion develops from analyzing Neus AI into a broader prediction about the future of the AI industry.
• Han and Su agree to attend the meeting with Ryan alone, signaling caution and trust.
• At the restaurant, Jiang Yixin’s playful claim that Han and Su are her “left and right arms” provides a lighter contrast to the serious strategic threat they have just discussed.

Key plot points and events

• Source Code’s staff celebrate the launch of a new feature at a Hunan restaurant. The meal is lively, with newcomers initially nervous, playful banter about spicy food, and a dramatic flaming fish-head dish that everyone enjoys.
• Although the celebration is underway, Han Luyi remains preoccupied with an email from Ryan. He concludes that Source Code’s code-focused model cannot support autonomous AI: the company needs a general-purpose foundation model.
• Han considers creating a separate company to develop that model, with its own team and financing, rather than stretching Source Code’s resources. He invites Su Niannian to discuss the idea privately.
• Han and Su leave the dinner early. At Han’s apartment, he explains his plan. Su asks the crucial question: will he lead the new model company or remain CEO of Source Code? Realizing he has already made up his mind, Han messages Zhao Wenyuan to meet him the next morning.
• As the rest of the group prepares to go to karaoke, Liang Wei approaches Jiang Yixin and Li Ting, saying he has something to tell them. The chapter ends before he explains.

Character developments

• Han Luyi shifts from worrying over a strategic challenge to recognizing that building a foundation model may require a separate company. Su’s question forces him to confront his own role in the decision.
• Su Niannian is supportive and perceptive. She listens without rushing to answer and identifies the personal leadership choice at the heart of Han’s proposal.
• Jiang Yixin remains energetic and impulsive, enthusiastically ordering food and lighting the fish-head dish despite Li Ting’s attempt to stop her.
• The newer employees become more at ease over the course of the meal, while Little Yang’s uncertainty about workplace customs provides comic relief.
• Liang Wei breaks his usual quiet, signaling that he has something important to say to Jiang Yixin and Li Ting.

Significant themes

• Strategic change and risk: Source Code’s existing expertise may not be enough for the next stage of AI development.
• Focus versus expansion: A separate company could protect Source Code’s attention while pursuing a new, expensive direction.
• Leadership and responsibility: The decision is not only what to build, but who will lead each organization.
• Workplace community: The dinner shows a team gradually relaxing and bonding, even as major business decisions weigh on its leader.

Character interactions

• Jiang Yixin’s jokes and enthusiasm help loosen the atmosphere; Li Ting gently reins in both her and Little Yang.
• Han declines the group’s karaoke invitation and leaves with Zhang Biao, while Su follows him to discuss the proposal.
• In their private conversation, Su moves beyond the business plan to ask what Han himself intends to do—prompting his next step.
• Liang Wei’s approach to Jiang Yixin and Li Ting sets up an unresolved conversation.

Chapter Summary: “Doing What Is Difficult but Right”

Key Plot Points and Events

• Han Luyi decides to establish a new company focused on developing a more powerful AI model and contacts Zhao Wenyuan to discuss joining him.
• Su Niannian challenges the decision, raising practical concerns:
  • Who will maintain and improve Heavenly Craft after Zhao leaves?
  • Whether Zhao will agree to abandon Source Code.
  • How the new company will secure independent funding.
  • What Source Code will gain from the new model.
  • Whether Han can manage both companies while Source Code is expanding rapidly.
• The chapter briefly shifts to a dinner at Spicy Realm, where Liang Wei shares a painful story with Jiang Yixin and Li Ting. He once worked for a successful startup founded by a married couple, accumulated valuable stock options, and expected financial security after the company went public. However, conflict between the founders caused the company to collapse, leaving his shares nearly worthless.
• Liang’s story makes Jiang Yixin worry about the dangers of a husband-and-wife-led company, though she privately believes Han and Su are different.
• Back at Han’s apartment, Han explains that Open Things has enormous value but that its potential is limited by Source Code’s existing model. He believes AI is entering a new era in which intelligence and comprehension—not conventional products or features—will determine success.
• Han argues that delaying until every risk is calculated could mean permanently falling behind. He sees creating the new company as a fight for survival rather than a simple business gamble.
• Su remains unconvinced that the risks are acceptable, but chooses to support Han anyway. She agrees to “gamble” with him and asks what she can do.
• Han plans to speak with Zhao Wenyuan and then meet President He in Beijing to seek investment. He resolves to find other backers if President He does not understand or support the vision.
• Su leaves, and Zhao later confirms that he will meet Han the next morning.

Character Developments

• Han Luyi: Reveals his strong conviction that the AI industry is approaching a fundamental paradigm shift. He accepts the risks of starting over because he believes failing to act would be more dangerous than taking the gamble.
• Su Niannian: Acts as the practical and cautious voice, carefully examining the financial, organizational, and strategic risks. Although Han does not fully persuade her, she ultimately chooses loyalty and shared commitment, agreeing to support his decision.
• Liang Wei: His past experience has made him deeply wary of founder conflict and unstable company leadership. His warning adds an emotional and practical dimension to the risks Han and Su face.
• Jiang Yixin: Begins the chapter playfully trying to bring Han and Su closer, but Liang’s story forces her to reconsider the stability of their partnership and the future of Source Code.
• Zhao Wenyuan: Though absent from most of the chapter, his willingness to meet Han suggests he may become an important member of the new venture.

Significant Themes

• Doing what is difficult but right: The central theme is the choice to pursue a risky but necessary path before certainty is possible.
• Innovation versus stability: Han believes protecting Source Code’s current success could ultimately lead to obsolescence if the company fails to embrace the new AI paradigm.
• Risk and responsibility: The decision affects investors, employees, existing products, and the future of Source Code, making the gamble far larger than a personal career choice.
• Trust and partnership: Su’s decision to support Han despite her doubts demonstrates the strength of their personal and professional relationship.
• The danger of internal conflict: Liang’s story serves as a warning about how disputes between founders can destroy both a company and the financial hopes of its employees.
• Survival in technological competition: Han views AI development as a flywheel race in which falling behind even briefly may make recovery impossible.

Important Character Interactions

• Han and Su: Their extended debate forms the emotional core of the chapter. Su questions Han’s plan from every angle, while Han explains why he believes inaction is the greater danger. Their conversation ends not in complete agreement, but in mutual commitment.
• Liang, Jiang Yixin, and Li Ting: Liang’s personal story warns them about the consequences of founder conflict and makes them reconsider the risks facing Source Code.
• Han and Zhao Wenyuan: Zhao’s agreement to meet marks the first concrete step toward assembling the new company.
• Han and President He: Their upcoming meeting represents Han’s next attempt to secure support and financing for his vision.

Key Plot Points and Events

• Zhao Wenyuan arrives early at work and reviews data, receiving a LinkedIn message from Neus AI about a potential job opportunity.
• He meets with Han Luyi, who proposes building a general-purpose foundation model and outlines a plan for a new company.
• Zhao Wenyuan decides to join the project, signaling the end of post-training and the beginning of a new path.

Character Developments

• Zhao Wenyuan: Struggles with the limitations of post-training and is drawn to the idea of building a foundation model from scratch. He shows loyalty and trust in Han Luyi's leadership.
• Han Luyi: Demonstrates vision and determination in pursuing a new project. He values Zhao Wenyuan's technical skills and trustworthiness.

Significant Themes

• Ambition vs. Pragmatism: Zhao Wenyuan grapples with his desire to build a foundation model versus the practical constraints of his current role.
• Leadership and Trust: Han Luyi's actions and decisions earn Zhao Wenyuan's trust, leading to his commitment to the new project.

Character Interactions

• Zhao Wenyuan and Han Luyi: Their meeting leads to a significant decision, with Han Luyi presenting a compelling vision and Zhao Wenyuan showing loyalty and commitment.
• Zhao Wenyuan and his team: He interacts briefly with his team members, demonstrating his role as a mentor and leader.

Chapter Summary

1. Key Plot Points

• Han Luyi meets investor He Yunshen, hinting he wants to enter AI foundation models / autonomous workflows after learning from Silicon Valley’s NeusAI.
• He Yunshen firmly opposes, using Chen Tianqiao / Shanda’s failure as a warning: the right direction at the wrong time still leads to ruin.
• Han Luyi tries a compromise: back an independent team instead of doing it himself. He Yunshen rejects it too — as long as Han is involved, it distracts from Source Code.
• Han Luyi leaves, internally disagreeing. He believes waiting for consensus will be too late.
• He Yunshen’s assistant probes Han’s bodyguard Zhang Biao about Han’s purpose in Beijing.
• He Yunshen decides to keep leading Source Code’s next round but orders post-investment monitoring to prevent Han from spending recklessly.

2. Character Developments

• Han Luyi: Patient on the surface, but ambitious and calculating. He uses his Sight ability to read He Yunshen’s emotions and motives. He is willing to consider a risky second front but refuses to fully reveal his plans.
• He Yunshen: Seasoned, cautious investor. He values Han but prioritizes safety, especially with his new fund’s fundraising pending. His stance is paternalistic: “steady first, gamble later.”
• Zhang Biao: Sharp but discreet; notices He Yunshen’s use of “personal matter” and answers vaguely.

3. Significant Themes

• Timing vs. vision: Is it better to move early or wait for the right moment?
• Application vs. underlying technology: Whether a startup should build models or focus on products/cash flow.
• Founder–investor trust: Han hides his real intentions; He Yunshen watches him closely.
• History as warning: Shanda’s box as a case of “right direction, wrong time.”

4. Character Interactions

• Han Luyi tests He Yunshen’s openness twice; both times He Yunshen blocks him politely but firmly.
• Han Luyi feigns agreement and ends the meeting gracefully.
• He Yunshen’s assistant casually interrogates Zhang Biao; Zhang gives nothing away.
• He Yunshen’s final instructions reveal he is protective but wary of Han’s ambition.

Summary: Borrowing Strength

1. Key Plot Points & Events

• Han Luyi returns to his room after meeting He Yunshen in the capital. He Yunshen advised waiting six months to a year to run more data before raising a Series B, arguing the valuation would multiply several times.
• Zhao Wenyuan texts asking how the meeting went; Han Luyi replies vaguely, "Still talking," fearing the truth would shake Zhao's resolve.
• Han Luyi decides not to wait. He plans to act first, report later—build the new venture, secure data/compute/team, then return to He Yunshen to present a fait accompli rather than ask permission.
• He realizes the new company's real bottlenecks are data and compute (GPUs), not money. Since money is just an intermediate step to buying GPUs, he decides to bypass cash financing entirely and find partners to trade compute for equity.
• He calls Liang Yu of Qingyue Cloud to ask about training-grade compute. Liang Yu says all H100/H200s are rented out, Q4 orders are booked into next year, and new capacity won't be available until at least April.
• Han Luyi concludes the whole market is short on GPUs—except the top-tier tech giants, implying his next target.

2. Character Developments

• Han Luyi: Becomes more decisive and risk-tolerant. He acknowledges his plan makes him a "Lü Fengxian" (a figure who betrays a benefactor), but justifies it strategically: better to show results than argue with a judgment already made. He turns an open-ended question ("Can I do this?") into a multiple-choice one for He Yunshen ("I've already done it. Do you want in?"). He also shows resourceful, unconventional thinking—cutting out money as the middleman and treating compute as the true currency.
• He Yunshen (off-page): Functions as the cautious, rational investor voice. His sound advice paradoxically pushes Han Luyi toward a riskier path.
• Liang Yu: Honest and pragmatic; willing to help but constrained by a GPU-scarce market.

3. Significant Themes

• The safe road is already crowded: Waiting for a "perfect" funding round is a zero-sum game; unconventional speed and action may be the only edge.
• Trust vs. initiative in founder-investor relationships: How to disagree with a backer who saved your company without destroying trust—by letting results speak instead of words.
• Resource scarcity as leverage: GPUs are the real bottleneck; capital is merely a means to an end. Whoever controls compute holds power.
• "Act first, report later" as a deliberate start-up strategy when traditional fundraising is too slow.

4. Character Interactions

• Zhao Wenyuan → Han Luyi (text): Zhao is anxious for news about the meeting. Han deliberately withholds his true plans, prioritizing Zhao's morale over full transparency.
• He Yunshen → Han Luyi (recollection): He Yunshen's advice (wait for Series B) is the catalyst for Han's entire internal conflict and eventual decision to go his own way.
• Liang Yu → Han Luyi (phone call): Han probes for GPU capacity; Liang candidly explains the market shortage and offers to ask around but can't promise training-grade compute before April. The call solidifies Han's conclusion that only tech giants can supply the needed compute.

Chapter Summary: "The Real Use of a Cheat Ability"

Key Plot Points / Events

• Han Luyi and Zhang Biao wait at Beijing Daxing Airport for a flight back to Haicheng.
• Han Luyi observes a young woman, Chen Jiujiu, struggling to use an AI chatbot to generate a Q4 performance report PPT. Her prompts are generic and ineffective.
• Using his cheat ability [Reality Code Sight], Han Luyi sees her real identity, emotions, and—crucially—her deeper need: she actually wants a proposal to spin off her data auditing segment into an independent business line, not just a routine PPT.
• On the plane, Han Luyi reflects deeply on how Sight works and how to apply it to building a large AI model.
• He realizes Sight is an information processor, not a creator — it can only structure existing information, not invent or predict.
• He re-examines why BugKiller succeeded: it replicated Sight’s ability to “see bugs” as a product.
• He concludes that the same logic should be applied to model training: use Sight to see users’ true intent behind surface instructions, turn that into high-quality training data, and fine-tune a model to understand intent better.
• He plans to use this as a bargaining chip in negotiations with Dingsheng.
• After landing, he messages Zhao Wenyuan to meet at the office, skips eating, and pushes for faster development than BugKiller’s two-month timeline.

Character Developments

• Han Luyi: Moves from tactical use of Sight (reading people, finding bugs) to a strategic insight — the real value of his cheat ability is turning it into a scalable product/data advantage. Shows growing clarity and ambition in the AI race.
• Zhang Biao: Loyal and practical; asks about meals but follows Han Luyi’s lead without question.
• Chen Jiujiu (minor): Illustrates the gap between users’ stated prompts and actual needs; fatigued, frustrated, and unaware that her request is misaligned with her real goal.

Significant Themes

• The real use of a cheat ability: Not using supernatural power directly, but translating it into a product or data moat.
• Information vs. creation: Sight cannot predict or invent, only process what already exists.
• The intent gap in AI: Models fail because they cannot see the true goal behind a user’s words; this is the next breakthrough point.
• Data as competitive advantage: High-quality, intent-labeled data is a moat others cannot replicate.
• Speed as strategy: The market demands faster iteration; past success (BugKiller) sets the pace.

Character Interactions

• Han Luyi observes Chen Jiujiu’s screen and uses Sight on her — no direct conversation, but a key demonstration of his data-gathering capability.
• Brief exchange with Zhang Biao about eating after arrival; Han Luyi prioritizes returning to the company.
• Indirect interaction via message with Zhao Wenyuan, signaling immediate work upon return.

Key Realization

Sight can’t build the model for him. But Sight can see what users really want — and that data, converted into high-quality annotations, can train a model that understands intent better than anyone else’s. That becomes his new moat and Dingsheng bargaining chip.

Chapter Summary: "I'll Count It as a Loan"

Key Plot Points

1. Han Luyi presents high-quality data annotations: He returns from an unsuccessful meeting with President He and shows Zhao Wenyuan a spreadsheet of 287 user-behavior entries from Open Things—each annotated with a new column, "Actual Intent," explaining why each user made the choices they did. Example: a salesperson who asked for a "customer management system" actually needed a CRM supplement tool (a follow-up log), and an HR user who requested a "shift schedule" actually needed an overtime-compliance calculator.
2. Purpose of the data: The annotations are not for Heavenly Craft (the existing code model). Han Luyi plans to use them to fine-tune a small open-source model as a proof-of-concept prototype—demonstrating that high-quality intent-understanding data can dramatically improve model performance—to bring to investors.
3. Feasibility discussion: Zhao Wenyuan verifies the annotations and finds them remarkably accurate. He says 5,000–8,000 such entries would suffice to show visible improvement. He cautions a small model still can't match GPT-4, but Han Luyi insists the prototype is only to prove feasibility. Han Luyi commits to delivering within 2–3 weeks. (Zhao assumes "done" means data ready; Han Luyi means the model ready—unnoticed miscommunication.)
4. Su Niannian arrives: She brings Han Luyi lunch (sandwich, rice ball, coffee). After learning the He meeting failed, she offers to personally lend 16 million yuan (post-tax money from selling her original shares) as a loan to the new company—no equity, no shares. She jokes about wanting an interest rate for her retirement.
5. Han Luyi hints at another funding path: He says he has his own way to handle funding (referring to Dingsheng compute) but doesn't disclose it, knowing his partners have had bad experiences with Dingsheng. Su Niannian doesn't press him.
6. Upcoming meeting: Su Niannian reminds Han Luyi of the video meeting with Ryan (arranged by Neus) next Wednesday at 8 a.m.
7. Chapter close: Han Luyi silently acknowledges his hidden assist—"Brother Sight, I'm counting on you"—implying an AI (Sight) enabled the annotation quality he can't explain.

Character Developments

• Han Luyi: Plays cards close to his chest—doesn't reveal Sight's involvement or his Dingsheng plan. Strategic, willing to use a POC to convince investors he couldn't convince directly.
• Zhao Wenyuan: Thorough, skeptical, empirically validating before believing. Grows impressed by Han Luyi but senses unexplained gaps. Restrains curiosity about Han–Su dynamic.
• Su Niannian: Bold and selfless—lends virtually her entire liquid net worth with zero equity, betting purely on Han Luyi's judgment. Trusts his timing and doesn't pry into details.
• President He (off-screen): Rejects funding, pessimistic about the venture—catalyst for the pivot to POC fundraising.

Significant Themes

• Data quality over data quantity: Precise annotation of intent (why users behave) is more valuable than raw behavioral logs.
• Bridging the investor gap: A pragmatic prototype may persuade where a vision pitch fails.
• Trust and personal capital: Loyalty expressed through unsecured financial risk; partnership built on mutual confidence, not contracts.
• Hidden leverage: Quiet reliance on non-human intelligence (Sight) as an unfair advantage.
• The "surface request vs. real need" problem: Users rarely ask for what they actually want—the core insight driving the annotation philosophy.

Character Interactions

• Han Luyi & Zhao Wenyuan: Data-focused, technical back-and-forth; Han reveals just enough, Zhao verifies and challenges but becomes convinced.
• Han Luyi & Su Niannian: Terse, comfortable, and deep-trust chemistry—she feeds him, offers millions, asks nothing; he accepts without ceremony, promising to return the favor if needed.
• Zhao Wenyuan & Su Niannian: Zhao is visibly stiff when she enters mid-confidential-discussion, exposing his unawareness of her involvement; Han covers with "she knows everything." Zhao's probing question about their relationship is deflected with "old classmates"—both men silently read between the lines.

1. Key plot points and events

• Han Luyi tells Zhao Wenyuan not to use Open Things’ data for training. The 200 annotated entries Han showed him were only a quality benchmark; the actual model needs broader, more diverse data.
• Zhao explains the industry-standard pipeline: massive collection, cleaning, and annotation. He says open-source corpora and outsourced annotation can’t match Han’s sample quality, and rejects the idea that Han can guarantee it.
• Han takes full responsibility for the data, promising annotation quality matched to his samples, and tells Zhao to focus on building the training pipeline.
• Han internally reveals his method: ordinary annotation extracts surface intent, but his “Sight” ability uncovers hidden context. Example: a user saying “My dad just retired last month” is not asking for persuasion techniques, but how to refuse a newly lonely father without hurting him.
• Han visits lawyer Gu Siyue to discuss a new independent model company, with core team member Zhao Wenyuan transferred from Source Code Technology.
• Gu explains Zhao’s legal position: non-compete clauses require Source Code to sue first, but his unvested options are a hidden risk. At a 2-billion valuation, Zhao’s 3% options are worth ~60 million; they become void if he resigns or is fired.
• Investor He Yunshen, as board member, could discover Zhao’s outside work, force him to stop, fire him, void his options, and question Han’s full-time commitment.
• Han’s own Series A full-time commitment clause could make simultaneous control of another company a technical breach, but Gu notes investors rarely sue over it in practice.
• Gu warns that the real danger is trust. She recounts the Ding Pay story: in 2011, Lü Yun transferred Ding Pay out of Prosperity Group into his own entity, citing regulatory reasons, without investor consent. He kept control but paid compensation and permanently fractured trust. Investors later always asked if every spin-off was “another move.”

2. Character developments

• Han Luyi: confident, secretive about his Sight-based advantage, and willing to move beyond technical work into corporate strategy and legal risk. He believes raw data, not cleaned corpora, is the gold mine.
• Zhao Wenyuan: experienced and pragmatic; he doubts Han’s guarantee but accepts the assignment. He is exposed to legal risk because of his options but trusts Han to lead.
• Gu Siyue: meticulous, strategic, and protective of Han. She gives precise legal analysis but also forces him to confront the long-term value of investor trust.

3. Significant themes

• Annotation as understanding: good data is not about surface labels, but reading hidden emotional and situational signals.
• Turning stone into gold: with superior insight, a small team can outperform corporate-scale data pipelines.
• Trust and control in founder–investor relations: legal permissibility is not enough; reputation and trust are fragile assets.
• Hidden conflicts and corporate risk: simultaneous roles, unvested options, board oversight, and full-time commitment clauses can all become weapons.

4. Character interactions

• Han and Zhao: a working dialogue about data strategy; Zhao explains industry reality, Han deflects doubt and assigns responsibilities. Their relationship is based on mutual technical respect but not full transparency.
• Han and Gu: a legal consultation that evolves into a cautionary lesson. Gu reviews contracts, lays out risks with He Yunshen, and ends with the Ding Pay story, implying Han must decide how far he is willing to go.

Chapter Summary

Key Plot Points

• Han Luyi clarifies he won't use Source Code's resources for his new venture; Gu Siyue warns the real risk isn't legal, but breach of trust with the investor who put in 200 million.
• Gu Siyue unexpectedly treats Han Luyi to McDonald's—her first time thanking a client.
• Han Luyi informs Zhao Wenyuan that forming a new company could force him to resign and void his ~60 million in options.
• They decide not to register a company yet, but build a prototype first using Han Luyi's personal cloud account, keeping no legal trail.
• In three days, Han Luyi delivers 8,100 annotated open-source corpus entries to Zhao Wenyuan—all annotated personally by Han Luyi using Sight.
• Zhao Wenyuan discovers the annotation quality is extraordinary (e.g., preserving supposedly "irrelevant" details that reveal true user intent).
• Chapter ends with Sight's long-stagnant experience bar suddenly showing 1% —Lv.3 may not be the cap after all.

Character Developments

• Han Luyi: Chooses transparency over self-interest by telling Zhao Wenyuan the risk he could have hidden. His strategy: results can replace prestige—once the model proves itself, no one will oppose it.
• Zhao Wenyuan: Reaffirms his loyalty, concluding "Following the right person mattered more than doing the right thing."
• Gu Siyue: Shows a rare thaw in her professional armor (blushing ears, McDonald's gesture), hinting at personal feelings.

Significant Themes

• Trust as the core currency: Both in Han Luyi's relationship with his investor and with Zhao Wenyuan.
• Integrity in leadership: Telling someone a truth that could cost you their loyalty, rather than manipulating through silence.
• Results over authority: Using a working model to win support, rather than relying on prestige Han Luyi hasn't yet built.
• Vision vs. obligation: Navigating the tension between a new venture and an existing company's claims.

Character Interactions

• Han Luyi & Gu Siyue: Starts as a legal consult; edges into something warmer through a shared, awkward McDonald's meal.
• Han Luyi & Zhao Wenyuan: A test of loyalty—Han Luyi presents the risk plainly, Zhao Wenyuan stays without hesitation, then immediately asks "When do I get the data?"—pragmatic commitment.
• Zhang Biao (brief): Silently eating at McDonald's nearby, bodyguard detail providing grounding comic relief.

Chapter Summary: “People Are Gullible and Rich—Come Quickly”

Key Plot Points and Events

• Han Luyi wakes at 6:15 and checks his Sight leveling progress. His experience remains at Lv.3, 1%, despite months of no change.
• He recalls how he previously leveled up:
  • From Lv.1 to Lv.2 by exposing Flash’s data-falsification and privacy-selling scheme.
  • From Lv.2 to Lv.3 by helping people solve technical problems for free and developing BugKiller.
• His experience unexpectedly increased slightly after several days of intensive Sight use, but he cannot determine whether the cause was Sight’s annotation function, his ongoing training, or the sheer amount of usage. Sight provides no explanation or leveling guidance.
• Han heads to work for an 8:00 video meeting with Neus AI. Su Niannian has prepared extensive research on the company and its CEO, Ryan Miller.
• The research reveals that Neus is a heavily funded Silicon Valley AI startup valued at $5 billion. Its flagship model, Neus Mercury, is highly regarded for code generation and earns revenue through APIs and subscriptions.
• Han and Su test Mercury by asking it to build a customer-management tool. Mercury generates a complete CRM system with a React frontend, Node.js backend, PostgreSQL database, permissions, exports, and marketing features.
• Using Sight, Han finds only minor bugs and confirms that Mercury’s technical quality is excellent. However, he identifies major usability problems: the system is overly elaborate, difficult for nontechnical users to deploy, and provides far more functionality than the user requested.
• Han concludes that Mercury’s weakness is not code quality but its inability to deliver a usable final product. Neus lacks the product experience and “end-to-end” usability of Source Code Technology’s Open Things.
• He realizes that if Neus could combine its powerful model with Open Things’ ability to understand user intent and produce directly usable results, it could access a vastly larger market.
• Han and Su agree to hear Neus’s proposal without making promises, intending to learn what Neus can offer and negotiate from a stronger position.
• At 7:58, Han joins the scheduled video meeting.

Character Developments

• Han Luyi remains analytical and strategically cautious. He uses Sight not only to inspect code but also to evaluate business opportunities and product design.
• Han shows increasing awareness of his own system’s unexplained mechanics and becomes frustrated by Sight’s lack of guidance.
• His business instincts sharpen as he recognizes that technical superiority alone does not guarantee commercial success.
• Su Niannian demonstrates diligence and preparation by studying Neus late into the night. She also shows sound business judgment, correctly predicting that Neus is interested in Source Code Technology’s product experience.
• Han and Su function as increasingly aligned business partners, sharing assessments and agreeing on a cautious negotiation strategy.

Significant Themes

• Technical ability versus user experience: Mercury can generate excellent code, but its complexity prevents ordinary users from benefiting from it.
• The importance of productization: The chapter emphasizes that successful software must solve users’ problems end-to-end rather than merely provide powerful tools.
• Strategic cooperation and negotiation: Han and Su refuse to be dazzled by Neus’s funding, reputation, or technology and plan to demand clear benefits before cooperating.
• Uncertainty and progression: Han’s unexplained experience gain introduces an ongoing mystery about how Sight develops and what conditions enable him to level up.
• Market opportunity through intent understanding: Han recognizes that software capable of interpreting what users truly want could serve a much broader market than traditional code-generation tools.

Character Interactions

• Han and Su exchange coffee and discuss their preparation for the meeting, establishing a cooperative and comfortable relationship.
• Su shares her research and business analysis with Han, while Han confirms and expands on her conclusions.
• Their playful jokes—such as yawning contagiously and Han’s remark that foreigners are “gullible and rich”—lighten the otherwise strategic discussion.
• Han and Su jointly decide to remain noncommittal during the meeting and extract as much information as possible from Neus before considering an agreement.

Summary: "A Smiling Tiger"

Key Plot Points & Events

• Han Luyi holds a video call with Ryan Miller, CEO of Neus AI. Ryan greets in Chinese, then switches to English.
• Ryan shows a demo of Neus's automated workflow prototype: a user requests a "user permissions system" and their AI—powered by the model Mercury—plans, codes, tests, and fixes the project autonomously in three minutes.
• Han Luyi notes the demo is impressive but purely developer-focused, unlike Open Things, which targets non-programmers from scratch.
• Ryan proposes a partnership: jointly develop an overseas version of Open Things, replacing Source Code's model Heavenly Craft with Neus's general-purpose model as the backend. Neus would bear model costs; Source Code handles productization and localization; revenue split would be negotiated. Ryan claims they could launch within a month.
• Han Luyi perceives Ryan's larger ambition: he wants Open Things to become a vertical application on Neus's future general-purpose platform, not just a code-generation tool.
• Han Luyi asks about data isolation and user behavioral data. Ryan says compliance requires local data processing, but anonymized usage statistics can be shared.
• Near the end, Ryan asks directly: "Is Source Code also going to build models?" He adds that any cooperation contract must include a clause: Source Code cannot use data/technology from the collaboration to develop or train its own large model.
• Han Luyi does not answer immediately. Internally, he recognizes Ryan as a "smiling tiger"—Ryan offers benefits but also prevents Source Code from training models, while Neus can learn from the data and eventually build a competing product.
• The chapter ends with Su Niannian's pen stopping when Ryan asks the model question, showing her tension.

Character Developments

• Han Luyi: Strategic and cautious. He uses his Sight ability (limited here, revealing no emotions), carefully dissects Ryan's proposal, and identifies both the opportunity and the hidden danger. He is secretly developing a general-purpose intent-understanding foundation model—a capability that could generalize Source Code's product-level edge beyond coding scenarios, forming a revolutionary advantage.
• Ryan Miller: Charismatic, polished, and shrewd. He appears warm and complimentary but systematically positions Neus as the indispensable model provider while locking Source Code out of model development. His "push-and-pull" negotiation style and hidden platform ambitions reveal him as a charming but dangerous counterpart.
• Su Niannian: Quietly observant, takes detailed notes, and senses the underlying tension. Her reaction signals her awareness of the stakes and her loyalty/attunement to Han Luyi's thinking.

Significant Themes

• Cooperation vs. Competition: The thin line between strategic partnership and long-term competitive threat.
• Data as leverage and weapon: Data sharing can fuel both sides, but asymmetric restrictions may determine who ultimately wins.
• Product layer vs. model layer: Ryan and Han Luyi represent two paths—model-first versus product-first—and the question is which one will dominate AI-driven software creation.
• The "smiling tiger": Politeness and charm masking competitive aggression; a negotiation where the real threat is hidden inside the most reasonable-sounding clauses.
• Nondisclosure and strategic ambiguity: Han Luyi must decide what to reveal about Source Code's future direction.

Character Interactions

• Han Luyi ↔ Ryan Miller: A polite but intense negotiation. Ryan compliments, demonstrates power, then shifts to pressure and a restrictive condition. Han Luyi listens, probes, and withholds his final answer—reflecting a patient, calculating approach.
• Han Luyi ↔ Su Niannian (implicit): They share an unspoken understanding; she tracks not just Ryan's words but what he omits. Her physical reaction underscores the gravity of the moment.

Chapter Summary

1. Key Plot Points and Events

• Han Luyi gives a carefully crafted pitch to Ryan of Open Things, arguing that Source Code Technology will not pursue general-purpose foundation models.
• He claims Heavenly Craft exists only as a means to support Source Code’s product closed-loop, not as the company’s ultimate goal.
• Han Luyi presents a historical analogy: the 2000 dot-com/ISP bubble. He argues that infrastructure collapses while application-layer companies ultimately capture value (Amazon, Google, Apple vs. Cisco, telecoms).
• Ryan challenges him directly: AI models have flywheel effects unlike ISPs; model capability ceilings will limit Source Code’s expansion.
• Han Luyi concedes Heavenly Craft’s English-language ceiling but insists the real moat is product, user habits, and trust.
• Ryan warns that Google and Meta are already working on similar automated workflows; time is short. He proposes launching an overseas Open Things within a month and co-developing automated workflow products.
• Ryan implicitly assumes Source Code’s next move would be building a general-purpose model—but Han Luyi hides the fact that Zhao Wenyuan is already training a prototype.
• The call ends cordially; Han Luyi says he needs internal alignment before deciding.

2. Character Developments

• Han Luyi: Demonstrates strategic deception and self-control. He deliberately plays the "application-layer only" card to lower Ryan’s guard, while secretly advancing a foundation-model prototype. He shows historical knowledge and business reasoning, but also reveals a willingness to manipulate perceptions.
• Ryan: Sharp, skeptical, and pragmatic. He presses Han Luyi on weak points, reveals market urgency, and clearly prioritizes speed and cooperation value over philosophical alignment. He is not fully convinced but respects Han Luyi’s logic.

3. Significant Themes

• Infrastructure vs. application-layer value: Whether the real long-term value lies in foundational models or in user-facing products.
• Strategic misdirection: Using a plausible, self-consistent narrative to hide true intentions.
• Speed as competitive advantage: The race to launch before Google/Meta, and the pressure of an "urgent" market window.
• Moat and flywheel effects: Product closed loops, user trust, and data versus model capabilities.

4. Character Interactions

• The entire chapter is a tense, high-stakes video-call negotiation.
• Han Luyi presents a unified, aggressive stance; Ryan probes and challenges, but ultimately acknowledges Han Luyi’s reasoning as coherent.
• Ryan reveals confidential competitive intel (Google/Meta moves) as a trust-building and pressure tactic.
• Su Niannian (present in the room) asks whether Han Luyi really believes what he said; Han Luyi reveals it was all an act—confirming the gap between his public stance and private plan.

1. Key Plot Points & Events

• Han Luyi and Su Niannian return to his office to discuss the recent video conference with NeusAI/Ryan.
• Zhao Wenyuan bursts in with stunning results from the first training round of their model, PX-7B-post trained-v0.1.
• Built on a 7B open-source base model, the model beats all major LLMs in intent understanding: score 83.7, versus GPT-4’s 78.2, Gemini’s 76.9, and Mercury 3.1’s 74.8.
• Zhao triple-verified the results across different frameworks and test sets; the margin of error was under 0.3 points.
• Examples show the model consistently grasps unspoken user intent, not just literal meaning.
• Han asks whether 100,000 such high-quality annotations could scale to a larger model. Zhao says theoretically yes—enough to push intent understanding beyond the industry ceiling and improve all downstream tasks, with caveats.
• Han shifts strategy: instead of raising money from investors, he proposes directly partnering with Dingsheng, using their computing power in exchange for equity.
• Su Niannian agrees; Zhao Wenyuan hesitates and asks whether it has to be domestic. Han insists it must, given large-model training constraints. Zhao reluctantly accepts.
• Han then privately messages Cheng Yuan of Dingsheng, proposing a new project and collaboration.

2. Character Developments

• Han Luyi: Decisive and pragmatic; willing to set aside past friction with Dingsheng for strategic advantage. Shows quiet confidence, but also relief and gratitude (“Brother Shi, you came through!”).
• Su Niannian: Sharp, business-focused; emphasizes retaining control in any partnership, and aligns quickly with Han’s plan.
• Zhao Wenyuan: Passionate and rigorous about the results; professionally cautious about claims. Wary of domestic-only constraints, but ultimately defers to Han’s judgment. Shows commitment by offering to take unpaid leave to finish tuning.

3. Significant Themes

• Breakthrough innovation: A small 7B model outperforming far larger models in a foundational capability is unprecedented.
• Intent understanding as a differentiator: Understanding people rather than words is seen as the key to raising AI quality across the board.
• Data quality over quantity: High-quality annotation data may be the real competitive moat.
• Strategic pragmatism: Business partnerships are about interests, not grudges—even with former rivals.
• Sovereignty/domestic constraints: Large-model work must happen domestically, limiting options but forcing focus.

4. Character Interactions

• Han and Su sense relief and excitement from the results together.
• Zhao Wenyuan presents the evidence meticulously, pushing past his initial assumption that the results were a bug.
• Han and Su exchange looks of mutual understanding; Su connects the breakthrough to Open Things’ mission.
• Han directly asks Zhao for a timeline; Zhao commits to a week, with Han approving his unpaid leave independently of Shen Congyun.
• Han explains his Dingsheng partnership idea; Su supports it, Zhao quietly voices hesitation, and Han firmly overrules with “large-model training has to be done domestically!”
• The chapter ends with Han cold-messaging Cheng Yuan—showing Han’s willingness to make the first move when it matters.

Chapter Summary: "Famous Overseas"

Key Plot Points & Events

1. Han Luyi initiates large-model cooperation talks with Cheng Yuan. Han contacts Cheng Yuan of Dingsheng's Strategic Investment Department about large model cooperation. Cheng Yuan responds with subtle disrespect—scheduling the meeting a week out, on his own turf, and implying Source Code may be overreaching ("I didn't expect Source Code to have intentions in this area"). Han Luyi remains unbothered, confident his prototype will speak for itself. Meeting set for next Tuesday at 2 PM.
2. Jiang Yixin's morning routine reveals her growth. In December's cold, she arrives at Source Code's Qiantan Center office. A flashback reveals former classmates mocked her for choosing an obscure startup over Celestial Pivot; after Source Code's Series A and Han Luyi's GAIS speech, those same people asked for internal referrals—none got interviews. She now posts occasional work photos on Moments.
3. Zhao Wenyuan takes an unexplained week of leave. Jiang Yixin encounters him leaving the office at 9:30 AM. She messages Little Niu (model team), who confirms Zhao has requested a week off for "personal matters"—noting he came in person to request it rather than using Lark. His departure follows the successful integration of his "Heavenly Craft" model into Open Things, which dramatically improved metrics.
4. Li Ting delivers the first "Overseas Public Opinion Weekly" report. Key data:
   • GAIS speech video: 1M+ views on X, 1M+ on TickTick
   • A former Google engineer's frame-by-frame breakdown video: 120K+ likes
   • Li Ting's own English tutorial for Open Things (unofficial, zero promotion): 410K views, 2,300 comments
   • Comment analysis shows three waves: shock at functionality, demands for an English version ("I am ready to pay!"—400+ likes, 100+ replies), and overseas users praising it despite using VPNs with heavy latency
5. Competitor discovered: AppFlow. On Product Hunt, a post titled "Why can't Open Things be used overseas?" gained 200+ upvotes. The top reply links to AppFlow—launched last week with the slogan "Create your APP with pure English, no coding, no drag nor drop either." Its features are significantly inferior to Open Things' (simpler features, rougher interface, worse outputs), but it has a native English interface, US servers, low latency, international payment support, and English customer service—and already has paying users.
6. Jiang Yixin storms to report the news. She bursts into Han Luyi's office, but he and Su Niannian are heading to lunch and invite her along—leaving her deflated: "Am I going to be the third wheel again?"

Character Developments

• Han Luyi: Shows strategic confidence and unshakeable self-assurance. He isn't offended by Cheng Yuan's dismissiveness because he knows the prototype will force Cheng Yuan to seek him out. Pragmatic and forward-thinking about large model expansion.
• Cheng Yuan: Displays the classic Dingsheng posture—hierarchical, passive-aggressive, territory-conscious. Deliberately makes Han Luyi wait a week and come to him, implying he sees Source Code as beneath Dingsheng.
• Jiang Yixin: Has visibly grown into her role—more confident, possessing product instincts, fiercely protective of Open Things. Her emotional response to AppFlow ("Isn't this a complete copy of us?") shows deep ownership of the product. Her character arc includes proving her former classmates wrong through her choices.
• Li Ting: Exceptionally capable and industrious—completed a full overseas market report in a week alongside daily operations. Proactive (self-published an English tutorial on her own initiative). Calmly delivers both good and bad news without drama.
• Zhao Wenyuan: Gains humanity in this chapter—revealed to have a life beyond work, taking personal leave. His "janitor monk in a suit" nickname from Jiang Yixin's earlier impression shows her growing respect for him.
• Su Niannian: Mentioned as mentor figure to Jiang Yixin (taught her user story maps) and constant companion to Han Luyi.

Significant Themes

• Overseas market opportunity vs. domestic neglect: The chapter's core tension—Open Things is generating massive overseas demand despite being Chinese-only, high-latency, and requiring VPNs, yet has no international presence. The "Why can't Open Things be used overseas?" post crystallizes both the opportunity and the gap.
• Competition and imitation: AppFlow represents the first real copycat threat. The chapter poses the question: does the first-mover's superior product win, or does the second-mover's localized experience win?
• Confidence in one's work: Han Luyi models this with Cheng Yuan; Li Ting models this with her self-published tutorial; Jiang Yixin models it with her angry possessiveness.
• Growth through struggle: Jiang Yixin's journey from mocked startup employee to workplace pride.
• The demands of speed: The competitive landscape moves fast—in the time Source Code focuses inward, a competitor launches overseas and acquires paying users.

Character Interactions

• Han Luyi ↔ Cheng Yuan: Formal, courteous texts laced with subtext. Cheng Yuan's message is deliberately slighting; Han Luyi's internal response is amused confidence.
• Jiang Yixin ↔ Li Ting: Warm, efficient colleague relationship. Li Ting's terse "Keep reading" and milk-tea-sipping nonchalance create a strong comedic contrast to Jiang Yixin's growing shock.
• Jiang Yixin ↔ Little Niu: Casual messaging about Zhao Wenyuan's leave—shows workplace camaraderie.
• Jiang Yixin ↔ Zhao Wenyuan: Brief nod of greeting; she's observant of his absence.
• Han Luyi, Su Niannian, Jiang Yixin: The closing interaction sets up a comedic "third wheel" moment, showing Han Luyi's casual, friendly leadership style and Jiang Yixin's mortified reaction.

Chapter Summary: "Silly Girl, You're Moving Up"

1. Key Plot Points & Events

• Jiang Yixin is summoned to a work meal with her supervisor Su Niannian and company president Han Luyi at a dim sum restaurant.
• She reports on an overseas threat: someone launched a copycat of "Open Things" on Product Hunt and it's gaining traction. Han and Su are unsurprised—imitators are expected.
• Test #1: Han Luyi asks what they should do. Jiang Yixin initially thinks of legal statements but recalls Han's first lesson (statements are useless). She argues they should ignore copycats, treat them as market validation, and enter the overseas market quickly themselves.
• Test #2: Su Niannian reveals a proposed partnership with Silicon Valley's Neus AI (owned of the "Mercury" model, stronger than their current backend). Neus AI offers free inference costs + revenue sharing, but with a clause: no using partnership data to train Source Code's own model.
• Jiang Yixin sees through the trap: "free things are the most expensive." The clause would lock them in permanently. She proposes a short-term partnership with core data protected and the no-training clause removed; if Neus AI refuses, it confirms malicious intent.
• Han Luyi approves her analysis, assigns her to write a one-page Lark document, and confirms the overseas project—she'll sit in on future coordination meetings.
• Li Ting congratulates her with milk tea: "Silly girl, you're moving up."
• Cut scene: Cheng Yuan (from rival company Dingsheng) submits an internal report on cooperation alignment with Source Code Technology—foreshadowing a larger competitive plot.

2. Character Developments

• Jiang Yixin: Demonstrates real growth from executing tasks to strategic thinking. She passes two on-the-spot leadership tests, showing she can analyze competitive threats and see through deceptive partnership terms. Gains official responsibility on a major overseas project—a meaningful career step up.
• Han Luyi: Continues his mentor role—quiet, testing, approving only based on substance. His "no reaction" approach pushes Jiang Yixin to self-validate.
• Su Niannian: Acts as Jiang Yixin's quiet advocate; her look at Han Luyi ("See? I told you so") implies she has been vouching for Jiang Yixin behind the scenes.
• Li Ting: Warm, senior-colleague figure who recognizes what the opportunity means for Jiang Yixin before she fully grasps it herself.

3. Significant Themes

• "Free things are the most expensive" — True cost in partnerships is strategic dependency, not money. The no-training clause represents a classic vendor lock-in trap.
• Copycats as validation — Competitors copying your product proves the market; the correct response is speed, not legal defense.
• Career mobility through demonstrated judgment — Jiang Yixin's advancement comes from passing tests of thinking, not just execution skill.
• Data as the real battleground — Whoever owns the data and the model owns the future; user experience is the front door, data is the fortress.
• Competitive surveillance — The Dingsheng subplot reveals the company is being watched at a higher strategic level.

4. Character Interactions

• Han Luyi & Jiang Yixin: Tutorial-by-interrogation dynamic. He poses questions, she works through answers, he validates with minimal approval ("Good thinking").
• Su Niannian & Jiang Yixin: Supportive supervisor, defers to Han but privately reinforces her confidence.
• Li Ting & Jiang Yixin: Playful but warm senior/junior rapport; Li Ting clarifies the significance of the opportunity and claims a treat.
• Cheng Yuan (solo): Internal corporate reporting, positioning Dingsheng's stance toward Source Code Technology—ominous B-plot.

Chapter Summary: "A Show Is Coming"

1. Key Plot Points & Events

• Cheng Yuan (VP, Strategic Investment at Prosperity Group) reviews his internal report on Source Code Technology, carefully toning down his negative assessment of the company after Han Luyi rejected his acquisition offer the previous month.
• Han Luyi has proactively reached out, expressing interest in "cooperation in the large-model field"—a surprising reversal that Cheng Yuan distrusts.
• Cheng Yuan submits the report on a Monday lunchtime; within two hours (during lunch break), President Zheng Xiaobo's secretary summons him to the 49th floor—proof that top leadership is intensely focused on Source Code Technology.
• In the meeting, Zheng Xiaobo dismisses Han Luyi's large-model ambition as absurdism ("Build a large model? Does he know the scale of investment that requires?"). He implies Han Luyi is overestimating himself.
• Zheng orders a "show of force": Dr. Liu Dahai (chief scientist, ex-OpenAI) will attend Tuesday's meeting and deliver a technical demo of Prosperity Group's foundation model—to prove the gap and disabuse Han Luyi of his ambitions.
• Strategy is set: large-model cooperation is off the table; acquisition can be negotiated at up to double the initial price (within Cheng Yuan's authority).
• Cheng Yuan coordinates via his secretary for Dr. Liu's demonstration. The chapter ends with Cheng Yuan watching the elevator numbers descend, anticipating "a good show."

2. Character Developments

• Cheng Yuan: Demonstrates deep corporate instincts—he masks resentment professionally, reads organizational signals precisely (e.g., interpreting "keep an eye on them" as a serious directive), and recognizes political dynamics (e.g., Zheng used the All Things Alive failure to sideline Wang Zhiyuan). He is an instrument of Prosperity's will, disciplined and calculating.
• Zheng Xiaobo: Revealed as ruthless, condescending, and methodical. He sees Han Luyi's application-layer products only as harvestable assets ("bring the person in and use him directly"). His plan to publicly humiliate Han Luyi with the technical demo showcases his arrogance and underestimation of his opponent.
• Han Luyi (off-page): Previously established as decisive and confident (cleanly rejecting acquisition). His proactive outreach signals either a strategic gambit or delusion—Prosperity Group assumes the latter. The narrative invites readers to suspect he holds a hidden advantage.

3. Significant Themes

• Overestimation vs. Underestimation: Prosperity Group dismisses a nimble startup's audacious ambition as ignorance; the story implies they're misreading the situation. Both sides are preparing for a power play.
• Corporate Hierarchy & Institutional Power: Detailed worldbuilding of how a big company works—report formatting standards, feedback processes for coffee, meeting rooms named after scientists, the weight of word choice, and authority structures.
• Wealth vs. Innovation: Prosperity's vast resources (hundreds on model teams, billions invested) are contrasted with Source Code's fifty employees and agility. The tension suggests scale isn't everything.
• Strategic Performance: The chapter title "A Show Is Coming" signals the staged demonstration, highlighting how corporate power operates through performance and ritualized dominance.

4. Character Interactions

• Cheng Yuan ↔ Zheng Xiaobo: The core interaction. Zheng questions Cheng to gauge his judgment, then overrides nuance with blunt assessment. He delivers directives (acquisition path, technical demo) and grants Cheng authority within limits. Cheng performs deference while absorbing strategic intelligence.
• Cheng Yuan ↔ Han Luyi (recounted): Cheng's pride was wounded when Han Luyi refused to even read the valuation proposal—"not given him an inch of face"—shaping his current caution.
• Zheng Xiaobo ↔ internal politics (implicit): His referencing the failed All Things Alive project signals a power shift and his intent to redeem it by absorbing successful application-layer talent.
• Cheng Yuan ↔ Secretary: Brief transactional exchange setting up Dr. Liu's involvement, revealing Prosperity's logistical efficiency and the orchestration of the Tuesday "show."

Chapter Summary: "The Person I Like Likes Me Too"

Key Plot Points

• Han Luyi wakes up early and checks his Sight ability; experience is at 7%. He deduces it only grows when Zhao Wenyuan's AI model training progresses.
• Su Niannian invites him to Taikoo Li to pick year-end employee gifts for Dingsheng. He accepts, taking the day off from his draining work pace.
• They browse stores: a home fragrance shop (Han Luyi can't distinguish scents), then an instant camera store where he buys a camera for employees. He casually takes a photo of Su Niannian, which she inspects and approves.
• Su Niannian impulsively buys him a light brown wool scarf and wraps it around his neck herself. When he thanks her, she jokes she'll deduct it from his salary.
• At a Japanese restaurant for lunch, Su Niannian makes an extended metaphor about detective fiction—the murderer's identity is hidden from the start, and observant people already know but "don't dare confirm" because the answer seems "too straightforward."
• Earlier, at the mall's Christmas tree wish display, Han Luyi wrote: "May Source Code go smoothly this year and next year too." Su Niannian hid hers, but as they left, he glimpsed her neat handwriting: "The person I like likes me too."
• At the table, she asks whether such wishes can come true. He deflects, saying "sincerity makes wishes come true" and redirects to Source Code.
• The chapter ends with Han Luyi noticing his Sight experience has silently ticked up to 8%.

Character Developments

• Han Luyi: Shows subtle growth in emotional awareness—he notices Su Niannian's hints but repeatedly ducks them, retreating into work-related deflections. His internal conflict between reciprocating and avoiding is palpable. Also notably affected by the scarf gift and her words.
• Su Niannian: Boldly and progressively pushes the romantic subtext—gifting him a scarf, framing their situation through the mystery novel metaphor, and writing a candid wish. She is clearly testing whether Han Luyi will "reason toward the obvious answer."
• Sight: Experiencing slow growth (1% → 7% → 8%), directly tied to external model training rather than his own usage.

Significant Themes

• Unspoken vs. Spoken Feelings: The core tension—Su Niannian practically confesses through metaphor and action while Han Luyi feigns obliviousness.
• The Hidden Answer in Plain Sight: The detective novel analogy is the chapter's thesis; Su Niannian describes exactly their situation—he already knows, but can't confirm because it's "too simple."
• Work vs. Personal Life: Time off is framed as necessary; Han Luyi's wish prioritizing the company mirrors his habit of burying himself in work to avoid emotional vulnerability.
• Reciprocity and Timing: Both characters "like the other who likes them," but neither will be the first to say it directly.

Character Interactions

• Han Luyi & Su Niannian: Warm, comfortable, increasingly intimate. The shopping trip is full of small domestic moments: his hands in his pockets following her like "a large dog led through the park," her adjusting his scarf, her teasing tone. She leads the conversation toward emotional honesty while he deflects with humor and work talk.
• The mystery novel exchange carries the heaviest subtext—her words are essentially a coded confession, and his "the author has the final say" is a coded refusal to take the leap.
• The wish-card scene silently exposes her true feelings to him while he pretends not to have seen.