Personality Profile
The two game engines in this story:
-
Lin Zhe, Male, 26, INTJ, algorithm engineer at Hangzhou Cloudinnov Technology, desk in zone 202, seventh row by the window
- Dominant functions: Ni-Te, excels at deconstructing all behaviors with logical frameworks, pursues system predictability. His resume’s top achievement: after optimizing the recommendation system, the CTO praised his code’s reusability at an all-hands meeting – but didn’t say his name, so no one knew it was him.
- Current goal: Verify whether Su Tang’s ambiguous signals are genuine interest or social gaming. His data sources: Slack message frequency, eye contact duration, WeChat reply intervals (precise to seconds), and whether she deliberately uses industry terms he doesn’t understand when starting a conversation (like “brand personification” – a phrase he wants to curse about).
- Current dilemma: He can’t predict her behavior with data models, challenging his basic assumptions about relationships. Last weekend he tried modeling her reply patterns with a Markov chain, only to conclude it’s a “chaos system” – he named the model file “useless_shit.py” and didn’t delete it because he was too lazy.
-
Su Tang, Female, 25, ENTP, marketing strategist at the same company, desk at the end of the hallway in zone 202. On her desk there’s always an empty Starbucks “Express” cup with her own doodles (last week it was a cat sticking out its tongue).
- Dominant functions: Ne-Ti, loves exploring relationships through language games and role-play, treats every interaction as an intellectual playground. Her previous job was at an acquired ad agency; when she quit, her boss said, “You’re a great brainstorming partner, but the execution end needs someone to weld the interface for you” – she took that as a compliment.
- Current goal: Test whether Lin Zhe has the intellectual depth she needs – ENTPs need an opponent who can keep up with their mental leaps, not a compliant pursuer. She’s already observed that when she threw out the idea of an “AI assistant in snark mode,” his first reaction wasn’t rejection but a feasible improvement plan – that’s number one on her screening checklist.
- Current dilemma: She enjoys triggering Lin Zhe’s genuine reactions, but isn’t sure she really wants a traditional relationship. Her last boyfriend was an ISFJ; they broke up because he couldn’t handle her late-night calls saying, “I think we could open a convenience store in Iceland and use the midnight sun for 24-hour coffee” – and Su Tang actually checked Reykjavík commercial rents on Fliggy.
Story
Act 1: The Trigger – That “Test” in the Meeting Room
Wednesday, 10 a.m., project review meeting. The projector was a three-year-old Epson, stuttering for three seconds on every slide change. The HRBP muttered “time to replace the bulb” every meeting, but nobody ever filed a repair request.
When the marketing team proposed the new quarter’s AI product user interaction plan, Su Tang stood up from her seat, her tone carrying the typical ENTP playfulness: “I have a crazy idea – what if we set the AI assistant to snark mode? Would users find it more fun?”
Sporadic chuckles in the room. The director was about to speak – he was holding a ballpoint pen gifted by a supplier, with the logo on the cap, ink skipping – when Lin Zhe’s voice cut in first: “Data says no.”
He opened his notebook, his tone calm, as if stating a known fact: “In Q2 2024 we did an internal A/B test. Negative interactions reduced retention by 27% after the first three uses. The novelty of snark mode lasts at most two conversation rounds; after that, user frustration outweighs entertainment – especially when the user curses at the AI and the AI replies, ‘Is your keyboard missing the enter key?’ The direct uninstall rate was 9%.”
Su Tang’s eyes lit up.
“But we could make it a switchable mode,” Lin Zhe continued. “When users actively choose snark mode, their expectations are set, and the retention drop curve flattens. Your intuition is right, but the implementation needs adjustment – I suggest making snark mode a ‘hidden achievement entry,’ requiring users to complete a specific action to unlock it, further filtering out non-core users.”
After the meeting, Lin Zhe had just returned to his desk when WeChat lit up.
Su Tang: “I love watching you seriously refute me.”
His heart skipped a beat – not because of the words themselves, but because of the sticker attached: a Shiba Inu peeking from behind a corner, captioned “sneaking a peek.jpg.” That sticker came from a photo she took at a cat café last weekend.
This was none of the conversation patterns he knew. In a work setting, affirmation usually came as “good proposal” or “valuable data,” not “I love watching you do things.” Ni went into overdrive: is this expressing interest? Or just playing an office flirtation game? Or just a casual remark? – He noticed the message timestamp was 10:37, exactly two minutes after the meeting ended, usually when she was on her way to the restroom.
He stared at the green bubble on the screen, typed three replies, and deleted them all.
First was “Thanks,” deleted – too businesslike. Second was “Yeah, your idea was also inspiring,” deleted – too formal. Third was “When are you free to discuss the mode switching details?” – but after typing it, he felt he was actively creating contact opportunities, too transparent.
Finally, he replied with just one word: “OK.”
But he knew something was starting to slip out of control – because he found himself saving that Shiba Inu sticker to his favorites.
Act 2: Escalating Conflict – Sticky Notes, Clinking Glasses, and Logic Failure
Lin Zhe decided to use his strongest approach: systematic observation and model building. He created a folder in iCloud Notes named “S_behavior_analysis,” containing four subdirectories: timeline, frequency stats, semantics analysis, and anomaly records.
He began recording Su Tang’s behavior patterns:
- Monday: At the water cooler, she deliberately stood next to him while making coffee, shoulder distance less than 15 cm. He noticed she used her own white mug instead of a disposable cup – she uses disposable cups 82% of the time, and her mug only when she’s in a good mood.
- Tuesday: She left a sticky note on his desk with just one line: “Guess what I just did.” He flipped the note over and saw it was cut from printer paper, edges uneven, obviously cut with scissors – if it were something serious, she would have used a formal notepad. This detail plunged him into deeper confusion.
- Wednesday: Team dinner at a Hunan restaurant downstairs. She clinked glasses with a male colleague – he said, “Su Jie, bottoms up,” but she only touched her glass and moved away – then she walked around half the table specifically to come to him. When their glass rims touched, she murmured, “This one is the real deal.” He noticed her beer was just poured, foam still bubbly, and after saying that, she didn’t drink, just held the glass looking at him, waiting for him to clink too.
Lin Zhe’s Te started failing in the face of this data.
He could model user behavior, predict recommendation system click-through rates – just last week he built a Transformer model for next-click prediction with 91.3% accuracy – but he couldn’t deconstruct with logic what those “six words on a sticky note” meant: is this a valid communication? Or just a stimulus to see his reaction? He tried sentiment analysis with BERT, but the message was too short; the output was “neutral (confidence 0.54).”
Ni started producing a byproduct: emotions.
One night at 11 p.m., lying in bed, he realized he’d been staring at a crack in the ceiling for twenty minutes – a crack left by last year’s typhoon leak, reported to maintenance three times and still unfixed. His mind looped: Su Tang’s voice pitch today was half a tone higher than usual – excitement or nervousness? Her blink rate when talking to him was higher than with other colleagues (he recorded ten seconds on his phone and slow-motion counted several times) – does that mean she cares more about his feedback? But the next second he remembered he’s an algorithm engineer doing this, and cursed himself, “You’re such an idiot.”
The greatest fear of an INTJ isn’t rejection – it’s the system outputting “unpredictable” – like writing perfect code that compiles with an error: “Invalid data type: human.”
He decided to take the most direct approach.
Act 3: Emotional Low Point – “You Make Me Find It Interesting”
Thursday afternoon, Lin Zhe waited by her desk for ten minutes. During that time, a product manager came to borrow her stapler. She wasn’t there. Lin Zhe scowled and said, “Come back later.” The product manager, an ESFJ, walked away confused.
Su Tang returned from the restroom and saw him standing by her seat. Her eyebrows lifted slightly – that ENTP curiosity mixed with an “you finally came” awareness. Her phone was in her hand, screen still on, showing a popular science video about quantum entanglement she hadn’t finished watching (at 0.75x speed).
“Su Tang, I need to say something.” His tone was professional, like requesting a code review: voice deliberately lowered, speech speed increased, avoiding any emotional vocabulary.
“Mm-hmm?”
“Your recent behavior has caused some… data deviation in me,” he said. “I need to verify – do you have feelings for me or not?”
Su Tang was silent for three seconds.
In those three seconds, Lin Zhe felt his logic system crash through three versions internally. He saw a flicker of complex light in her eyes – not embarrassment, not shyness, but genuine thought. He noticed she unconsciously twisted a strand of hair around her finger, then released it.
“I’m not sure yet,” she said, her tone surprisingly serious – no rising intonation, no playful pause, just a declarative statement. “But your reaction makes me find it interesting.”
Interesting.
Lin Zhe’s brain stopped on that word.
In the INTJ value system, “interesting” is a neutral word, an evaluation metric for a project, a litmus test for liking – but it’s not an answer. He felt doused with cold water: he could accept rejection, but not being treated as a variable.
He lowered his eyes, his voice colder than intended: “So I’m on your experiment list.”
Su Tang’s expression changed – her mouth slightly open, as if about to say something, but nothing came out. She lowered her lashes, then turned her head, staring at the cold coffee on her desk.
Lin Zhe turned and walked away.
For the next week, he executed a perfect avoidance strategy: no longer sending messages proactively, not going to the water cooler at the same time as her (he marked her usual water break times on the calendar to avoid encounters), and during meetings he spoke but avoided eye contact. He used Te to draw a clear boundary, then stood inside it, telling himself this was “rational loss prevention.”
But he had insomnia for four nights. On the third night, he pulled up that Shiba Inu sticker and stared at it five times, then threw his phone to the foot of the bed, cursing, “This is so messed up.”
Act 4: Key Reversal – The 3 a.m. Breakdown
Saturday, 10 p.m., Lin Zhe went back to the office to pick up his forgotten laptop – a ThinkPad X1 Carbon he bought on Singles’ Day last year, battery already fading, a scratch on the screen from a subway rush last week.
The entire floor had only one light on – the IKEA desk lamp at Su Tang’s station, warm bulb, contrasting sharply with the cold white LEDs in the hallway.
Su Tang sat at her desk, a stack of proposal drafts covered in red pen marks spread out before her. Her eyes were red, with traces of tears still on her lashes – Lin Zhe noticed she wasn’t wearing makeup today, a few pimples from staying up late, her hair thrown into a messy ponytail instead of its usual polished style.
Lin Zhe’s steps halted.
This was the first time he saw Su Tang without her mask – no playful smile, no testing words, just a 25-year-old girl broken down because her proposal was rejected. Her coffee had gone cold, a brown ring stained the coaster.
“Proposal got rejected?” he asked.
Su Tang looked up, saw it was him, and her first reaction was to quickly wipe her eyes – with the back of her hand, not a tissue – then her voice came out a bit hoarse: “What are you doing here?”
“Picking up my laptop.”
Then ten seconds of silence.
Su Tang lowered her head, her voice for the first time devoid of that ENTP frivolity: “My boss said my market positioning was completely off and told me to redo it over the weekend. But I’ve already revised it three times – the first version was criticized as ‘too vague,’ the second as ‘too conservative,’ and the third he said, ‘Are you messing with me?’”
Lin Zhe stood there, his body reacting before logic – he pulled out the chair next to her and sat down, the chair legs screeching against the floor. He picked up the stack of proposal drafts and flipped through them page by page.
Reading speed extremely fast, the frown lines between his brows deepening with each page. Fifteen minutes later, he looked up: “Your core problem isn’t the proposal itself, it’s the narrative direction – the ‘Project Background’ on the first page takes up half a page of fluff, the actual user persona data is buried on page six, and the boss doesn’t have the patience to scroll that far.”
Su Tang froze.
“Your analysis data is good – though the competitive comparison section uses outdated sources, a 2023 report – but it’s placed wrong in the PPT.” Lin Zhe pulled her keyboard closer, quickly opened her folder, and typed as he spoke: “The first block should be a visual analysis of user pain points – I can use your user interview recordings transcribed into text to make a word frequency heatmap – not competitive comparison. Because the boss’s focus is ‘can this fly,’ not ‘has anyone else done it.’ We can rebuild the logic chain from three angles: first, borrow from your snark mode idea and add a fun data module – like ‘Top 10 Things Users Want to Curse at the AI’; second, use scenario immersion to increase persuasion – open with a typical late-night user need story; third…”
As he talked, he didn’t notice Su Tang watching him.
He didn’t notice the change in her eyes.
He didn’t notice her going from crying to a silent laugh – a light laugh, tears still hanging at the corners of her eyes, but the corners of her mouth turning up.
Three hours later, they had created a brand new proposal framework. Lin Zhe’s laptop screen was full of open tabs, some with scripts he had just written, one with Python code to generate a word frequency chart (because the design department wouldn’t work weekends). Su Tang leaned back in her chair, let out a long breath, and turned to look at him: “You know what?”
“What.”
“When I was emotionally breaking down, your first reaction was to give me a solution, not comfort,” she said, her tone without teasing, only a serious wonder. “This confirms to me – you’re the person I most want to keep talking to.”
Lin Zhe’s fingers stopped on the keyboard.
For the first time, he felt his logic system stop running.
His mind went blank.
And at the center of that blankness, only one phrase looped: She meant it. Not an experiment, not a test – an ENTP expressing in her most honest way: she’s interested in you.
Act 5: New Equilibrium – “Let’s Do Something Real”
Monday, Su Tang pushed open the partition at Lin Zhe’s desk.
“I want to start a side project with you.”
Lin Zhe looked up and saw her handing him a handwritten note of ideas – scribbled, logically jumpy, all Ne-style flights of fancy, but next to each idea there was an empty column labeled: “Waiting for Lin Zhe to add implementation direction.” The notebook was a Muji recycled paper one, and on the cover she had drawn a cat wearing glasses.
He flipped through two pages, then looked up: “Is this serious, or an experiment?”
“Serious,” she said, then paused. “And I have a positive feeling for you – not ‘like you,’ but a good feeling – meaning I want to stay on the same frequency with you a bit longer.”
This was the first time Lin Zhe heard an ENTP express emotions with a declarative sentence instead of a rhetorical question.
He felt a certain string at the back of his neck loosen.
“Okay,” he said, lowering his head to continue flipping through the proposal. “If we’re doing this, your ‘user persona’ section needs a data analysis interface – we can adapt the user behavior clustering API I wrote last week.”
Su Tang smiled, pulled her chair a little closer – the chair legs screeched again, but this time Lin Zhe didn’t find it grating.
From then on, a new communication protocol emerged between them: Su Tang no longer used ambiguous signals to test, but directly said “I want to do this with you” or “I need your input”; Lin Zhe learned to ask directly when uncertain: “Is this suggestion serious or an experiment?” – he even wrote a Slack bot for himself that auto-replied “Please clarify intent” whenever a message contained words like “guess.”
They started meeting once a week late at night in the office to work on the project face-to-face, with occasional silences, occasional arguments, occasional – Lin Zhe found himself unconsciously curling the corners of his mouth when she laughed, though he didn’t realize it.
For NT types, “I love you” is a proposition that needs verification. But “I want to solve this problem with you” is an already-proven answer – because the cost of trust is higher than that of liking.
Psychological Analysis
What is Su Tang’s “flirting”?
In the story, every “ambiguous signal” from Su Tang is essentially ENTP’s Ne exploring boundaries: she creates cognitive dissonance to trigger Lin Zhe’s Ni-Te loop, quickly assessing whether he has the intellectual depth she needs. The horoscope screenshot is the opening door, the “snark mode” in the meeting is a target setter, “you guess” is a way to detour direct answers – each step is a “test” designed to gather more data. According to feedback from multiple ENTP users in MBTI communities, this kind of “tentative flirtation” is standard operating procedure for ENTPs before establishing a relationship — not malicious, but a screening mechanism akin to an “interview.”
This is not love; this is screening.
And when she cried about her rejected proposal, ENTP’s tertiary function Fe (Extraverted Feeling) was in a state of breakdown — she didn’t need emotional comfort; she needed “the problem to be solved.” Lin Zhe’s instinctive offering of a solution instead of a hug perfectly hit ENTP’s value point: problem orientation trumps emotional resonance. In real-life product reviews, this kind of “rational support” is more effective than “I understand you” in moving ENTPs of any gender — provided the solution itself is robust enough.
What is Lin Zhe’s “coldness”?
Lin Zhe’s Ni makes him overinterpret every word and action of Su Tang, trying to construct a unified behavior model — this is standard INTJ behavior around strangers. But when the model fails, his Te chooses to “cut the connection” (avoidance strategy) rather than “modify the model” (re-evaluation). In the daily management of startup CTOs, this “either-or” decision-making pattern is often complained about as “binary thinking” by subordinates — but for INTJs, it’s self-protection to maintain cognitive system stability.
What truly shifts him is not Su Tang’s confession, but when she shows vulnerability, he “without thinking gives a Ne-style solution” — this shows that his attention to her has already exceeded the control of his logic system, entering the realm of “subconscious priority.” This kind of conditioned reflex of caring is the most typical signal that an INTJ has developed emotional attachment to someone (though they won’t admit it).
Root of Conflict
INTJ believes “ambiguity is for confirming the relationship” (efficiency-oriented), while ENTP believes “ambiguity itself is the process of building the relationship” (process-oriented). Lin Zhe’s “seeking results” and Su Tang’s “enjoying the process” are out of sync in timing, causing both to find the other “unreasonable.” According to a project director who once worked cross-departmentally at a major company in Shenzhen, this rhythm mismatch is extremely common in NT-type relationships — often requiring one party to proactively define a “communication protocol” to avoid internal friction.
The solution is not for one side to compromise, but to establish a shared “experimental mode” — turning ambiguity into a two-person puzzle-solving project, something NT types have in common. Just like at the end of the story, Su Tang uses “project” as the relationship framework, Lin Zhe uses “API interface” as a way to participate, and the two NTs finally find the same channel.
FAQ
How to tell if an ENTP girl genuinely likes me or is just experimenting?
Look at her reaction after you give a solution: if she absorbs your point and challenges you further (e.g., asks “what if the user persona data is insufficient?”), she’s seeking an intellectual partner; if she’s just satisfied with your reaction itself (e.g., laughs happily and tells others, “look how serious he is, his speaking manner is so cute”), it’s likely an experiment. The former is the starting point of trust, the latter is entertainment material.
Should an INTJ directly call out an ENTP’s testing?
Yes, but approach it as “I’ve noticed a pattern,” not accusation. For example: “I find it hard to judge your real intent when you speak in an uncertain tone. Next time, could you just say ‘I need your input on this idea’?” This respects her logic while setting boundaries — in real product reviews, INTJs using this “error attribution + behavior suggestion” speaking style effectively reduces ENTP defensiveness.
ENTPs often find INTJs boring; how to maintain attraction?
INTJs don’t need to become interesting; they need to become “solvable.” Occasionally toss out your own unfinished logical loopholes and invite the ENTP to fill them. The ENTP’s Ne will be drawn to the “opportunity to solve,” which is a hundred times more effective than telling jokes. For example, Lin Zhe’s “snark mode improvement plan” was the perfect completion of Su Tang’s thrown “loophole” — this complementary intellectual interaction is far more addictive to ENTPs than any ambiguous signal.
If an ENTP girl is hot and cold, how should an INTJ handle it?
Accept that her Ne needs to explore multiple threads. You can proactively set fixed “focus times” (e.g., one deep discussion per week), and keep normal interactions the rest of the time. She will value these conversations more because of their “scarcity,” and your Te will gain a sense of control. A real case: an INTJ programmer and his ENTP girlfriend agreed on Wednesday nights as “project time” and left each other alone otherwise. A year later, their side project was live and they secured angel investment — and their relationship stabilized.
In the story, Su Tang says “I’m not sure yet.” Is she really uncertain or just trying not to hurt Lin Zhe?
For ENTPs, this is a genuine logical judgment: she needs more data. If she said “yes,” it would be more likely a lie — because the ENTP’s deep fear is making a wrong decision that can’t be reversed. An INTJ could respond: “I respect your pace. When you’ve gathered enough information, please tell me your conclusion.” This makes the ENTP feel trusted while subtly hinting “your testing time is limited, don’t drag it out” — this kind of bounded trust is the best gift an INTJ can give.
Related Content
- INTP Male Hesitates Four Minutes, ESFJ Female Waits 127 Seconds: She Sent a Friend Request When an INTP homebody receives a friend request from an ESFJ girl, his rationality vs her expectations — what sparks fly? See how two opposite personalities test each other through a simple friend request.
- Subway Sketch: How an ENFP Male Melts an ISTJ Female’s Heart with One Pen Subway sketch: How an ENFP male melts an ISTJ female’s heart with one pen.
- ESTJ Female Interviewer Adds Me on WeChat? INFP Male’s Interview Twist After a failed interview, ESTJ interviewer Lin Yue proactively adds the INFP male candidate on WeChat. A drama of personality clash, from cold evaluation to a late-night friend request, the story unfolds.
Author: Shu Yan
Updated: June 2026