One sentence: This is not a chicken-soup values playbook, but an operating system for "how to build products"operating system for "how to build products" that Duolingo refined over 14 years of trial and error. Its core can be compressed into one line —find great people → let them test quickly → judge with real data → scale what works, decisively stop what doesn't; when all tradeoffs conflict, use the "mission" (to make the highest quality education universally accessible to the world) as the sole arbiter.
Between the five principles there is intentional tension(Luis explicitly states in his letter: they contradict each other internally, and may not apply in some scenarios). For example, "Raise the Bar" demands careful, meticulous work, while "Ship It" demands fast release — these two clash every day. The brilliance of the playbook is not that it resolves the contradiction, but that it provides a measuring stick (mission + retention first). For PMs building gamified / AI interactive narrative products, Duolingo is the only living example that has stitched "education" and "addictive entertainment" together and built a ten-billion-US-dollar listed company; how it walks the tightrope between "useful" and "fun" is almost exactly the problem you face every day.
Five principles + one implementation engine:Take the Long View (retention > growth > revenue) → Raise the Bar (build V1 not MVP) → Ship It (weekly releases + hundreds of experiments) → Show Don't Tell (numbers are the story) → Make It Fun (compete with TikTok for attention) → the 6-step Green Machine strings all of the above into an executable process.
0. Foundation: Mission is the sole arbiter, principles are "alive" and will clash with each other
EN: Our mission is to develop the best education in the world and make it universally available. The five principles have internal tension and won't always apply.
Core: Duolingo's mission has two dimensions —quality (best) + universal access (universally available). This is not a slogan on the wall, but a knife actually used to slice priorities. When "Raise the Bar (be the best)" clashes with "Ship It (cover more people fast)", ask one question:Which choice lets more people learn things faster? The answer becomes clear.
Why it matters: Most companies' values are slogans packaged after the fact; Duolingo does the opposite — principles are"living ideas" summarized from mistakes, and Luis directly admits in his CEO letter that "there is tension between them, they may not apply in some situations, and I hope you question them". This means: don't recite any of them as dogma, understand under what scenarios one yields to another.
Transfer for you (gamified AI product PM): First define for yourself a mission that can act as an arbiter(not fluff like "build the best AI interactive narrative", but something like "let users be willing to come back every day to experience a story with emotional tension" — the kind that can directly judge wins and losses). From then on, use it to cut through all debates about "should we build this feature / is the launch quality good enough". Without this knife, the team will endlessly burn energy between "polish it" and "hit the deadline".
1. Take the Long View: Retention > Growth > Revenue
EN: If it helps in the short-term, but hurts in the long-term, it's not right. ("take the long view" = prioritize long-term impact over short-term gains when making decisions)
Core: Duolingo has treated this as something "to build for decades" from day one, and all decisions pass through the filter of"beneficial short-term but harmful long-term = wrong". The sharpest priority ranking is:Retention > Growth > Revenue.
Why this ranking: The value of language learning (and any product requiring long-term commitment) is delivered when users "stick with it long enough"— the longer users use it, the more value they get, and the product only truly works. So even if it doesn't make money in the short term, if the product can't deliver on its value promise, users will eventually churn. Retention is the foundation, growth is building on the foundation, revenue is collecting rent on the building; if the foundation collapses, everything behind it is empty.
Real implementation at Duolingo:
- Notifications = textbook case of long-term vs short-term: Sending more notifications immediately boosts short-term DAU; but too much harassment means users won't stay long-term.User trust > short-term activity, so they exercise restraint in sending notifications.
- Ads also enforce a "retention threshold": Testing shows "show 1 ad after each lesson" does not hurt retention,any more and churn starts— so ad volume is nailed to this threshold, never crossing the line for a little extra revenue.
- Betting on Technology: They used machine-generated speech from the start instead of human recordings. It sounded terrible early on, but they bet "technology will get better sooner or later". Long-termism is not just "avoiding stupid things", but also proactively choosing a path that is ugly now but strong in the future.
- Linear Path (2022): Replaced the original flexible "skill tree" with a linear path that must be completed in order. Short-term DAU and revenue did not improve, and users may have disliked it — but they believed "actually learning" was more important than "the satisfaction of grinding progress", and pushed it live. Hindsight proved the Path became a key to success.Long-termism sometimes means making bold decisions that seem counterintuitive.
- Mascot Duo = competitive moat: The character is not decoration, it is a strategic asset. Even if someone copies the entire product exactly, users will still come back because of their emotional connection to Duo. That is retention no one else can copy.
- Hundred-year brand + no focus on quarterly earnings: If quarterly revenue misses expectations the stock price drops, but they "do not overfocus on quarterly performance", and only watch long-term user value. Do the right things, and company value will rise naturally.
Transfer for you:
- Rank your own product with a priority order like "retention > growth > revenue", and actually use it to reject short-term temptations. The biggest short-term temptations for gamified AI products are "add gacha / add paywalls / add push" — every single one should first pass the test of "will this damage long-term retention and trust".
- Treat "characters / worldbuilding" as a moat to operate, not as art assets. The core barrier of AI interactive narrative is precisely "users' emotional connection to characters" — that is something competitors cannot copy even with stronger models. It deserves outsized investment.
- Dare to launch major overhauls where short-term metrics don't rise but the direction is more correct(like Path). The prerequisite is you have a data baseline + long-term conviction, otherwise it easily becomes "self-indulgent redesign".
⚠️ Counterexample (don't learn this wrong): Decisions like Path where "you launch even if short-term metrics don't rise"carry extremely high risk; Duolingo dares to do this because it is backed by a massive experimentation system + founder-level long-term conviction. If small teams copy "counterintuitive bold decisions" without a data baseline, they will easily mistake self-indulgence for long-termism.Long-termism is not a get-out-of-jail-free card for "stubbornly pushing ahead on gut feel".
2. Raise the Bar: Excellence is the baseline, not the goal
EN: Excellence isn't some lofty goal—it's our baseline. Excellence does not get easier. But we did not sign up for easy.
Core: Demote "excellence" from "a lofty goal worth pursuing" to"the daily passing grade". It is not about pursuing perfection, but about improving a little bit every day; when mistakes happen, don't investigate "whose fault it is", only dig for root causes.
2.1 Four standards for product and design (the most transferable section)
Every new feature must pass four gates before launch,none are optional:
| Standard EN/CN | Meaning | Red flag (indicates standard not met) |
|---|---|---|
| Useful | Learners get real value | Only adds app complexity, distracts attention |
| Intuitive | Users can focus on learning,no need to figure out how to use it | "If a feature still needs explanation, that means the design is flawed"— a 75-year-old Android user in India and a 16-year-old iPhone user in New York must both understand it at a glance |
| Delightful | Every feature carries a bit of fun and "magic feeling" | The first version can skip the most complex animations, but cannot lose the core "magic" that users love |
| Polished | Looks complete, smooth, and consistent | Having both a "back" and a "close" button that do the exact same thing |
Why: These four are ordered from "foundation → upper layers" — first "Useful" (otherwise it's wasted work), then "Intuitive" (otherwise users can't learn to use it), then "Delightful" (otherwise they won't stay), finally "Polished" (otherwise it looks incomplete and untrustworthy). It breaks down the vague "make it a bit better" into a checklist you can tick off item by item.
2.2 Build V1, not MVP (key difference)
EN: V1s Not MVPs.
- MVP (Minimum Viable Product): Often low quality, "sometimes even becomes an excuse for building bad products".
- V1 (First version of the official product):: Can have fewer features, but must be polished to meet the bar. Slower, but never hand half-finished work to users.
Why: MVP culture easily lets teams excuse shoddiness with "it's just an MVP". V1 nails the acceptance bar in advance —fewer features is fine, lowering the quality of individual features is not.
2.3 Other implementation mechanisms
- Dogfooding (using your own product): Everyone uses the app daily, and they built a"shake to report" tool— shake your phone to take a screenshot + file a bug, making everyone responsible for quality.
- Be Hard on the Work, Easy on People: Give honest, helpful feedback, but only target "what the problem is (what)", not "who is responsible (who)". This preserves both high standards and trust.
- Ownership (clear individual accountability): The most important projects must have a clear owner. "Only things that someone owns get done well." As companies grow, a task spanning 4 departments and 7 teams with unclear accountability is a quality killer.
- No lowering hiring standards: Better to leave a role empty than hire the wrong person. They once rejected an executive candidate they had spent over a year searching for because the candidate "was disrespectful to the pickup driver" —excellence is not just about strong ability, it's also about how you treat others.
Transfer for you:
- Copy that four-standard table directly as your feature acceptance checklist. For AI interactive narrative products in particular, you must keep a close eye on "Intuitive" — AI features are inherently complex; if a feature requires an instruction manual to teach users how to use it, the design has failed.
- Ship with a V1 mindset, not an MVP mindset: Instead of shipping 5 half-baked features, ship 1 feature polished to the point of being "delightful + refined". The "magic feeling" of gamified products dies precisely from the cheapness of half-finished work.
- Assign a real, single owner to every key feature, don't let "everyone is responsible" turn into "no one is responsible".
- Build a low-friction internal feedback channel(similar to a shake-to-report feature) — the team uses their own product every day, and they are the fastest radar for spotting what is "not delightful / not refined".
3. Ship It: Speed itself is compound interest
EN: Moving with urgency means choosing what matters and pursuing it with a ridiculous amount of focus. Go, Go, Go!
Core: To turn good ideas into reality you must increase speed. Duolingo They update iOS / Android every week, while running hundreds of experiments at the same time. Speed is not for rushing; it is for creating more opportunities to "learn and improve"— every experiment, whether it succeeds or fails, brings you closer to the right answer, and these small improvements compound and stack like a snowball.
3.1 Clock Speed (borrowed from chip terminology)
EN: Clock Speed = shortening the time between every step (from decision to execution, from receiving feedback to finishing the fix).
Why: The shorter the interval, the faster you ship, the faster you learn. "You shouldn't have to wait a whole month just to see a new version."Note: This is not about blindly rushing progress; it is about eliminating wasted waiting time caused by "process interruptions".
3.2 Ruthless Prioritization
For every project, ask:How does this help the goal? What concrete result will it deliver? If it is useless, cut it,no matter how much effort has already been spent on it(fighting sunk cost fallacy)."One high-impact result beats many scattered small results." Abandoning what doesn't work is just as important as creating new things.
3.3 Experiment culture + 99 bad ideas
EN: 99 Bad Ideas.
- There is a tradition among executives called"99 bad ideas": everyone deliberately comes up with all kinds of exaggerated, absurd ideas (make Duo do silly things, change the stock ticker, can we chat with characters, what to do with 5 seconds of Super Bowl airtime, can exams be taken at home).
- Many hit features came from these "absurd questions".Shutting down an experiment is not failure; it is a learning opportunity. Only about half of the experiments they run actually ship to users.
3.4 Use "mission" to resolve the contradiction between "good and fast" (the most subtle point in this playbook)
"Raise the bar (slow, careful work for quality)" and "Ship fast (get it out now)" are inherently in conflict.The referee is the mission: Does this make education better and more accessible?
The real choice for generative AI: A = wait 6 months to bring AI content close to human-level quality; B = ship now at 70% quality but cover more languages and reach more people. Using the mission (let more people learn faster) to judge everything,clearly points to B.
How does this not contradict "polish it well before shipping"? Think of it as two stages:First ensure "nothing below the bar can ship" (hold the minimum passing line); above the "qualified" level, then decide what to do first (trade off between speed and better quality).
3.5 Appropriate process: Use PR (Product Review) to quantify decisions
As a company grows, "making decisions in casual meetings" causes chaos (some people are out of the loop, no one can tell a formal decision apart from an offhand comment, reasons cannot be traced back). They borrowed from Code Review to build Product Review (PR):
- All changes are "seen" → solves information opacity
- Clear decision points → distinguishes formal decisions from casual thoughts
- Records exist for traceability → makes it possible to understand why decisions were made
- Multiple people participate in rotation → introduces different perspectives, avoids single points of failure
- Structured process → meetings are no longer chaotic
Anyone can sit in on PR. The marketing team built a Marketing Review the same way.The criterion for a good process: it makes work easier, quality higher, and decisions faster — otherwise it is just red tape.
Migration for you:
- Shorten the "decide → ship → get feedback" cycle, instead of simply forcing everyone to work overtime. The numbers and fun factor of gamified products are almost impossible to get right on the first try; you can only find them by testing in quick, small steps.
- Hold your own "99 bad ideas" meeting— the breakout hits of AI interactive narrative are often hidden in the "that's so absurd" brainstorms (that's how Duolingo's Super Bowl meme came about). Make it clear that "stopping an experiment = learning, not a shame".
- Use your mission as the referee for "good and fast": First lock in the hard passing line of "nothing below the bar ships" (for example, the plot cannot break, cannot break immersion), and above that passing line, prioritize by "which advances the core goal more".
- As soon as the company grows, introduce a lightweight PR mechanism, so that key decisions are "seen + traceable + multi-perspective", but always test it against "does this make decisions faster" to make sure it doesn't turn into bureaucracy.
4. Show Don't Tell: Let results speak — data is the story
EN: "Show Don't Tell" means we lead with the results of our work, not the story of its effort.
Core: Communication should be clear, concise,speak with data and actual results, not about how hard you worked. School and the workplace train us to persuade people by "telling stories and giving presentations" — but that approach cannot help you find real problems or build good products.
4.1 TL;DR = Conclusion + Evidence + Action (the most practical section)
Lead with the conclusion. A good TL;DR is just three sentences:What happened / what should we do? + What is the data / fact? + How will we do it?
| Dimension | Telling (Tell) | Show(TL;DR) |
|---|---|---|
| Order | Process → Conclusion | Conclusion → Evidence |
| Core | What I did | What the result is |
| Basis | Narrative / feelings | Data / facts |
| Listening experience | Lengthy | Efficient |
| Risk | Self-persuasion | Closer to the truth |
❌ Tell: "We did a lot of optimizations, user feedback is pretty good, we suggest continuing to push forward..." ✅ Show:Conclusion: This feature is worth shipping.Data: 7-day retention +12%.Action: Prepare for full rollout and continue optimizing.
One sentence to drive the point home:Telling stories = proving "I worked hard"; TL;DR = proving "this thing works".
⚠️ Counterexamples / boundaries (data can also lie): Over-fixating on data will lead you astray.Some things are very hard to measure accurately with data(such as "learning effectiveness" itself). Data is important but incomplete; in such cases you must rely on experience and judgment. (Note: Portfolios actually need storytelling — because they need to reflect your thinking and value through the process, and the audience cannot understand it without context. Show Don't Tell is a rule for internal communication, not a universal law.)
4.2 Other implementation mechanisms
- Prototypes, Not Pitch Decks: Instead of just talking about an idea, build it directly. Prototypes get everyone on the same understanding and move things forward faster — especially for complex problems like "adding AI to existing features", prototypes are critically important.
- A good product needs no explanation: Instead of relying on a pile of prompts to teach users, use design and animation to let users understand on their own. Even the learning method follows this principle —don't lecture on grammar, let you learn it yourself through practice(the best learning is doing, not reading instructions).
- Disagree and Commit: Resolve disagreements by "taking action", don't get stuck in gridlock. Both sides first commit to moving forward based on a decision,and let the result prove who is right. The leaderboard was a core feature that Luis had doubts about, but which came out of "let's just try it first".
- Ideas Before Egos (ideas matter more than ego): Look at results, not at who is right. A good idea doesn't need to be sold; what it needs is a chance to be validated.
- The Trust Battery: Trust is not given by default,it is earned through results. Every contribution you make charges your "trust battery" — whether you are an intern or an executive, you first have to prove yourself with results.
Migration for you:
- Lead every report with a TL;DR (conclusion + evidence + action), and delete "how hard I worked" from internal communication. This is also exactly the antidote to the common complaint that PMs are "wordy / have no conclusion".
- When you hit arguments over complex AI features, don't meet to argue — build a prototype first — especially for interactive narrative experiences, words can never describe it clearly; running a demo works better than 100 pages of PPT.
- Build a culture of "try it first": When opinions differ, design a small experiment that can produce data, and let the result decide, instead of letting whoever is loudest win.
- But hold the boundary of data: In gamified AI products, core values like "fun / emotional resonance"are very hard to quantify, don't let "data-only thinking" kill the things that data can't see but which are the soul of the product.
5. Make It Fun: Competing with TikTok for attention
EN: Learning doesn't have to be boring, and neither does work. We're at our best when we're subverting normal expectations people would have about an education company.
Core: Weave humor, joy, and imagination into everything. This is not a nice-to-have, it is a survival necessity.
5.1 Why: Your real competitor is TikTok, not other education products
Duolingo puts it very bluntly: they are competing for user attention with TikTok, Instagram, and online games, so learning must be as fun as those platforms to retain people.The difference is: those platforms are "mindless scrolling", while Duolingo users come with the clear goal of "learning" — it is purposeful, efficient use of time. But it is precisely those fun, unexpected designs that keep users coming back.
"A product built through play": From day one it was built as gamified, with lessons like variety shows or games. Absurd sentences like "Your bear is drinking beer", the character Lily "supporting" you with sarcastic slow claps — these are not just fun,they are the engine that keeps users learning.
"You don't have to be only one thing": Duolingo is not a game, and not just an education product —this blurry boundary itself creates magic.
5.2 Committing to the Bit: Take a joke all the way (the most counterintuitive point)
EN: In marketing, the worst thing you can be is boring.
- At first they tried to please everyone, and ended up with no character. Then they pivoted: weird, fun, unexpected.
- Accept that "humor cannot please everyone": it is subjective, and even divisive."It doesn't matter if not everyone gets the joke; what really matters is that the people who do like it will love it, and serve that group well."
- The duality of Duo: The mascot was originally meant to urge you to study, but netizens turned it into a complex character who "will threaten to take your family away if you don't study (jokingly)" —it was precisely by taking this bit all the way that Duo became a hit(buff Duo, anime Duo...).
- April Fools' + 5-second Super Bowl ad: The more absurd the April Fools' stunt, the more it goes viral (the 2024 "Duolingo on Ice" got 100 million social impressions). Instead of buying 30 seconds at the Super Bowl, they bet on "can 5 seconds blow up?" — spending very little money to get the same attention as everyone else."It's fun, unexpected, and unpretentious — that is exactly the secret to success."
5.3 Joy by Design (culture layer)
- The product is fun because the people building it are fun: There is a life-sized Duo working out in the office, puppy parades, all kinds of clubs.
- Open Doors communication: Glass-walled meeting rooms, anyone can sit in on PR, executives can be contacted directly —this openness makes people willing to take risks and do things differently(the soil for fun is psychological safety).
- Respect Boundaries: The prerequisite for being free to joke around is basic care and respect internally — the team is diverse, and some jokes may not be funny to others,"We are all adults here, respect each other's boundaries".
Migration for you:
- Recognize that your real competitors are entertainment products (short videos / games), not similar products. AI interactive narrative naturally stands at the intersection of "learning / useful" and "addictive / fun"; Duolingo is the most successful model on this path — you must treat "fun" as the core engine of retention, not an add-on.
- Don't please everyone, Commit to the Bit: Pick a distinct tone / character persona,and take it all the way. "The worst thing is to be boring" — mediocre, people-pleasing content will die for sure in the attention war. Serve the group of people who truly love it well.
- Treat "fans running wild with the character" as an opportunity, not a loss of control: Duo being remixed by users into a viral love-to-hate meme is exactly what made it successful. Leave room for your characters to be memed and remixed by users; that may be the cheapest growth there is.
- Fun must be "built in" through design, and requires team soil with psychological safety: A team that dares not take risks and passes the buck cannot build fun products. Open communication + respect for boundaries are prerequisites for fun.
VI. The Green Machine: Turning principles into an executable 6-step process
EN: The principles are theory. The Green Machine is how they actually get put into practice.
Core: Principles are theory; the Green Machine is the implementation method. In essence it is constant small improvements, getting better continuously— do more of whatever small change works. In one sentence:Find great people → let them try → double down on what works. When a project is stuck, use these 6 steps to check against them and diagnose where the problem is(usually stuck at "wrong team" or "no feedback mechanism").
| Step EN/CN | What to do | Corresponding principle / key point |
|---|---|---|
| 1. Staff It with Great People | Keep the team lean at first,add more people only after the process works | Excellent experiments require teams that can quickly generate ideas, execute, and pivot → high hiring standards |
| 2. Define Success | Set clear, measurable goals (DAU/revenue/impressions). If mature metrics don’t exist in the early stage,set the direction first, clarify as you go, and observe signals | Metrics-driven, but allow ambiguity in early stages |
| 3. Set Guardrails & Think Long-Term | Use rules to ensure long-term benefits for both users and the company | "Guardrails" = use the company mission as the standard for prioritizing decisions ← echoes long-termism |
| 4. Build the Thing & Set Up Feedback Loops | Don’t rely on discussion,just build it, and set up feedback at the same time: quantitative (A/B testing) + qualitative (internal use).Rely more on qualitative feedback early on, then shift to data-driven | Echoes "no PPT for prototypes" and "Dogfooding" |
| 5. Execute with Urgency & Excellence | Small improvements compound, so slowness is a cost; but moving too fast drops quality. Continuously deliver up-to-standard results quickly | Hold the line on both Ship It and Raise the Bar |
| 6. Double Down on What Works, Stop What Doesn’t | Invest more resources in what works and keep pushing it forward (sometimes for years);decisively shut down what doesn’t work, even kill the product | Accept that "termination" is part of the engine |
Tips: It does not replace planning. Green Machine It is not meant to replace OKRs / annual goals / long-term financial planning. Duolingo still has quarterly OKRs and annual goals — it just doesn’t rely on multi-year fixed product roadmaps. The engine handles "how to keep testing things out", planning handles "which big direction to go", and the two coexist.
Migration for your use:
- Treat these 6 steps as your "stuck project diagnostic tool": when a gamification feature/storyline gets stuck, check item by item — do you have the right people? Is the success standard defined? Are the guardrails (mission as referee) clear? Did you actually build it + set up feedback loops? Is the execution pace fast enough while meeting standards? Did you promptly cut what clearly doesn’t work?Bottlenecks almost always lie in Step 1 (team) or Step 4 (feedback mechanism).
- Step 6 "decisively shut down" is the hardest and most valuable: gamification products most easily fall into the sunk cost fallacy of "we’ve invested so much in this mechanic, don’t waste it". Explicitly write "shut down what doesn’t work" as a normal action of the engine, not a failure.
- Don’t use "agile trial and error" as an excuse to skip planning: OKRs and long-term direction still sit above the Green Machine.The engine is responsible for "how to move fast", mission and OKRs are responsible for "where to go"— you need both.
Appendix: Quick Glossary of High-Frequency Terms
| Term | Meaning |
|---|---|
| DAU / MAU | Daily active users / monthly active users. DAU is one of Duolingo’s most important metrics, used to judge product performance |
| Bookings | Subscriptions + in-app purchases + DET exam purchases + ad network revenue |
| DET | Duolingo English Test,the second-largest product. An adaptive certification exam, mainly used for overseas university admissions (positioned as "authoritative/standard") |
| Dogfooding (internal self-testing) | Employees use their own product to find problems |
| Experiments / A/B testing | Tests to verify what works; A/B tests run 2+ variants simultaneously for comparison |
| Juicy | Duolingo’s visual design system, fuller, friendlier, and more energetic than before (making learning feel like a game) |
| OKRs | Objectives and Key Results (proposed by Intel in the 1970s). Duolingo has quarterly team OKRs + annual company OKRs |
| PR (Product Review) | A meeting that approves all changes to the app, led by the product team,open to all to observe |
| Go, Go, Go! | A catchphrase Luis shouts when you do the right thing (= the slogan for Ship It) |
| Trust Battery | Trust is charged repeatedly through results, not given by default |
| Duoversary | An annual anniversary celebration in Pittsburgh marking the public launch on 2012-06-19 |
Overview: Five Principles + Engine One-Page Table (can be posted directly at your workstation)
| Principle | One-sentence core mindset | Most important counterintuitive point to remember | Strongest migration for gamified AI products |
|---|---|---|---|
| Take the Long View (Long-termism) | Good short-term but bad long-term = wrong | Retention > growth > revenue; dare to launch major overhauls that don’t lift short-term metrics (Path) | Operate characters/worldview as a moat, set notifications/paywalls against retention thresholds |
| Raise the Bar | Excellence is the baseline, not the goal | Build V1, not MVP; fewer features are acceptable, lower quality is not | Copy the four-standard acceptance checklist (useful/intuitive/delightful/polished), lock in on "intuitive" |
| Ship It (Launch Fast) | Speed compounds, slowness is a cost | Cut low-impact items no matter how much has been invested; use mission to resolve the tradeoff between good and fast | Shorten the decide→feedback loop + hold "99 bad ideas" meetings |
| Show Don’t Tell (Let Results Speak) | Data is the story | Telling stories = proving I worked hard, TL;DR = proving this works | Lead reports with TL;DR; use prototypes instead of PPT for complex AI features |
| Make It Fun | Learning and work don’t have to be boring | The real competitor is TikTok; don’t please everyone, Commit to the Bit | Pick a distinct persona and play it to the hilt, leave room for user fan creation/memes |
| Green Machine | Find people → let them test → scale what works | "Stopping what doesn’t work" is a normal action, not a failure; it does not replace OKRs | Use it as a "stuck project diagnostic tool" to check step by step; bottlenecks are mostly in team/feedback loops |