Talks & slides

Different work calls for different ways to use AI.

Explore the materials behind team trainings, strategic briefings, and public workshops. Each entry shows the audience, the problem, and the original deck or recording.

12enterprise programs & briefings12public talks & workshops2working languages

DoorDash team offsite · Seattle
DoorDash team offsite · Seattle
The library

Find the material you need.

The collection includes enterprise programs, briefings, and selected public workshops. Original deck language is preserved.

24 items

Data & decisions

How AI changes the value of analysis, experimentation, and professional judgment.

Strategy & knowledge work

Turn judgment, context, and repeated research into systems that compound.

  • 刘润直播间2026.10Public session

    假工作的终结

    Fake work swaps value that is hard to measure for actions that are easy to see. AI makes looking busy nearly free, so fake work will first explode and then lose its price. We could never tell real from fake because the visible layer of reports, plans and decks was costly to produce and served as a proxy for the skill below the waterline. That proxy has now failed, so we have to look directly at judgment and at what it grows from: a world model shaped by the scientific method, and craft shaped by practice. What most people lack is the ability, the habits and the mechanisms to see it.

    Professionals and managers who stay busy all week but can't say what their work changed, and anyone asking what still counts once AI can write the reports · Liu Run's livestream · talk on the book 《真本事》

  • 课代表立正2026.09Public session

    AI会让假工作先爆炸,再贬值

    AI first makes reports, plans, and other work evidence abundant. As those signals lose scarcity, value moves back toward problems, judgment, responsibility, and real-world results.

    Knowledge workers and leaders rethinking what remains valuable when AI makes work evidence abundant · Long-form video argument · public cue-card edition

  • Tencent IEG2026.07Enterprise

    什么是好的战略分析?

    A strong strategic analysis challenges the starting view, clarifies what is tangled, and turns judgment into action. AI can build the 60-point factual base; people must push the judgment to 90.

    Strategy, business analysis, and AI leaders · Delivered · Judgment Engineering strategy briefing

  • 腾讯游戏战略2026.05Enterprise

    战略研究的第三界面

    Strategy teams do not lack information. The scarce work is turning information into organizational judgment—and preserving the context so the next cycle starts further ahead.

    Strategy researchers and leaders · May 2026 session edition

Engineering & organization

Move beyond adoption metrics and redesign how teams build, evaluate, and learn.

From user to builder

Help non-engineering teams move from isolated prompts to end-to-end work.

  • 课代表立正2026.10Public session

    AI都会教你了,为什么还要上课?

    AI has made answers nearly free. What a class still has to teach is what AI can't hand you: the judgment below the waterline, and the drive to build it. Judgment comes from a world model, shaped by the scientific method, and from craft, shaped by practice; the parts that can be said, we turn into lessons. The drive leaks into watching videos and into anxiety; small concrete loops and someone beside you bring it back.

    Chinese-speaking learners deciding whether a class is still worth it now that AI can teach, and people who have watched plenty but still can't get results · Open class · long-form video

  • 通往AGI之路2026.10Public session

    差的不是人,是市场

    Between the US and China, the people are on par and the markets are far apart: US companies pay for software, so AI becomes productivity, while in China much of the compute goes to content such as short dramas. You can't change the market, but you decide how you use AI: practice the verbs, install an agentic tool tonight, and manage AI by its results.

    Chinese-speaking AI practitioners and learners following the US–China AI gap · Livestream: What can you actually see in Silicon Valley now?

  • Columbia University2026.10Public session

    The End of Fake Learning

    What it means to learn when AI can do the work: why your feeling of learning is often backwards, why AI is most dangerous when it’s right, and how to use it as a spotter instead of a forklift.

    Undergraduate and graduate students building with AI · Guest session · October 2, 2026

  • 课代表立正2026.08Public session

    AI副业,别越做越不值钱

    A side project is not only an income bet; it allocates time, reputation, and career capital. Use three ledgers—cash, commercial evidence, and durable assets—to decide when to stop, pivot, or double down.

    Programmers and AI builders exploring side projects · Public talk deck · Career capital edition

  • 课代表立正2026.08Public session

    AI副业,从做出来到持续赚钱

    AI makes building cheaper, but durable income still depends on demand, distribution, trust, retention, margin, and the willingness to keep taking responsibility.

    Programmers and AI builders exploring side projects · Public talk deck · Commercial gates edition

  • TSVC×网易新闻2026.06Public session

    普通人应该怎么学AI

    The useful leap is not from one chatbot to another. It is from isolated conversations to a personal AI work system that compounds.

    Professionals and investors without a technical background · Joint public online session

    • Public cut in progress
    • Chinese
  • 美团2026.06Enterprise

    从User到Builder:AI时代的个人进化指南

    When non-engineers can build, the ceiling moves from using a tool well to redesigning an entire workflow.

    Company-wide, cross-functional teams · Company-wide AI session

  • 小红书2026.06Enterprise

    从30%到x10:AI时代的个人生产力跃迁

    When production gets cheaper, judgment, context, trust, and coordination become more valuable.

    Content, product, operations, and business teams · Company-wide productivity session

  • XWZ OPC Founder Alliance2026.05Public session

    用AI提效30%VS提效10倍,本质差别到底在哪儿?

    A 10x outcome rarely comes from a better prompt. It comes from changing the operating system around the work.

    One-person-company founders · Founder workshop and live demo

  • Seattle (AI) Startup Summit2026.04Public session

    Learning to Code When AI Does the Heavy Lifting

    When AI writes much of the code, beginners should learn to specify, inspect, debug, and own the result.

    Founders and first-time builders · Hands-on workshop

  • Superlinear × AceMode2026.03Public session

    从ChatGPT到Cursor:普通人如何抓住AI实战红利

    The important move is not from one tool to another, but from asking for answers to building work that can be inspected and reused.

    Knowledge workers and new builders · Public online session

  • 斯斯·小私董2026.03Enterprise

    从“会用AI”到“用AI创造价值”

    The gap between using AI and creating value with it is a better problem, a complete workflow, and a reusable asset.

    Creators and small-business operators · AI internal workshop

Product & enterprise delivery

Use the customer frontier to improve both enterprise delivery and the product behind it.

  • AI Engineer Shanghai2026.08Enterprise

    把全球AI工程师网络与上海产业生态连接起来

    Introduce the joint operating team, show how AIE creates a two-way bridge between Shanghai and the global AI engineering network, and define a focused first cooperation package with the association.

    Industry associations, technology companies, and ecosystem partners · Partnership briefing for the Shanghai AI industry association

  • Superlinear Academy2026.06Public session

    FDE模式:从Statsig Enterprise Engineering看产品化的大客户交付

    FDE is not merely onsite delivery. Done well, it turns the customer frontier into a high-speed product learning loop.

    Founders and product leaders · Public product and organization talk

  • 腾讯游戏IEG2026.03Enterprise

    腾讯游戏AI Intensive Bootcamp

    Enterprise AI training should start with real role-level constraints, then connect capability progression, project practice, and reusable organizational assets.

    Game development teams and learning leaders · Three-track custom training proposal · public-safe edition

Bring a real team problem

What does your team need to do differently?

Tell me who is in the room, what they are responsible for, and what should become possible afterward. That is enough to start designing the right session.