Signal to BuildSIGNAL TO BUILD中文Owner sign in ↗Find the mechanisms, cases and conditions in each episode. Start with a topic, follow a guest, or read chronologically to see how perspectives change.
Dates refer to original episode releases. An episode may have multiple topics; roles and claims retain their historical context.
Guests / narrator：逯雨鑫
The official claim concerns Hugging Face’s Model Trending popularity list, not first place in overall performance. Lu post-trained an existing model rather than pretraining a general model from scratch. Improvement on a task does not establish stronger capability everywhere.
Guests / narrator：孟繁青
Meng used a mature training platform inside a model company. Starting a business meant building platforms and finding a complementary opportunity. Turning training into a service simplifies data submission and execution without automatically making the data valuable.
Guests / narrator：曲凯
Listen: 01:03–05:34. Qu recalls successive swings: uncertainty about model–application relationships in 2023, cooling and recovery in 2024, reasoning and agents sustaining expectations in 2025, and stronger models alongside colder application funding in his circles during 2026. He says this is the first time he has heard firms declare the category dead.
Guests / narrator：朱宁
Listen: 01:27–03:59. Zhu identifies overconfidence, linear extrapolation, and herding. After prices rise, people remember having predicted them and credit their own ability. This encourages larger subsequent commitments.
Guests / narrator：张津剑
The introduction uses the listings of 智谱 and MiniMax as a three-year milestone, inviting the podcast’s first guest to review AI again. Investment outcomes and market descriptions remain contemporary perspectives. A listing cannot alone validate products, returns or every early hypothesis.
Guests / narrator：Jenny Xiao
Jenny describes a Japanese financial company finding document organization and analysis insufficiently reliable. Even mostly correct output needed item-by-item checking, adding work. This turns a strong-model claim into a delivery question: how much work disappears, and what responsibility remains for errors?
Guests / narrator：莫傑麟
Asked whether AI has a bubble, the opening answer is yes. The ensuing conversation repeatedly distinguishes technical value, expectations inside prices, infrastructure payback, who ultimately receives money and whether founders survive until another financing round.
Guests / narrator：莫傑麟
Listen: 00:59–09:27. The guest recalls monthly token consumption growing more than 20% over the previous month, without supplying the statistical institution, sample, or consistent measurement definition. The number explains his starting observation; it is not a verified industry growth rate.
Guests / narrator：朱亦博
Listen: 01:07–07:11; 08:40–11:20. Zhu reduces the underlying problems to computation, communication, and storage. GPUs need networks, switches, storage, operations, resource scheduling, and training and inference frameworks. His comparison with cloud service layers is an explanatory framework, not the only possible industry taxonomy.
Guests / narrator：莫傑麟
After reaching the United States, Qu noticed rapid changes in how people discussed Chinese and American assets. He asks why two years of pessimism could reverse suddenly and how long new optimism might last.
Guests / narrator：曲凯
The title sounds certain, but the argument contains many conditions. Applications might grow differently if models cross a usability threshold, if agents can deliver results and if organizations adapt to a new business model. An agent here is a model system that uses tools and advances tasks, rather than a name for every chat window.
Guests / narrator：曲凯
Listen: 01:08–03:16. Qu first rejects equating the event with the failure of conversational companionship. In his description, model researchers and people continuing the product receive different arrangements. A technical team's decision to enter a large company does not establish that users no longer want character conversations.
Guests / narrator：汪华
Listen: 09:18–17:25. Wang's first layer covers language and multimodal interfaces, together with simple tasks that still involve people. Users can express needs through text, voice, or images without learning every menu. People remain responsible for judgment, handoffs, and completion. Wang compares this change to mobile internet and expresses strong confidence in it.
Roles reverse in this episode: Mo asks and Qu mainly answers. Qu favors participating early while warning against sprinting just after the start. These positions are compatible. Time, attention and learning can receive serious commitment while money, staffing and scale require staged evidence.
Guests / narrator：季宇（Mackler）
Ji combines chip-architecture experience with work on AI compilers at HiSilicon. He explains Nvidia through relationships among hardware, software and developers. This is a practitioner’s strategic interpretation from 2024, not independently verified company history.
Guests / narrator：谢羽佳
Xie brings experience in consumer electronics, crowdfunding and distribution, while developing a children’s entertainment terminal. The introduction says her previous team served more than forty million-dollar-scale crowdfunding projects. That supplies background rather than audited project profit or fulfillment.
Guests / narrator：莫傑麟
The episode places Sora’s first demonstrations within a larger question: why can compute and model narratives advance before many mature applications exist? The guests distinguish technical progress, competitive position and application demand.
Guests / narrator：曲凯
The title identifies 2024 and 2025 as crucial years, while much of the episode addresses immature technology. Its central question is how to believe in a long-term opportunity while obtaining customers, revenue and learning time through work deliverable now.
Guests / narrator：胡渊鸣
An image with depth or a video rotating around an object may not provide an asset that can be edited, animated and placed in a game. Hu moves from graphics fundamentals through representations to Meshy’s product and startup choices. The central thread is that outputs need to enter existing workflows to move beyond demonstrations.
Guests / narrator：莫傑麟
The conversation begins with a contrast: people sound pessimistic while working harder. Perhaps effort has not disappeared; familiar rewards simply arrive more slowly. If so, the task extends beyond changing mood to recognizing feedback cycles, revising evaluation and preparing capabilities.