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：范皓宇
Fan describes approximately 25 years in products and design across industrial design, phones, games, media, operating systems and cars. He identifies a common capability: observe, understand, combine and compress, turning numerous aspects of life into a coherent solution people can directly experience.
Guests / narrator：任川
Listen: 02:04–09:25 and 40:57–41:53. Palona AI then provided telephone services for American restaurants, including reservations, orders, and inquiries. Ren reports 95–98% reliability without defining the task set, failures, or human transfers. The figure cannot establish a ranking against other agents.
Guests / narrator：虞快
Listen: 01:57–06:39 and 18:48–19:43. Yu lists doctors, lawyers, engineers specializing in particular languages, and people combining unusual disciplines or languages. A model company may know something is wrong without knowing how to evaluate medical or legal answers systematically. It needs judgment as well as labor.
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：雷磊
An agent performing tasks needs more than a capable model. It must run code, browse websites, preserve login, wait for confirmation and return to its previous position. Lei calls the supporting environments Agent Infra, or agent infrastructure. He wants Grasp to handle common browser engineering so application developers can concentrate on industry tasks.
Guests / narrator：丁丁
A benchmark is a set of test tasks and corresponding standards. Evaluation is the process of using standards to analyze and judge performance. Dingding begins with this distinction: public tasks reflect some capabilities without necessarily covering a product’s actual inputs and processes.
Guests / narrator：王文锋
Wang explains agents through models, environmental feedback and tool-based action. In a product, these form a continuing process: observe the current situation, invoke a tool, obtain a result and use the new information to decide what to do next.
Guests / narrator：吴翼
Listen: 01:55–06:14. Classifying a cat or dog involves a labeled answer and one decision. Playing a game requires many choices—left, right, or jump—and several paths may win. Reinforcement learning, or RL, improves an action policy using rewards received after acting. It addresses processes with multiple possible choices whose quality becomes clear only later.
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：Joe
The title connects three months in the United States, conviction and ten-million-dollar ARR. Joe is more specific: ACE Studio then earned about US$800,000 monthly, compared with a previous level above US$80,000. Growth was slow for five weeks before a change from week six. Multiplying US$800,000 by twelve gives about US$9.6 million, a monthly run rate rather than proof of full-year collections or revenue entirely meeting recurring-revenue definitions.
Guests / narrator：Binson
逗逗游戏伙伴 is an AI desktop companion that watches gameplay, accompanies players and converses. Binson’s approach asks users to allocate less additional time to AI content: participate in something they already want to do. The game continues while the companion acts like interactive commentary beside it.
Guests / narrator：朱哲清 Bill
A business trip can appear to require only an itinerary. Across cities, days and meetings, however, hotels, transport and timing constrain one another. An early choice can make the eleventh step impossible. Bill uses this example to explain why an agent needs to consider consequences, search alternatives and recover when a path fails.
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：Ted
Open conversation has no single correct answer. A user may enjoy a response, rewrite it, delete it or request another. Reflecting on approximately a year at Character.AI, Ted discusses turning such behavior into data, training and experiments. The central subject is a system that continually discovers mistakes.
Guests / narrator：Tristan
Tristan wants EVE to build a continuing relationship around the real user’s life: understanding earlier experiences, connecting them with present goals and participating in new daily conversations. He distinguishes this aim from jointly creating stories about characters. The host points out that extended use can also come from consuming content.
Guests / narrator：张海龙
Qu Kai asks Zhang Hailong, who has more than twenty years of programming experience, to explain the excitement around Cursor. A useful distinction emerges: helping a person operate faster and accepting a task for independent completion are different product promises.
Guests / narrator：封诚
If a company connects AI throughout its processes and reports impressive coverage, has its business improved? Feng Cheng examines acquisition, conversion and delivery, then concentrates on leads already obtained: potential customers who receive slow responses or inadequate explanations.
Guests / narrator：Albert
Founders often ask whether a model upgrade will erase their feature. Albert moves the question earlier: can you deliver defined value to a defined audience now, and identify which change improved the result? If product quality remains difficult to judge, discussing permanent defenses can lead a team into waiting without feedback.
Guests / narrator：陶芳波
An AI can retrieve information or fill out forms without knowing which result matters to you. Tao uses this distinction to explain the move from MindOS to Me.bot: first build AI that understands an individual’s needs, then let it call tools providing services. He also wants AI to keep learning from daily experience rather than depend permanently on its original workflows.