LeituraPesquisaRadarFramework de investimento
Entrar / Cadastrar
← Voltar para a leitura
NotíciasTópico · 新颖性预算开发法Visão rápida2 min de leituraEditorial DisconfirmAI

AI Engineer Hariprasad Sudharshan on the 'Median' and Novelty Budget

In a video talk, Hariprasad Sudharshan shares insights from two years of solo building an AI calculator-like app. He notes that identical models and tools often lead to converging solutions, termed the 'median.' He proposes a 'novelty budget' to define unique value and suggests leaders evaluate demos by asking who has already solved the problem, what distinct assets the team holds, and where users will be reached.

2026-10-15 02:00 GMT+8

In his video talk titled "I Spent Two Years Building the Median," Hariprasad Sudharshan explores the homogenization of software construction in the AI era. Drawing from his experience of independently developing a calculator-like AI application over two years, he observed that other builders, driven by the same models, tools, and familiar problems, often produce strikingly similar solutions without copying each other. He refers to this shared starting point as the "median," noting that while faster development makes it easier to ship useful software, a functioning demo does not establish a durable product advantage.

Sudharshan distinguishes between the model's capability, its harness, and the final product received by the customer. He describes a "build first" reflex where developers immediately open a coding agent instead of searching for existing solutions. As an alternative, he introduces the concept of a "novelty budget": naming the one distinctive thing before starting, including it in the roadmap, and reviewing it as deliberately as other project spending.

For leaders evaluating demos, Sudharshan proposes three critical questions: Who has already solved this? What does this team have that a competitor does not? Where will the application reach users? In later discussions about iterating on the calculator, he addresses the difference between running open model weights inside an enterprise versus purchasing inference, touching on utilization and infrastructure costs. His core lesson is to use readily available starting points where they help, then spend judgment on a specific problem, distribution, and a defensible difference.

Entrar / Cadastrar
LeituraPesquisaRadarFramework de investimento

Fontes

  • youtube.com↗
Tópico

新颖性预算开发法

“新颖性预算”是一种限制技术栈中未知风险的开发策略,其核心逻辑是:团队对不稳定技术的容忍度有限,必须将创新资源集中在用户可感知的核心价值上,而其余部分应保持“无聊”的稳定。

4 etapas1 peçasAtualizado 2026-10-18 15:45

Como chegamos aqui

  1. 2018-08-25Shimweasel博客首次提出“Novelty Budgets”概念,主张在软件工程中限制未经验证的库的使用,以降低生产环境风险。shimweasel.com ↗
  2. 2026-05-17Tian Pan发表文章指出LLM评估套件常沦为“博物馆”,呼吁建立“新颖性预算”以资助针对未知失败模式的探索性测试。tianpan.co ↗
  3. 2026-08-14Hitesh撰文论证“无聊技术”是AI战略,指出AI组件因概率性特性永久占用创新代币,要求其他基础设施保持极致稳定。hitesh.in ↗

O que ainda está indefinido

  • 如何在实际项目中量化“探索性测试”与“回归测试”的资源配比?
  • 当底层模型更新时,如何界定哪些功能属于“核心新颖性”从而保留预算,哪些应迅速转为“无聊”的标准组件?
  • 是否存在通用的工具或框架能自动识别并标记代码库中的“高新颖性”依赖项?
Ler o tópico completo →
Continuar pesquisando →