Lately, when chatting with friends about what's happening in the software industry, I often compare it to short videos. At first it felt like just an analogy, but the more I think about it, the underlying logic is almost identical.
How Short Videos Took Off
Watching videos used to cost something. Data was expensive—a single video could be dozens of megabytes. Who dared to keep scrolling when they were running low on data at the end of the month? Then 4G spread, 5G followed, and data plans kept getting cheaper. Watching videos anytime, anywhere became effortless. The barrier to entry was erased.
The production side was changing at the same time. Making your own video used to feel like building a rocket, to put it dramatically. Then CapCut came along, and with live streaming, you could just point the camera at your face, hit record, and have an audience. Production costs were pushed close to zero.
When both ends dropped toward zero simultaneously, the platforms in the middle exploded.
I used to see the industry data at my company. In 2017, the short video ad market was roughly 50–60 billion RMB. A few years later it soared past 200 billion, and with e-commerce and live streaming added in, the whole ecosystem surged past 400 billion. The bulk of the profits were captured by Douyin and Kuaishou.
The Software Industry Is Replaying the Same Story
In the past, downloading and installing an open source project wasn't trivial for ordinary people. The documentation was mediocre, environment setup was full of pitfalls, and non-technical folks mostly couldn't get it to work. Now it's different: when you hit a problem, you just ask ChatGPT or DeepSeek, and it walks you through step by step. Most open source projects can be made to run. Acquiring software has become much easier.
GitHub's data tells the story. In 2025, there were over 180 million developers worldwide, with more than 36 million new sign-ups in a single year—roughly one new registration every second. India added 5.2 million in one year; Indonesia grew from 900,000 in 2020 to over 4.3 million. Eighty percent of new users tried Copilot in their first week. Many of them aren't programmers in the traditional sense. AI has lowered the barrier to entering the open source community.
After our team started using open source community infrastructure, we found that many colleagues had already dabbled with it before, or at least had an account. With the spread of AI, people can now simply ask, "How do I install this software? How do I use it?" and follow the steps to get it working.
The production side is changing even faster. Coding agents are pushing the cost of writing software toward zero. Building an MVP used to take a small team several months; now one person plus a coding agent can ship something usable in a few days. Cursor has over a million daily active users, and its annualized revenue surpassed $2 billion in early 2026. Eighty-four percent of developers are using or planning to use AI coding tools, and 41% of code already has AI involvement in its creation.
Recently I've seen people sharing that they knew nothing about tech before, yet now they've open sourced a project on GitHub and garnered thousands of stars. MIT Technology Review listed "generative coding" as one of the ten breakthrough technologies of 2026. Entrepreneurs, marketers, and designers are all starting to build software directly with AI tools. One growth marketer used ChatGPT plus Lovable to build a crypto visualization app without knowing how to code. This was unthinkable before.
Production costs are approaching zero, and acquisition costs are approaching zero too. Open source communities, as the platforms in the middle, are facing the same situation short video platforms did. This is the same force at work behind the collapse of the software bidding system.
OpenClaw
OpenClaw deserves its own section.
It launched only at the end of 2025, and has already surpassed 250,000 stars, making it the fastest-growing open source project in GitHub history. Linux took years to reach that number.
The reaction it sparked in China is even more interesting. In March, nearly a thousand people queued outside Tencent's headquarters in Shenzhen, asking engineers to help install OpenClaw. In line were students, retirees, office workers—not just programmers. "Raising a crayfish" became a meme. The Longgang district of Shenzhen is offering subsidies of up to 10 million RMB to startups based on OpenClaw; Wuxi followed with 5 million RMB. Alibaba Cloud, Tencent Cloud, ByteDance, JD.com, and Baidu have all launched their own versions.
Viewed in isolation, OpenClaw might look like just another viral project. But placed in the framework I just described, it's actually a microcosm of the platform effect beginning to manifest. Creators use coding agents to iterate rapidly; users get up to speed with the help of AI; and GitHub distributes it at near-zero cost. A good piece of software going from birth to usage by hundreds of thousands of people used to take a long cycle; now the speed is completely different.
Why You Should Pay Attention to Open Source Communities
Looking ahead, I think open source communities will mean to the software industry what short video platforms mean to entertainment.
GitHub added 121 million new repositories in 2025—230 new projects every minute—and merged over 500 million pull requests in a single year. AI-related projects grew 178% year over year, with more than 1.1 million repositories using LLM SDKs.
Open source communities are no longer just places for programmers to exchange code. Production happens here, distribution happens here, collaboration happens here, and influence grows from here. Having open source projects to show already unlocks significant traffic and resources.
If you're in the software industry, whether in engineering or product, I suggest you take open source communities seriously. Not the "keep an eye on it" kind of serious, but treat it as a core channel to actively cultivate.
The driving forces behind this are AI and coding agents. A coding agent is not just a tool for writing code; it is a meta-ability—one that lets people without a technical background participate in open source projects. But looking at the phenomenon itself, what you'll see is good software spreading through open source communities at an astonishing rate, just as short video platforms exploded after 4G and smartphones became ubiquitous infrastructure.
For those of us already in the software industry, this is worth watching closely.