Liang Wenfeng Debuts on Hurun Rich List at No. 7 with 280 Billion Yuan Fortune

It was only after I finished reading the Weibo posts Liang Wenfeng wrote at age 26 that I understood: DeepSeek's success has nothing whatsoever to do with its technical approach. At 26, Liang Wenfeng drove a Fiat alone into the Ali no-man's-land in Tibet. He went 400 kilometers deep, was stranded for a week, he walked out, the car stayed there forever. Twelve years later he built DeepSeek; if you'd read this story, you wouldn't be surprised at all. At the end of 2011, he drove a beat-up Fiat to Tibet. His companions reached Lhasa, had their fun, and flew back. He didn't go back. Alone he kept heading west—Yamdrok Lake, Everest, Shishapangma—all the way 400 kilometers into the Ali no-man's-land. He used an entry-level Nikon D50 with a kit lens, but he could wait half a day just for a beam of light, and the photos he took looked like a professional photographer's. The captions read "The wind sighs and the Yi River runs cold," "On the ancient road, the west wind, a lean horse," "One rider traverses the world." The last Weibo post: Stranded in the no-man's-land for a week; fortunately I got out. But my Fiat didn't. It stays forever in that boundless alpine grassland. That's the END. That's the whole story, and when we look back at DeepSeek, it's all the same person playing the same game. Others turn back once they reach Lhasa; he goes another 400 kilometers in alone. Doesn't that look like now? Everyone is building applications, chasing commercialization, doing things that can be monetized quickly, while he goes it alone inward, building foundation models, pursuing AGI, not raising funds for a long time, not socializing, not showing up at all kinds of conferences. An entry-level camera, but waiting half a day for the light, capturing the best shot— doesn't that look like DeepSeek? No piling up resources, no burning cash, but pushing efficiency to the extreme, achieving something rivaling others at one-tenth the cost. The car was left in the no-man's-land, the man walked out, and wrote "That's the END." Doesn't that look like his style? Dare to bet, dare to lose, accept the loss, then keep going. Open source is open source, sanctioned is sanctioned, release the model and release it; the car is gone but the man is out, and next trip he'll go again. Everyone is analyzing DeepSeek's technical approach, cost structure, and national-luck backing, but look at what this man did at 26—the answer is all there. A man who dares to drive alone 400 kilometers into a no-man's-land won't follow behind others when running a company. A man who left his car on the plateau, walked out, and could still write "That's the END" won't care about short-term stock prices and funding news. A man who waits half a day for light with an entry-level camera will squeeze every bit of compute power when building models. So none of this is some business strategy; this is one person's character. DeepSeek is not the result of a business plan; it is one person's nature, naturally grown out twelve years later. That young man who walked for a week in the no-man's-land and the man who has turned the global AI industry upside down are the same person—and always have been.

View original

我翻完梁文锋26岁写的微博才明白DeepSeek的成功，跟技术路线半毛钱关系都没有。 梁文锋26岁那年,一个人开着菲亚特进了西藏阿里无人区。 深入400公里， 被困了一周， 人走出来了， 车永远留在了那里。 12年后他做了DeepSeek， 你要是读过这个故事,就一点都不意外。 2011年底,他开一辆破菲亚特去西藏， 同行的人到了拉萨玩完就飞回去了， 他没回。 一个人继续往西,羊卓雍措,珠峰,希夏邦马,一直扎进阿里无人区400公里。 用的是入门级尼康D50,套头， 但他能等大半天就为了一束光， 拍出来的照片像专业摄影师的。 配文写的是"风萧萧兮易水寒",写的是"古道西风瘦马",写的是"一骑走天下"。 最后一条微博: 被困无人区一个星期，所幸我出来了。但我的菲亚特没有出来。它永远地留在了那片漫无边际的高山草原。That's the END。 就这么个事， 我们回头看DeepSeek,全是同一个人的同一个打法。 别人到拉萨就回去了， 他一个人再往里走400公里。 像不像现在? 大家都在做应用、做商业化、做能快速变现的事情， 他一个人往里走,做基础模型,做AGI,很长时间都不融资,不社交,不出来参加各种大会。 入门级相机,但等半天光线,拍到最好， 像不像DeepSeek? 不堆资源,不烧钱,但把效率做到极致,用别人十分之一的成本做出匹敌的东西。 车扔在无人区了,人走出来了,写一句That's the END。 像不像他的风格? 敢赌,敢输,输了认,然后继续走。 开源就开源,被制裁就被制裁,模型放出去就放出去,车没了但人出来了,下一趟再来。 所有人都在分析DeepSeek的技术路线、成本结构、国运加持， 但你看看这个人26岁干的事,答案全在那。 一个敢独自把车开进无人区400公里的人,他做公司就不会跟在别人后面。 一个把车扔在高原走出来还能写That's the END的人,他不会在乎短期股价和融资新闻。 一个用入门相机等半天光线的人,他做模型就会把每一分算力都榨干。 所以这些都不是什么商业策略， 这就是一个人的性格， DeepSeek不是一份商业计划书的结果， 是一个人的本性,在12年后,自然长出来的东西。 那个在无人区里走了一周的年轻人,和那个把全球AI行业搅翻天的人,是同一个人，一直以来都是。