Altman: An 8-year-old kid makes a 3D game in 10 minutes
In a conversation with Stripe co-founder John Collison, Sam Altman used the example of an 8-to-10-year-old child building a complete 3D game on the spot in ten minutes to illustrate that the way this AI wave most resembles the early internet lies with the next generation. He also mentioned that when OpenAI's internal Codex briefly went down, engineers fell into a panic for a while, realizing they could hardly imagine going back to the days of handwriting code line by line.
View original postIn the interview, Altman said something highly controversial: everyone is occupying precious memory chips in their personal computers, yet using them only 5% of the time—that's a huge waste, it's terrible, and chips should be pooled together to maximize utilization. This is like a ride-hailing company urging you to smash all your private cars, on the grounds that cars sitting in garages waste too much steel. On the surface it's about computing utilization, but what it really wants to take away is your ownership of hardware. Or rather, it's a top technocrat feeling sorry for global semiconductor capacity. But if we dig one layer deeper into this calculation, what he really wants to reclaim is your ownership of personal hardware. This is like the CEO of a ride-hailing platform standing on stage preaching to everyone: Each of you buys a private car and leaves it sitting in the garage gathering dust 90% of the time—this is a huge waste of global steel capacity, and everyone should sell their cars and ride in my ride-hailing service. Efficiency is indeed maxed out, but freedom is confiscated too. That 95% of idle time on a personal computer was never waste. It is the freedom premium humanity pays in exchange for being able to boot up at any time, zero censorship, data entirely belonging to oneself, and still working after the network cable is unplugged. Do the math for the giants, and we can see the real calculation behind it: when everyone's Mac or PC is stuffed with 64G or 128G of unified memory and can smoothly run high-quality open-source models locally, the giants will face two nightmares: First, they can't grab enough memory chips to expand cloud data centers. Second, you no longer need to pay $20 a month in rent to the cloud. AI has developed to this day, and the real watershed is no longer the high or low of model benchmark scores, but the ownership of computing power: Is it toward a digital tenant where all computing power belongs to the cloud, or retaining a piece of offline silicon that belongs to yourself? Once all intelligence must connect to the giants' servers to work, what you own is no longer a personal computer, but just a cloud remote control that can be cut off at any time. Never easily hand over your local memory. Even if it sits idle 95% of the time, that is your last private territory in the digital world. https://t.co/YTTfLlKs2P
View original
奥特曼在访谈里说了一句极具争议的话：每个人都在个人电脑里占着宝贵的内存芯片，却只有 5% 的时间在用是巨大的浪费，这太糟糕了，芯片应该集中起来把利用率拉满。 这就像网约车公司劝你把私家车全砸了，理由是车子停在车库里太浪费钢铁，表面讲的是算力利用率，背后想剥夺的其实是你的硬件所有权。 或者说是顶级技术官僚在替全球半导体产能心疼。 但咱们把这笔账往深处算一层，他真正想收回的，其实是你对个人硬件的所有权。 这就像一家网约车平台的 CEO 站在台上向所有人布道： 你们每个人都买一辆私家车，90% 的时间都停在车库里吃灰，这是对全球钢铁产能极大的浪费，所有人都应该把车卖了来打我的网约车。 效率确实拉满了，但自由也被收缴了。 个人电脑那 95% 的闲置时间，从来都不是浪费。 那是人类为了换取随时开机、零审查、数据完全属于自己、拔掉网线依然能运转而支付的自由溢价。 替巨头算一笔账，我们就能看懂背后的真实算盘： 当每个人的 Mac 或 PC 都塞满 64G、128G 统一内存，本地能顺畅跑起高质量开源模型时，巨头会面临两个噩梦： 第一，他们抢不到足够的存储芯片来扩建云端数据中心。 第二，你不再需要每个月掏 20 美元向云端交租。 AI 发展到今天，真正的分水岭已经不是模型跑分的高低，而是算力所有权的归属： 是走向一切算力归于云端的数字租客，还是保留一块属于自己的离线硅片。 一旦所有智能都必须连上巨头的服务器才能工作，你拥有的就不再是一台个人电脑，而只是一块随时可以被断供的云端遥控器。 永远不要轻易交出你的本地内存。 哪怕它 95% 的时间在闲置，那也是你在数字世界里最后一块私人领地。 https://t.co/YTTfLlKs2P
Body · AI translation
The founders of two hundred-billion-dollar companies, @stripe co-founder John Collison and OpenAI CEO Sam Altman, sat in a bar drinking beer and got to talking about a rather interesting question: when exactly in this wave of AI does it most resemble the early internet era?
Altman didn't bring up the trillion-dollar compute race, nor did he discuss the capital stories of Wall Street, but instead told a real detail about a kid eight to ten years old.
That kid spent just ten minutes and whipped up a complete 3D game on the spot, and afterward looked completely calm, as if the world had always been meant to be this way.
Adults are used to treating large models as office plugins for boosting efficiency, calculating code completion rates and the percentages of cost reduction and efficiency gains.
But the real technological rupture tends to happen with the next generation, who haven't even had time to learn what syntax errors and compilation environments are, and whose intuition, before any professional skill, has already grown directly into a world they can play with at will.
This irreversible sense of dependence is happening to top engineers too. Altman mentioned that once inside OpenAI, Codex had a failure and briefly went down, and the engineers across the whole office suddenly fell into an almost instinctive panic. Everyone abruptly realized that they could hardly imagine going back to the days of typing out raw code line by line.
Actually, the most fascinating thing about the early internet was never the later monopoly of the giants, but that wild era when no one needed anyone's approval, and a single person in a bedroom could connect their imagination to the whole world.
When the threshold for generating code is completely flattened to within ten minutes, skill recedes into the lowest layer of background noise, and the only thing left that can set people apart is the pure desire you originally had to make something happen.
The full interview background, the original video source, and the highlights of the conversation are compiled in the first floor, help yourself if you need them.