Dan Luu 写系统问题时，习惯先测量、再对照、最后才下结论。把「Keyboard latency」放到智能体、评测和线上系统里，真正要问的是：默认做法会不会系统性失败。下面用中文整理成可执行的工程笔记，去掉原站导航和无关链接。

Does keyboard latency matter?

If you look at “gaming" keyboards, a lot of them sell for $100 or more on the promise that they’re fast. Ad copy that you’ll see includes:

Despite all of these claims, I can only find one person who’s publicly benchmarked keyboard latency and they only tested two keyboards. In general, my belief is that if someone makes performance claims without benchmarks, the claims probably aren’t true, just like how code that isn’t tested (or otherwise verified) should be assumed broken.

Are gaming keyboards faster than other keyboards?

The situation with gaming keyboards reminds me a lot of talking to car salesmen:

Salesman: this car is super safe! It has 12 airbags! Me: that’s nice, but how does it fare in crash tests? Salesman: 12 airbags!

Conclusion

Sure, gaming keyboards have 1000Hz polling, but so what?

A year ago, if you’d asked me if I was going to build a custom setup to measure keyboard latency, I would have said that’s silly, and yet here I am, measuring keyboard latency with a logic analyzer .

Appendix: where is the latency coming from?

It all started because I had this feeling that some old computers feel much more responsive than modern machines . For example, an iMac G4 running macOS 9 or an Apple 2 both feel quicker than my 4.2 GHz Kaby Lake system. I never trust feelings like this because there’s decades of research showing that users often have feelings that are the literal opposite of reality, so got a high-speed camera and started measuring actual keypress-to-screen-update latency

Anyway, at the time I did these measurements, my 4.2 GHz kaby lake had the fastest single-threaded performance of any machine you could buy but had worse latency than a quick machine from the 70s (roughly 6x worse than an Apple 2), which seems a bit curious. To figure out where the latency comes from, I started measuring keyboard latency because that’s the first part of the pipeline. My plan was to look at the end-to-end pipeline and start at the beginning

2022 update

Let’s look at the measured keypress-to-USB latency on some keyboards:

The latency measurements are the time from when the key starts moving to the time when the USB packet associated with the key makes it out onto the USB bus. Numbers are rounded to the nearest 5 ms in order to avoid giving a false sense of precision. The easterntimes i500 is also sold as the tomoko MMC023 .

Appendix: counter-arguments to common arguments that latency doesn’t m

The connection column indicates the connection used. USB FS stands for the usb full speed protocol, which allows up to 1000Hz polling, a feature commonly advertised by high-end keyboards. USB is the usb low speed protocol, which is the protocol most keyboards use. The ‘gaming’ column indicates whether or not the keyboard is branded as a gaming keyboard. wireless indicates some kind of keyboard-specific dongle and unifying is logitech's wireless device stan

We can see that, even with the limited set of keyboards tested, there can be as much as a 45ms difference in latency between keyboards. Moreover, a modern computer with one of the slower keyboards attached can’t possibly be as responsive as a quick machine from the 70s or 80s because the keyboard alone is slower than the entire response pipeline of some older computers.

值得单独记下的观察

• a custom designed keycap that has been made shorter to reduce the time it takes for your actions to register
• 8x FASTER – Polling Rate of 1000Hz: Response time 0.1 milliseconds
• Wield the ultimate performance advantage over your opponents with light operation 45g key switches and an actuation 40% faster than standard Cherry MX Red switches
• World's Fastest Ultra Polling 1000Hz
• World's Fastest Gaming Keyboard, 1000Hz Polling Rate, 0.001 Second Response Time
• Does keyboard latency matter?
• Are gaming keyboards actually quicker than other keyboards?

落地时建议先做的 5 件事

1. 用自己的真实负载测，而不是只用公开榜或厂商数字。
2. 把评测设计成能抓到失败模式：平均分好看但尾部崩溃，仍然算失败。
3. 智能体默认不会好好用测试；要写进流程，而不是写在口头规范里。
4. 性能和正确性都要有基线，改模型或改语言前后必须能对比。
5. 结论写成可回滚的决策：哪一版配置、哪一版评测集、谁签字。

和智能体产品怎么接

龙虾PRO做 OpenClaw 落地时，同样吃「先测量再扩面」这条纪律：技能、数字员工和网关都要有可复现评测，而不是只看一次演示通过。

本文侧重全链路风控方法论。落地时请用自身业务单据做回放验证，不要把示例阈值直接当生产策略。 相关：风控体检 · 方案资源