Aaron Levie：AI 智能体降低启动成本，反而让人更忙
One of the biggest reasons why we’re working more than ever is because the pace of change is so fast that we’re working in overdrive just trying to keep up with
Box CEO Aaron Levie 认为，AI 智能体降低了启动任何任务的门槛，人们因此不断启动更多智能体，结果为自己制造出更多新工作。智能体把研究和新想法验证变得相对简单，但流程最终会撞上人的瓶颈，人处理完瓶颈后又回去启动智能体，这个循环不断重复，何时放缓尚不清楚。
查看原帖Quick reflection staring at my computer at 10pm: if AI is taking my software job, why have I never worked so hard in my life? Two options: 1/ The narrative is wrong. There’s so much to build that there’s more work than ever. 2/ I mostly work with AI agents now, so I’m the bottleneck. My human speed just can’t keep up with them. Thoughts?
正文
当前显示原文，中文全文译文尚未就绪。
One of the biggest reasons why we’re working more than ever is because the pace of change is so fast that we’re working in overdrive just trying to keep up with everything. But that likely only explains part of the demands on our time.
We’re also working like mad because AI agents have lowered the cost to getting anything started.
In a resource constrained system, there are just lots of limits to how fast you can move. You can’t spin up every project you think of because you’d have to change everyone’s priorities. You barely can even test a hypothesis without changing course materially on an existing worksteam.
Even getting a couple people peeled off from what they’re doing for a couple days is a fairly thrashy and difficult process in most organizations that have limited surplus resources.
Enter agents. Anything you want to research, you just have agents do it. Testing out a new idea becomes relatively trivial. So now, we just spin up more and more agents to do work for us that we wouldn’t have been able to contemplate kicking off before.
As a result of that work, inevitably now we just create more new work for ourselves. Obviously if you could automate everything then this might be moot. But eventually you run into some bottleneck that is a human constraint. So we go spend our time on that part of the workflow. And when we’re done, we’re back to spinning up agents to do work again.
This cycle just repeats on and on and it’s not obvious when and how it starts to slow down. The opportunity cost in a world where compute can be thrown at any problem becomes massive to not figure out what you can be throwing compute at.
So this is one of the reasons we’re working more than ever, among others.