@guansi: Recently, I chatted with a few colleagues who handle review processes, and it feels like everyone’s got this pent-up fr…

Recently, I chatted with a few colleagues who handle review processes, and it feels like everyone’s got this pent-up frustration inside, not quite ready to let it out. Now, some of the proposals submitted by newcomers are pretty damn impressive. Cutting-edge tech, solid logic, tossing around all sorts of jargon like it’s nothing. You can tell right away they’ve used AI, and after a couple of follow-up questions, it’s clear a lot of the content goes way beyond their own grasp. But once the proposal’s in, someone’s got to review it. The reviewers have to go through it section by section, checking sources, hunting for gaps, figuring out if these techniques really fit the current project. The more it sounds legit, the less you can just wave it through. Some parts read smoothly at first, but only after mulling it over do you spot what’s off—and just explaining the issues clearly can take half the day. After the review, you call the newbie over and walk them through it point by point. They’re pretty cooperative, jotting it all down diligently, then feeding it back to the AI, and soon enough, they hand in a revised version. Then you review it again. You’ve got to double-check if the last round’s issues got fixed right, and whether any new ones have cropped up this time—all from scratch. They revise fast, but you can’t keep up that pace. After a few rounds, the proposal looks sharper and sharper, while the reviewer’s brain is throbbing. You’ve poured a ton of effort into dissecting one document, probably more than they spent generating it in the first place. Finally, you sort out the problems, send it back—and before long, here comes another plate served up. As for whether they actually understood what you pointed out, you’ll have to ask again to find out. But you can’t just straight-up criticize the newbie. If you tell them not to use AI, what they write on their own might be even worse; if you make them review it themselves first, they genuinely can’t spot the issues. Some of that tech they’ve never even touched before, and now it’s showing up for the first time in a proposal with their name on it. This whole thing’s starting to feel trickier and trickier to me. Back in the day, when newbies wrote proposals, their skill level was limited, and the problems were usually pretty straightforward. Now AI’s elevating their proposals, which ramps up the review difficulty right along with it—but the person themselves might not be keeping pace. I’ve got this vague sense that this is going to be a big issue down the line.