The Talk That Almost Didn't Make the Cut

Every program committee I've sat on has a version of this moment. Someone reads a submission out loud, there's a pause, and then someone says some version of I don't know, this doesn't sound like much. The abstract is thin. Sentences run long, or clipped in a way that suggests someone translating from a language that organizes ideas differently than English does. Maybe there's a typo in the first line, which somehow reads as a signal about the whole forty-five minutes to come, even though it plainly isn't.

I've watched this scene play out enough times, with different committees, different topics, different years, that I no longer think of it as bad luck. It's a pattern. And the submissions that trigger it aren't always weak. Some of the strongest talks I've sat through at these events started life as an abstract a committee nearly threw out.

The mechanics are pretty ordinary once you look at them. A program committee works through maybe eighty or a hundred submissions in a sitting, against a deadline, with everyone half-distracted by whatever they were doing before the call started. An abstract has to do a lot in two hundred words: state the topic, find an angle, explain why anyone should care, and convince a stranger that this particular speaker is the one to do it. That's a genre. Writing well inside a genre is its own skill, and it has nothing to do with whether the underlying work is any good.

Some people can do both. A lot of good engineers can't. Someone who spent eighteen months grinding through a hard problem is frequently the worst person to summarize why it mattered in a paragraph, because they've lived inside it so long that the interesting part looks obvious and the boring part looks like the part that needs defending. Someone writing in a second or third language is translating not just vocabulary but the shape of an argument, and technical nuance is exactly what gets flattened first. Someone nervous about the whole process hedges every claim and buries the one line that would have made a reviewer sit up. None of that predicts whether the talk itself will be any good.

What gets lost in the process is a distinction committees don't always spell out. An abstract tells you something about a talk. It doesn't settle the question. Treating it like a verdict is efficient, and that efficiency is exactly why the mistake keeps happening: reading a paragraph takes two minutes, and getting a borderline submitter on a call takes twenty, and a committee with a hundred submissions and four volunteers doesn't have twenty minutes to spend on all of them.

But it doesn't need to spend it on all of them. It needs a way to flag the maybes and treat the maybes differently than the clear rejections. I've seen committees do this without ever writing it into a process: a member notices a topic worth hearing about buried under a flat pitch, and instead of quietly rejecting it, sends the submitter a direct question. Not a form email. Something closer to can you walk me through what you actually found, written by a person, answered by a person. Ten minutes of that conversation usually tells you more than the abstract ever could, because you're listening to someone talk about their own work instead of reading their attempt to write about it.

Committees that skip this step aren't lazy, most of the time. They're running on the assumption that an abstract is a decent proxy for the talk, and most of the time it roughly is. The failure only shows up at the edges, on the borderline submissions, which is exactly where it costs the most, because those are the ones a short conversation could have saved.

I don't think the fix is complicated. It's a norm more than a process: when a committee member has a nagging feeling about a weak-looking abstract on a topic that seems worth hearing, that feeling is worth a message before it's worth a rejection. The talks that come out of those conversations aren't always great. But often enough they're the ones people are still talking about a year later, and nobody in the room ever finds out how close the whole thing came to not happening.