It seemed that nothing could slow down YouTube, the online-video upstart bought by Google ( GOOG) last year for $1.65 billion.

YouTube has thrived against all odds and despite predictions of its imminent demise. Observers thought it would be smothered under Google, that it would no longer be seen as a trendy venue for posting and watching videos.

Others predicted certain doom as big content owners sued the deep-pocketed Google for posting unauthorized snippets of their programming.

In the nearly four months since the buyout was announced, none of this has hurt YouTube. Traffic on the site continues to steam steadily upward, according to Alexa.com. Media giants are talking with Google about revenue sharing.

Google even rolled YouTube into its own video search, but rather than hurting YouTube, it's provided a superior way to search its sprawling video inventory.

Last week, 20th Century Fox subpoenaed YouTube for names of users who posted clips of its television shows 24 and The Simpsons. But so far, this has hardly caused the scare that critics predicted. And in the Kabuki world of media copyrights, where tough actions are often gestures aimed at another agenda, the move could be aimed at giving News Corp. ( NWS) leverage over revenue sharing.

YouTube could still suffer a crippling lawsuit, but for now at least, that seems much less of a threat than a decision that seems to be coming from inside YouTube itself: Over the weekend, company founder Chad Hurley said the video site would start sharing revenue with users. Blogger Jeff Jarvis broke the news in a clip that is hosted, appropriately enough, over at YouTube .

While sharing revenue with users is laudable -- given its inevitability -- what's troubling about it is how YouTube will generate the revenue it plans to share: the dreaded pre-roll, according to a report from the BBC.

A pre-roll is video advertisement that you'll have to watch before you can see the video you came to YouTube to see in the first place. All evidence to date suggests that pre-roll ads are powerfully repellent to viewers, driving them to other sites. YouTube's pre-rolls may be only three seconds long, according to Hurley, but that may be enough to turn users off from them.

Last fall, Forrester Research announced a survey that said 80% of people are annoyed by pre-roll ads, and that 75% simply ignore them. Those numbers were far higher than other ads, such as banners adjacent to videos or the text link favored by Google in its search results.

Text-link ads in Google's search results work for three reasons: they aren't mandatory, they aren't intrusive, and they're more likely to be relevant. Pre-roll ads are none of the above.

They are an especially bad fit for YouTube, too, because of how people watch videos there. Browsing videos on YouTube is like channel-surfing on cable TV. There's an element of risk that the next video you click on -- just like the next cable channel you switch to -- won't be any good.

Yet people keep flipping ahead to the next video clip, believing that one of them, at just the right moment, will have something worth watching.

Now imagine if you were channel-surfing and you had to watch a three-second commercial every time you changed the channel. Would that cure your channel-surfing addiction? You bet it would.

Also, a three-second ad isn't yet proven enough to lure in advertisers. Three seconds is time enough to flash a logo. Advertisers for whom the 30-second spot is a highly polished, almost scientific process may not take to the brutally truncated format. It's like asking an advertiser to rent an entire billboard but use only a tenth of its space.

Some ad agencies will experiment with three-second format, however. And they will surely come to the conclusion that the way to make a three-second ad stand out is to jam it as full as possible with blaring, flashing content. The result? Obnoxious ads, the equivalent of the two-stepping cowboys, every time you watch a YouTube video.

YouTube and Google are smart enterprises, so there may be something else going on here. Maybe the pre-roll idea is just a gambit to keep the media companies at bay. Maybe the media giants are insisting on pre-rolls, knowing full well it will sabotage YouTube's success, or maybe YouTube is floating this trial balloon knowing full well it will fail, so it can say to media companies, "Well, we tried."

Whatever the reason behind this idea, it's not a positive for Google. The company's success is based on advertising that is useful, or at least nonintrusive. Its future in online video rests on an ad format that is similarly user-friendly, so it can share revenue with content owners large and small.

But this idea, however nascent, feels like a drastic misstep in the making.