No, your difficulty wasn't too little space. If you had known precisely what to say you could have said it. The way to make the draft better is to reduce the name-dropping and tighten the description of the problem, then write about what to do using the space gained.

"Polarization" is a red herring, I think. Whether people want to argue or agree about politics or policy is not determined by the efforts of advertisers to affect their brand choices. Pattern-recognition (grandiosely described as machine learning) may make advertising targeting efficient: you haven't shown it doesn't, or that efficient advertising is a problem that inefficient advertising isn't.

But there doesn't have to be Facebook. Why not explain how to replace it? Surely modifying how we receive certain basic services in the Net is easier than modifying the US Constitution?