I'm not a very good speaker. I say "um" a lot. Sometimes I have to pause when I lose my train of thought. I wish I were a better speaker. But I don't wish I were a better speaker like I wish I were a better writer. What I really want is to have good ideas, and that's a much bigger part of being a good writer than being a good speaker.
Having good ideas is most of writing well. If you know what you're talking about, you can say it in the plainest words and you'll be perceived as having a good style. With speaking it's the opposite: having good ideas is an alarmingly small component of being a good speaker.
I first noticed this at a conference several years ago. There was another speaker who was much better than me. He had all of us roaring with laughter. I seemed awkward and halting by comparison. Afterward I put my talk online like I usually do. As I was doing it I tried to imagine what a transcript of the other guy's talk would be like, and it was only then I realized he hadn't said very much.
Maybe this would have been obvious to someone who knew more about speaking, but it was a revelation to me how much less ideas mattered in speaking than writing. 
A few years later I heard a talk by someone who was not merely a better speaker than me, but a famous speaker. Boy was he good. So I decided I'd pay close attention to what he said, to learn how he did it. After about ten sentences I found myself thinking "I don't want to be a good speaker."
Being a really good speaker is not merely orthogonal to having good ideas, but in many ways pushes you in the opposite direction. For example, when I give a talk I usually write it out beforehand. I know that's a mistake; I know delivering a prewritten talk makes it harder to engage with an audience. The way to get the attention of an audience is to give them your full attention, and when you're delivering a prewritten talk your attention is always divided between the audience and the talk—even if you've memorized it. If you want to engage an audience it's better to start with no more than an outline of what you want to say and ad lib the individual sentences. But if you do that you could spend no more time thinking about each sentence than it takes to say it. Occasionally the stimulation of talking to a live audience makes you think of new things, but in general this is not going to generate ideas as well as writing does, where you can spend as long on each sentence as you want.
If you rehearse a prewritten speech enough, you can get asymptotically close to the sort of engagement you get when speaking ad lib. Actors do. But here again there's a tradeoff between smoothness and ideas. All the time you spend practicing a talk, you could instead spend making it better. Actors don't face that temptation except in the rare cases where they've written the script, but any speaker does. Before I give a talk I can usually be found sitting in a corner somewhere with a copy printed out on paper, trying to rehearse it in my head. But I always end up spending most of the time rewriting it instead. Every talk I give ends up being given from a manuscript full of things crossed out and rewritten. Which of course makes me um even more, because I haven't had any time at all to practice the new bits. 
Depending on your audience, there are even worse tradeoffs than these. Audiences like to be flattered; they like jokes; they like to be swept off their feet by a vigorous stream of words. As you decrease the intelligence of the audience, being a good speaker is increasingly a matter of being a good bullshitter. That's true in writing too of course, but the descent is steeper with talks. Any given person is dumber as a member of an audience than as a reader. Just as a speaker ad libbing can only spend as long thinking about each sentence as it takes to say it, a person hearing a talk can only spend as long thinking about each sentence as it takes to hear it. Plus people in an audience are always affected by the reactions of those around them, and the reactions that spread from person to person in an audience are disproportionately the more brutish sort, just as low notes travel through walls better than high ones. Every audience is an incipient mob, and a good speaker uses that. Part of the reason I laughed so much at the talk by the good speaker at that conference was that everyone else did.
So are talks useless? They're certainly inferior to the written word as a source of ideas. But that's not all talks are good for. When I go to a talk, it's usually because I'm interested in the speaker. Listening to a talk is the closest most of us can get to having a conversation with someone like the president, who doesn't have time to meet individually with all the people who want to meet him.
Talks are also good at motivating me to do things. It's probably no coincidence that so many famous speakers are described as motivational speakers. That may be what public speaking is really for. It's probably what it was originally for. The emotional reactions you can elicit with a talk can be a powerful force. I wish I could say that force was more often used for good than ill, but I'm not sure.
 I'm not talking here about academic talks, which are a different type of thing. While the audience at an academic talk might appreciate a joke, they will (or at least should) make a conscious effort to see what new ideas you're presenting.
 That's the lower bound. In practice you can often do better, because talks are usually about things you've written or talked about before, and when you ad lib you end up reproducing some of those sentences. Like early medieval architecture, impromptu talks are made of spolia. Which feels a bit dishonest, incidentally, because you have to deliver these sentences as if you'd just thought of them.
 Robert Morris points out that there is a way in which practicing talks makes them better: reading a talk out loud can expose awkward parts. I agree and in fact I read most things I write out loud at least once for that reason.
 For sufficiently small audiences, it may not be true that being part of an audience makes people dumber. The real decline seems to set in when the audience gets too big for the talk to feel like a conversation—maybe around 10 people.
Thanks to Sam Altman and Robert Morris for reading drafts of this.