https://www.gamesindustry.biz/articles/2016-07-18-obsidian-entertainment-survival-success-and-independence said:
Q: For me, you were doing really interesting work in that period, on Fallout: New Vegas and Alpha Protocol in particular. Nevertheless, and I'm sure you're aware of this, Obsidian ended up with a reputation for bugs and technical issues. Is that a necessary consequence of towing the line, and working to imposed budgets and time-scales? These aren't simple games you're making.
FU: So, with Alpha Protocol, the challenge was that we weren't even totally sure what we wanted to make until, like, way into the game - and that's bad. You can do that with your own money; when you're doing that with someone else's money they're just getting mad, they're getting mad at you more and more and more. The story's bigger than that, but I think I'd go back to what I was just saying: you've gotta cut it early. For that reason, as a developer, you have to take it on yourself to prove your ideas quickly. You need to show yourself that you can make the game, that you understand the game.
It's too easy to not be critical. To not say, 'Okay, that all sounds wonderful, but what's the plan? Like, really, what is the plan?' That's where we've made big mistakes in the past: not holding to our plans. That doesn't sound like sexy development, right? But if we haven't proven it, let's know that now. Let's look at it and go, 'Oh, it's like an ugly little child. That's not good. So what do we do with it?' And it's too easy at that point to add just a little bit more here, and it'll be good, and everything will be fine. No. 'Should we cut it?'
I think publishers should kill way more games way earlier, but if we do that ourselves it makes us more reasonable about what we can actually accomplish early on. We weren't doing that, and that put pressure on our publishers.
Q: Isn't that the publisher's call anyway? You say more games should be killed earlier, but surely Obsidian wouldn't decide that.
FU: No, it wouldn't be my call, but we always should look at the fact of, 'Would we kill this game?' It's nice if the publisher keeps on paying us for another year, but if we would kill it then it really should be killed. So back up from that: what are we doing today to make sure we can do what we want and have the game not be killed? We're now doing that on day 2, and not day 430.
That's what we did too much of before. We just kept on going, sweeping this and that problem under the rug - sweep, sweep, sweep, it's all fine. And you get to a point and it's screwed.