Comment by underlipton
11 hours ago
This. These platforms are asking to be trusted with unprecedented amounts of the public's data and, unprecedentedly itself, the public's reasoning and decision-making. It's an awesome responsibility that requires a singular approach that smaller platforms with less responsibility don't necessarily have to devote resources to. OpenAI, Anthropic, Google, Facebook, they're the big dogs. They can't do the things the small guys can get away with. They're the 18-wheelers, and when you're an 18-wheeler, you HAVE to act differently. You stay in the middle lane, you do not speed, you always yield, because when you make a mistake, when you drive aggressively, you can kill dozens and blow up and interstate and stop traffic for hours, if not days. You don't get to do shit like this; the cost of everyone giving you their data is that you give up every opportunity to use it for your own interests, even though you technically have the capability to exploit it.
Nice analogy! This seems to me the main difference between Google and the other companies you list -- In regards to LLMs Google is the only one mostly acting like an mature, established organization and not rushing to market as soon as possible. For their responsibility they often get labelled as having fumbled some imagined race. Maybe they actually do just lack the talent to make better models but it's not like they aren't making advances in other areas of AI.
What a great analogy, can't believe I haven't heard it before