You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I think @NirantK's point is that as a field advances, it gets more difficult to make a scientific contribution. The easy problems have been solved; the hard ones are hard.
Is it getting harder though? Folks in 2014-2015 had to come up with new ideas, folks in 2021 mostly train larger Transformers on larger datasets. The latter is objectively easier
But also BERT, GPT-3 etc require hundreds of engineers, massive server farms, millions of dollars, etc these models are not something a single person can do from scratch. It's like going from the Wright brothers cobbling together a glider to building an A-380.
But just scaling the data and network size is not a research contribution (although studying scaling laws certainly is).
That's a strong statement. There are many ways of scaling data and network size. There are plenty of counterarguments on why doing this can bring many research contributions. Not to mention the fact that solving engineering challenges can also lead to research contributions.
People training larger models on larger datasets is a super small subsets of all researchers. Even if my university is lucky to have access to a decent amount of GPUs, I cannot go in that direction without much more comp. power
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
