A user argues that obvious ways to get more training tokens undercut concern about a “data wall.”
A user dismisses the idea that AI's “data wall” is a problem, saying there are “so many obvious solutions” to getting more training tokens. They do not give examples.
262
3 Sources, first seen ago
@code_starThere are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a problem.
@kalomazeanyways the argument is "no two distinct problems should be expected to be wholly disjoint in the representational sense", and "there's no serious reason to believe that ensembled synthetics can't powerfully overlap the span exercised by your specific domain that 'lacks data'"
@yacineMTBRT @code_star: There are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a…@code_starThere are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a problem.@kalomazeanyways the argument is "no two distinct problems should be expected to be wholly disjoint in the representational sense", and "there's no serious reason to believe that ensembled synthetics can't powerfully overlap the span exercised by your specific domain that 'lacks data'"@yacineMTBRT @code_star: There are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a…@code_starThere are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a problem.@kalomazeanyways the argument is "no two distinct problems should be expected to be wholly disjoint in the representational sense", and "there's no serious reason to believe that ensembled synthetics can't powerfully overlap the span exercised by your specific domain that 'lacks data'"@yacineMTBRT @code_star: There are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a…@code_starThere are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a problem.@kalomazeanyways the argument is "no two distinct problems should be expected to be wholly disjoint in the representational sense", and "there's no serious reason to believe that ensembled synthetics can't powerfully overlap the span exercised by your specific domain that 'lacks data'"@yacineMTBRT @code_star: There are so many obvious solutions to getting more training tokens, it's crazy that people even act like the data wall is a…—
Not ranked yet
—
Not ranked yet
