I've heard this take on machine learning models that they're benefiting from uncredited or uncompensated works, with the implication that this is bad and wouldn't be permitted if the model were replaced with, say, a human. that's obviously mistaken though.. the way almost all of these models work, we'd have to be similarly upset at humans who consume heaps of images, text, sounds, etc then produce new artifacts under the influence of those, but we're not. we celebrate the people who produce such things for their ingenuity, artistry, or analysis.
#machinelearning
I love the PureCSS portraits by Diana Smith (Cyan Harlow)
What gets me is how relatively few rules there are -- I mean, it's hundreds or thousands of lines in some cases, but it's not 10s of thousands. Besides, the rules use names that you can actually understand to get a sense for how the piece, as a whole, works together.
twitter link but worth it
"By far, the most accurate representation of machine learning pipelines in the real world 🍿 😀" -- this video is hilarious in that context.
"Don't say I never gave you anything"/"Don't say I never did anything for you" are among the phases I most hate to hear from someone. it pretty much always means the person saying it *knows* what they've given or what they've done is insufficient but they still want to demand that you protect their image. it's just so weak
A capable software engineer and aspirating (sic) cook. Also posting about space stuff (mostly NASA) occasionally
pronouns: he, him