>>192602, >>192604, >>192605, >>192606, >>192607, >>192608, >>192609, >>192610, >>192611, >>192612, >>192613, >>192614, >>192615, >>192616, >>192617, >>192618, >>192619, >>192620, >>192621, >>192622, >>192623, >>192624
DataRepublican (small r) @DataRepublican - Accurate.
Parent Post:
MOMof DataRepublican @data_republican - OK I will translate for the rest of us:
I think she said that she has been working on a project having to do with computer stuff for 10 years that may or may not be useful. And AI just rebuilt it from her notes in 10 minutes. She doesn't know if that's a good thing or if she should feel dumb.
You're welcome.
https://x.com/DataRepublican/status/2097415694543646814

DataRepublican (small r) @DataRepublican - Very much so. I stared at hex dumps to figure out why an inference that reduced computations by 85% wasn't getting the promised speedup (tl;dr - it came down to needing to optimize cpu caching as well).
Parent Post:
Arthur Morgan @ArthurMorganTX - The scary part isn't that it did it in 10 minutes, it's that management will see this and assume a single junior with a prompt can replace a decade of domain/design expertise. We're watching hard-won technical moats get bulldozed in real time.
https://x.com/DataRepublican/status/2097398651991797955

DataRepublican (small r) @DataRepublican - Had a 10-year-old project on data indexing that may or may not find application for vector databases (it optimizes exact kNN over medium dimensionality, a rarity).
Just watched GPT Astra rebuild the entire project in 10 minutes from an unpublished paper.
Not sure how I feel about that.
https://x.com/DataRepublican/status/2097386373099753552