3 ms·
Why the Great AI Backlash Came for a Tiny Startup You’ve Probably Never Heard Of
- sharemywin 3y agoEven if he could have gotten away with it using fair use. probably not a good idea to upset the people your trying to get as customers.
- turtleyacht 3y agoIt's interesting that one can share a dashboard and see it interpreted in completely opposite ways. To one, useful to compare the centrifuged distillate; to another, theft of labored craft, unwilling participants of weighted metrics. "Don't measure us," says one party. For they regard their product unique, an aggregate of history and (non-)choice, which could never be reproduced in one age or the next. "Please measure," says the other. For they are fans and consumers first, and seek to compare and consume. And somewhere in this, we also find the germ of humans and Key Performance Indicators (KPIs), of accounting and finance, of measuring things to make decisions against limited time and attention, with competitive urgency, while balancing folklore, aesthetics, and "don't datamine us." Supposing math or another level of indirection would solve it, I wonder if there is a more general means to unite both sides. For example, the only couple chapters of accounting in a business textbook, I see in the line items between assets and liabilities a narrative of something--but we can only write quantitative amounts, doubly, under certain account types. Yet that is the form of communication taken. Somehow, all activity and insight must be ground down and shaped up from a powder into a specific currency.
- greenie_beans 3y agohis thing actually looks pretty cool (screenshots in the blog post) https://blog.shaxpir.com/taking-down-prosecraft-io-37e189797121 https://blog.shaxpir.com/taking-down-prosecraft-io-37e189797... hopefully he figures out a way to bring it back.
- sterlind 3y agoThe surge in anti-AI clauses reminds me of robots.txt, albeit with more (gnashing of) teeth. But robots.txt mostly served to entrench Google and Bing, and frustrated attempts to build Common Crawl (a project that's also in the crossfire now!) I'm concerned that like with robots.txt, these clauses will end up centralizing who has access to the training corpus, more than it will meaningfully restrict what the corpus is used for. I'm sure Google already had a license to train ML models on the Google Books corpus, and that they've fed it all into Bard. I'm sure Google's more than happy to be the trusted, sole custodian of these datasets.
- ilaksh 3y agoSounds like he made a useful tool but "oversold" it with a bad marketing approach. And then a few rightfully offended writers turned into a social media mob.