▲ 215 ▼ The hidden time bomb in the tax code that's fueling mass tech layoffs (qz.com) submitted 1 year ago by MuskyMelon@lemmy.world to c/technology@lemmy.world 33 comments fedilink hide all child comments
[–] Endmaker@ani.social 20 points 1 year ago (8 children) How is it that we are only learning about this now💀 permalink fedilink source hideshow 8 child comments replies: [–] Eknz@lemmy.eknz.org 24 points 1 year ago (4 children) I remember talking about this at the time. It did get some attention, although the industry was quick to blame other things instead, like AI. permalink fedilink source parent hideshow 4 child comments replies: [–] WhatAmLemmy@lemmy.world 10 points 1 year ago (3 children) Meanwhile I would argue that most efficiency gains from AI are outweighed by the cost of false positives wasting time and energy, learning prompt engineering, and other testing/fiddling adding overhead. permalink fedilink source parent hideshow 3 child comments replies: [–] krashmo@lemmy.world 5 points 1 year ago LLMs work in some limited use cases with good data to draw from. Most companies, especially large corporations, have shit data. Current AI cannot fix bad data. They will never do what these guys are promising they will do. permalink fedilink source parent [–] Eknz@lemmy.eknz.org 3 points 1 year ago You're not arguing against anything I've said given I said blame, as in, scapegoating AI for mismanagement and fundamental changes. permalink fedilink source parent load more comments (1 reply) [–] Maeve@kbin.earth 17 points 1 year ago (2 children) We were, but being set to expire in a decade and redundant 24 hour news cycles means they were designed to be forgotten. permalink fedilink source parent hideshow 2 child comments replies: [–] tburkhol@lemmy.world 15 points 1 year ago (1 child) At the time, it was part of the whole poisonous structure of the 2017 tax bill, where everything would expire after Trump's presumed 2nd term to sabotage his Democratic successor, confident that no one has long enough memory to realize where it came from. permalink fedilink source parent hideshow 1 child comment replies: [–] Maeve@kbin.earth 3 points 1 year ago Exactly. permalink fedilink source parent
[–] Eknz@lemmy.eknz.org 24 points 1 year ago (4 children) I remember talking about this at the time. It did get some attention, although the industry was quick to blame other things instead, like AI. permalink fedilink source parent hideshow 4 child comments replies: [–] WhatAmLemmy@lemmy.world 10 points 1 year ago (3 children) Meanwhile I would argue that most efficiency gains from AI are outweighed by the cost of false positives wasting time and energy, learning prompt engineering, and other testing/fiddling adding overhead. permalink fedilink source parent hideshow 3 child comments replies: [–] krashmo@lemmy.world 5 points 1 year ago LLMs work in some limited use cases with good data to draw from. Most companies, especially large corporations, have shit data. Current AI cannot fix bad data. They will never do what these guys are promising they will do. permalink fedilink source parent [–] Eknz@lemmy.eknz.org 3 points 1 year ago You're not arguing against anything I've said given I said blame, as in, scapegoating AI for mismanagement and fundamental changes. permalink fedilink source parent load more comments (1 reply)
[–] WhatAmLemmy@lemmy.world 10 points 1 year ago (3 children) Meanwhile I would argue that most efficiency gains from AI are outweighed by the cost of false positives wasting time and energy, learning prompt engineering, and other testing/fiddling adding overhead. permalink fedilink source parent hideshow 3 child comments replies: [–] krashmo@lemmy.world 5 points 1 year ago LLMs work in some limited use cases with good data to draw from. Most companies, especially large corporations, have shit data. Current AI cannot fix bad data. They will never do what these guys are promising they will do. permalink fedilink source parent [–] Eknz@lemmy.eknz.org 3 points 1 year ago You're not arguing against anything I've said given I said blame, as in, scapegoating AI for mismanagement and fundamental changes. permalink fedilink source parent load more comments (1 reply)
[–] krashmo@lemmy.world 5 points 1 year ago LLMs work in some limited use cases with good data to draw from. Most companies, especially large corporations, have shit data. Current AI cannot fix bad data. They will never do what these guys are promising they will do. permalink fedilink source parent
[–] Eknz@lemmy.eknz.org 3 points 1 year ago You're not arguing against anything I've said given I said blame, as in, scapegoating AI for mismanagement and fundamental changes. permalink fedilink source parent
[–] Maeve@kbin.earth 17 points 1 year ago (2 children) We were, but being set to expire in a decade and redundant 24 hour news cycles means they were designed to be forgotten. permalink fedilink source parent hideshow 2 child comments replies: [–] tburkhol@lemmy.world 15 points 1 year ago (1 child) At the time, it was part of the whole poisonous structure of the 2017 tax bill, where everything would expire after Trump's presumed 2nd term to sabotage his Democratic successor, confident that no one has long enough memory to realize where it came from. permalink fedilink source parent hideshow 1 child comment replies: [–] Maeve@kbin.earth 3 points 1 year ago Exactly. permalink fedilink source parent
[–] tburkhol@lemmy.world 15 points 1 year ago (1 child) At the time, it was part of the whole poisonous structure of the 2017 tax bill, where everything would expire after Trump's presumed 2nd term to sabotage his Democratic successor, confident that no one has long enough memory to realize where it came from. permalink fedilink source parent hideshow 1 child comment replies: [–] Maeve@kbin.earth 3 points 1 year ago Exactly. permalink fedilink source parent