3. "Understanding which occupations have tasks within the current frontier of AI capability is important because these jobs are most likely to see productivity boosts or a shift in their core tasks."
4. This whole paragraph:
4. This whole paragraph:
February 11, 2026 at 2:17 PM
3. "Understanding which occupations have tasks within the current frontier of AI capability is important because these jobs are most likely to see productivity boosts or a shift in their core tasks."
4. This whole paragraph:
4. This whole paragraph:
When one looks at the details, the article makes basically the opposite argument as claimed. Some key quotes:
1. "AI action completion x score is consistently lower than on the user goal side [...], indicating that AI can help users with a broader fraction of their work than it can perform directly"
1. "AI action completion x score is consistently lower than on the user goal side [...], indicating that AI can help users with a broader fraction of their work than it can perform directly"
 by the product of their completion rate and the share of conversations where Bing Copilot demonstrated at least a moderate scope of capability. All panels show researcher-generated IWA groupings to highlight patterns."
The highest score in completion x scope is ~0.6, with all work activities showing gaps between user goal and action. From my estimation, the smallest gap is in "analyze data using mathematics," where AI action is at about 0.2 and user goal is 0.3. The largest gaps are in "interview people to obtain info" and "purchase goods or services" (bottom IWAs) and "implement procedures/processes" and "explain medical information" (top IWAs), with most of the top IWAs showing gaps of at least ~0.2](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:jpwkj5g56sld5nfvwbshibr2/bafkreie4vrrcvmm3e6ve3mbwhfseunffajulxuey57lxns5tqbnxjq6ptq@jpeg)
February 11, 2026 at 2:17 PM
When one looks at the details, the article makes basically the opposite argument as claimed. Some key quotes:
1. "AI action completion x score is consistently lower than on the user goal side [...], indicating that AI can help users with a broader fraction of their work than it can perform directly"
1. "AI action completion x score is consistently lower than on the user goal side [...], indicating that AI can help users with a broader fraction of their work than it can perform directly"
So many people (including a ton of self-identified historians, who I would have expected would check primary sources) uncritically sharing this misleading skeet.
#AI is, for the most part, bullshit and not something that should have been released, much less foisted upon the general population. But:
#AI is, for the most part, bullshit and not something that should have been released, much less foisted upon the general population. But:
February 11, 2026 at 2:17 PM
So many people (including a ton of self-identified historians, who I would have expected would check primary sources) uncritically sharing this misleading skeet.
#AI is, for the most part, bullshit and not something that should have been released, much less foisted upon the general population. But:
#AI is, for the most part, bullshit and not something that should have been released, much less foisted upon the general population. But:
The broad AI push is terrible, and hopefully will be stopped and/or massively side-eyed by future generations, but this actual paper is *actually* trying to demonstrate which occupations could benefit from AI assistance, not which will be replaced by AI.
February 11, 2026 at 12:42 PM
The broad AI push is terrible, and hopefully will be stopped and/or massively side-eyed by future generations, but this actual paper is *actually* trying to demonstrate which occupations could benefit from AI assistance, not which will be replaced by AI.
I'm sorry, drug companies are now selling cars into the US? It's this like that 30 Rock bit about NBC being a subsidiary of a wig company?
Is it something like the BMW/Pfizer conglomerate or Renault/Sanofi corporation? And, for curiosity's sake, are there any wig companies involved?
Is it something like the BMW/Pfizer conglomerate or Renault/Sanofi corporation? And, for curiosity's sake, are there any wig companies involved?
May 22, 2025 at 9:20 PM
I'm sorry, drug companies are now selling cars into the US? It's this like that 30 Rock bit about NBC being a subsidiary of a wig company?
Is it something like the BMW/Pfizer conglomerate or Renault/Sanofi corporation? And, for curiosity's sake, are there any wig companies involved?
Is it something like the BMW/Pfizer conglomerate or Renault/Sanofi corporation? And, for curiosity's sake, are there any wig companies involved?
You should really know this, but... Principal, guys, not principle.
March 20, 2025 at 9:20 PM
You should really know this, but... Principal, guys, not principle.
Shoot, that's what I get for not double-checking my sources. Looks like it was less, though not by much (6.4M Texans vs. 6.1M Californians).
Hard to see the distributions in electoral maps that don't account for population density.
Hard to see the distributions in electoral maps that don't account for population density.
March 4, 2025 at 11:29 PM
Shoot, that's what I get for not double-checking my sources. Looks like it was less, though not by much (6.4M Texans vs. 6.1M Californians).
Hard to see the distributions in electoral maps that don't account for population density.
Hard to see the distributions in electoral maps that don't account for population density.
More Californians voted for Trump than Texans, and that is especially true in their wine growing regions (Sorry to Sonoma and Napa, though).
March 4, 2025 at 10:32 PM
More Californians voted for Trump than Texans, and that is especially true in their wine growing regions (Sorry to Sonoma and Napa, though).
It is disappointing but not totally surprising that the NYT would carry water for a fascist by conflating popular support with legality (or morality). I suppose their mealy-mouthy approach to Hitler's rise didn't cost them enough. Worth remembering that this time, the fascism isn't across the pond
February 9, 2025 at 10:16 PM
It is disappointing but not totally surprising that the NYT would carry water for a fascist by conflating popular support with legality (or morality). I suppose their mealy-mouthy approach to Hitler's rise didn't cost them enough. Worth remembering that this time, the fascism isn't across the pond