Writer: nytimes, NatGeo, sciam, etc., Disruptive tech editor @bulletinatomic. Author of THE ALMOND IN THE APRICOT | saragoudarzi.com
Signal: saragoudarzi.19
Loading...
“If we stop testing right now, we’d have very few cases, if any.”
yup, that about sums it up
this is the way
Company that uses humans to label data for training AI spends more on tokens for said AI than on humans for training said AI. I'd say this is too Meta for me but....
Hey folks, I'm starting to think AI might be a tad problematic for the integrity of science publishing!
A study published as Correspondence in @thelancet.com found 4,046 fabricated references across 2810 papers, presumably generated by AI
By @bruceylee.bsky.social
www.forbes.com/sites/brucel...
Anthropic Confidentially Files for What Could Be the Largest IPO Ever
www.wired.com/story/anthro...
I might actually be the youngest person quoted in this (long, well-written and well-sourced) article.
thebulletin.org/2026/06/agis...
"An honest look at where the industry stands with regard to achieving artificial general intelligence—however that’s defined—shows a vast and perhaps permanent gulf between machine and human intelligence," writes Sara Goudarzi @saragoud.bsky.social.
incredible news: @saragoud.bsky.social shows it *is* possible to write a feature about AGI that doesn't weirdly anthropomorphize the LLMs and robots or ask them "are you alive?" and then reprint the answer verbatim. major mainstream magazines, please take note!
I would pay for a podcast player that bans AI-generated podslop. I don't need to hear two chatbots in overly enthusiastic, uncritical, unimaginative conversation.
AGI gets thrown around a lot. So, for the @thebulletin.org I spoke to researchers and took a look at the term and reality of AI that could rival human intelligence. Turns out it's a better tool for raising capital than a scientific milestone.
Jessica McKenzie
A Lancet correspondence described how over a three-year period, 4,046 references in 2,810 published scientific journal articles had been fabricated, many likely by AI.
Without data about what has happened in the past, and is happening in the present, we are committed to navigating a perilous and uncertain future while flying blind.
"Right now we're spending more on tokens for our internal agents than we are on employee head count," Mercor's CEO Brendan Foody said. https://bit.ly/4a6pJAE
Business Insider
"I would bet that in five years the average enterprise spends more on compute than headcount," Mercor's CEO Brendan Foody said.
