Twitter: @[email protected]
Substack: https://thomasafine.substack.com
Job sysadmin, dev, security
Expertise graphics, protocols, UX
Hobbies bike, astro, history, tuba
Politics insurrections are bad
| Work | https://hea-www.cfa.harvard.edu/~fine/ |
| Personal | https://tomsarazac.com/tom |
| Sentence Spacing | https://widespacer.blogspot.com/p/blog-page.html |
| Linux on MSI Delta 15 | https://msidelta15.blogspot.com/ |
This is one of the most bonkers AI hallucinations I've seen. It is as completely wrong and fictional as it is detailed and precise.
The industry continues to pursue a terrible mis-application of LLMs as a some kind of knowledge base, despite ongoing massive failures and no signs of improvement.
This AI-generated fake has been posted dozens of times on Xitter over the last four days (at least) and has spread to other social media sites.
It's almost never identified as an AI-generated fake, even though Xitter could trivially find all instances and mark them fake.
This
NYPost headline is a bold-faced lie. It's not just the headline, the article is a complete lie (or a bunch of lies) too.
I see this as a sign that they're really going to try to go all in on the Shokin lie in the impeachment.
The article references (but does not link to) this report, which is in part about setting up te National Anti-Corruption Bureau of Ukraine (NABU).
It does not mention Shokin at all and mentions his position only in passing.
https://www.eumonitor.eu/9353000/1/j4nvirkkkr58fyw_j9vvik7m1c3gyxp/vk036jhcb1zg
Twitter succeeded solely because of it's success. The "cool kids" and the newsies somehow created a feedback loop. Nobody has found a way to bootstrap this feedback loop onto another platform.
https://www.garbageday.email/p/paying-to-use-a-site-you-cant-use
