In case you’re curious, The Guardian and Juno News both have the same rating for accuracy (“mixed factuality”) on @gandersocial.ca
July 2, 2026 at 3:51 AM
In case you’re curious, The Guardian and Juno News both have the same rating for accuracy (“mixed factuality”) on @gandersocial.ca
I was there can confirm factuality.
September 30, 2025 at 3:57 AM
I was there can confirm factuality.
lol “Media Bias Fact-Check” self-identifies as centrist and gives itself a “high factuality” rating, apparently indicating it is a more reliable source of information than The Guardian
July 2, 2026 at 12:45 PM
lol “Media Bias Fact-Check” self-identifies as centrist and gives itself a “high factuality” rating, apparently indicating it is a more reliable source of information than The Guardian
Again tapping the "the bar for factuality should be *higher* for opinion pieces than for reporting" sign bsky.app/profile/band...
Also this isn't Ai generated. they gave her laser eyes. Does the Washington Post not understand you can do all of that client side and you've been able to since the 90s at least?
July 17, 2026 at 3:11 AM
Again tapping the "the bar for factuality should be *higher* for opinion pieces than for reporting" sign bsky.app/profile/band...
the problem with large language models is what they do, not some collateral issue unrelated to the fact that they produce an arbitrary quantity of coherent text of questionable factuality
March 11, 2025 at 2:44 AM
the problem with large language models is what they do, not some collateral issue unrelated to the fact that they produce an arbitrary quantity of coherent text of questionable factuality
This is a hugely important story. I wrote about its consequences already in "Road to Unfreedom" and we know much more now. If we want to be a democracy we have to create structures of factuality rather than selling our future to the loudest yeller in another country.
Although the scale of this is still to be assessed, this is the inevitable outcome of the monetization of the algorithm. Rage-bait becomes a personal revenue model. Nothing about this should be a surprise.
NEWS: I spent the past 24 hours going through major MAGA accounts on Twitter, and a bombshell development has become clear: most of the right-wing ecosphere is being fed propaganda from foreign actors.
This impacts elections. This impacts discourse. This is major. Subscribe to support my work:
This impacts elections. This impacts discourse. This is major. Subscribe to support my work:
November 24, 2025 at 5:43 PM
This is a hugely important story. I wrote about its consequences already in "Road to Unfreedom" and we know much more now. If we want to be a democracy we have to create structures of factuality rather than selling our future to the loudest yeller in another country.
They quite literally cannot afford to indulge in post-factuality.
September 14, 2025 at 3:49 PM
They quite literally cannot afford to indulge in post-factuality.
Like you will have a better idea of what is actually occurring in reality reading a deranged lib medias touch type blog that's like "TRUMP CRASH OUT OVER POOL BOTCH" than reading the New York Times carefully report the President's claims as if they deserve the presumption of factuality
June 21, 2026 at 1:34 PM
Like you will have a better idea of what is actually occurring in reality reading a deranged lib medias touch type blog that's like "TRUMP CRASH OUT OVER POOL BOTCH" than reading the New York Times carefully report the President's claims as if they deserve the presumption of factuality
This is another fakie for sure. Let's try to apply basic rules of factuality to keep Bluesky rational. If you can't find the original, don't repost comments about something purporting to be real.
May 19, 2026 at 12:41 PM
This is another fakie for sure. Let's try to apply basic rules of factuality to keep Bluesky rational. If you can't find the original, don't repost comments about something purporting to be real.
Factuality: high
June 18, 2026 at 5:52 AM
Factuality: high
THE BALL ULTIMATELY CONTAINS FACTUALITY AND RENDERS OUTCOMES AS SUCH
July 2, 2026 at 1:49 AM
THE BALL ULTIMATELY CONTAINS FACTUALITY AND RENDERS OUTCOMES AS SUCH
ive said this before, but the dissonance of marketing AI as an absolute source of unlimited intelligence (the implication being some sort of factuality or neutrality), while a few very wealthy humans manually carve it to say the things they want, should be ringing alarm bells for everyone
The administration “is currently pressuring OpenAI and other AI companies to make their models more conservative-friendly.”
OpenAI is trying to clamp down on ‘bias’ in ChatGPT
GPT-5 is better at resisting liberal ‘pressure,’ the company says.
www.theverge.com
October 11, 2025 at 2:11 PM
ive said this before, but the dissonance of marketing AI as an absolute source of unlimited intelligence (the implication being some sort of factuality or neutrality), while a few very wealthy humans manually carve it to say the things they want, should be ringing alarm bells for everyone
yeah, this is the basic problem. what Anthropic is selling, to a substantial extent, is downstream of the floating point number representing estimated factuality in posttraining.
And, of note, facts are specifically excluded from being “owned” in the sense of IP (for very obvious reasons.)
June 2, 2026 at 4:50 AM
yeah, this is the basic problem. what Anthropic is selling, to a substantial extent, is downstream of the floating point number representing estimated factuality in posttraining.
Community Notes don't work by themselves. So many things don't get labeled because people don't reach "consensus" on them, which is really what Community Notes optimize for, not factuality
What an utter failure. So disappointing
What an utter failure. So disappointing
"We're going to get rid of fact-checkers and replace them with community notes, similar to X, starting in the U.S," Mark Zuckerberg announces
He asserts that "the fact-checkers have just been too politically biased, and have destroyed more trust than they've created, especially in the U.S."
He asserts that "the fact-checkers have just been too politically biased, and have destroyed more trust than they've created, especially in the U.S."
January 7, 2025 at 12:49 PM
Community Notes don't work by themselves. So many things don't get labeled because people don't reach "consensus" on them, which is really what Community Notes optimize for, not factuality
What an utter failure. So disappointing
What an utter failure. So disappointing
Any time I have imposter syndrome issues, I just remember people get jobs doing this. Zero historical context, zero factuality, zero awareness, just pure nonsense.
Gen Z has brought us the 6pm dinner – and I am not ready for it | Polly Hudson
Gen Z has brought us the 6pm dinner – and I am not ready for it | Polly Hudson
Early evening restaurant bookings are on the rise. Great for commuters and sleep purists, but what about the rest of us?
www.theguardian.com
August 5, 2025 at 7:32 PM
Any time I have imposter syndrome issues, I just remember people get jobs doing this. Zero historical context, zero factuality, zero awareness, just pure nonsense.
They would significantly improve the factuality of the Telegraphs reporting if they did.
February 27, 2026 at 3:00 PM
They would significantly improve the factuality of the Telegraphs reporting if they did.
“But LLMs aren’t necessarily the best approach to always get at factuality, “
LOL THEN WHY ARE YOU USING THIS BULLSHIT
LOL THEN WHY ARE YOU USING THIS BULLSHIT
I can't get over this interview from The Verge with Sundar Pichai - he made Google integrate generative AI into search despite hallucinations being "unsolved," because it makes them "very creative." I don't want Google to be creative!!! I NEED AN ANSWER!!!
www.theverge.com/24158374/goo...
www.theverge.com/24158374/goo...
May 25, 2024 at 6:46 PM
“But LLMs aren’t necessarily the best approach to always get at factuality, “
LOL THEN WHY ARE YOU USING THIS BULLSHIT
LOL THEN WHY ARE YOU USING THIS BULLSHIT
This is… not good. At the same time these models are being sold as universal answer machines, they’re being tuned for both-sides-ism instead of factuality.
The administration “is currently pressuring OpenAI and other AI companies to make their models more conservative-friendly.”
OpenAI is trying to clamp down on ‘bias’ in ChatGPT
GPT-5 is better at resisting liberal ‘pressure,’ the company says.
www.theverge.com
October 11, 2025 at 12:45 PM
This is… not good. At the same time these models are being sold as universal answer machines, they’re being tuned for both-sides-ism instead of factuality.
They spent 10X more resources and manpower on making that stupid poster board than they did on investigating its factuality.
September 29, 2026 at 11:35 PM
They spent 10X more resources and manpower on making that stupid poster board than they did on investigating its factuality.
these are people just looking to dunk, factuality be damned
April 18, 2026 at 3:42 AM
these are people just looking to dunk, factuality be damned
Gotta respect the chaotic energy of bots distorting discourse with fake pleas for civility, factuality and inclusivity. It’s toxic af but it has a fresh 2025 feel to it
January 19, 2025 at 8:17 AM
Gotta respect the chaotic energy of bots distorting discourse with fake pleas for civility, factuality and inclusivity. It’s toxic af but it has a fresh 2025 feel to it
This is how you report on these people. You state their claims and then state their factuality in the same post
September 16, 2025 at 1:49 AM
This is how you report on these people. You state their claims and then state their factuality in the same post
And of course there’s zero problem with the factuality of the claim that “iq is 50% genetic” as opposed to “intelligence is a polygenic trait and there is dispute about what exactly an IQ score measures” which would admittedly be less catchy
December 14, 2025 at 2:17 PM
And of course there’s zero problem with the factuality of the claim that “iq is 50% genetic” as opposed to “intelligence is a polygenic trait and there is dispute about what exactly an IQ score measures” which would admittedly be less catchy
AI is decadent, as least as it’s deployed in the liberal arts.
It’s extraordinarily expensive tech, producible only by exploiting vast amounts of unseen labor, to produce things that are not only ugly but with utter indifference toward their factuality. It exists solely to exist. That’s decadence.
It’s extraordinarily expensive tech, producible only by exploiting vast amounts of unseen labor, to produce things that are not only ugly but with utter indifference toward their factuality. It exists solely to exist. That’s decadence.
December 15, 2024 at 10:00 PM
AI is decadent, as least as it’s deployed in the liberal arts.
It’s extraordinarily expensive tech, producible only by exploiting vast amounts of unseen labor, to produce things that are not only ugly but with utter indifference toward their factuality. It exists solely to exist. That’s decadence.
It’s extraordinarily expensive tech, producible only by exploiting vast amounts of unseen labor, to produce things that are not only ugly but with utter indifference toward their factuality. It exists solely to exist. That’s decadence.
The first one is a common understandable mistake. The second one is baffling, bizarre even by Abbott standards and another damning indictment on the non-editing of op eds for factuality. AI?
July 19, 2026 at 4:08 AM
The first one is a common understandable mistake. The second one is baffling, bizarre even by Abbott standards and another damning indictment on the non-editing of op eds for factuality. AI?