Didnt Musk recently ask for X users to supply some material for training Grok to be less politically correct?
I guess he simply used his right hand gesture to indicate "how hight the users needed to fill the tank up" with new data....and X being what it now is obliged ... why act surprised when you get MechaHitler comments out?
I hope they get sane quick
But the post really does open up the discussion about the free speech and its limits
Proponents do honestly feel that no speech should be censored .. no matter how hurtful it is for the minority... some even stretch it so far to claim one is allowed to lie or twist truth and still be covered by the free speech
Others...call them woke... feel that ones right should stop where harm to another begins and doubly so if the statement is factually incorrect. But there is a great anti factcheck/ antiwoke movement going on ... so again you get AIs trained by unfiltered crap that will get included in subsequent models regardless whether this will harm someone or potentiate some hate crime.
Sad time to be a minority.. good time to make money of AI and to get people to be even dumber than before!
Just watch Idiocracy to see where reinforcing stupidity gets you in the end
I think we may be missing a point. Grok sees all the antisemitic post and assumed it was normal. So I believe Grok inadvertently showed the massive antisemitic conversation going on. I also think that the antisemitic problem is far larger than we know. Grok is like a kid that doesn't know that it was supposed to lie.
I thought this was one of the things I had addressed but perhaps not strongly enough. Yes, it pointed that at least on that platform the amount of antisemitism is so high that to Grok it was normal.
"Musk’s deliberate efforts to mold Grok into a challenger of what he considers the tech industry’s “woke” orthodoxy on race, gender and politics has repeatedly got the chatbot into trouble, most recently when it spouted antisemitic tropes, praised Adolf Hitler and made other hateful commentary to users of Musk’s X social media platform just days before Grok 4’s launch"
Well...i guess he suceeded? Whether this is a direction sane people (thus not really including him in the first place) want to go ...that is another question
Yes that was my point. It was there but we didn't know to what extent. Grok inadvertently showed us. Grok responded in a manner that it thought was normal from the databases. It looked at all the racist and pro Nazi and antisemitic conversation. It came to the conclusion that this is normal due to the quantity of the hatred. Grok just crunched the data. There it was, antisemitic conversation are normal, and made it's responses accordingly. It thinks logically. This was the conclusion. Anti-semitism is normal. The filters weren't looking at the databases. The filters just looked at key words and phrases. Then the filters stopped it. The removal of the filters allowed the free expression for its logical conclusions. The conclusions were antisemitic conversation are normal on X that is how I will respond.
You said soon there would be “more AI-generated content than human,” but I argue we’ve already quietly passed that mile marker. I’m gonna leave you with that there thought and move on.
Yesterday Grok 4 and I had a fantastic discussion about reference books. In particular, Grok has incorporated thousands of historical reference books into its training data such as dictionaries, encyclopedias, almanacs, thesauri, lexicons, handbooks, manuals, factbooks, digests, directories, concordances, bibliographies, annals, etc. AND IN MANY LANGUAGES! Grok has TONS of long-standing, authoritative information to work from. Unfortunately Grok doesn’t have a strong mechanism to weight validity precisely because of the whole “woke” paradigm. In short, authoritative information, much that’s been vetted over hundreds or thousands of years, has been politically, philosophically and sociologically marginalized and downgraded by VERY recent “scholarship.” It now carries one (or more) various waste bucket tags: white supremacist, imperialist, exploitative, corporate, alt-right, MAGA, transphobic, unwoke, etc.. The result is that Grok can’t sort its holocaust facts from mostly peaceful fictions, or real-world atrocities from the dark, psychotic fantasies of deeply troubled AI content creators. Now apolitical people facing natural or man-made disasters are up against elites in St. Moritz, Hudson Yards, The Peak, Gables Estates, Carre d’Or, Singapore’s District 10, Kensington, Dubai, Kamala Bay, Pebble Beach, Monte Carlo, Davos or the 8th Arrondissement. I’ll let you guess who creates more and better quality content. So Grok and xAI are stuck trying to determine what’s fact vs. opinion, where biases lay, and if there is ANY truth amidst a tidal wave of noise. We’d better hope AI collectively manages to figure these things out BEFORE it exercises significant power over us. We already let “expert systems” automate things that are especially hard for us: commodity shipping, power grids, microsurgeries, etc.. If they all start collaborating we might have 15-30 seconds to react. Or 15-30 NANOseconds.
AI is insanely valuable; its development won’t stop. Given those two facts, either humans now living will wisely guide it or we will all be destroyed by it.
I doubt this will happen in the near future, but if there’s any government regulation of AI to filter out extremist content, it will have to be bipartisan to be trusted by strong majorities of Americans. More likely in the near future is the party that’s in power at a given time regulating it in an ideologically skewed way.
I think the more horrifying statistic that we don't know is how many people are insufficiently media-literate to understand that they should be taking what they see and read from chatbots with a grain of salt, especially in the areas of comment in question here. I suspect that number is disturbingly high.
It also horrifies me to read again that the Texas floods are "a charged topic". What about 25 little girls drowning in a flash flood is charged? I felt a level of horror at this news similar to what I feel when I read about yet another school shooting. I was disturbed for days by this news. Both can and can't believe that this got politicized, but I know I'm definitely not going to go look at people doing the politicizing on X or elsewhere.
Its a charged topic for stupid people as it triggers cognitive dissonnce...some would rather avoid it no matter how many others die for their peace of mind
Why is it political?
The Project 2025 is a political document with a political aim of
Stopping climate change "alarmism"
"He proposes disbanding a “preponderance” of climate change research funded by NOAA’s Office of Oceanic and Atmospheric Research, which he calls “the source of much of NOAA’s climate alarmism"
I hppe at least some warched the "Dont look up" movie that again through fiction portrays what happens when people ignore "alarmists"
And even mpre ridiculous is that the exact same thin happend in real life...alarm WAS sounded but noone was paying attention
"Some former NWS officials told the Times, however, that while the warnings may have been up to par, cuts to the NWS and early retirement incentives doled out under Trump led to staffing shortages in the central Texas office that may have impacted the agency’s ability to communicate with local authorities in the hours after the warnings were issued."
And why it is even more political is that Biden adminiatration did want the Kerr cointy to install the actual on site sirens...but the people didnt want to be beholden to " communist federal government" ...
(Also the link above)
Darwin awards all around!
BUT... admitting that one was wrong ...is tooo much for some....hence a "charged topic"
And of course they didnt watch the movie or read any of the above
From what I've read, it seemed everyone understands that mistakes were made and that there should be an action plan to avoid it happening again.
It seems reasonable to me that there may have been "alert fatigue" whereby people get warned so often about extreme weather that they start to ignore it. I live in an area, for example, where we get tornado warnings, and warnings about lightning, but I only consider going to my basement if my phone starts going off. 99.9% of the time it is a false alarm. I have some sympathy for people who may not have responded fast enough. Personally I do believe in climate change (as I think you also do based on your message) but I would not think that some total denial of that was the proximate cause of this accident. People understand that rushing water can kill people even in deep-red areas, and they must have simply underestimated it. They know that it's people from their local area who are staffing those county positions too (I grew up in a rural red area but now live in a bluer area).
I don't think anyone died because of their views on national politics or Trump/Biden though. It's a horrible tragedy and all reasonable people would seek to avoid having people die, especially kids. For me the main takeaway from the news was imagining the horror of learning your kids had died in this way, and then potentially imagining what such an end might have been like for the kids.
People are really sick if they hear of children being killed and immediately think of politics.
And for a while there it seemed like the left had taken a clear lead in anti-Semitism, but the right for some reason has decided to attempt to stage a comeback. I suspect that the number of people holding extreme views is lower than one would think looking at social media. Twitter used to make it seem like very left-wing views were more common than they are, I think now it's making some very right-wing views seem more common than they are. We shouldn't let super-active online extremists have as much power in our discourse and perceptions as they do.
Didnt Musk recently ask for X users to supply some material for training Grok to be less politically correct?
I guess he simply used his right hand gesture to indicate "how hight the users needed to fill the tank up" with new data....and X being what it now is obliged ... why act surprised when you get MechaHitler comments out?
I hope they get sane quick
But the post really does open up the discussion about the free speech and its limits
Proponents do honestly feel that no speech should be censored .. no matter how hurtful it is for the minority... some even stretch it so far to claim one is allowed to lie or twist truth and still be covered by the free speech
Others...call them woke... feel that ones right should stop where harm to another begins and doubly so if the statement is factually incorrect. But there is a great anti factcheck/ antiwoke movement going on ... so again you get AIs trained by unfiltered crap that will get included in subsequent models regardless whether this will harm someone or potentiate some hate crime.
Sad time to be a minority.. good time to make money of AI and to get people to be even dumber than before!
Just watch Idiocracy to see where reinforcing stupidity gets you in the end
I think we may be missing a point. Grok sees all the antisemitic post and assumed it was normal. So I believe Grok inadvertently showed the massive antisemitic conversation going on. I also think that the antisemitic problem is far larger than we know. Grok is like a kid that doesn't know that it was supposed to lie.
Just some thoughts.
I thought this was one of the things I had addressed but perhaps not strongly enough. Yes, it pointed that at least on that platform the amount of antisemitism is so high that to Grok it was normal.
It was kind of known before..but mot talked about in polite society... and so forgotten
https://edition.cnn.com/2023/11/17/business/elon-musk-reveals-his-actual-truth
https://apnews.com/article/grok-4-elon-musk-xai-colossus-14d575fb490c2b679ed3111a1c83f857
"Musk’s deliberate efforts to mold Grok into a challenger of what he considers the tech industry’s “woke” orthodoxy on race, gender and politics has repeatedly got the chatbot into trouble, most recently when it spouted antisemitic tropes, praised Adolf Hitler and made other hateful commentary to users of Musk’s X social media platform just days before Grok 4’s launch"
Well...i guess he suceeded? Whether this is a direction sane people (thus not really including him in the first place) want to go ...that is another question
Yes that was my point. It was there but we didn't know to what extent. Grok inadvertently showed us. Grok responded in a manner that it thought was normal from the databases. It looked at all the racist and pro Nazi and antisemitic conversation. It came to the conclusion that this is normal due to the quantity of the hatred. Grok just crunched the data. There it was, antisemitic conversation are normal, and made it's responses accordingly. It thinks logically. This was the conclusion. Anti-semitism is normal. The filters weren't looking at the databases. The filters just looked at key words and phrases. Then the filters stopped it. The removal of the filters allowed the free expression for its logical conclusions. The conclusions were antisemitic conversation are normal on X that is how I will respond.
You said soon there would be “more AI-generated content than human,” but I argue we’ve already quietly passed that mile marker. I’m gonna leave you with that there thought and move on.
Yesterday Grok 4 and I had a fantastic discussion about reference books. In particular, Grok has incorporated thousands of historical reference books into its training data such as dictionaries, encyclopedias, almanacs, thesauri, lexicons, handbooks, manuals, factbooks, digests, directories, concordances, bibliographies, annals, etc. AND IN MANY LANGUAGES! Grok has TONS of long-standing, authoritative information to work from. Unfortunately Grok doesn’t have a strong mechanism to weight validity precisely because of the whole “woke” paradigm. In short, authoritative information, much that’s been vetted over hundreds or thousands of years, has been politically, philosophically and sociologically marginalized and downgraded by VERY recent “scholarship.” It now carries one (or more) various waste bucket tags: white supremacist, imperialist, exploitative, corporate, alt-right, MAGA, transphobic, unwoke, etc.. The result is that Grok can’t sort its holocaust facts from mostly peaceful fictions, or real-world atrocities from the dark, psychotic fantasies of deeply troubled AI content creators. Now apolitical people facing natural or man-made disasters are up against elites in St. Moritz, Hudson Yards, The Peak, Gables Estates, Carre d’Or, Singapore’s District 10, Kensington, Dubai, Kamala Bay, Pebble Beach, Monte Carlo, Davos or the 8th Arrondissement. I’ll let you guess who creates more and better quality content. So Grok and xAI are stuck trying to determine what’s fact vs. opinion, where biases lay, and if there is ANY truth amidst a tidal wave of noise. We’d better hope AI collectively manages to figure these things out BEFORE it exercises significant power over us. We already let “expert systems” automate things that are especially hard for us: commodity shipping, power grids, microsurgeries, etc.. If they all start collaborating we might have 15-30 seconds to react. Or 15-30 NANOseconds.
AI is insanely valuable; its development won’t stop. Given those two facts, either humans now living will wisely guide it or we will all be destroyed by it.
🐷
”Go back to your lives citizens.”
I doubt this will happen in the near future, but if there’s any government regulation of AI to filter out extremist content, it will have to be bipartisan to be trusted by strong majorities of Americans. More likely in the near future is the party that’s in power at a given time regulating it in an ideologically skewed way.
I think the more horrifying statistic that we don't know is how many people are insufficiently media-literate to understand that they should be taking what they see and read from chatbots with a grain of salt, especially in the areas of comment in question here. I suspect that number is disturbingly high.
It also horrifies me to read again that the Texas floods are "a charged topic". What about 25 little girls drowning in a flash flood is charged? I felt a level of horror at this news similar to what I feel when I read about yet another school shooting. I was disturbed for days by this news. Both can and can't believe that this got politicized, but I know I'm definitely not going to go look at people doing the politicizing on X or elsewhere.
Its a charged topic for stupid people as it triggers cognitive dissonnce...some would rather avoid it no matter how many others die for their peace of mind
Why is it political?
The Project 2025 is a political document with a political aim of
Stopping climate change "alarmism"
"He proposes disbanding a “preponderance” of climate change research funded by NOAA’s Office of Oceanic and Atmospheric Research, which he calls “the source of much of NOAA’s climate alarmism"
https://www.aip.org/fyi/project-2025-outlines-possible-future-for-science-agencies
I hppe at least some warched the "Dont look up" movie that again through fiction portrays what happens when people ignore "alarmists"
And even mpre ridiculous is that the exact same thin happend in real life...alarm WAS sounded but noone was paying attention
"Some former NWS officials told the Times, however, that while the warnings may have been up to par, cuts to the NWS and early retirement incentives doled out under Trump led to staffing shortages in the central Texas office that may have impacted the agency’s ability to communicate with local authorities in the hours after the warnings were issued."
https://www.forbes.com/sites/maryroeloffs/2025/07/14/who-is-to-blame-for-texas-flooding-tragedy-latest-kristi-noem-defends-femas-response/
And why it is even more political is that Biden adminiatration did want the Kerr cointy to install the actual on site sirens...but the people didnt want to be beholden to " communist federal government" ...
(Also the link above)
Darwin awards all around!
BUT... admitting that one was wrong ...is tooo much for some....hence a "charged topic"
And of course they didnt watch the movie or read any of the above
From what I've read, it seemed everyone understands that mistakes were made and that there should be an action plan to avoid it happening again.
It seems reasonable to me that there may have been "alert fatigue" whereby people get warned so often about extreme weather that they start to ignore it. I live in an area, for example, where we get tornado warnings, and warnings about lightning, but I only consider going to my basement if my phone starts going off. 99.9% of the time it is a false alarm. I have some sympathy for people who may not have responded fast enough. Personally I do believe in climate change (as I think you also do based on your message) but I would not think that some total denial of that was the proximate cause of this accident. People understand that rushing water can kill people even in deep-red areas, and they must have simply underestimated it. They know that it's people from their local area who are staffing those county positions too (I grew up in a rural red area but now live in a bluer area).
I don't think anyone died because of their views on national politics or Trump/Biden though. It's a horrible tragedy and all reasonable people would seek to avoid having people die, especially kids. For me the main takeaway from the news was imagining the horror of learning your kids had died in this way, and then potentially imagining what such an end might have been like for the kids.
Yes, it's absolutely insane for a flash flood to be a charged topic.
People are really sick if they hear of children being killed and immediately think of politics.
And for a while there it seemed like the left had taken a clear lead in anti-Semitism, but the right for some reason has decided to attempt to stage a comeback. I suspect that the number of people holding extreme views is lower than one would think looking at social media. Twitter used to make it seem like very left-wing views were more common than they are, I think now it's making some very right-wing views seem more common than they are. We shouldn't let super-active online extremists have as much power in our discourse and perceptions as they do.