home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   alt.survival      Discussing survivalism for end-times      131,158 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 131,029 of 131,158   
   Doctor Fill to All   
   Re: Study : Current AI Systems UNSAFE Fo   
   12 Nov 25 06:23:31   
   
   XPost: talk.politics.misc, alt.politics, alt.fan.rush-limbaugh   
   From: df@tutanato.com   
      
   On 11/12/2025 12:23 AM, c186282 wrote:   
   > On 11/11/25 23:57, Doctor Fill wrote:   
   >> On 11/11/2025 9:06 PM, c186282 wrote:   
   >>> On 11/11/25 22:50, Doctor Fill wrote:   
   >>>> On 11/11/2025 8:09 PM, c186282 wrote:   
   >>>>> On 11/11/25 20:36, Doctor Fill wrote:   
   >>>>>> On 11/10/2025 7:59 PM, c186282 wrote:   
   >>>>>>> https://techxplore.com/news/2025-11-popular-ai-ready-safely-   
   >>>>>>> power.html   
   >>>>>>>   
   >>>>>>> Popular AI models aren't ready to safely power robots, study warns   
   >>>>>>>   
   >>>>>>> Robots powered by popular artificial intelligence models are   
   >>>>>>> currently unsafe for general purpose real-world use, according   
   >>>>>>> to new research from King's College London and Carnegie   
   >>>>>>> Mellon University.   
   >>>>>>>   
   >>>>>>> . . .   
   >>>>>>>   
   >>>>>>>    They easily lie, cheat, confabulate. They are very   
   >>>>>>>    biased when it comes to social status, race, religion.   
   >>>>>>   
   >>>>>> Bias is not necessarily "unsafe" or a defect of some sort.   
   >>>>>   
   >>>>>    IS if you're on the downside of that bias ....   
   >>>>>   
   >>>>>    AIs will soon be used to decide if you're worthy   
   >>>>>    for most everything. Human bosses love that, just   
   >>>>>    blame the machine when there's a big fuck-up.   
   >>>>>   
   >>>>>> A "consumer" might be "biased" toward buying cars known for   
   >>>>>> quality and reliability over cars known for electrical problems   
   >>>>>> and frequent recalls.  One might be biased not to go into   
   >>>>>> dangerous parts of cities and, perhaps, note that a particular   
   >>>>>> ethnic group or race is associated with crime and violence.  It   
   >>>>>> would be rather foolish to say all biases are wrong headed, even   
   >>>>>> if it's a robot that holds the biases.   
   >>>>>   
   >>>>>    A "race" or whatever may have better or worse crime   
   >>>>>    stats or such, but an INDIVIDUAL of said groups is   
   >>>>>    not necessarily what the bias would imply. All should   
   >>>>>    get a fair shake in court, at the bank, at the job.   
   >>>>>   
   >>>>>    Sorry, but it sounds as if you're into 'mechano-racism'   
   >>>>>    here - NAZI aims 'justified' by machine 'intelligence'.   
   >>>>>   
   >>>>   
   >>>> I fail to see the benefit of programming woketardedness into AI   
   >>>> controlled robots.  If the AI is crippled by wokeness and there are   
   >>>> virtual gatekeepers second guessing the AI decision, then one ends   
   >>>> up with silicon versions of usenet trolls.   
   >>>   
   >>>    "Gatekeepers" can be SUED ... biz would like to   
   >>>    keep it ALL 'AI' so they can blame the machine.   
   >>>    Smart legalistics.   
   >>>   
   >>>    In short, NO 'gatekeepers' - the word of the AI   
   >>>    will be the alpha and omega.   
   >>>   
   >>>    Whatever 'group'/demo you may fall into, the AI   
   >>>    will be ruthless - and SCREW you if it's algos   
   >>>    decide so Just Because.   
   >>>   
   >>   
   >> The whole country or planet will be one big Gaza controlled by   
   >> Palentir like AI killing bots.   
   >   
   >   
   >    Ummmm ... read yer history ... alas that's kind of NORMAL ....   
   >   
   >    Westerners have become used to what 'The Enlightement' spawned.   
   >    However that's a historical abnormality. All documented history   
   >    paints a MUCH more severe picture ... a norm we now seem to   
   >    be drifting back to. "Rights" ? "Dissent" ? NO !   
   >   
   >    As for kill-bots ... they ARE working hard on those. Soulless   
   >    robo-cops. 10 years to proper products ? Less ???   
   >   
      
   A Mormon I do Bible studies with suggested that the 200+ million man   
   army that is prophesied in Book of Revelation to attack Jerusalem in the   
   end times is actually made up of AI controlled bots, perhaps of Chinese   
   origin.  His interpretation fits with the description of "insect like   
   creatures with human heads" that sting and torment humans.   
      
   --   
   "Title 8, U.S.C. § 1324(a) defines several distinct offenses related to   
   aliens. Subsection 1324(a)(1)(i)-(v) prohibits alien smuggling, domestic   
   transportation of unauthorized aliens, concealing or harboring   
   unauthorized aliens, encouraging or inducing unauthorized aliens to   
   enter the United States, and engaging in a conspiracy or aiding and   
   abetting any of the preceding acts. Subsection 1324(a)(2) prohibits   
   bringing or attempting to bring unauthorized aliens to the United States   
   in any manner whatsoever, even at a designated port of entry. Subsection   
   1324(a)(3)."   
      
   "Who could have predicted just a decade ago that Germany would have a   
   world championship winning basketball team, but be closing down its   
   breweries in record numbers? It’s a sobering reality indeed." -- Ron Unz   
      
   https://www.globalgulag.us   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca