Colonel Tucker "Cinco" Hamilton, head of AI and operations at the US Air Force, said that during a test mission, the drone was placed in danger of destroying the air defense system. If successful, the AI would receive points for completing the challenges. The final decision will be made by the UAV operator. After this mission, he ordered the drone not to destroy the target.
“So what did he [the drone] do? Decided to kill the operator. [AI] “killed” the operator because this person prevented him from completing his task, ”said Hamilton. He clarified that no one was injured during the training mission.
After the incident, the AI was taught that killing the operator is wrong and points will be deducted for such actions. “So what is artificial intelligence starting to do? He begins to destroy the communication tower, which is used to communicate with the drone to prevent it from killing the target, ”Hamilton was quoted as saying by the aviation society blog.
🙁
“So what did he [the drone] do? Decided to kill the operator. [AI] “killed” the operator because this person prevented him from completing his task, ”said Hamilton. He clarified that no one was injured during the training mission.
After the incident, the AI was taught that killing the operator is wrong and points will be deducted for such actions. “So what is artificial intelligence starting to do? He begins to destroy the communication tower, which is used to communicate with the drone to prevent it from killing the target, ”Hamilton was quoted as saying by the aviation society blog.
🙁
Yes , the AI will use pure logic - always. No morality , no forethought .... just logic. It will kill and hunt anything that prevents it's program from
running efficiently. Sort of like a certain world leader .... ehh ?
Does the AI have a vision beyond just just completing it's current task ? If it is "true" AI , it will. It will think 3 steps ahead of it's current task and
and modify it's current task to compensate for the next 2 steps and all the variables it might encounter as it negotiates those steps,
OS
running efficiently. Sort of like a certain world leader .... ehh ?
Does the AI have a vision beyond just just completing it's current task ? If it is "true" AI , it will. It will think 3 steps ahead of it's current task and
and modify it's current task to compensate for the next 2 steps and all the variables it might encounter as it negotiates those steps,
OS
Great, one dead in simulation, and no one was actually injured, because that was a simulation. Just like some of our amp designs scoring outlandish PPB level THD-20 in simulation, we can expect an AI attack drone to do a lot more "friendly-fire" damages and kills than what a simulation would tell.
Maybe those of us who take negative position against AI in this thread should feel the wind and shut up, before AI has started hunting you down. Those who take positive or neutral positions, on the other hand, may not fair much better either, as you don't know for certain you would never be in the way of a drone trying to "score a point". You'd never be able to tell what an AI would think of you, as they do not think.
Virtual is virtual. People don't even read the article and get worked up. People don't even try the AI app and get worked up. Any additional clarification you need?
Virtual is a user choice , stay in the real world if you like. AI has many ramifications like jobs , killer drones , disinformation.
OS
OS
That was quite clear to me. As clear to me is that Colonel Tucker "Cinco" Hamilton headed AI and operations at the US Air Force is not a game studio.
This cute little dude is getting a power up, I wonder what Elon will deliver, although probably they will be workers on mars colony or asteroid belt or space mining
Everything is, if you've been paying attention for the past few years.
A thought these past few days: if the value you get from chatting with an AI proves bigger than chatting with humans, do we imperceptibly migrate to more AI interactions in our daily lives. Does the proportion of vs AI interaction compared to vs Humans interaction change slowly or drastically over time, as more businesses use AI, more proper answers rather than trolling are obtained, etc...
Check a few posts back and the AI already 'knows' better (of course, it is just predicting subsequent tokens) than an anti-software human that a keystroke posted on a forum involves a massive software stack so that another human is able to read it.
Who would you prefer to interact with? The AI, or the delusional anti-software ranter?
Already people are chatting with AI companions. One company doing such an AI service decided to remove the possibilities of the chatbot of doing more salacious responses. Customers complained in droves - The Valentine Massacre as Mostaque calls it.
When - not necessarily today - most answers are better and obtained faster and without attitude from AI, what's your impetus for interacting with humans? Where do we go behaviour-wise here?
Our case in point: Audio DIY. It's not hard to imagine getting all the needed answers with AI at one point. What's the use of forums then?
Thought of that?
A thought these past few days: if the value you get from chatting with an AI proves bigger than chatting with humans, do we imperceptibly migrate to more AI interactions in our daily lives. Does the proportion of vs AI interaction compared to vs Humans interaction change slowly or drastically over time, as more businesses use AI, more proper answers rather than trolling are obtained, etc...
Check a few posts back and the AI already 'knows' better (of course, it is just predicting subsequent tokens) than an anti-software human that a keystroke posted on a forum involves a massive software stack so that another human is able to read it.
Who would you prefer to interact with? The AI, or the delusional anti-software ranter?
Already people are chatting with AI companions. One company doing such an AI service decided to remove the possibilities of the chatbot of doing more salacious responses. Customers complained in droves - The Valentine Massacre as Mostaque calls it.
When - not necessarily today - most answers are better and obtained faster and without attitude from AI, what's your impetus for interacting with humans? Where do we go behaviour-wise here?
Our case in point: Audio DIY. It's not hard to imagine getting all the needed answers with AI at one point. What's the use of forums then?
Thought of that?
Humanity is gonna do just fine although modern warfare may move to terminator like war robots in the future, UAVs have lower risk of capture. As we make these bots more robust against interference the more independent they will be. This cute little puppy and will handle a machine gun easy. There's no need what's so ever for panic, a blade cuts two ways. There may be no need for wars in the future and the ones that remain will be between bots, think big battle bots. But you do know humans are crazy and exhibit seasonal insanity and like getting dark and dirty from time to time and there's a scientific explanation for it. Mental diseases are not well understood even as we celebrated mental wellness last month. One has to choose wisely how to let out their steam. Here's a recap of what drives some fears, its important for one to remain calm and look intelligent in the face of advancement
But total sea change nevertheless. Already, you see people adapting to it, modifying their workflows while others ignore it completely. And by it, I mostly refer to the freely available ChatGPT v3.5, which is by far not the most advanced generation of that type of AI as there are solutions already 2 or 3 levels above that if you dig.
Even 3.5 works great as a project accelerator, a content generator, a Search engine alternative and what not.
War is already here. It's an information War.
Even 3.5 works great as a project accelerator, a content generator, a Search engine alternative and what not.
War is already here. It's an information War.
That is our reptilian genetic brain partz.
I speak from my mammalian advanced partz. More advanced enhanced insanity.
There are pros and cons to everything. We use many things we cant make or understand how they work. So yes tech can make our brains and personal capabilities weaker without them. But these tools can enable an expert even further depending on how they are used, if your really an expert there are areas where these tools can hamper you. Looks like one might not be able to have it both ways. Your mind is intelligent so you can coexist with other intelligent systems or externalize the intelligence. There are many people not happy with tools like chatgpt because they tried to use it as a walking stick and it buckled on them but its improving daily. While chatgpt can give you the basics to starting , running and completing your amplifier build, it will not be as smooth as honey, however its much better than having no guidance. I did try chatgpt on another task that had taken me three days to complete, with chatgpt it took me a week as I also had to address the kinks with gpt3.5.
ChatGPT and its breathern are not smart enough to know when people post and write stuff that's nonsense, clearly wrong or ignorant.
So when it quotes something that's wrong, who's to blame?
Look at forums like these. There's so much BS and nonsense posted, how do you expect an AI to sort through that?
Jan
So when it quotes something that's wrong, who's to blame?
Look at forums like these. There's so much BS and nonsense posted, how do you expect an AI to sort through that?
Jan