AI and GB's of prepper data on an old phone?

[Moab]

"I wonder if we are at the point where you can load an AI client in an old phone?

That would be pretty powerful coupled with 11gbs of source prepper data to search and use within seconds.

That's always been the one limiting factor about all those prepper files. Finding what you need and using it. AI could turbo charge that into a preppers dream."

I write that in a response to a topic about e readers. And I could go ask AI right now. But for the sake of conversation. Thought I would post this.

If that's possible. Even in a tablet or laptop. That would make all those gb's of prepper data we've been hoarding really usable. 

The thing with the phone is you would probably have plenty of room to load the gbs of source data. But if not, the older phones with external SD cards would be a valuable choice. You could even load YouTube videos.

[Uomo Senza Nome]

I see AI more like Cliff Claven then something useful.

[Moab]

I see AI more like Cliff Claven then something useful.

Have you spent any time with it?

[Uomo Senza Nome]

Every day. I ask it questions. It is invariably wrong every time in some regard. Some of the stories are told well, explanations seem legitimate. But just like Cliff Claven, very often wrong.  Usually it does the best when keeping an answer short. But it seems built to want to impress us with its detailed knowledge. That is where it stumbles.

In criminal statements it is often the details that trip people up. People often try to make something appear more credible by putting extra details into their statements when they are lying. All this does is raise red flags and make it more suspect. 

I'm not saying it is never right, just wrong way more often, 70-80% of the time, about something.

[Moab]

Every day. I ask it questions. It is invariably wrong every time in some regard. Some of the stories are told well, explanations seem legitimate. But just like Cliff Claven, very often wrong.  Usually it does the best when keeping an answer short. But it seems built to want to impress us with its detailed knowledge. That is where it stumbles.

In criminal statements it is often the details that trip people up. People often try to make something appear more credible by putting extra details into their statements when they are lying. All this does is raise red flags and make it more suspect.

I'm not saying it is never right, just wrong way more often, 70-80% of the time, about something.

I would look into your queries.

I used it recently to negotiate a new vehicle purchase in a foreign country - Phillipines. It read, analyzed and understood Phillipine consumer law, common dealer practices here, and can read an analyze any contract or agreement you screenshot and upload. It was like having 5 experienced Phillipine attorneys and a secretarial pool of 29 in my living room. Except faster. I could have not done this without it. And after using it will never sign any document again without screenshotting it and having AI analyse it. I wish I had used it when I signed my lease.

Does it make mistakes? Yes absolutely. You still have to double check it's works. But if I added the time it would take me to research and analyze a given subject. Versus having AI do it. It's easily hundreds of hours to one.

Grok for instance is excellent at locating manufacturers in Alibaba. Selling specific products. It's a terrible, giant database. Being able to identify 3 manufacturers of a gas can I was looking for today, in about 39 seconds, cut out hours of research on my own.

Limiting it's access to data like within NotebookLM or running an AI client on a set of data you control. Like prepping data would refine it even further. This takes much of the "slop" out. Random runs in information maybe not related to your focus. If it set it up on the 11gb if data I have collected in the prepper realm your own ability to do anything with that vast amount of info has been turbo charged. Can you imagine how long it would take just to catalogue all that information? A ridiculous amount of time. AI can do that in a matter of seconds. That might take you and I years. At least many many months.

The other key attribute I have found extremely useful is its ability to write. Each step thru my vehicle purchase usually included an incorrect agreement. Chatgpt could quickly point out the key inconsistencies. Which I could verify without having to over analyse the third version(?) of a document. Chatgpt would then ask if I wanted it to write a follow up email and what tone I wanted to take. It would right a page and a half email within seconds, I could adjust anything including tone or specific words or phrasing. The level of writing was far above my own. And with tweaking it could come to the axact tone I wanted to use. This is within seconds. It would have taken hours to write something this polished.

I think you might want to reaxamine your use cases and queries. To say it's wrong 70% of the time. Seems your missing something.

AI is not perfect. But perfect enough that anyone can tell most jobs will be gone in 5 years. Probably far less. Hundreds of corporations worldwide are having massive layoffs already. Once this is coupled with robotics which is already rolling out at Amazon and other large corps. The question is going to become how do we feed and house the millions of people that had previously been involved in the work economy?

According to the International Labour Organization (ILO):

Roughly 70% of jobs are mental and 30% are physical labor.

AI already reads medical scans as well as a doctor. How long before you think it takes over half of the worlds jobs?

I think we're going to see a future full of dormatory type apartment buildings with cafeterias and medical centers. Funded by large corporations or governments taxing those corporations. Just to take care of a populace that is otherwise starving and rioting.

Pretty bleak. But I don't see anyone stopping corporations from using AI. When it never gets sick, never has to go the bathroom, never sexually harrasses anyone, never needs to sleep or take a day off, and costs little to nothing compared to human time.

I think you should go back and look at AI again. I don't think your seeing it clearly.

[Uomo Senza Nome]

Fair enough . This morning I asked it:

"Tell me about the safety of the Sig P320 Pistol and the US Airman who was killed?"

This is an easy slow ball since there has been a lot of current media coverage about it.


It correctly identified the airman killed, calls for it's recall, the US Air Force suspense of its use and the allegations of the pistol being defective as well as Sig denials that there was anything wrong with the pistols.

If I wanted that answer I could have gone to a gun shop and listened to people hanging around talking.

Missing from the AI story was the suicide of the Airman who killed the other Airman and the court martial of two Airmen who had written false statements and then recanted.

The real story being that the suicide airman had pulled his pistol out and pointed it at the soon to be dead Airman as a joke and the pistol "went off".  Also that the Air Force had returned the pistol to service was missing. That there was shown to be nothing wrong with the P320 after all. The Airman had fabricated the story to cover it up what was basically a negligent homicide.

Lots of articles out there mention these things but AI doesn't do the work and gives and very incomplete answer. So yep, Cliff Claven or even "gun shop guy". Take your pick.


https://taskandpurpose.com/news/airmen-guilty-statements-fe-warren-m18/

Really this was a lost opportunity to yet again go over the rules of gun safety and why it is the worst idea ever to break them. Here we have the death of two Airmen and the career destruction of another two.

[Moab]

Fair enough . This morning I asked it:

"Tell me about the safety of the Sig P320 Pistol and the US Airman who was killed?"

This is an easy slow ball since there has been a lot of current media coverage about it.


It correctly identified the airman killed, calls for it's recall, the US Air Force suspense of its use and the allegations of the pistol being defective as well as Sig denials that there was anything wrong with the pistols.

If I wanted that answer I could have gone to a gun shop and listened to people hanging around talking.

Missing from the AI story was the suicide of the Airman who killed the other Airman and the court martial of two Airmen who had written false statements and then recanted.

The real story being that the suicide airman had pulled his pistol out and pointed it at the soon to be dead Airman as a joke and the pistol "went off".  Also that the Air Force had returned the pistol to service was missing. That there was shown to be nothing wrong with the P320 after all. The Airman had fabricated the story to cover it up what was basically a negligent homicide.

Lots of articles out there mention these things but AI doesn't do the work and gives and very incomplete answer. So yep, Cliff Claven or even "gun shop guy". Take your pick.


https://taskandpurpose.com/news/airmen-guilty-statements-fe-warren-m18/

Really this was a lost opportunity to yet again go over the rules of gun safety and why it is the worst idea ever to break them. Here we have the death of two Airmen and the career destruction of another two.

I can tell I'm not going to convince you. Lol. And I can't say I think AI won't have a devastating affect on our society worldwide. I am honestly concerned about its missuse. And the sheer lack of forethought about safety and political indoctrination. Chatgot has already had various liberal politics programmed into it. Grok is perhaps more balanced. 

But I use neither of these tools for political research. Or anything beyond practical tasks. Like deciphering contracts. Figuring out mechical problems. Sourcing product solutions. Things of a more logicical, scientific or practical activity that would take me much more time to do manually. If I could do them at all. 

I also use different AI for different things. I'm a pragmatist. This is a tool. And if you look at it just as a tool. It's uses are powerful enough to drastically change industry and business if nothing else. 

I would not turn to AI for a nuanced opinion about gun control. It's not a person. But if you wanted more nuanced information. Say the additional facts you mention about the Sig case. That could have easily been handled with follow up questions. Follow up or refining questions are essential to it's use. To say "it missed this important thing" is sort of like admitting you haven't fully understood how to use it. Because it's only as good as the questions you ask. Like a professor sitting in your living room. If you want to know what he knows - it's all about the questions. You can't fault a professor for his answers if you are the uneducated one asking the questions. 

Not meant as a slam at all. Don't get me wrong. Just trying to point out your basing an opinion on AI not being any good. Based in an answer to a limited question you asked. Without - at least it sounds like - many follow up questions. 

But again. I understand your anymosity. AI is better than a certain lower percentage of the human populace at most of the things they do for a living. That sucks. Because industry now has a replacement for the human. The perfect slave. They don't have house, feed or pay it. And it never stops working. That is going to rearrange the world. 

I'm not saying AI is a good thing. In alot if ways it's not. But you have to be able to see it's very powerful uses. Otherwise it's kind of like saying "The telephone is bullshit.  You should get off your ass and write a letter!". Or "this internet thing is a kids play toy! Get back to work!".  Lol! 

I'm an AI neophyte too. Keep that in mind. My uses and arguments for it are lame. I'm sure. I know jack shit about its full spectrum of uses. I'm using it to buy a car, find good deals, do research. Real users are using it to write software, read brain scans, I'm an idiot by comparison. Lol. So take my words with a huge grain of salt. For and against. 

But back to the subject at hand. If I had more time right now. I'd download an AI agent. Plug it into my very loose, very unorganized, probably full of duplicates prepper data dump. And ask it some questions. 

Because that pile of data is just that - a pile. Ya. Maybe if you had a first aid problem is be able to find a few first aid manuals in that 11gbs. And find an index and a page that explains what I should do in whatever situation we are in. 

But AI could catalogue that entire collection, remove dupes, and start analysing it for any number of complex issues you might come across in a paw. How should I build a large water treatment system given the supplies I have on hand? Based in whatever documents you can find in that 11gbs? 

Can you imagine how long that would take you and I to do? 

And it's "you and I"! Lmao! We're a couple of slow, stupid humans.

Even if you had a professor of biology next to you. How long would it take him to sift thru 11gb of data? Analyze it, read it, develop ideas and come up with a solid plan that is better than most? Days, weeks, months? 

AI could do that in a matter of minutes at most. Further refining it with additional queries and tweaks. You might have a very doable system within a few hours.  

Imagine all the other tasks just like that? Imagine greater databases of scientific information that you might be able to give it access too?

The difference between my old cellphone holding an SD card of prepper data. Versus a laptop running an AI agent and just that same 11gb of data is monumentally more effective as a tool after a major disaster. 

Forget about how good it is or how bad it is for society. A prepper AI seems easily within our grasp. And light-years ahead of saving pdf's. 

Don't you agree?

[Uomo Senza Nome]

I wasn't asking a gun control question. I was asking whether a certain gun was safe and specifically in regards to a particular case. It provided a wrong answer. There isn't much else to talk about. Except that AI always gives wrong answers. 

If I didn't already know the answer How would I know I needed to ask it more questions? If I were dumb and lazy I would run with whatever it told me. If I needed to ask it a bunch of questions I may as well do my own research. The truth is I know how to be accurate and concise and AI doesn't and isn't. Therefore it shouldn't be relied on. 

Just because there is large volume of room for improvement doesn't mean it will always suck.  But right now? I deem it very unreliable. I'm not the only one either. I work in education environment and the teacher always knows when AI was used for a product. They usually get very low grades. Maybe at some future date things will improve. Till then it is just entertaining and that is as far as it goes.