I don't understand 98% of what was just outlined above. I don't understand how to do the things that are described, my brain does not work at that level. But what I do understand is that Commercial AI has so many guardrails and "programed in" biases, it make the sort of use that TWW outlines impossible. If you do not own and run an application locally, you really do not have a useful or durable tool. Right now, when the internet goes down, many are just screwed. "You will own nothing and be happy" fits this perfectly.
If I didnât have a local AI system, I would not be able to do this work anymore. The corporate AI ecosystem is so guardrailed now that if you even mention âIsraelâ â you get 3 paragraphs about anti-semitism and the holocaust. I tried asking Claude to research Jewish officers in the Nazi military a few months back and I am pretty sure a data center somewhere in California caught on fire. It freaked out about it. Kept insisting I was wrong, tried showing me some Snopes-style âfake newsâ articles. When I listed actual names and academics that have written about the topic, instead of agreeing it just ended the conversation and locked itself down.
This is how you know something is fishy. If corporate AI refuses to touch it, chances are itâs because itâs true.
You should see some of the âdiscussionsâ I have had with âMr. GPTâ. Same experience here, mention any number of âprotected classesâ and the LLM instantly loses its collective mind.
One positive is that I have become quite adept at identifying âAI Bullshit.â I sometimes bait the LLM deliberately just to see what comes out. One thing is certain, the output is 99.9% predictable. If nothing else the adversarial mental gymnastics keep the mind sharper.
Crawling into a cave is not good self-defense. Anybody younger than about 60 who fails to keep up with this technology is going to find themselves socially isolated or worse. Don't give up.
Wolf you wrote exactly what I need to read. I will definitely go through this whole article. Because I have been looking for a home AI set up, even considered building one, for months now. Thank you for sharing your expertise and experience. We do need to take control before they take it from us.
do it now while you still can because a few years from now who knows if we will even have access to uncensored ai anymore. this technology has the power to bring down the entire epstein class hegemony. they know it.
I have one of the most modern computers in the world. It runs Windows XP offline. I use a version of Maximizer (CRM) that doesn't charge a monthly subscription fee and uses btrieve which is still one of the fastest dbase engines. For graphics and writing I use Wordperfect which still beats Word hands down every day of the week and twice on weekends. Boot up is super fast. 30 seconds from start to fully operational OS. The best part is that it does everything I need it to do and with a rate of efficiency that makes $10,000 computers blush.
Thank you for this article. Conceptually I understood a good chunk of what you shared; however, the details of starting to go down the road of setting up my own localized open source AI machine feels a little overwhelming.
Can you recommend a resource or article link that gives a little bit more of a step by step guide for someone is wanting to start?
I appreciate your work and I'm proud to support it! Blessings to you
the easiest route now is just get a decent nvidia graphics card with at lest 16gb of vram and just run qwen 3.8 27b locally. it scores at 70% efficiency compared to models that require a 50 grand investment to run locally.
itâs what i use now. the strix halo system is overkill honestly but with it i can run 5 asynchronous instances of this model for agentic development so itâs still pretty useful f or me but if youâre just looking to have local ai superpowers at a hobbyist level, get a decent gpu. you can find them used on ebay for around 1200 bucks. still pricey but itâs way cheaper than paying for a graphic designer or programmer!
This was written for coders and corporations and I understand the need and the concern. As a consumer, it affects me peripherally so I hope y'all succeed. In the meantime, what does the average consumer do to defend against what you predict? Pray?
Thank you for this. I grabbed Qwen and a couple other models. I have been sitting on my butt wrt, but your article was the kick in the pants that I needed.
Which one(s) do you think are best for research and web crawling?
i guess AI is now SI - but like SMART devices which are pretty dumb all things considered, SI is hardly super and not real intelligent - just intelligent for those who lack intelligence i'd say since man can't even make another man yet so they hardly understand the brain let alone the interaction between the body, soul, and spirit. SI just runs faster than our brains so it seems better in that way but our brainâs power comes less from one neuron being fast and more from approximately 86 billion neurons operating simultaneously with a combined parallel activity that seemingly enables our perception, memory, language, and decision-making. what's funny is they're using living neurons in experimental biocomputers (organoid intelligence; biohybrid computing, etc.) showing by doing so that the human brain is still way more advanced than SI or even so-called sentient AI or their nirvana of AI singularity. professing themselves to be wise they're still fools and can't even compare to the human beings capabilities and will never given the link between body, soul, and spirit which they can't figure out and will never attain.
Well here I go agin with my comment or is it okay to state my opinion that is normal.. if you let the genie out of the bottle you canât out it back in . Wise wolf I do agree with a lot of your work but I also have my opinions thatâs human just to let you in on that little side remark of mine ..
you said starving kids should have parents that should work.
anyone with any level of empathy is going to see an issue with that but clearly you feel guilty about saying it since youâre still bringing it up 6 hours later so i suppose that makes up for it.
I don't understand 98% of what was just outlined above. I don't understand how to do the things that are described, my brain does not work at that level. But what I do understand is that Commercial AI has so many guardrails and "programed in" biases, it make the sort of use that TWW outlines impossible. If you do not own and run an application locally, you really do not have a useful or durable tool. Right now, when the internet goes down, many are just screwed. "You will own nothing and be happy" fits this perfectly.
If I didnât have a local AI system, I would not be able to do this work anymore. The corporate AI ecosystem is so guardrailed now that if you even mention âIsraelâ â you get 3 paragraphs about anti-semitism and the holocaust. I tried asking Claude to research Jewish officers in the Nazi military a few months back and I am pretty sure a data center somewhere in California caught on fire. It freaked out about it. Kept insisting I was wrong, tried showing me some Snopes-style âfake newsâ articles. When I listed actual names and academics that have written about the topic, instead of agreeing it just ended the conversation and locked itself down.
This is how you know something is fishy. If corporate AI refuses to touch it, chances are itâs because itâs true.
You should see some of the âdiscussionsâ I have had with âMr. GPTâ. Same experience here, mention any number of âprotected classesâ and the LLM instantly loses its collective mind.
One positive is that I have become quite adept at identifying âAI Bullshit.â I sometimes bait the LLM deliberately just to see what comes out. One thing is certain, the output is 99.9% predictable. If nothing else the adversarial mental gymnastics keep the mind sharper.
I love your reply rules, honestly I checked them cuz I thought wtf THIS guy has REPLY RULES~~~~!!!
Yup ~
The Wise Wolf reply rules
I could care less what you say in the comments.
Whelp time to bring back the age of piracy
Would love to subscribe, but I stopped using substack paywall a year ago at least.
I likely am going to sell my computer by end of Dec.
Time to stop the access to my hard head and the social engineering defense needed is ramping up higher every week!
Crawling into a cave is not good self-defense. Anybody younger than about 60 who fails to keep up with this technology is going to find themselves socially isolated or worse. Don't give up.
I'm not younger than 60 :) well passed that marker...
Thanks for the pep talk though.
Is X, formerly known as Twitter, almost totally AI propaganda? It seems so to me and I would guess it's around 80% unreliable bull.
Wolf you wrote exactly what I need to read. I will definitely go through this whole article. Because I have been looking for a home AI set up, even considered building one, for months now. Thank you for sharing your expertise and experience. We do need to take control before they take it from us.
do it now while you still can because a few years from now who knows if we will even have access to uncensored ai anymore. this technology has the power to bring down the entire epstein class hegemony. they know it.
Eli, I love the âfor dummiesâ summaries in boxes you always leave at the end of your articles!
I have one of the most modern computers in the world. It runs Windows XP offline. I use a version of Maximizer (CRM) that doesn't charge a monthly subscription fee and uses btrieve which is still one of the fastest dbase engines. For graphics and writing I use Wordperfect which still beats Word hands down every day of the week and twice on weekends. Boot up is super fast. 30 seconds from start to fully operational OS. The best part is that it does everything I need it to do and with a rate of efficiency that makes $10,000 computers blush.
I will stick with my own brain đ§ and CRITICAL THINKING ability đ
Thank you for this article. Conceptually I understood a good chunk of what you shared; however, the details of starting to go down the road of setting up my own localized open source AI machine feels a little overwhelming.
Can you recommend a resource or article link that gives a little bit more of a step by step guide for someone is wanting to start?
I appreciate your work and I'm proud to support it! Blessings to you
the easiest route now is just get a decent nvidia graphics card with at lest 16gb of vram and just run qwen 3.8 27b locally. it scores at 70% efficiency compared to models that require a 50 grand investment to run locally.
itâs what i use now. the strix halo system is overkill honestly but with it i can run 5 asynchronous instances of this model for agentic development so itâs still pretty useful f or me but if youâre just looking to have local ai superpowers at a hobbyist level, get a decent gpu. you can find them used on ebay for around 1200 bucks. still pricey but itâs way cheaper than paying for a graphic designer or programmer!
This was written for coders and corporations and I understand the need and the concern. As a consumer, it affects me peripherally so I hope y'all succeed. In the meantime, what does the average consumer do to defend against what you predict? Pray?
i wish i had an answer for that
The all seeing AI. Chat G-sus gpt.
Controlled by the all grabbing Trump messias
Thank you for this. I grabbed Qwen and a couple other models. I have been sitting on my butt wrt, but your article was the kick in the pants that I needed.
Which one(s) do you think are best for research and web crawling?
qwen 3.8 27b can do all of that but you need to have an agentic harness plugged into it to give it tools access.
In the near future there will be no personal computers, only terminals that connect to a master computer.
This has already begun.
read thru twice, only a glimpse of understanding . only sure thing is
once newest tech escapes pandoras confines it is certain never to return.
best to be well versed on its peccadillos, or have a resource who does.
Thanks Wolf, for all you do
i guess AI is now SI - but like SMART devices which are pretty dumb all things considered, SI is hardly super and not real intelligent - just intelligent for those who lack intelligence i'd say since man can't even make another man yet so they hardly understand the brain let alone the interaction between the body, soul, and spirit. SI just runs faster than our brains so it seems better in that way but our brainâs power comes less from one neuron being fast and more from approximately 86 billion neurons operating simultaneously with a combined parallel activity that seemingly enables our perception, memory, language, and decision-making. what's funny is they're using living neurons in experimental biocomputers (organoid intelligence; biohybrid computing, etc.) showing by doing so that the human brain is still way more advanced than SI or even so-called sentient AI or their nirvana of AI singularity. professing themselves to be wise they're still fools and can't even compare to the human beings capabilities and will never given the link between body, soul, and spirit which they can't figure out and will never attain.
Well here I go agin with my comment or is it okay to state my opinion that is normal.. if you let the genie out of the bottle you canât out it back in . Wise wolf I do agree with a lot of your work but I also have my opinions thatâs human just to let you in on that little side remark of mine ..
you said starving kids should have parents that should work.
anyone with any level of empathy is going to see an issue with that but clearly you feel guilty about saying it since youâre still bringing it up 6 hours later so i suppose that makes up for it.