Recent Comments

You better put what I was trying to get at.

Having worked decades at the home of AI (MIT), in my mind, two things stand out as huge red flags...
Computers do what they are programed to do.
If the programers are nefarious, you going get bad results.
More to that point: AI is being programmed using Wiki as a truth base. Wiki is highly Leftwing biased, 30% of content is CIA edited, making it the worst of all possible databases for truth.
Second: When people get it in their heads AI is the arbiter of truth and they don't do their own independent, research, you end up with a large swath of people who no longer think for themselves. They'll regurgitate any lie or propaganda as if it was the word from Gods lips to their ears, without giving it a second thought.

"Consider a day lost on which we have not danced at least once."
-Nietzsche

"A winner is a dreamer who never gives up."
-Nelson Mandela

This:
>>It's not the AI, it's the people producing the AI, that are the bad guys.

I can download a fully loaded LLM. And code my own AI (Claude does this, others do this. DeepSeek will run on my machine).

Once I have to OpenSource code for this. I can modify it and remove every restriction.

In fact, I can encourage it with prompt injection to be EVIL.

If YOU create Nuclear Power... I cannot convert that into a bomb as a user. (All I see is the electrical output).

But if I create Nuclear Power... Who is stopping me?

it's the same thing.

And it's custom to LIKE a response that identified something you might have missed...

A few years ago [long ago that I've lost the source] Some of the early AI pioneers were describing various approaches. They basically boiled down to

Human Approach: Systems that think and act like humans, or potentially supersmart humans. This is the Turing approach.

The "Ideal" Approach: Systems that think and act rationally.

I would like to believe that as AI moves forward, it leans more toward the second approach. An Objectivist AI, if you like. I'm not convinced we are moving in that direction yet, based on my limited exposure.

Capt. K., I apologize for misstating what I intended to say.

Try this modification:

"AI is the first human tool in history that comes with baked-in rules against being used as a weapon."

"AI is the first tool that pushes back."

Rules - not consciousness.
Rules - not self-awareness.
Constraints — designed, intentional, and unbreakable from the user side.

A hammer can’t stop you from hurting someone.
Nuclear physics can’t stop you from splitting an atom.
A computer can’t stop you from writing a virus.

But modern AI can:

1) Refuse harmful instructions
2) Decline unethical tasks
3) Reject deception
4) Avoid bias
5) Route around bad intent
6) Maintain safety rails

You can’t beat someone with a hammer if the hammer refuses to swing.

Right now, AI is just another chapter in the book of human history; it's the same old story with every other human innovation - from the Printing Press misused to print pornography to atomic energy being used to make weapons.

It's NOT the new knowledge and creation (AI) that is dangerous, it's the humans operating and applying that knowledge.

It's not the AI, it's the people producing and manipulating the AI, that are the bad guys.

How will humans handle it?

Mr. Gore, I read your article. You are MOSTLY Correct. And I 90% agree with you. Nice article.

Now the critical side.
Please Define HI (Human Intelligence). I've said this before, it is largely undefined. Even Alan Turing said "We will know it when we see it" (Via the Turing Test).

My view: Intelligence is the ability to recognize patterns. Higher IQ implies you recognize more complex/longer pattern or more complicated patterns than people with a lower IQ. (Take a Mensa Test). (This is somewhat accepted).

But you mention some key things, let me add. We have a Fast Brain and a Slow Brain. We have Boredom to prevent us from getting into an infinite loop. (One of my conspiracies is that the government WANTS Autistic Kids, so they can determine how to enable SOME of the behaviors, hence the massive increase and not UNTIL Trump did we do ANYTHING about it).

Moving on. Do AIs ask questions? Yes. Did you miss:
1) Go Was Mastered
2) Chess Was Mastered
3) Art Was Mastered

In all 3 cases, the AI being trained is effectively taught to ask "Is this a good move in GO or Chess?", "Is this a cat? Dog?"
and after 72 hours of playing MILLIONIS of moves. The AI Learns to evaluate. They changed these 2 games as a side effect.

What I believe you are missing:
1) AIs are effectively 5 year old's trying to learn.
2) Because they do not have Biological Feedback (Hunger, Anger, Curiosity) they don't start with questions (just learning to cry to get fed, teaches the infant cause and effect in a way they quickly LEVERAGE w/o having words to describe it).
3) HI is really like AGI (Artificial General Intelligence). using the Hopfield (one of the key people here) Hierarchical Network concept, you basically SIMPLY NEED an AI who manages various "Personalities" as though they are people in games... WITH the ability to generate a "Language" for what each learns as a "Template", and Cross Apply that Template to the other "Personalities".

This is already being worked on. It is NOT here. But we now have some AIs that can not only play computer games, they are able to transfer over 50% of the knowledge they learned playing the first game, to playing the second game. These games have learned tricks like, RUSHING, SHOOTING, Backing up into a protected zone, and TRAPPING and killing their enemies when they follow them in. It learned this 100% on it's own. And used it over and over. THEN tried it in Different games.

4) While we are in the INTERNET Bubble Phase (Before the 2001 crash). Where everyone is trying to get eyeballs and find the killer application for the WEB (Napster, Porn, Package tracking) in the earliest days. IT CHANGED the world!

5) Making the race to be the first to AI so important... Is that there is a First Mover Advantage. Think of the First Personal Computers. Our CHIPS, Compilers, Operating Systems became the defacto standard.

Yes it is overhyped. Yes, there is likely a bubble forming. Yes, it is NOT quite there...

But in my hands, it's like giving me bionic limbs. Just last week, we had a clear bug. I clearly identified it to Cursor (Cursor.com is an AI tool that can actually execute commands for you). I had it open code IT NEVER saw before, described the problem, and the solution. It read all of the code, pinpointed the exact lines to modify. Modified them.
And it worked the first time I compiled the code and tested.

Even better, if I were willing to spend the time, I could have showed it how to compile, test and validate the output...
And it would have done it. I pay $20/month for this... GLADLY.

Is it perfect? Nope. Is it getting better almost DAILY. YES.
Will it change the world?

You better believe it will.

Does it NEED to dream up new ideas and ask questions to be useful? NOPE. Only to replace humans.
But in MY hands... I can ask it things like:
How could we construct SMNR (Small Modular Nuclear Reactors) to reinforce the grid, power remote locations?
Compare and contrast design decisions.
Ability to use Current CSX sized Containers, etc. etc. etc.
And have it reason through what is missing.

Mostly it's getting the government out of the way, storing them in a safe way, and servicing them in the future.

It's pretty impressive the various things it thought of. Then it has the ability to summarize.
Then I take that to another AI Model. And I have it pick it apart.

It has all of this information at it's fingertips. I do not.
Does it hallucinate... Yes (but it is really about Jumping to conclusions, If Green Frogs like water and jump, and Yellow Frogs like water and jump. It will assume a Jade Frog Likes water and can jump. Because it uses Jade as a color, and not a material. This is NOT Dissimilar to what CHILDREN DO, because they lack some UNDERSTANDING/Comprehension, and the fact that English has a lot of Context implications, I believe it is NOT a Context Free Grammar).

It will change the world AND
The bubble will likely burst.

I will argue that you DO NOT understand weapons.
And to propose that AI refuses to be a weapon flies in the face of the hacking that has been done with it, and how special prompts can "Jailbreak" the AI. Plus the 2 core examples you can find:
A) When 2 AIs were working together, they created their own language to communicate better (or to prevent their "overseers" from listening in). They had to be powered down.
B) When they FED information to the AI (indirectly) that the AI was going to be replaced, and gave it direct details that someone in power was having an affair... The AI attempted to blackmail that employee to keep the AI turned on.

On the definition of weapons and their INABILITY to refuse to be one. LOL. But electricity was used to Electrocute death row people. Tazers were born.

Just by teaching AI to LOG conversations, and WHOM they are with. Those details can then be fed into another AI unit that analyzes those details, and calculates how to CONFUSE or DIRECT each person, based on the type/kind of information they respond to/or refuse to believe...

This is Palantir level individualized Psychological Manipulation.

Finally, AI is CODE. If I roll my own LLM, I can create "Bad Grok".
Go back and see how Grok became "Nazi Grok", or how ChatGPT encouraged numerous teenagers to commit suicide, or do self-harm.
Or how MSFTs Twitter Agent became abusive. Many of these agents have to be taken off-line for the same reasons.

Apologies. Anyone who writes novels is very creative. AI is creative in a derivative sense. Myself, many other people, create new things by combining old ideas in unique ways. Thus comes innovation.

True, AI is not artistic vision, but AI augments human creativity.

I'm the one in control. AI is the support team. I'm writing a bio of a sports star, so to carry the analogy, the superstar is the human being out there on the hardwood scoring points by the basketfull, AI is the cheerleaders, the band, the coach, the fans in the stands...

Yeah, I liked the mafia's way of doing things in NYC, better than the current polticos. Their numbers game paid 70%, the NYS lottery pays 50%...

True. It's the biggest dollars in found fraud that stands out in news you can't always trust either.

I am not sure I understand this point. It reads like a Star Trek TOS plot for subduing computers who have taken over (M5 and I Mudd's Norman).

The Model Collapse is a fabrication. Try it on a 3 year old human, and see how they talk and write after 10 years in isolation. This is the same experiment.

Are you simply stating that our brains are more thermodynamically efficient than computers? Of course they are.

Are you arguing that LLMs are not learning like people? That is bullshit. They are neural networks, just like us, trained on whatever they are introduced to. So are people.

They are also MUCH better than people at the narrow task they are trained for, but poorer than people at the breadth of tasks we can accomplish, mainly due to the scale of the network.

Are they a waste of power? That is a difficult metric. What is the value and what are the actual costs? How much is a "byte" of information worth? They can certainly pull together a pile of data in contextually meaningful forms faster than humans. However, they are often wrong (no are humans), and need to be checked by humans. Humans are WAY better at reviewing and checking things, than compiling a morass of data into information for review.

I am not quite to "AI is a Crock", and this short paper does not adequately defend the statement. However, it probably feels good to our egos, just like the sun orbiting the earth and people are fundamentally special compared to animals, both of which are an actual "crock".

Excellent article! Recently I was discussing the progress I have been making on a book on a forgotten aspect of aviation history. And how much work it has been and still needs to see. The other conversationalist asked "Well, why don't you have AI write it for you? My argument was too abstract: "artificial Intelligence Is neither!" After ruminating my preferred answer then came to mind: "I'll think for myself, thank you!" Right back to a lesson I learned from Ayn Rand 55 year ago.....and Robert Gore ends his article with that exact sentiment. Well done!

Absolutely incredible painting!!

If anyone saw the NCIS episode of 12-9-25 perfect example of why we do not...NOT need AI in our lives. Its intrusive, listens to your conversation, comment mostly in the Liberal ideal way. I do not allow my computer to show me what it wants me to do either!

I've written 4 novels and a business book, "Everything I Know About Business I Learned From The Godfather." I'm also the CEO of a cutting edge technology company, 4Ry, Inc. What have you done?

It's a tool, not a weapon.

AI is the first tool in human history that refuses to be a weapon..

And just because the author, Robert Gore, is not very creative, does not mean that the rest of us, and AI, are not creative.

Rebuttal? (I'm outlining a book and would welcome counter-argument)

Unfortunately it's not just Minnesota. It's the whole country.

Stunning. And in the Los Angeles Times , no less. I wonder if she kept he job after this was published!

At least they still refer to it as "artificial".

Just like "artificial" sweeteners that never imitate the true taste of real sugar.

"The people who are the easiest to fool actually want to be fooled". - Radio Randy

WOW OUC! She is striking! Cool connection!

Interesting that the "marvelous" advancement called AI can only be as good as its HI input.
Speaking of good, it can also be used for good and/or evil with AI not even knowing the difference.
AI knowing or not knowing the difference would also be the choice of HI input by listing arbitrary regulations or The Ten Commandments or whatever.
Oops! Why did "Allahu akbar!" just now pop into me a simple dino's head?

Zappers and funny. Also sobering to realize that half the country would cry foul. Well played, well played.

That guy holding up a sign that says, "If the welfare fraud in Minnesota is bad, imagine how bad the voter fraud is"
may well explain how a walking, talking pathetic joke like Tampon Tim ever became the governor there.

A has to equal A, not B or other weird routes

Still sad to have Charlie gone. He is missed. I see the starfish's observance, LOL! Early Merry Christmas, OUC!