Leo's amazing predictions

[thenondualtankie]

Just ask real developers if AI is increasing productivity or not:

https://survey.stackoverflow.co/2025/ai#sentiment-and-usage-ai-sent-ai-sent-prof

[Uddi]

@OBEler I think he means regular use of LLMs. Good programmers will thrive with AI because they know how to use it properly. However if you are learning coding, AI can hurt your learning as you will be tempted to take shortcuts instead of actually properly studying the materials.

[Ero]

@Uddi I guarantee you he does not understand the difference between the LLM chatbots, the diffusion models used for image generation, the LRMs used for long-context reasoning in Claude Code and Codex.

It's all one thing to him which is why he makes fundamental category errors when he talks about AI. 

[Leo Gura]

7 hours ago, Ero said:

@OBEler His hubris behind the statements while simultaneously being this out of touch is what did it for me. 

He can’t be claiming to have the most sophisticated epistemology and then engage on a topic as thorny and important as this one with the same way a flat-earther/anti-vaxxer does.

Dude, I posted in this thread several times that I am open to hearing counter-arguments.

Don't try to portray me as some kind of bigot.

I could be wrong about AI. I am just thinking out loud here.

Epistemology does not guarantee you will have the correct view on every topic. I can be wrong about stuff.

My arguments are not comparable to anti-vax arguments at all.

It is too hard to be right about every damn thing.

Edited 2 hours ago by Leo Gura

[Leo Gura]

In other news, the price of gold is relatively good right now. Gold looks like a great investment opp right now.

Now is an ideal time to cash out tech stock and crypto into gold. If you wait too long it might get too late. Don't get distracted by all the AI debate. The market is historically overpriced and AGI is not coming soon.

Edited 2 hours ago by Leo Gura

[Joshe]

18 hours ago, Ero said:

https://sketchplanations.com/ is badass. Gonna buy the book. Reminds me of Visualize Value by Jack Butcher.

 

[Ero]

15 minutes ago, Leo Gura said:

Epistemology does not guarantee you will have the correct view on every topic. I can be wrong about stuff. My arguments are not comparable to anti-vax arguments at all.

(1):  Everyone is getting paid off to hype AI, and the scientists do bad science so you can't trust their AI studies.

(2): "You expect me to trust science on AI? What kind of fool do you take me for? "

(1) + (2) = can't trust the earth is round/ vaccines are safe because all is paid by NASA/Big Pharma. In the last 2 years of commenting on AI here or on the blog you have used ONLY Youtube videos, blogs or memes. How is that serious epistemology?

16 minutes ago, Leo Gura said:

Epistemology does not guarantee you will have the correct view on every topic. I can be wrong about stuff.

That's why I called it hubris. I don't speak on topics I haven't done my research on. The problem isn't with this just thread or topic, it's systemic. 

Instead of being more careful in your commentary, you keep on lallygagging and default to this on pushback. And when people really press you, because they are fed up with the excuses, you un-mod or ban.

[Leo Gura]

3 minutes ago, Ero said:

(1):  Everyone is getting paid off to hype AI, and the scientists do bad science so you can't trust their AI studies.

(2): "You expect me to trust science on AI? What kind of fool do you take me for? "

(1) + (2) = can't trust the earth is round/ vaccines are safe because all is paid by NASA/Big Pharma. In the last 2 years of commenting on AI here or on the blog you have used ONLY Youtube videos, blogs or memes. How is that serious epistemology?

Not comparable to anti-vax.

AI is over-hyped, over-invested, and the science on it sucks.

3 minutes ago, Ero said:

That's why I called it hubris. I don't speak on topics I haven't done my research on. The problem isn't with this just thread or topic, it's systemic. 

I have done plenty of research on AI. I know its fundamental limitations despite all the hype.

3 minutes ago, Ero said:

because they are fed up with the excuses, you un-mod or ban.

I ban people who are too closedminded to learn.

[thenondualtankie]

Leo is criticizing science from a post rationalist POV. You're conflating it with a pre rationalist criticism.

No amount of science or finite benchmarking can actually give you an accurate perspective on AI. It's too complicated of a thing.

[Ero]

2 minutes ago, thenondualtankie said:

Leo is criticizing science from a post rationalist POV. You're conflating it with a pre rationalist criticism.

No amount of science or finite benchmarking can actually give you an accurate perspective on AI. It's too complicated of a thing.

Gotta pass through "rational" to get to "post-rational" and he doesn't seem to have done so. 

[Jodistrict]

In evaluating AI, people tend to by confused in their analysis because of a lack of knowledge of the culture.  The term “Artificial Intelligence” is a lofty goal that has acted historically more as an inspiration and an ideal than a statement of reality.   

For example, AI researches initially assumed that if they could get a computer to play chess that would be intelligent.  But once they actually got computers to beat world champions they realized that they had developed a program that could only do one narrow task.  A program that plays chess can’t recite a poem.  So, the final verdict is it is not Artificial Intelligence. 

This same cycle repeats, and the claim can also be leveled against LLMs.  They are highly effective in certain tasks, but ultimately they can only do one narrow task and are not Artificial Intelligence. 

In a sense, when an AI project becomes successful, it is then judged no longer AI.  It spins off and becomes its own field of engineering.  So the failures of AI are frequently engineering successes.

So there are two evaluations going on:  1) how close is it to artificial intelligence, and 2) is it successful engineering that solves a problem? 

Another example.  LLMS have essentially solved the natural language problem.  The natural language symbolic processing of the 1980s is basically obsolete.  I am always amazed when I put in complex convoluted paragraph to an AI assistant and it actually understands what I said.  If you did natural language based on understanding, like a Chomsky grammar,  you would only be able to get It to understand perfectly formed toy sentences.   So Chomsky criticizes LLMs saying they use brute force and thus are not intelligent.  However, his system would only work in toy worlds.

I know people are confusing issues when they talk about the capabilities of an AI program and in the same breath speculate whether it is conscious or has emotions.   These issues are irrelevant to analysis of the commercial impact of an engineered product.

[something_else]

22 minutes ago, thenondualtankie said:

Leo is criticizing science from a post rationalist POV. You're conflating it with a pre rationalist criticism.

No amount of science or finite benchmarking can actually give you an accurate perspective on AI. It's too complicated of a thing.

I agree. Direct experience, however, will give you an accurate perspective on AI.

Anybody who has used this tech beyond consumer level has direct experience of what it is capable of.

[Leo Gura]

11 minutes ago, something_else said:

Anybody who has used this tech beyond consumer level has direct experience of what it is capable of.

But what does this say about AGI and the bubble?

Does your direct experience explain whether the bubble will pop?

[Leo Gura]

3 minutes ago, UnbornTao said:

You've got to be kidding me. Here we go again.

I am not kidding at all.

If I see that someone is too closedminded to learn from me, I will stop letting them have access to me. My time is valuable and I will not waste it debating with closedminded.

If I was not on this forum, I would not care. But since I am here, there must be limits. Any other teacher would do the same.

No serious teacher will endlessly argue with closedmindedness. If you cannot learn what is being taught, the class is too advanced for you.

Edited 1 hour ago by Leo Gura

[UnbornTao]

Just now, Leo Gura said:

I am not kidding at all.

If I see that someone is too closedminded to learn for me, I will stop letting them have access to me. My time is valuable and I will not waste it debating with closedminded.

You know, @zurew could just as easily said the same about you. And he did. And the closed-mindedness accusation is a pretext in this case.

[Leo Gura]

11 minutes ago, UnbornTao said:

You know, @zurew could just as easily said the same about you. And he did. And the closed-mindedness accusation is a pretext in this case.

Sure, that is called relativity.

But the difference is that I created this platform and teachings. I am responaible for leading this place. I have to make decisions about who is suitable to be here and who is not.

If he wants to teach his own thing, I am happy for him to do so in a space of his own making. But I am not going to be here every day listening to him.

Edited 1 hour ago by Leo Gura

[Joshe]

"I can be wrong" is said from above the error. It's the master looking at his fallibility in the abstract, granting the error is possible. The self stays elevated the whole time. Nothing is lost. Whereas owning a live error where others are right and you are wrong means climbing down off your perch and standing level with the people who corrected you. It means being an ordinary person who got it wrong and got told so before there's any story available that turns the error into wisdom or the correction into something you already actually knew, or the critics into people who missed your larger point. 

People don't want "I can be wrong". They want you to demonstrate in real-time you're capable of owning  your error. That tends to matter to people with integrity. When truth and self-image collide, can you put truth first?

Instead of saying "I can be wrong", say "It appears I am wrong and haven't even explored a large part of this space where the actual value is, let me reconsider all this stuff you guys showed me".

This is all people want you to demonstrate, because the inability to demonstrate what everyone considers a basic virtue points to something much larger than being right or wrong. 

But, I've been trying to tell everyone. It's not going to happen bro. It's just how Leo is and there is no path to correction because it's deeply structural and load-bearing.

Edited 1 hour ago by Joshe

[Leo Gura]

@Joshe The problem I have with the arguement you guys are making here on AI is that you are trying to win the argument on narrow technical ground: "Leo hasn't used LLMs to code, so he doesn't see how useful it is." But this muddles a larger point. The larger point is that even that AI will underperform in the long-run when all factors are taken into account.

I concede the point that I have not coded with AI. I am looking at the big picture.

Edited 1 hour ago by Leo Gura

[UnbornTao]

5 minutes ago, Leo Gura said:

the difference is that I created this platform and teachings. I am responaible for leading this place. I have to make decisions about who is suitable to be here and who is not.

If he wants to teach his own thing, I am happy for him to do so in a space of his own making. But I am not going to be here every day listening to him.

It is certainly your platform, and you didn't listen to him much, to be frank. You did your Tier 2 moves of spiritually ignoring him with the usual mechanisms.

Why do you pretend that your pretext for banning him is true? He was probably the most rigorous and intellectually open-minded user here. He could run circles around you and everyone else.

Many of your responses boiled down to "you don't know what you're talking about™," and similar.

Whenever you mention open- or closed-mindedness, I like to pay attention to the subtext of how those terms are used. My assessment is that, usually, to put it naively, they are based on your intuition of whether you are believed - or not.

You're not going to like that you do the same thing - that of being "closed-minded." It just takes the form of a couched, spiritual, condescending Actualized veneer.

[UnbornTao]

47 minutes ago, Leo Gura said:

No serious teacher will endlessly argue with closedmindedness. If you cannot learn what is being taught, the class is too advanced for you.

You edited it, read above. 

Did you learn to admit your failure? Start doing that yourself. 

If you cannot learn to admit whenever you're plain wrong after several years, you're fooling yourself about the class you're in.

Edited 47 minutes ago by UnbornTao