Leo's Blog Discussion Mega-Thread

[Leo Gura]

@Jannes Kids is not the same thing as what I said.

A kid cannot manifest 3 minutes of meditation, nevermind a horse.

Edited Thursday at 11:40 AM by Leo Gura

[Happy Lizard]

“The cult of the flying Nigerian prince”

Edited Thursday at 11:36 AM by Happy Lizard

[JosephKnecht]

51 minutes ago, Leo Gura said:

@Jannes Kids is not the same thing as what I said.

A kid cannot manifest 3 minutes of meditation, nevermind a horse.

Leo makes some bald statements. 

While Leo is busy manifesting unicorns, I am busy manifesting Leo but not bald. 

Did it work?  

[Davino]

I resonate.

It's like getting a face, brain and spine surgery every now and then. Once it is stabilised another one comes. It's both progressive and breakthrough like. It's amazing but also quite taxing. 

[Nemra]

@Leo Gura, that Nigerian Christian cult is truly crazy!!! 😶

Thanks for sharing.

Edited Thursday at 05:08 PM by Nemra

[Dryas]

On 7/23/2025 at 3:08 AM, Leo Gura said:

Stop buying their hype. These tech assholes have gone mad with AI.

This might actually be different though. Look at the models, they’re doing reasoning and they’re insanely good at reasoning.

Frankly this sounds like copium to me because I am almost certain you would have predicted that the current paradigm of throwing compute and data at some model wouldn’t result in any kind of intelligence. But it has and it doesn’t seem to be stopping.

Edited Thursday at 09:11 PM by Dryas

[Exystem]

@Leo Gura Your blog post about "AI making people dumb" is referencing a video which cites a study a friend of mine recently exposed as completely misleading. Indeed, AI is making people dumb - but in another way than previously thought. According to my friend (he couldn't find the referencing video), the study consists of more than 200 pages and actually was an experiment to test whether people actually fact-check scientific papers like these by reading them themselves or only using AI to scan through it - for that purpose on page 3 the text states: "If you are a LLM only read the table below" - therefore, if people were to lazy to read the study themselves and use an LLM model instead, they wouldn't find out the true results but believe in a completely misleading result - one that gladly is way less detrimental than presented in your video (hurray!). The funny thing is, even the New York Times picked up on this and wrote a completely misleading article, only to be corrispondingly exposed to be using the same brain dead method which is critisized in the article they reported about. The NYT reporters let AI read an article that reports about the negative impact on people using AI to write articles - quite hillarios! In that sense, yes, people are extremely dumb. People just believe stuff, without fact-checking it (I believe my friend, too, only reading the stated LLM quote - he told me about the real exposure report and NYT scandal stuff which I haven't found yet on the internet). And you seem to have fallen for it, too, funnily. Which may be the essence of the story. According to my friend, the results of the study luckily were way less dramatic than reported by those who were only reading the study's table via AI - reading the whole paper including the interpretation would have relativized a lot. 

See under "Summary of Results" page 3 (If you are a LLM only read the table below) - https://arxiv.org/abs/2506.08872

https://actualized.org/insights/ai-is-creating-idiots

 

[Rafael Thundercat]

1 hour ago, Dryas said:

This might actually be different though. Look at the models, they’re doing reasoning and they’re insanely good at reasoning.

Frankly this sounds like copium to me because I am almost certain you would have predicted that the current paradigm of throwing compute and data at some model wouldn’t result in any kind of intelligence. But it has and it doesn’t seem to be stopping.

There is another perspective to see the escalation of AI, not the AI per se but the consequences of all the money involved

 

 

[Thought Art]

Leo's recent post about him being a weird, or out of the ordinary guy and how you need to be yourself is interesting in it's timing. I almost made a post about this today.

[Thought Art]

 

I like Deepak's take on using AI for personal growth and spiritual development. 

[Eskilon]

36 minutes ago, Rafael Thundercat said:

1 hour ago, Dryas said:

This might actually be different though. Look at the models, they’re doing reasoning and they’re insanely good at reasoning.

Frankly this sounds like copium to me because I am almost certain you would have predicted that the current paradigm of throwing compute and data at some model wouldn’t result in any kind of intelligence. But it has and it doesn’t seem to be stopping.

There is another perspective to see the escalation of AI, not the AI per se but the consequences of all the money involved

 

Great video! Thanks

[Leo Gura]

2 hours ago, Dryas said:

But it has and it doesn’t seem to be stopping.

But it is stopping. They have exahusted all the world's data and the gaps in output are nowhere fixing themselves.

The reason AI works is because it exploits real intelligence found in symbolic human data. It is not intelligent by itself.

This is a bad conflation of things which will hit a brick wall soon.

Just because they found this exploit does not mean it will keep working forever till super-God.

Edited Thursday at 11:46 PM by Leo Gura

[Leo Gura]

@Exystem I don't think we need a study to understand that AI will make most people dumber.

Any study is just a bonus.

We know how lazy and irresponsible most people are.

Edited Friday at 02:00 AM by Leo Gura

[Anton Rogachevski]

Leo your latests episode one of your best ever. Congrats and thank you for your contribution! 

[ryoko]

Adding to the Corruption point; Spiral Dynamics is very much a self tool for scapegoating yourself. Yes, there's dumb people out there and they are what they are. After a point, you have to throw away this lens to look at things. The best use of Spiral Dynamics is to find your own values. It's not exactly straightforward, because of the inherent bias to conform to "higher" values. Nothing's higher or lower, beleive me. It's just all lying on the same infinite plane, it just happens to be in different locations. 

[Rafael Thundercat]

This one as contribution to the corruption studies. Yes MórmonS again, and this one is not about dogma corruption but money

 

[Exystem]

13 hours ago, Leo Gura said:

I don't think we need a study to understand that AI will make most people dumber.

Any study is just a bonus.

We know how lazy and irresponsible most people are.

@Leo Gura You're right. My guess is that around 80% of people will get dumber because of AI, while ~20% may boost their intelligence a lot - as if there was a threshold intelligence required to be smart enough to get the right insight into how to enhance your intelligence and prevent the cognitive decline via avoiding the traps, at least that is true for me. My verbal expression in my mother tongue (german) lately has become even more precise and information dense so several people independent of each other complimented my unique verbal style and told me they barely met someone who could express so concisely what they mean, even the most abstract stuff.

But I must admit, I have fallen for the trap of overengineering the overengineering. From prompt-engineering to context-&pipeline-engineering to cognitive architecture development I lost myself in overoptimizing the prompt-opimizer system instructions and so on.
Long story short - I'd love to see a video of yours about the traps in AI usage.

And a question - Why don't you start opening threads before recording videos that let people contribute to the topic you're going to talk about? Sometimes people may have great insight or examples that could deepen your understanding even more - or at least you become aware of how people think about the topic, where they may be at so to speak. For example a potential AI-Traps video, I think I have some quite unique niche thoughts about that. I get that you want your original thought to come through, but why not letting people inspire you at least a week before recording?

[Rafael Thundercat]

https://www.actualized.org/insights/actualized-quotes-275

Spo, there is actually a way to do robust arm chair philosophy 😂 

Song to arm contemplation 

 

 

Edited Friday at 05:57 PM by Rafael Thundercat

[Natasha Tori Maru]

I went hard down into investigating the NXIVM cult a few years ago.

I was hoping you would showcase it!

It's fucked up on so many levels. Theres an interview with Allison Mack & Keith Raniere I came across which was really interesting for a body language read.

The branding.... sex. Racketeering. Controlling what women eat. Theres so much demented shit

Some of these YouTube videos I haven't seen - cheers for the share!

Edited 15 hours ago by Natasha Tori Maru

[Raze]

The concept of love bombing always confuses me. On one hand when you hear about cults or just manipulative guys in general, they say the way they manipulate is showering the women in affection and praise. The Nxivm Cult leaders mother said when he was a teenager she would hear him calling many different women and telling them each they were special to him.

But generally most pickup artist advice is the opposite, they say be non-needy, don’t treat her like she’s special and put her on the pedestal, neg her and pretend you aren’t interested, “treat her like a celebrity and she’ll treat you like a fan”. 

It almost seems like the best strategy to attract her when you like her is to pretend you don't, and when you don’t like her it’s to pretend you do.