AI Anti-Hype Examples Thread

[Butters]

I really enjoyed Leo's videos on this, let me share them here again.

 

 

And another interesting one about Ai replacing programmer jobs. 

My experience vibe coding is: I've learned how all this works and will be working with a professional developer on actual commercial projects.

Share more! 

[Twentyfirst]

Do you think vibe coding is just a tool to work between you and the developer?

[Salvijus]

I don't think Ai is a fake hype after watching this video. Robots and Ai are already replacing everything and it's only 2025 still. 

 

 

Edited Thursday at 08:00 PM by Salvijus

[Butters]

On 18-12-2025 at 8:02 PM, Twentyfirst said:

Do you think vibe coding is just a tool to work between you and the developer?

No. You can do simple things right now for free with Google Ai Studio. You can make a cool app. But now imagine that you added 10 other complex integrations to do custom things inside an organization. Vibe coding won't be able to do that properly and just write tons of low quality code. 

For a project like that, I'd hire the dev from day 1 and not interfere at all with my vibe coding. I'd use Ai to understand the code maybe or help teach best practices, but not touch the code. 

But for learning, understanding and making simple apps for yourself it's great. 

For actual programmers, vibe coding does replace writing long code, but it doesn't replace the human architect.

That being said, I've seen complex n8n setups with multiple LLMs. Very impressive, those seem to function like legit agents. 

Edited Saturday at 06:46 PM by Butters

[Thought Art]

 

[Twentyfirst]

On 12/20/2025 at 9:32 PM, Butters said:

No. You can do simple things right now for free with Google Ai Studio. You can make a cool app. But now imagine that you added 10 other complex integrations to do custom things inside an organization. Vibe coding won't be able to do that properly and just write tons of low quality code. 

For a project like that, I'd hire the dev from day 1 and not interfere at all with my vibe coding. I'd use Ai to understand the code maybe or help teach best practices, but not touch the code. 

But for learning, understanding and making simple apps for yourself it's great. 

For actual programmers, vibe coding does replace writing long code, but it doesn't replace the human architect.

That being said, I've seen complex n8n setups with multiple LLMs. Very impressive, those seem to function like legit agents. 

How many months/years until AI vibe coding will be able to act as a middle man agent between someone who doesn't code at all and their developer team?

[AION]

This is porn for stupid people. AI is not a hype. Even if AI stops in its tracks right now which is not the case. It has changed everything forever.

[Leo Gura]

1 hour ago, AION said:

AI is not a hype.

It is certainly hype.

But it will probably do some great things too in 10 years.

The way hype works is by mixing some kernels of truth with a bunch of BS, so that people are totally confused and stop thinking.

Edited yesterday at 12:42 PM by Leo Gura

[Butters]

6 hours ago, Leo Gura said:

It is certainly hype.

But it will probably do some great things too in 10 years.

The way hype works is by mixing some kernels of truth with a bunch of BS, so that people are totally confused and stop thinking.

Your take on Ai really confuses me. So we can never get far with the LLM model, yet you say there could be agi. You mean off an entire new tech that nobody is currently working on? And do they start working on this from scratch or off the LLM technology? Or do you mean the existing infra of all those datacenters gives us the ability to propel into agi? 

Are these your own ideas or did you get them somewhere else? 

I'd really love a video on this. 

Edited yesterday at 07:35 PM by Butters

[Leo Gura]

13 hours ago, Butters said:

You mean off an entire new tech that nobody is currently working on?

Yes.

But of course a few people are working on entirely new tech. We just don't know when it will be viable.

Ilya Sutskever is working on an entirely new approach that isn't just another LLM. But who knows how long that will take? Could be never.

Quote

And do they start working on this from scratch or off the LLM technology?

Could be both.

I think it needs to be entirely different from LLM, using no training data. It needs to be an abstract reasoning model. Not based on pattern-matching human data.

Quote

Or do you mean the existing infra of all those datacenters gives us the ability to propel into agi? 

No way. AGI is not a hardware problem, it is a software problem. We don't need more datacenters.

Quote

Are these your own ideas or did you get them somewhere else? 

My own reasoning. But there are AI reaearchers who explain why LLMs cannot do AGI. Yann Lecun is one example. He is Facebook's top AI scientist. He says no way can LLMs do AGI. He is correct. But I am not just believing him, I am using my own understanding of the issue.

Edited 13 hours ago by Leo Gura

[Basman]

The financing of MLMs is pure hype and a circlejerk between various AI companies, Nvidia, datacenters and the US government. AI hasn't made a single dollar in profit and I'm not optimistic if that is even possible. It's way too expensive relative to the output, which is itself often dubious in quality.

https://www.wheresyoured.at/why-everybody-is-losing-money-on-ai/

Quote

Generative AI has a massive problem that the majority of the tech and business media has been desperately avoiding discussing: that every single company is unprofitable, even those providing the models themselves. Reporters have spent years hand-waving around this issue, insisting that "these companies will just work it out," yet never really explaining how they'd do so other than "the cost of inference will come down" or "new silicon will bring down the cost of compute."

Neither of these things have happened, and it's time to take a harsh look at the rotten economics of the Large Language Model era.

Generative AI companies — OpenAI and Anthropic included — lose millions or billions of dollars, and so do the companies building on top of them, in part because the costs associated with delivering models continue to increase. Integrating Large Language Models into your product already loses you money, at a price where the Large Language Model provider (EG: OpenAI and Anthropic) is losing money.

I believe that generative AI is, at its core, unprofitable, and that no company building their core services on top of models from Anthropic or OpenAI has a path to profitability outside of massive, unrealistic price increases.

The only realistic path forward for generative AI firms is to start charging their users the direct costs for running their services, and I do not believe users will be enthusiastic to do so, because the amount of compute that the average user costs vastly exceeds the amount of money that the company generates from a user each month.

As I'll discuss, I don't believe it's possible for these companies to make a profit even with usage-based pricing, because the outcomes that are required to make things like coding LLMs useful require a lot more compute than is feasible for an individual or business to pay for.

Every time you've used AI it has been at a loss. AI is actually more expensive (and worse in terms of quality) than plain human effort. That cost is just obfuscated and not talked about (and financed by billions of dollars of investments and loans). 

[Leo Gura]

The definition of investment is losing money until years later you start making it.

How long do you think it takes a new gas station to break even?

Break even on a nuclear powerplant is about 20 years.

Edited 11 hours ago by Leo Gura

[Butters]

10 hours ago, Leo Gura said:

No way. AGI is not a hardware problem, it is a software problem. We don't need more datacenters.

And is this future AGI that you're describing conscious? What is your definition of an actually intelligent AI? And how can something like that run on a computer? God can run on Windows? 

[Leo Gura]

47 minutes ago, Butters said:

And is this future AGI that you're describing conscious?

Unknown.

Quote

What is your definition of an actually intelligent AI?

AGI is able to do all the intellectual work of a human.

Quote

And how can something like that run on a computer? God can run on Windows? 

I don't think they will create it any time soon.

A computer exists within God. That's how.

Edited 3 hours ago by Leo Gura

[Raze]

https://www.nytimes.com/2023/03/08/opinion/noam-chomsky-chatgpt-ai.html

[Rick1]

True AGI seems unlikely with present technology, but a potent form of 'artificial general ingenuity' is emerging. These systems solve complex issues not through intelligence, but through massive computation and pattern matching. While this approach is error-prone and could result in baseless outputs, it is incredibly effective when combined with oversight. It frequently achieves significant success across a growing range of tasks, operating at a scale that exceeds individual human capability.