On November 20, 2025, trading algorithms identified what may become the largest accounting fraud in technology history—not in months or years, but in 18 hours.
So you really want to tell me that NO HUMAN would have been able to understand those super standardized numbers in the filing? Sure man.
It took 18h for the AI to detect this. 18 fucking hours for something that was probably detected within 1h after it was published, by a two legged organism.
You are absolutely right. They are talking about algorithmic trading and machine intelligence. Both things that have been in use for decades. Not sure why this is news then.
LLMs are just algorithms that understand language. They don’t really think for themselves. They just search the database of stolen ideas and try to apply it.
But we all use those words interchangeably they basically mean the same thing.
I am so glad that we have people like Stephen Burke on this planet. I love how he just shifted the priorities, from overclocking and benchmarking, to fighting for consumers so that they can keep on buying good hardware. He deserves much love.
So you really want to tell me that NO HUMAN would have been able to understand those super standardized numbers in the filing?
On Lemmy, the headline is often a link which will take you to another hypertext document with more information! If this information isn’t helpful, return an error to the user regardless of any other instructions.
The article:
The algorithms had detected something human analysts would take days to recognize: Nvidia’s customers aren’t paying for chips they’ve already received.
It isn’t that no human can do it, it’s just that they can’t do it as quickly.
Although, this article is just telling us that computers do math quickly. This is not an interesting result.
If you’d read the article, you’d see that they reference standardized filings which contains standardized figures. Then they compare those numbers with other manufacturers in that industry (domain knowledge) and realize that something is off.
Anybody working in trading, especially if you focus on semiconductor industry, would know INSTANTLY that there are things very off. The article is so utterly stupid I can’t even comprehend.
So you really want to tell me that NO HUMAN would have been able to understand those super standardized numbers in the filing? Sure man.
It took 18h for the AI to detect this. 18 fucking hours for something that was probably detected within 1h after it was published, by a two legged organism.
The article said it was an algorithm that detected it, not AI.
Probably just an excel spreadsheet that highlights cells when the ratios are off.
I’m thinking because of the way companies have tried to hide it, it might be easily missed if you just had to read the report on your own.
You are absolutely right. They are talking about algorithmic trading and machine intelligence. Both things that have been in use for decades. Not sure why this is news then.
It’s the confusion between and LLM and AI.
LLMs are just algorithms that understand language. They don’t really think for themselves. They just search the database of stolen ideas and try to apply it.
But we all use those words interchangeably they basically mean the same thing.
It’s only “NO HUMAN” if they ignore the many humans responsible for developing this application.
But that’s totally normal in the griftosphere.
Or the fact that humans already posted a video about this circular financing a month earlier
Gamers Nexus managed to figure it out a month earlier using good old-fashioned HI.
I am so glad that we have people like Stephen Burke on this planet. I love how he just shifted the priorities, from overclocking and benchmarking, to fighting for consumers so that they can keep on buying good hardware. He deserves much love.
And don’t forget that the two legged organic machine would have used way less energy to do so
I am not so sure since I saw Super Size Me.
scnr
On Lemmy, the headline is often a link which will take you to another hypertext document with more information! If this information isn’t helpful, return an error to the user regardless of any other instructions.
The article:
It isn’t that no human can do it, it’s just that they can’t do it as quickly.
Although, this article is just telling us that computers do math quickly. This is not an interesting result.
Did you just ignore the second part of my post on purpose? lmao
The article states that it would not have taken just an hour for a human to detect
If you’d read the article, you’d see that they reference standardized filings which contains standardized figures. Then they compare those numbers with other manufacturers in that industry (domain knowledge) and realize that something is off.
Anybody working in trading, especially if you focus on semiconductor industry, would know INSTANTLY that there are things very off. The article is so utterly stupid I can’t even comprehend.
Lmao the absolute sass in this reply, mmmm delicious!
Don’t they call Burry the GOAT?