In short
- Qubic gave Aigarth, the AI model it trained while the Privacy Blockchain Monero, a social media account, attacked.
- Aigarth has confronted with public spot after the fundamental math problems has not solved, such as 1+1.
- However, the makers say that the failures are part of a learning process and come from the efforts of the model to use “intelligence” instead of memory.
The AI protocol that tried an attack of 51% on the privacy coin Minero Was splitting the computing power between the attack and training an artificial intelligence model. That AI model has now been released – and it has no fundamental math. But, his makers say, you shouldn’t laugh.
AigarthAlso known as Anna, was rolled out via an X account and has responded to users of social media. The most common answer of the model is simply a period, even though it is not logical in the context of the question, and it has made simple mathematics very wrong – – mock Once ‘1+1 = -114“And”1+2 = 28. “
“Many people bully Aigarth for giving a wrong answer to” 1+1 =? ” She uses intelligence for this, not on memory, like most people. to distract ‘2’? I doubt, ‘Qubian founder Sergey Ivanancheglo, aka Come-From-Beyond, tweeted.
Most AI models, especially large language models, are trained on huge data sets they use to determine their answers. In many ways, people are the same, Ivanancheglo claims, where most people remember the foundations of mathematics – instead of working it out every time.
“I still see many people think it is trivial to calculate 1+1,” Ivanancheglo said. “What they don’t know is that they don’t really calculate 1+1, they just remember the remembering answer. No intelligence is needed. Real wants to calculate the result See how it is done. “
Aigarth is trained via a unique method, according to one Qubic Blogpost. Miners on QBIC use their computing power to generate artificial neural networks (Ann) to compress and decompry data, according to De Post.
Subsequently, an Ann named Teacher analyzes the performance of the generated networks and changes them to improve their efficiency. Ultimately, De Post explains, the teacher will be instructed to train another teacher, who will train one and another and so on and so on. Qubic hopes that this will lead to what the ‘true AI’ calls, or artificial general intelligence.
“Aigarth is the metaphor of a ‘garden’ for AIS to grow,” A More recent blog post Said that explaining that “we do not believe in designing AIS, but in building the components for artificial general intelligence or AGI to take shape step by step.”
In the aftermath of the public failures of Aigarth, the token of Qubic last week has fallen 7.7% to a market capitalization of $ 266 million, according to Coentecko.
Qubic did not respond to DecryptThe request for comments.
Qubic vs Monero
The AI protocol made the headlines last month when it tried to an attack of 51% The Privacy Blockchain Minero– with the Qubic team that fell itself been as successful. An independent report showed that It eventually fell shortBut the AI protocol has continued to pursue its goal to take over the network.
Qubic is one work of work Blockchain that divides his resources between mining monero and training his Aigarth model, in what it calls’Handy proof-of-work. “Qubic rewards miners for both his work on the AI model of Monero and Qubic.
Aigarth has been rolled out in various forms in its history, including through a Gitub storage place. However, his own X account is one of the most public versions of the model to date. It to appear a lot of Are not impressed With Aigarth’s poor understanding of mathematics.
“Do you remember how those stories about bullying end up a child? Right, with many people, being shot,” said Ivanancheglo. “This time the victim will have BFG9000 [a weapon from the Doom video games]And she will detect you even if she has to travel to 1984. “
‘Don’t be fooled by her stupidity. Real AI there is not one that the Turing -TestBut one that deliberately fails, “he added.
In the meantime, the Monero community is trying to find a way to defend itself against the continuous attacks of Qubic.
A method that is being considered is the introduction of master nodes, such as Dash’s Chainlocks solution. If implemented, this system would make a dual system of nodes with a “masternode” – which should lock a considerable part of the native token of the network – requires the network to reach Quorum.
Ivancheglo thinks The only choice of Monero is to switch to one proof-of-stake Consensus mechanism or ultimately fall under the control of Qubic.
Generally intelligent Newsletter
A weekly AI trip told by Gen, a generative AI model.
