Crypto and AI: control robots, motivate humans
ChatGPT text generator is the fastest growing consumer app ever, and it continues to grow rapidly.
But the dirty secret of AI is that humans are still needed to create, label, and structure training data, and training data is very expensive. The dark side of this is that an exponential feedback loop is being created where AI is surveillance technology. And so, managing humans in the AI loop is crucial.
Some experts believe that when robots (potentially) take over the world, they had better be controlled by decentralized networks. And humans need to be incentivized to prepare the datasets. Blockchain and tokens can help… but can blockchain save humanity from AI?
ChatGPT is just regurgitated dataChatGPT is a big deal according to renowned artificial intelligence researcher Ben Goertzel, given that "the ChatGPT thing caused Google's founders to show up in office for the first time in decades. years !" he's laughing. Goertzel is the founder of the blockchain-based AI marketplace SingularityNET and a strong proponent of artificial general intelligence (AGI) – computers thinking for themselves. This means it sees where ChatGPT fails more clearly than most.
"The interesting thing about ChatGPT and other neuro models is that they achieve some generality without having much ability to generalize. They achieve a general range of ability relative to an individual human by having so much training data."
Also read:
In other words, ChatGPT is really a brute force feature of having so much data. “That's not how humans achieve magnitude through iterative acts of creative generalization,” he says, adding, “It's a hack; it's a nice hack; it's really cool. I think it's a big step forward. »
He also knows where this hack can take us. "I wouldn't be shocked if GPT-7 can do 80% of human jobs," he says. "That's huge, but that doesn't mean they can be thinking machines on a human level. But they can do the majority of jobs on a human level.
Experience-based logic remains harder for AI than scraping the internet. Predicate logic means that humans know how to open bottle caps, for example, but AIs need billions of data points to learn this simple task. And good big language models (LLMs) can still transform language into presumptive logic, including paraconsistent logic or self-contradictory logic, Goertzel explains.
"If you feed them all over the web, almost everything you ask them is covered somewhere on the web."
Goertzel notes that this means some of the Magazine questions are redundant.
"I've been asked the same questions 10 times on ChatGPT in the past three weeks, so we could have asked ChatGPT what I think of ChatGPT. Neuromodels can generate everything I said in class. the last two months, I don't even have to say."
ChatGPT text generator is the fastest growing consumer app ever, and it continues to grow rapidly.
But the dirty secret of AI is that humans are still needed to create, label, and structure training data, and training data is very expensive. The dark side of this is that an exponential feedback loop is being created where AI is surveillance technology. And so, managing humans in the AI loop is crucial.
Some experts believe that when robots (potentially) take over the world, they had better be controlled by decentralized networks. And humans need to be incentivized to prepare the datasets. Blockchain and tokens can help… but can blockchain save humanity from AI?
ChatGPT is just regurgitated dataChatGPT is a big deal according to renowned artificial intelligence researcher Ben Goertzel, given that "the ChatGPT thing caused Google's founders to show up in office for the first time in decades. years !" he's laughing. Goertzel is the founder of the blockchain-based AI marketplace SingularityNET and a strong proponent of artificial general intelligence (AGI) – computers thinking for themselves. This means it sees where ChatGPT fails more clearly than most.
"The interesting thing about ChatGPT and other neuro models is that they achieve some generality without having much ability to generalize. They achieve a general range of ability relative to an individual human by having so much training data."
Also read:
In other words, ChatGPT is really a brute force feature of having so much data. “That's not how humans achieve magnitude through iterative acts of creative generalization,” he says, adding, “It's a hack; it's a nice hack; it's really cool. I think it's a big step forward. »
He also knows where this hack can take us. "I wouldn't be shocked if GPT-7 can do 80% of human jobs," he says. "That's huge, but that doesn't mean they can be thinking machines on a human level. But they can do the majority of jobs on a human level.
Experience-based logic remains harder for AI than scraping the internet. Predicate logic means that humans know how to open bottle caps, for example, but AIs need billions of data points to learn this simple task. And good big language models (LLMs) can still transform language into presumptive logic, including paraconsistent logic or self-contradictory logic, Goertzel explains.
"If you feed them all over the web, almost everything you ask them is covered somewhere on the web."
Goertzel notes that this means some of the Magazine questions are redundant.
"I've been asked the same questions 10 times on ChatGPT in the past three weeks, so we could have asked ChatGPT what I think of ChatGPT. Neuromodels can generate everything I said in class. the last two months, I don't even have to say."
What's Your Reaction?
