GPT-5 Expected this Year, Could Make ChatGPT Indistinguishable from a Human
Remember all those robot movies we used to watch growing up? Some of the robots would become so clever they would start taking over the world. Well, we have robots now although they aren’t taking over they are definitely of great help. ChatGPT is one such chatbot. I never thought I’d live to the time when I’d see robots doing things that human beings can’t do and this is definitely something to enter into the history books.
As we get ready for GPT 5 we can’t help but wonder what she has for us.
Siqi Chen, tech entrepreneur, and developer claims that GPT-5 will achieve Artificial General Intelligence (AGI) by the end of the year. This has caused quite a stir in the AI world considering this type of technology we’re talking about is nothing short of amazing.
In a tweet on Monday Chen said GPT-5 is set to complete training in December as OpenAI said it will achieve AGI. This is mind-blowing as many wonders if AGI is possible. With the use of AGI, it means that GPT will become more like a person as it’ll be indistinguishable from a human.
Chen also went on to say that AGI will be in GPT 5 is a consensus belief within Open AI as some believe it’s possible. AGI is when an AI learns and comprehends tasks or concepts commonly done by humans. AGI is also considered much higher than AI as the tasks and functions aren't limited.
The one thing that developers say is a plus with AGI is the ability to boost productivity because it accelerates any AI-enabled process which means humans don’t have to perform repetitive tasks.
Could this be giving AI too much power? There are some repercussions and negative effects associated with giving AI so much authority. This means human-like bots can covertly spread propaganda, or enable harmful misinformation on social media platforms.
Chen goes on to give an example of a recent tweet by Yohie Nakajima where he describes an AI that’s beyond what any mind can comprehend. In this experiment, an AI is given the task to continue generating and reordering its own task list as it executes them.
This particular experiment is currently connected to search and Chen says that by also using the power of chatbot plugins and GPT-5 it means they are closer to achieving AGI.
Although this could contribute to a major revolution in the AI industry there are some concerns that AI is being given way too much power.
Elon Musk and a thousand other tech leaders and researchers signed an open letter on Wednesday warning that accelerating AI development to GPT-5 isn’t a great idea and halting progress beyond GPT-4 is in the best for human interest. In the letter, they request that AI labs put a pause for at least six months in training any AI Systems more powerful than GPT-4. They also request that the halt be made verifiable and public and also include all key actors. Governments were also requested to act quickly by instituting a moratorium.
Advertisement
Stirs the mud. GPT-5, etc. won’t be doing much of anything.
POET may arrive at some distant point in the future:
https://www.technologyreview.com/2021/05/27/1025453/artificial-intelligence-learning-create-itself-agi/
But most researchers agree with the following:
https://www.bbc.com/reel/video/p0fc9tlx/chatgpt-why-we-re-still-smarter-than-machines
GPT-5 will only be another iteration of a common algorithm with much more data from various sources, so it will “appear” to be quantum when in reality it will be the same old story. Using algorithms to play GO–Mu-Zero self-taught based initially on machine learning–still can’t function as a humanoid.
But it will “seem” so–people will be impressed, dazzled, amazed. But behind the scenes . . . blah . . . while 10% of the work force is eliminated, starting with lead programmers and bloggers.
Hi zakhi, were rumours that amazon fixed bug due to which alexa laughs at night .
Btw, please keep us informed on this anti AI Open letter