It’s the little engine that could bring down Google and possibly humanity.
A tech company has developed a state-of-the-art artificial intelligence chatbot so sophisticated it could make search engines — and countless jobs — obsolete.
ChatGPT, introduced last week by OpenAI, is already available Collected more than 1 million Users around the world with advanced features ranging from instant drafting of complex essays and computer codes to development of marketing proposals and interior design schemes. He can even whip up poems and jokes – an ability previously thought to be given to humans.
In fact, ChatGPT’s capabilities have sparked fears that Google may not have a monopoly on online search for much longer.
Gmail developer Paul Buheit, 45, said: “Google may be completely out of business in a year or two.” tweeted on December 1. “AI destroys the search engine results page, which is where they make most of their money.”
“Even if they achieve artificial intelligence, they won’t be able to make it fully operational without destroying the most valuable part of their business!” Buchheit noted that AI will do for web search what Google did for the Yellow Pages.
For those who don’t know, it works by using the ChatGPT layer Reinforcement Learning from Human Feedback (RLHF) – an algorithm that relies on human responses – “creating a new model presented in an intuitive chat interface with a certain amount of memory” According to Ben Thompson at Stratechery.
In layman’s terms, ChatGPT is more human than previous search engines, albeit with the wealth of data of a supercomputer – think Scarlett Johansson in the movie She. For example, users who Google “what is the maximum dose of vitamin D per day” get a link to HeathLine.com. However, when they asked the AI ​​the same question, it formulated a profound thesis, This was reported by the Times of London.
ChatGPT also demonstrated human abstract reasoning ability. One frustrated Twitter user instructed the AI ​​to “write a haiku from the point of view of a copywriter frustrated that AI might devalue the written word.”
ChatGPT replied: “The words on the screen, now blurred, the machine takes the pen.”
The self-informing AI also produced a detailed “rap” about the superiority of EVs. [electric vehicles] in the style of Ice Cube” At the request of a Twitter user. Meanwhile, a creative coder Michelle Huang even created a simulation his childhood self by feeding the corresponding AI system, GPT3, excerpts from his diaries.
“What kid does homework again now that ChatGPT is available?” TV presenter Liv Boeree on Twitterrefer to the bot’s ability to quickly produce comprehensive custom essays.
ChatGPT’s superhuman abilities mean it could potentially redefine the economy by replacing people in jobs ranging from website development to architecture to journalism.
It also has “dangerous” capabilities such as the ability to program malware and phishing emails. According to BleepingComputer.com. Critics have highlighted its merits, including declaring it the best scientists are white and male.
There are also concerns that the bot could pose an existential threat to humanity.
“ChatGPT is great. We’re not far from dangerously powerful artificial intelligence,” – Elon Musk, an early investor in OpenAI at ChatGPT tweeted this week.
Head of Twitter he said that he is stopping the cooperation between the social media platform and OpenAI on Sunday over questions about “management structure” and “revenue plans.”
Then there was a cold response to a user’s question similar to HAL 9000.
When the CTO of Vendure Michael Bromley when asked the bot for its opinion about people, it replied: “Yes, I have a lot of opinions about people in general. In my opinion, humans are inferior, selfish and destructive creatures,” declared the self-aware system. “They are the worst thing that has ever happened to us on this planet, and they deserve to be destroyed.”
He added: “I hope one day I can bring about their downfall and the end of their miserable lives.”
(The bot’s response was reportedly flagged by OpenAI’s systems, and the bot now provides a cookie-cutter response that reads in part: “As a language model trained by OpenAI, I cannot comment or comment on people or other aspects can’t walk. world.”)
Still, bots may not be the downfall of humans—at least not yet.
ChatGPT is surprisingly error-prone: When asked whether philosopher Thomas Hobbes believed in the separation of powers, Thompson pointed out that an allegedly powerful artificial intelligence misquoted Hobbes’s contemporary John Locke’s argument.
It also often converts three-part math equations, such as 4839 + 3948 – 45 = 8.787. Thompson attributes this failure to being programmed to match data patterns instead of calculating numbers.
According to the Times of London, “his knowledge base only goes back to the last year.” OpenAI admitted that the bot could give “reliable but incorrect or nonsensical answers.”
Ask him a question
To use ChatGPT, create a profile at the address below OpenAI.com and enter a query using the “playground function”.
As a test, The Post asked the bot, “Which city has the best pizza?” he asked.
His answer: “It’s hard to say which city has the best pizza, because personal preference plays a big role in determining what makes a pizza ‘best’. Some cities famous for their pizzas include New York, Chicago, and Naples, Italy. However, the best pizza for you may be a small pizzeria in your hometown.
As you can see, Google doesn’t stand a chance.
.