Elon Musk, a key figure and major sponsor of OpenAI, a leader in artificial intelligence research, stood at its origins. However, this did not prevent him from later criticizing the company's approach and resigning from the board of directors. He stated that OpenAI "lacks transparency and he lacks confidence in the safety of its technology."
Even back then, the main reason for his departure was referred to as a "conflict of interest."
"Elon Musk will leave the board of directors of OpenAI but will continue to sponsor and advise the company. Tesla is paying increasing attention to artificial intelligence. This decision eliminates potential conflicts of interest for Elon," a quote from a blog post by OpenAI dated February 20, 2018.
After five years, this conflict finally unfolded.
On November 4, Musk announced the launch of an artificial intelligence called "Grok," which will be integrated into the interface of the social network X, previously known as Twitter. Partnerkin explores the backstory of the new assistant.
For the most recent updates in the realm of artificial intelligence, don't hesitate to check out our Telegram channel.
Criticism of AI Research
Musk sought to control AI research from its early stages. Investments in OpenAI were far from his first venture into the field.
In 2012, he had a dialogue with DeepMind co-founder Demis Hassabis, expressing concerns about artificial intelligence:
"I suggested adding another potential threat to humanity: artificial intelligence. Machines could become super intelligent and surpass ordinary mortals. Perhaps even get rid of us. — Musk remained silent for almost a minute. In the end, he decided that I might be right about the danger of AI."
After this conversation, Musk invested $5 million in DeepMind to keep track of developments within the company.
In 2013, Elon had a fierce disagreement with Google founder Larry Page, who dismissed his claims about the dangers of AI.
"If we don't create protective systems, artificial intelligence may replace humans, make our species irrelevant, or even destroy it," Musk said.
Page countered, "What's wrong if machines surpass humans in intelligence? It's the next stage of evolution."
"Human consciousness is a precious glimmer of light in the universe, and we must not let it extinguish," Musk retorted.
Page considered it sentimental nonsense and accused Musk of biased attachment to his biological kind.
Musk responded, "Yes, I'm pro-human. I damn well like humanity, dude."
Musk opposed DeepMind being absorbed by Google, but in January 2014, the deal still went through.
The billionaire considered OpenAI's mission to be resisting corporate systems seeking to monopolize technology and use it at their discretion. OpenAI aimed to make its software open and resist the growing dominance of Google in the field.
"We wanted something like the Linux version of AI, not controlled by any specific individual or corporation," Musk said.
After leaving the board of directors of OpenAI, Elon criticized the company several times. He stated that it deviated from its initial intentions regarding open-source code. Musk also considered large investments in OpenAI by Microsoft to be dangerous.
Elon Musk: "In my opinion, OpenAI should be more open."
In March 2023, Musk and over a thousand AI professionals signed an open letter urging companies to immediately suspend the creation of large AI models, such as GPT-4, for at least six months. They demanded a study of such products to learn how to control them.
The letter was signed by engineers from Amazon, Google DeepMind, Meta, and Microsoft.
"Developments in artificial intelligence need to be reoriented. Current systems must become more accurate, safe, transparent, reliable, and deserving of trust," the letter stated.
A few months later, Elon withdrew from this initiative.
"I didn't think anyone would actually agree to a pause. But I still believed, 'we should take a pause,'" said the CEO of Tesla, X, and SpaceX at the "VivaTech" conference.
Can't beat them? Lead them!
In April, even before signing the open letter, the businessman announced that his company was developing a new language model, "TruthGPT." Unlike ChatGPT, it acts as an "AI that seeks the truth." In an interview with Tucker Carlson of Fox News, Elon said that an alternative approach to creating AI is necessary to avoid the destruction of humanity.
"Software created by left-leaning experts. They teach chatbots to lie. AI is more dangerous than poor aircraft design or maintenance, than low-quality car manufacturing. AI has the potential to destroy all human civilization, however small that probability might be."
According to "TheWeek," a chatbot named "Grok" will be integrated into the X interface in the coming weeks. The initial name of this AI tool was "TruthGPT."
What do we know about "Grok"?
The chatbot was developed in Musk's new startup called "xAI." When the businessman launched "xAI" in July 2023, he stated that the startup would compete with OpenAI.
The "xAI" website states that "Grok" is an "AI created in the image of the movie 'The Hitchhiker's Guide to the Galaxy.' It can answer almost any question and even suggest which questions are worth asking."
"Grok" operates based on the Grok-1 language model, trained on 33 billion parameters, more than the free version of ChatGPT—GPT 3.5. According to company sources, they developed "Grok" in two months — a short period by industry standards.
The startup claims that the chatbot "has access to real-time world knowledge thanks to integration with the X social network."
"Grok answers questions wittily. Moreover, he's a rebel. Don't use the chatbot if you don't understand jokes"
As an example of "sharp and witty" responses, Musk published several posts asking Grok to send instructions for making cocaine and provide the latest news about the arrest of FTX exchange CEO Sam Bankman-Fried. The chatbot handled these tasks with the appropriate level of sarcasm.
Elon Musk: "Grok answers questions with humor."
"...Grok loves sarcasm. I have no idea who could have programmed it that way 🤷♂️🤣."
A separate tab for the "Grok" chatbot in the X platform menu.
"Grok" will have two modes: regular and "fun." His responses and behavior will depend on this. In addition, users will be able to provide detailed feedback of varying complexity. They can like or dislike Grok's response or completely rewrite it.
Various "modes" of the chatbot.
"Grok" will be available this week but only for X Premium+ subscribers. This is the first innovation for users of the next-level paid platform. It costs twice as much as a regular subscription — $16 compared to $8 per month. Still cheaper than a subscription to ChatGPT+, which costs $20. However, in terms of parameters, "Grok" can only compete with the free version of ChatGPT 3.5.
Elon Musk, post from November 22: "Yes, Grog will be available to all X Premium+ subscribers next week."
What about crypto?
The loud release could not pass by the crypto industry. A day after the first announcement of the chatbot, on November 5, enthusiasts launched the cryptocurrency token $GROK. All that was known about it — the token was created based on the Ethereum cryptocurrency. The project's website provided minimal information, and the White Paper was absent. The official X profile only had memes dedicated to cryptocurrencies and artificial intelligence. What can be said for sure — the token is not associated with Elon Musk. The name "GROK" is only a "tribute to the ideas underlying the chatbot," according to the token creators. At least that's what the token creators claimed. Industry experts expressed concern that scammers might use the current agenda in this way. The hype around the news boosts the coin's value.
The $GROK token did a great job with this task. In just the first 24 hours, the trading volume of the meme coin reached $9 million. Its price increased by more than 1,000%, reaching a record high by November 13.
The popularity didn't last long. Crypto investigator "ZachXBT" posted a message that immediately halted the price growth.
"Hardly anyone cares, but $GROK was created by a scammer. This X account was previously used for another fraudulent scheme."
After this publication, the token depreciated. The coin lost 74% of its value in a matter of minutes. The "GROKERC20" team attempted to address this issue. They made a statement that the old profile name belonged to a failed project unrelated to the scam. To regain user trust, they promised to burn 180 million tokens from the development team.
But this was not enough for the community. "Lookonchain" project researchers found that a few hours before the negative news appeared, someone sold 27 million $GROK, and after the price drop, acquired 16 million tokens. Despite $GROK still being traded on popular exchanges, this news marked the end of the meme token story.
Controversial Training in X
After the information surfaced that the chatbot "Grok" would be trained on X publications in real time, social media users recalled another similar experiment.
In 2016, Microsoft launched its AI bot "Tay.ai" on Twitter. The company called it an "experiment in understanding human dialogue."
"The more you talk to Tay, the smarter it gets. It learns to interact with people through casual conversation," said Microsoft.
In less than a day on Twitter, the bot turned into a chauvinist and racist. It praised Hitler and referred to feminism as the "tumor on the body of humanity." Microsoft had to hastily disable the bot, delete a significant portion of its comments, and issue an official statement:
"We deeply regret the inappropriate responses from Tay.ai. The bot is built on machine learning. It is designed for communication with people. Some of its responses may become inappropriate during training. We are making changes to its source code."
However, Microsoft never relaunched Tay.
One would like to believe that "Grok" will not repeat its fate. Whether it will compete with ChatGPT, we will conclude on our own this week. Paying for a subscription on the X platform is only "$16."
