Artificial general intelligence or AGI is the great goal that AI gurus keep mentioning. From Sam Altmanpassing through Zuckerberg and his superintelligence teameven of course Elon Musk. The problem is that they are already beginning to repeat themselves and this whole thing sounds more and more like a huge déjà vu.
Today he doesn’t trust, tomorrow he does. Surely you have come across this nice sign in some of those authentic bars or shops. The AGI is starting to sound exactly the same. They count in Gizmodo that Elon Musk has set a date for the arrival of the long-awaited AGI: 2026. He recently said that Grok 5, which will be launched next year, had 10% chance of getting the AGI and it seems like he’s now upping his ante.
During an xAI meeting, Musk stated that he is confident that the company’s ability to scale its computing power will help them achieve AI that surpasses human intelligence. Everything good, except because in 2024 he said that the AGI would arrive in 2025.
He hype and the calm. What happened to Musk is further proof of the disconnect between the discourse of the “sellers” of AI and the experts who make AI. Altman, Musk and Zuckerberg start from the idea that the more AI scales (that is, the more investment is made), the sooner AGI will arrive. From there exorbitant investment in data centerssome the size of entire cities.
On the other side we have AI researchers and developers, whose speech is much more realistic. Yann LeCun recently saidconsidered the godfather of AI, that the path to achieving AGI is not language models, but world models. Research also points in that direction and recently we talked about how language is not the same as intelligenceso the current path seems more like a dead end. Andrej Karpathy, co-founder of OpenAI, has also spoken out, and in his opinion the AGI will arrive, but it will take at least another decade.
Musk’s other predictions. According to Business Insiderin the meeting with xAI employees, Musk also talked about the construction of data centers in space, an idea with which several companies are flirting in view of the energy problem. Musk related this infrastructure to his plans to colonize Mars and pointed to the possibility that Tesla Optimus robots were the operators of these infrastructures.
It hasn’t always been so optimistic. In 2017 issued a warning: “It is urgent to regulate artificial intelligence before it becomes a danger to humanity.” Maybe 2017 sounds very far away, but we don’t have to go that far back. In 2023 signed a letter with other personalities from the technological world in which he called for AI laboratories to pause model development for at least six months due to the immediate danger of a replacement of humanity.
Today he believes that AGI is imminent and defends that AI will do everything for us and that “working will be optional.” Musk’s speech on AI has taken a radical turn, especially now that he has an AI company. What things.
Image | Gage SkidmoreFlickr
In Xataka | Implanting a chip in your hand to perform magic tricks sounded spectacular. Until you forgot your password
GIPHY App Key not set. Please check settings