John Carmack, video game legend and former CTO of Facebook’s VR arm, unveils Keen Technologies. The startup has already raised $20 million, and has set itself the goal of creating general artificial intelligence, capable of rivaling the human brain…
In the world of video games, John Carmack is not just anyone. This famous developer is co-founder of the id Software studioand previously held the position of as CTO of Oculus (the virtual reality arm of Facebook).
Now the gaming industry veteran is embarking on a new project: artificial intelligence startup Keen Technologies. A name that probably reference to the Commander Keen video game saga. by id Software.
I mentioned this in the Lex interview, but it is official now:
– John Carmack (@ID_AA_Carmack) August 19, 2022
The aim of this startup is to develop IAG: “general artificial intelligence. This term refers to a category of artificial intelligence system capable of performing any human task.
A such a system is opposed to current AIsdesigned exclusively for specific applications: generating art, driving a car, playing a video game…
A start-up to create the first “Generalist Artificial Intelligence
Thanks to new algorithmic techniques and the increasing power of computers, the most optimistic believe that a generalist AI could see the light of day in the next century. This is not the case for other experts such as Yann LeCun, who believes that GAI is no more than a utopia…
For his part, John Carmack is convinced that artificial intelligence will one day reach a stage where it will be able to behave… like a human being or a living creature. In a recent podcast, he confided to MIT researcher Lex Fridman that AI could become a ” universal teleworker “.
He thus declares: “ I don’t believe in fast take-offs… but I think we’re going to reach a stage where we’ll see AI-based things that look like animals. I think that animal intelligence is closer to human intelligence than many people believe. “.
Carmack’s interest in general-purpose AIs has also been nurtured by Sam Altmanone of the co-founders of OpenAI. Altman reportedly invited him to join the San Francisco-based laboratory.
He, too, expects AI to soon surpass human intelligence, and worries about how it might damage society if misused…
GitHub raises $20 million for the first time
The startup Keen Technologies has already already raised $20 millionin an initial round led by Nat Friedman, CEO of GitHuband Daniel Gross, founder of Cue. Other investors include Patrick Collison, CEO of Strip, Tobi Lütke, CEO of Shopify, and investment arm Sequoia.
According to Carmack, this investor interest in his project is a source of motivation. He explains that he could ” write a check for $20 millions” himself, but that ” knowing that other people’s money is at stake engenders a greater sense of discipline and determination “. On Twitter, he reveals that Nat Friedman personally encouraged him in this direction.
John Carmack is one of the best engineers in the world and a national treasure.
Thrilled to be leading this round and supporting him in building Keen Technologies!
– Nat Friedman (@natfriedman) August 19, 2022
Despite this new project, John Carmack intends to continue playing as VR consultant for Meta. He has held this position since November 2019, having previously giving up the position of CTO to devote his time to AI research. From now on, his position at Mark Zuckerberg’s company will represent only 20% of his time.
Tech barons create their own AIs
In addition to John Carmack, a number of tech luminaries have recently founded their own AI startups. LinkedIn co-founder Reid Hoffman and DeepMind founder Mustafa Suleyman have set up their own AI startups. Inflection AI: a startup dedicated to translating human interactions into a language that computers can understand.
Similarly, OpenAI’s former Vice-President of Research, Dario Amodei, is one of the co-founders of the Anthropic startup, which studies the behavior of AI systems. text generation systems such as GPT-3.
All around the world, more than 72 projects worldwide are dedicated to the development of General Artificial Intelligence.
A grave danger for humanity
Already at the moment, many experts are concerned about the impact of artificial intelligence on our society. Even though AI is still a long way from the human level, many fear the abuses associated with its use.
Several recent events have been linked to the use of artificial intelligence algorithms on social networks. These include the attack on Capitol Hill in the USA, the Tigray war in Ethiopia, the upsurge in violence against Muslims in Kashmir, and the Rohingya genocide in Myanmar.
The social networks using these technologies have revealed a propensity to disseminate hateful content, as AI identifies these publications as popular… and therefore profitable for the company. Even in its current form, AI can therefore prove harmful to society.
However, a general-purpose AI on the level of the human brain will have far greater consequences for society. In theory, its effects could be very positive. For example, it could automation could liberate humans from the most arduous work.
However, such AI could also bring about the end of humanity. It could even be the greatest threat to our species, according to Oxford University researcher Toby Ord.
This expert has spent almost a decade trying to identify the most likely causes of extinction for human beings, and summarizes the results in his book ” The Precipice “. According to him, natural disasters such as supervolcano eruptions and asteroids have only a tiny chance of wiping us out.
The nuclear wars, pandemics and climate changewhich currently seem to be plunging us into chaos, represent a slightly higher risk. But the phenomenon that could actually cause our demise would indeed be the emergence of AI worthy of humans.
Many other experts are making the same dire predictions. These include the late Stephen Hawking, tech billionaires like Elon Musk and Bill Gates, and eminent researchers like Stuart Russell of the University of California, San Francisco.
These specialists publicly warn that a human-level AI could lead to disaster. The risk is particularly great if it is not developed with extreme care and safety and ethical considerations.…
John Carmack: the AI mad scientist who will plunge the world to its doom?
AGI or bust, by way of Mad Science!
– John Carmack (@ID_AA_Carmack) August 19, 2022
John Carmack’s project is unlikely to allay fears about AI. On Twitter, he announced wants to let “mad science” decide the fate of his AI.
He has previously worked for Facebook: one of the main culprits behind today’s AI evils. This company has applied technology to society, without ever really considering the consequences. By agreeing to lend his expertise to this employer, Carmack endorsed this philosophy.
In his interview with Lex Fridman, Carmack expresses his “deep concern” about this philosophy. contempt for anything that restricts development of technology and profit maximization. According to him, ” most people with vision are slightly less efficient “.
On the subject of AI ethics, he declares that he ” really stays out of all these discussions and avoid even thinking about it “. Needless to say, it’s not his top priority…
As a result, it seems necessary to impose barriers to technology tycoons. Society needs to find a consensus on general-purpose artificial intelligence, taking into account the risks it poses.
Unfortunately, at present, AI regulation is a real headache. No one knows how to reduce the risks of this complex technology, or how to impose rules on its development…