The article argues that current AI systems are far from AGI, using examples like creating manga or game characters to illustrate that AI cannot yet match human-level creativity and integration, criticizing figures like Elon Musk and Sam Altman for suggesting otherwise.
AGI = Artificial General Intelligence = Human level intelligence. A human, given a computer ability to focus 100% on a task, have it's memory capabilities, never getting tired, never giving up in the pursuit of his goal, the ability to duplicate himself and assign tasks to those duplicative could, as a minimum: Recreate with high fidelity a comic book or manga, and continue from the current chapter so it's undistinguishable for a human. Like you take One Piece, read it fully, practice drawings and create a new unpublished chapter matching the author (a human). Then people familiar with the manga could read that chapter and following ones without realizing it's from another author. Current AI can do exactly 0% of this task, meaning 100 out of 100 humans asked could tell the difference. Add a new character to league of legends that seamlessly integrates into the roaster and is indistinguishable from current ones. Meaning create the artwork and concept, create the high poly model for movies and the low poly for gameplay, be available to play and be balanced out, sound effects and voice. All this has been done, by humans. Current AI can do 0% of this, no human tested will be confused that played the game at least once. I'm not even asking for a much harder task: create a manga better than one piece or a game better than league and make them both more popular, which is more like ASI. I'm asking it to follow current trend of things done before, recently and with decent amount of data, done by multiple humans already repeatedly (many league programers / artists, many people drawing the anime, creating filler episodes...) How is it even debatable if current systems are AGI? An intelligent human trapped in a box, focused on a task, with the passion of 1000 GPUs, working 24/7 could do these no problem. What we currently have is good interpolators of data and decent extrapolators, but that's about it. How can someone like Elon or Altman have any doubts at all about this, even some engineers think it's close...
A debate on whether AGI is inevitable or facing a wall, weighing AI self-improvement and reasoning against issues like lack of understanding, power constraints, and shifting goalposts.
The author argues that credible anti-AI positions require understanding current frontier model capabilities, citing benchmarks like GDPval and models such as Opus 5 and ChatGPT 6 to show AI surpassing most humans on bounded tasks.
Discusses the advanced capabilities of frontier AI models like GPT-5.6 Sol, questioning whether they constitute AGI and reflecting on AI becoming a commodity for complex cognitive tasks.
The article explores the potential impact of AGI on generative art and the possibility of real-time world simulation, questioning the future relevance of current AI art tools.