Skip to content

Artificial intelligence is nine years and nine months away

And now for something completely different. In my youth I decided to try my hand at writing short stories, and my very first effort revolved around a pair of researchers from another world who spent a few thousand years guiding human cultural and scientific development. I wrote this in 1983, so I probably stole the idea from David Brin.

Our intrepid researchers help humans develop mathematics, physics, evolution, and finally artificial intelligence. Here's how the story ends:

On January 24, 2033, the IBM 4300 at the University of Seoul woke up. On January 25 it held a press conference during which it gave opinions on subjects ranging from abstract mathematics to comparative theology. On January 26 the ambassador of the Union of Islamic States delivered an ultimatum to the government of the Pan Asian Confederacy which, among other things, stated that the 4300's contention that it considered itself one of God's creations to be an insult to more than one billion of Allah's servants.

In the North American Republic President Hartning at first declared his unstinting support of the IBM 4300's freedom to say anything it wished. The next day, after an anxious meeting with representatives from Dallas and Los Angeles, he told reporters that he had misstated himself and was opposed to the obviously pernicious falsehoods emanating from Seoul.

In Rome, Cardinal Alfredo Sanchez told a throng in Vatican square that God was testing them and that it was time for them to show their Lord that they had not forsaken Him.

In Moscow, the First Citizen, with the primate of the one true religion at his side, issued no proclamations and spoke to no crowds. Instead he picked up his telephone and ordered the immediate launch of a MIRVed 15 megaton warhead targeted for 37 degrees 33 minutes north latitude, 126 degrees 56 minutes east longitude. Nine minutes later the IBM 4300 at the University of Seoul became the first casualty in the Last World War.

I pulled this up last night because I was curious about what date I had chosen for the emergence of true artificial intelligence. The answer turns out to be 2033, which is pretty close to what I still believe today. So I think I'll go ahead and adopt this as my official guess: our first true AGI will be unveiled on January 24, 2033. Take it to the bank.

14 thoughts on “Artificial intelligence is nine years and nine months away

  1. royko

    I personally think that's optimistic. I think in 10 years we'll be deep into the machine learning revolution and dealing with the many consequences of that, but some kind of AGI won't be seen for at least 20 years, maybe more. But that's just a guess.

  2. Jasper_in_Boston

    Young Kevin's predictions with respect to technology may prove fairly accurate, but his prognostications concerning geopolitical developments seem a bit off! Lol.

    1. name99

      Not at all.
      He gets the point that AI, by itself, is no threat at all, but human idiocy is a perpetual threat; and AI is just one more thing for it to be idiotic about.

  3. azumbrunn

    The newest news about IA: It is using so much energy that it is unsustainable. We will have to limit it seriously if we want the planet to stay habitable.

    Like every other exponential growth it hits its limit when it has grown for a while.

    The present route to AI--throwing a million times the computing power of the human brain at simple problems (like writing a letter) is unsustainable and we will need major break throughs before being able to move forward on AGI.

    1. name99

      Get back to me when you're willing to say the same thing about human growth.
      Until then, enough with the virtue signaling.

  4. D_Ohrk_E1

    You don't think AGIs will be banned within the next 5 years? I don't think most people really want to know if it's possible; I think the average John and Jane will prefer to limit AI to being very useful tools.

  5. Adam Strange

    The writing is amazingly good for a first effort. I'd also say that the writing is better than most of the sci-fi I've read by established authors.

  6. rick_jones

    Take it to which bank? Silicon Valley Bank, or First Republic?

    In https://jabberwocking.com/exponential-growth-is-messing-with-our-minds/ Kevin referenced something he write in 2013:

    In other words, just as it took us until 2025 to fill up Lake Michigan, the simple exponential curve of Moore’s Law suggests it’s going to take us until 2025 to build a computer with the processing power of the human brain.

    which then suggested we should be half way there today.

    Now he is telling us another nine years.

    Perhaps the only thing growing faster than the development of AGI is Kevin’s date for when it will arrive.

  7. pjcamp1905

    I'll take that bet. You always ignore experts in the field in favor of breathless tech journalists. Let's see Waymo handle some place where the weather is shitty and the streets are not mapped to the last inch. Take a crack at Toronto. Or New York.

    I'm putting up a box of Krispy Kreme lemon filled. What do you got?

  8. Goosedat

    When China achieves the singularity and imposes sanctions on the US, that is when the leader of America's righteous unrepentant sinners will give the order to launch a nuclear sneak attack on Beijing.

Comments are closed.