Life 3.0

By Max Tegmark - Read: July 11, 2023 - Rating: 7/10

This is a general introduction to AI. As a physicist, Tegmark has the ability to discuss many concepts from a scientific standpoint.

He also explores possible pathways and ethical considerations as the technology advances. Some parts were repetitive, others were outdated.

My Notes

There are many ways to define life. For the sake of not limiting life to what we know, Tegmark defines life as a process that can retain its complexity and replicate.

  • Life 1.0 is the biological stage – life where both the hardware and software are evolved rather than designed.

  • Life 2.0 is the cultural stage – life whose hardware is evolved, but whose software is designed.

  • Life 3.0 is the technological stage – life where both hardware and software can be designed.

  • Intelligence is the ability to accomplish complex goals.

  • Narrow intelligence refers to accomplishing only a specific task (playing chess, playing games).

  • Broad intelligence can be seen as the ability to accomplish various goals (learning, eating, memorizing).

  • The current human intelligence is far above machines intelligence since an average human can perform many narrow goals.

  • The holy grail of AI research is to build AGI that is maximally broad.

  • As humans, it's relatively easy for us to rate the complexity of a task based on how hard it is for us to perform the task.

  • For example, it's much harder to multiply 385,093 x 273,982 than to recognize a friend in a photo (for a computer, it's the opposite).

  • An atlas containing information about the world reveals the state of the book (the position of particles and molecules that give the letters and images their color) and the state of the world (for example, the locations of continents).

  • The molecules represent the world based on the real location of the continents. If the continents would change place, then the molecules of the book would as well.

  • We can think of bits as atoms of information—the smallest indivisible chunk of information that can't be further subdivided.

  • Information can take on a life of its own, independent of its physical substrate.

  • Our brain is different from a computer not only in the way it works but also in the way it's used.

  • We retrieve memories from a computer by specifying where they are stored, while we retrieve memories from our brain by specifying what they are.

  • Computation is the transformation of one memory state into another.

  • Computation takes informations and transforms it, implementing a function.

  • The fact the same computation can be performed on any computer means that computation is substrate-independent.

  • Computation can take a life of its own, independent of any physical substrate

  • It's not the particles but the pattern that really matters.

  • The substrate independence of computation implies that AI is possible: intelligence doesn't require flesh, blood or carbon atoms.

  • To get from today, to AGI powered world takeover requires 3 logical steps:

    1. Build human level AGI
    2. Use AGI to create superintelligence
    3. Unleash it to take over the world
  • The most dramatic scenarios at the end of the spectrum are the most valuable to explore. If we can’t convince ourselves that they are extremely unlikely, we need to understand them well enough to take precautions before it’s too late.

  • One of the challenges ahead is to align the goals of a superintelligent AI with our own. This issue splits into three tough subproblems:

    1. Making AI learn our goals
    2. Making AI adopt our goals
    3. Making AI retain our goals
  • Consciousness is the way information feels after being processed in certain complex ways.

  • Consciousness is subjective experience.

  • Understanding intelligence is different from understanding three distinct problems of consciousness.

    • the pretty hard problem of predicting which physical systems are conscious
    • the even harder problem of predicting qualia (subjective experience of perception)
    • the really hard problem of why anything is conscious
  • Tegmark suggests to name our species homo sentiens (ability to subjectively experience qualia) instead of homo sapiens (ability to think intelligently).