This is a collection of 192 short (1-4 page) essays in response to the 2015 Edge question.
Here are some broad trends within the essays:
- Those who think we will have artificial general intelligence(AGI) relatively soon point out how much faster technology progresses than biology. Those who are skeptical of AGI in the near future point out that there is no one discovery or lightbulb moment that will make it possible, but an accretion of minor discoveries.
- AGI proponents rely on some faulty statistics, such as the number of neurons in the human brain (a popular estimate is off by orders of magnitude) and what current supercomputers can do. This is like saying that because a car engine has more horsepower than a human it could run to the moon–they are qualitatively different pursuits. Furthermore, AGI is a software and not a hardware problem (at least for now).
- Another major divide between essayists is that some say AGI is “far away” (whether that’s 100, 1,000, or more years) while others counter that if AGI is (eventually) inevitable, we should give serious consideration to how to govern and relate to it now. This is a question of discounting
- The question of AGI’s humanity (human-ish-ness?) comes up in several essays. Some argue that if AGI is designed by humans then it is more like a prosthetic brain than it is a peer (an appendage and not a bet), while others counter that if AGI is evolved by large amounts of data, then even its human creators cannot understand it well.
- Plenty of authors in this collection question the premise (sure, machines compute, but do they think…) and even more try to move the goalposts (sure, they think, but do they have culture or emotions or desires). The essays that go down this route are mostly skippable.
Some of the best essays in the collection are the following:
- Peter Norvig: “Ask Not Can Machines Think, Ask How Machines Fit Into The Mechanisms We Design” (probably the best in the whole collection)
- David Christian: “Is Anyone in Charge of this Thing?”
- Scott Atran: “What Neuroscience And Machine Models Of The Mind Should Be Looking For”
- Brian Eno: “Just A New Fractal Detail In The Big Picture”
- Bart Kosko: “Thinking Machines = Old Algorithms On Faster Computers”