Dave Cullen released a video on his YouTube channel, Computing Forever, entitled “Why the Singularity is Bullshit“. He makes a lot of claims in it that are themselves, well, bullshit. Here’s my summary.
First Dave doesn’t understand what the technological singularity is:
- Dave claims the singularity is an “era”. It is not. Rather, it is the conjecture that there will be a point where learning machines (i.e. artificial intelligences) become more intelligent than humans, and their intelligence will then increase exponentially, making them the dominant intelligence on the planet.
- Dave states that to create the singularity, we will reverse engineer the human brain and create machines as intelligent as we are. This is incorrect. The conjecture is that by creating machines that are capable of learning, they will make themselves more intelligent than humans. Also, learning systems are not based on reverse engineering the human brain. Modern neural networks are mathematical constructs only loosely based on living organisms.
- Dave conflates the transfer of human consciousness into machines with the singularity, implying that the conjecture states that first causes the second. Most people who find the singularity conjecture plausible don’t seriously believe that will be the case. Rather, things like transfer of human consciousness might result from the AI singularity.
As to Dave’s claims about why it’s “bullshit”:
- He appeals to the authority of Sam Harris, who is not an expert in artificial intelligence.
- He compares the singularity to religion, based on his mistaken belief that the singularity is somehow the same as the idea of transfer of human consciousness.
- Even the claim that the transfer of human consciousness to machines is a religious belief is false, since it is based in science (though not science we have today) rather than superstition. Dave is confusing speculation or futurism with religion.
- He claims that people will shun technology, yet we see huge demand for data science in business software. The AI behind intelligent search, which is driving enhanced productivity in many white collar industries, making it enormously attractive to businesses that need to remain competitive, is exactly what is proposed to be likely to lead to an AI singularity.
- He states that we don’t have the knowledge to alter the human genome, yet we are seeing that happen already. Machines will begin to outpace humans in intelligence even if there is no AI singularity. Even if they remain mere tools, we will use them to understand far more about the genome than we currently do. Like nuclear science, what has been learned can’t be unlearned, and the idea that knowledge of the genome and how to modify it won’t be used seems extremely naive to me. I agree with Dave that we don’t have the wisdom, but that never stopped us in the past.
- Dave’s “nature knows best” claim seems ludicrous, knowing what we do about the evolutionary process. The best that one can say is that it is slow and operates on a small scale. Mutations like sickle cell anemia, which confers some immunity against malaria at the expense of the general health of those who carry it, seem pretty crap compared to an intentional modification to the immune system that would give safe immunity to a wide range of illnesses.
- He claims that we never question what technology is doing to us, yet there are plenty of articles about GMOs, the impacts of automation, and other incremental changes that technology is bringing. If he means longer term thinking, well, that is exactly what the AI singularity conjecture is. The very thing he’s arguing against is evidence of long term thinking.
In summary, Dave argues that we need to put the brakes on. Who is this “we”, Dave? Individuals make advances, and knowing the outcomes enabled by any given advance is impossible. Are you saying we should stop trying to improve things? Who is going to stop us? You are starting to sound a lot like one of the priests of the Temples of Syrinx in Ayn Rand’s communist dystopian future.