Sometimes it takes technology a little while to catch up, but we get there.
I wanted to share a sudden shift in technology in my profession that I think is relevant to the AI discussion. It is a fine example of how an entire field of science can be flipped on its head in only a matter of years. This content is part of a talk I give to non-scientists, rotary clubs, motorcycle forums, etc... Hopefully you all might find it interesting.
The story begins with The Human Genome Project (1990-2003). This was a 13-year, 2.7 billion dollar international effort to sequence the 3 billion bases (i.e. letters) of the human genome.
These are traditional DNA sequencers. They worked non-stop in rooms like this all across the world to sequence the first human genome. Then, a sudden shift in technology
radically altered how DNA is analyzed. Enter Next-Generation DNA Sequencing (NGS).
Moore’s Law was developed in the early days of Intel by Gordon Moore. He noted that the number of transistors on a processor was doubling every year while the price was halved. This law has generally held true for the semiconductor industry, as well as many others, including the cost of genetic analysis. Well, it was holding true, until the development of NGS in 2008…
Within the span of only a few years the ability to sequence a human genome moved from an international effort to a single benchtop. This instrument can sequence 12 human genomes in a single run. It doesn’t just sequence them once, either,
it sequences each one ~100 times to ensure accuracy. That is 3.6x10^12 data points in an afternoon. Think about that… a traditional instrument may be able to get something like 2x10^5 data points. That is a difference of seven orders of magnitude…!
All of the magic happens on a silicon chip about the size of a single microscope slide. It is an amazing fusion of silicon-based electronics and biology. It should also be pointed out that there are MANY types of NGS, all appearing within a few years of each other. Pyrosequencing, reversible terminator, H+ ion, all of the sudden it was easy to conduct biological functions on semiconductors. A good example of multiple discovery.
If anyone is curious how exactly this works, this
video provides a great explanation of H+ ion sequencing.
My point here is that you never know when a sea change in technology is going to occur. In a single stroke an entire field of science (and medicine) is forever changed. I think the example of NGS is especially relevant to AI, as it is literally the coupling of semiconductors and many of the same molecular processes that govern our own biology. After seeing this I don’t think it is unreasonable to assume that one day a computer and software could be created that is so radically different from anything that we know of now that it could support a genuine intelligence.