I am of the belief that we will experience a Technological Singularity around 2050 or shortly thereafter. Many top futurists all arrive at prediction dates between 2045 and 2075. The bulk of Singularity debate revolves not so much around 'if' or even 'when', but rather 'what' the Singularity will appear like, and whether it will be positive or negative for humanity.
To be clear, some singularities have already happened. To non-human creatures, a technological singularity that overhauls their ecosystem already happened over the course of the 20th century. Domestic dogs and cats are immersed in a singularity where most of their surroundings surpass their comprehension. Even many humans have experienced a singularity - elderly people in poorer nations make no use of any of the major technologies of the last 20 years, except possibly the cellular phone. However, the Singularity that I am talking about has to be one that affects all humans, and the entire global economy, rather that just humans that are marginal participants in the economy. By definition, the real Technological Singularity has to be a 'disruption in the fabric of humanity'.
In the period between 2008 and 2050, there are several milestones one can watch for in order to see if the path to a possibile Singularity is still being followed. Each of these signifies a previously scarce resource becoming almost infinitely abundant (much like paper today, which was a rare and precious treasure centuries ago), or a dramatic expansion in human experience (such as the telephone, airplane, and Internet have been) to the extent that it can even be called a transhuman experience. The following are a random selection of milestones with their anticipated dates.
Technological :
Hours spent in videoconferencing surpass hours spent in air travel/airports : 2015
Video games with interactive, human-level AI : 2018
Semi-realistic fully immersive virtual reality : 2020
Over 5 billion people connected to the Internet (mostly wirelessly) at speeds greater than 10 Mbps : 2022
Over 30 network-connected devices in the average household worldwide : 2025
1 TeraFLOPS of computing power costs $1 : 2026
1 TeraWatt of worldwide photovoltaic power capacity : 2027
1 Petabyte of storage costs $1 : 2028
1 Terabyte of RAM costs $1 : 2031
An artificial intelligence can pass the Turing Test : 2040
Biological :
Complete personal genome sequencing costs $1000 : 2020
Cancer is no longer one of the top 5 causes of death : 2025
Complete personal genome sequencing costs $10 : 2030
Human life expectancy achieves Actuarial Escape Velocity for wealthy individuals : 50% chance by 2040
Economic :
Average US household net worth crosses $2 million in nominal dollars : 2024
90% of humans living in nations with a UN Human Development Index greater than 0.800 (the 2008 definition of a 'developed country', approximately that of the US in 1960) : 2025
10,000 billionaires worldwide (nominal dollars) : 2030
World GDP per Capita crosses $50,000 in 2008 dollars : 2045
_________________________________________________________________
Each of these milestones, while not causing a Singularity by themselves, increase the probability of a true Technological Singularity, with the event horizon pulled in closer to that date. Or, the path taken to each of these milestones may give rise to new questions and metrics altogether. We must watch for each of these events, and update our predictions for the 'when' and 'what' of the Singularity accordingly.
Related : The Top 10 Transhumanist Technologies
"Interactive human-level AI in 2018" and "AI can pass the Turing test in 2040" are mutually contradictory. If we have convincing human level AI in 2018, it isn't going to take 22 more years to prove that we have convincing human level AI.
So why are any of these statements and dates meaningful? Did you just make up the dates?
You should pull together some of your technology estimates for storage costs, ram costs and cpu costs. The "3M" machine (1 megahertz, 1 megabyte and 1 megapixel) was widely touted leading up to 1984; and this was a breakthrough combination of technology ushering in the GUI.
One could have constructed a similar "3G" concept for 2004. (1 Gigahertz cpu and 1 Gigabyte ram were reasonably available. But my widescreen display only has 125megabits (32*2560*1600). Gigabit local area networking was potentially available. Presumably the 3G computer ushered in some user visible innovation -- probably the widespread internet; maybe video editing.
For 2024, the "3T" computer could well be defined: 1 Teraflops of cpu, 1 Terabyte of RAM, 1 Terabit/sec of networking, 10s to 100s of terabytes of mass storage, 10s of gigabits of display. I'm not sure what the user experience breakthrough will be. But this powerful system is not the AI system.
The AI system would show up around 2024 in a "3P" computer: 1 petaflops of cpu, 1 petabyte of ram, 1 petabit/sec of networking, 1 petabyte of mass storage, 1 terabit of display.
Posted by: ces | December 01, 2008 at 09:54 AM
ces,
Not really, because we are talking about video games, vs. passing a test where a human interrogator is actively trying to find non-human flaws.
The best analogy would be the fact that Toy Story in 1995 had a certain level of graphics, but video games even in 2008 don't quite match that.
Why would the 3P and 3T emerge in the same year of 2024? Or do you mean a later year for the 3P?
Posted by: GK | December 02, 2008 at 01:13 PM