Oh, but will... I don't think many people are aware of what's going on now- and that's just what we know about. There's certainly much more research that is hidden from view. Machine learning is a stunningly powerful thing... Paired with the sorts of hardware advancement that's going on, it's likely to happen in the next ten years, if not sooner.

@StylisticIdiot this is my profession and the field I am published in. It will not happen in our lifetimes, not the singularity as it has been described by kurzweil and others. The technological process will be impressive but true machine intelligence, no. Not at 100x current cloud compute density, maybe not even 1000.

@jperlow I'm humbled, and sort of hope you're right. Still, there's a whole lot of chatter from people in your ranks who would disagree. In any case, thanks for your insight.

@StylisticIdiot I came out of the part of IBM that created Watson and I worked at the part of Microsoft that created Azure. machine intelligence is not just a function of the hardware. To create something like Alexa you need the power of multiple Amazon datacenters filled with compute nodes, and Alexa is prettty stupid. Google dedicates a lot more compute power to its voice assistant and that too, is pretty stupid. As automation engines it’s impressive, as AI for generalized purposes, no. If we see major advances in quantum computing that can speed up compute cycles by a factor of say, 10,000 in the cloud, we could have something maybe as intelligent as a eight year old child, maybe. But there’s little chance of making a robot that isn’t dependent on cloud intelligent like that, you need massive amounts of compute. And not just compute an entirely new way of programming and representing data that does not exist today. We will see smarter expert systems, cars that can drive us places, robots that will do our dishes, cook our meals and fold our laundry, intelligent agents that will organize our lives for us and free us from many mundane things, but creating something like HAL or the Enterprise computer on Star Trek or Skynet is very unlikely in the next 50 years. Powerful clouds, with connected robot drones, yes. C3PO or Terminator, or “She” or the Blade Runner replicant AI no. Neural networks are impressive things but we are not even close to being able to replicate the human brain. We will likely be able to stage missions to mars and nearby star systems before such a machine intelligence is ever built. Nuclear fusion perfected in 50 years, finding evidence aliens in the next 50 years, yes. Artificial minds, no. That’s assuming we do not annhilate ourselves first

@jperlow you just blew my mind.

@jperlow I'm glad to make your acquaintance today.

@jperlow I tend to agree with you. I have followed this field for a long time and there are still major shortcomings in our understanding of how to implement a mind as software and hardware both.

I would estimate a lower bound of 50 years myself, with no upper bound. I think it is inevitable that it will be done someday though.

Human brains are substantially more able at general tasks than even our closest relatives, probably due to some mutation that increased our ability for cognition.

Like any futurist prediction, we can't account for unknown developments in technology.

I call shenanigans

@jwd45244 Sure, go ahead and trust the government. They always tell the truth.