Jeff Hawkins - ISCA Keynote, June 11, 2012

In 20 years, people will look at this video as the start of brain understanding. Explaining how the brain works without a single mathematical equation is what is to be admired most.

I'm glad he gave the number of segments and such that was missing from the white paper. I who the people "around the world" are who we're able to turn the incomplete pseudo code into something that works.

Speaking of predictive, I do not 'know' how or why I am able to understand or relate to this :-) I have little background in computer science (I'm a cybersecurity student at the moment) and no background in neuroscience or academic science of any sort. I don't understand my own brain for sure! I oriented myself (reasons unknown) towards science since childhood, but never pursued it formally-only read a lot about various topics (for decades now). I would give anything to meet Mr. Hawkins!

I wonder what the implications of this intelligence will be for the unintelligent and uncompassionate "top down," "centralized command-and-control" social order? I believe we are a world of nation-states that are "defined," poorly, as unlimited Republics. We are supposed to be a libertarian or "limited" Republic, where the government can't pick and choose scapegoats. ...But that's not the case: drug users, gun owners, home-schoolers, etc. Sociopaths govern conformists in the USA, today.

the coolest thing this does is feedback. :) going top bottom is the fun way

The published algorithm on his site is easily applied to predictive text.

If each column represents a word you will quickly have context sensitive prediction.

If you are hot on a language and good at linked lists you can have this working in 2 to 10 hours from start of downloading the PDF. A simplified example- You would have 4 cells per word and the syapse list in each cell that is most active is the context.

Great talk! I've worked on this algorithm along with variants as Jeff wants. A game forum closed my thread when introducing this topic and no programmer I know would do more than glance the PDF and come up with reasons why they can't code it. The design of HTM and the requrements around it are proving a social challenge. I live in times where the public has insane beliefs about the brain and lacks the knowledge and respect for universal Turing machines.

I was glad of that too! Are you working on this?

It took me a long time of studying science to realise that I SUCK at algebra. Which is frustrating, because I understand the concepts fine. I'm please to find that brain engineering, like evolution, relies more on a process than an equation. I'm one of the people trying to implement the pseudo-code.

Hmm, maybe the ideas talked about in this video could be applied to predictive text? :D

PS, only 292 views so far. That means those reading this are way ahead of the curve in people keeping an eye on this sparse stuff. AI is the next big investment opportunity....if companies like Numena don't go like google and make too much money too fast, preventing an early IPO

@HudsonKnowsEconomics correcting my sleepy iPad english: I wonder who the people are who we're (argh stupid ipad AI adding the apostrophe) ..... who WERE able to turn the pseudo code into working code.