Thread by @bengoertzel

@wooldridgemike 1) I'm not going to try to explain why you're almost surely dead wrong about AGI being far away, in the inadequate format of a series of tweets, however in this thread I will give some links for folks who want to do some more in-depth reading/listening on the topic ...

@wooldridgemike 2) Re ur historical analogies, I won't insult the intelligence of the Twitterverse by giving a bunch of links regarding the concept of exponential accelerating change. But it's a real thing. The amount of change that used to take a century can (sometimes) now just take years.

@wooldridgemike 3) Indeed the bulk of commercial AI dev (including AGI-oriented dev) is not grounded in any deep theory of general intelligence. But there is a lot more depth in the AGI research community if you care to look. Deep NNs are not the most interesting stuff going on AGI-wise.

@wooldridgemike 4) My own thinking on the "General Theory of General Intelligence" is in a scattered bunch of papers and posts linked from here --
multiverseaccordingtoben.blogspot.com/2020/05/gtgi-g… -- and partially walked through here --

-- in time I'll write this stuff up more systematically

@wooldridgemike 5) I covered some of these ideas more informally and with more broad context in my marathon interview w/ @lexfridman on AGI earlier this year,

@wooldridgemike @lexfridman 6) My current practical effort to create human-level AGI is centered in the TrueAGI project
trueagi.io, a spinoff from SingularityNET. TrueAGI is working on scalable AGI systems based on the OpenCog Hyperon architecture, wiki.opencog.org/w/Hyperon

@wooldridgemike @lexfridman 7) To get slightly closer to the devilish details of current actual AGI-oriented work, see the talks by myself and Alexey Potapov and others on OpenCog Hyperon at the OpenCogCon event earlier this year,

@wooldridgemike @lexfridman 8) In short, yeah by looking at corporate deep NN work it's easy to get the idea that AGI is super far away and nobody has any fundamental idea how intelligence works. But there is a lot more going on in the AGI R&D world than that.

@wooldridgemike @lexfridman 9) Beyond my own work see the various presentations and ideas from the AGI-20 event this June
agi-conf.org/2020/schedule/ ... underline.io/conferences/23… ...
.. the work of Arthur Franz, Kristinn Thorrison, Pei Wang, so many more ...

@wooldridgemike @lexfridman 10) While Big Tech is working on DNNs that maximally leverage their advantages in data and compute power -- and pundits wax lyrical and historical about how far off AGI is -- meanwhile some of us are building the real deal and accelerating fast!