The talk covers
One of the early Skinner linear behaviour teaching machines from the 1950's (with a link to FutureLearn remedial system).
A first example of a touch screen multimedia EdTech from the 1960's (including a blueprint from 1966 that shows an approach to sequenced multimedia tele-education offerings as can be seen in most major MOOC platforms, instructivist pedagogy: inform, test, explain).
A 1970's reference to the first AI driven tutorial machines (including the Dyna book, the inspiration of the tablet 40 years prior to the ipad; the Plato IV which was the forerunner of the one laptop per child idea - which also was the first network computer system), and also looking at Papert's first use of the logo programming for children idea to increase problem-solving skills (pdf of Mindstorms book). Interesting remark from Mike Sharples on Papert being dead against any guidance for children's learning, bit similar to what Sugata Mitra concluded from his Hole in the Wall project, illustrating the need for guidance to support exploration (done by children with technology).
1980's sees the launch of micro-computers (reference to BBC's first micro computer), commercial teaching coming in with multimedia personal computers.
1990's spread of online learning (VLE's, intelligent agents, integrated learning environments).
2000 search of mobile learning with contextual and seamless learning coming up (with a reference to handler, one of the first personal mobile learning devices, paper describing HandLeR here).
After this first part of the talk, the focus shifts to old/new learning starting in 1990 to 2010.
An important paper from Meltzoff, Kuhl, Movellan and Sejnowski (2009) describes the foundations for a new science for learning (pdf here). This paper is seen as a manifesto to see learning as a science, and no longer as a craft. It also brings together multiple disciplines that are all part of learning (neurology, machine learning, education, psychology).
Looking at 3 important learning theories: John Dewey with instrumentalism (inquiry led learning), Yrjo Engestrom with expansive activity theory (socio-constructivist learning) and Gordon Pask with conversation theory (coming to know through conversation and mutual adjustment, and a nice reference to wikipedia as an example of conversational theory). With a nice hint to the usefulness of the Experience and Education book by Dewey that you can use in almost any learning situation (pdf here). Part of the FutureLearn learning approach was based upon Pask's theory, which is why each bit of content in FutureLearn actually has an option to interact with others (or yourself) as part of a learning conversation.
Last bit of the talk: evaluating learning
Visible learning by Hattie (more info here), Hattie used meta studies to look at the effect size (whether a learning intervention has an effect) of different factors coming out of around 800 studies. He concluded that if you can make learning more transparent, more visible, the learning will have more effect.
Another nice resource mentioned here is the map of learning theories created by Richard Millwood (map link)
Link to the video below, and there is an upcoming talk - a follow-up on this talk that will be given this Thursday (20 October 2016) at the OU, UK.