Steve Ardire is a software startup advisor with a focus in AI, machine intelligence and cognitive computing, and will be speaking at the Virtual Assistant Summit in San Francisco, on 28-29 January 2016.In 'The current state of machine intelligence 2.0', Shivon Zilis stated one of biggest changes seen over past year is startups shifting away from building broad technology platforms with “machine intelligence as magic box” to focusing on solving specific business problems to deliver real value. So as more enterprises are becoming “machine intelligence literate”, with machine intelligence players having figured how to speak the language of solving a business problem, we have a perfect storm for tremendous upsides. Shivon does a nice job delineating the many ways to go to market (Machine Intelligence In The Real World) but it’s no slam dunk because there’s three significant hurdles to overcome:
Unlike recent past just having right algorithms does not guarantee winning because startups also need access to right data sources where again big players like Google, Facebook, Amazon and have the advantage so startups face a serious chicken and egg problem.
Big players are rapidly pushing AI towards new horizons and now open sourcing machine Intelligence designs like Google Tensorflow and Cloud Vision API, Facebook Big Sur, IBM System ML, Microsoft Distributed Machine Learning Toolkit and now Open.AI with $1B in funding in addition to their commercial offerings like @ibmwatson.
Most categories Shivon delineates fall into Artificial Narrow Intelligence (weak AI), which excels at one function or task but is pretty much everywhere now so competitive differentiation and upsides are limited. The biggest upsides are in fast emerging Artificial General Intelligence (strong AI), like Google DeepMind Google, along with select startups mostly currently flying under the radar.
So can startups successfully play in the machine intelligence field?
Absolutely if they can jump all three hurdles (or at least two) but must have a 'why now’ solution that solves a problem better and faster to quickly get significant customer traction, revenue, build a differentiated brand.
So what are hottest machine intelligence areas?
Here the writer is excited about “lasers” that collect a focused dataset and “magic wands” that seamlessly fix a workflow because they can turn new types of data into actionable intelligence right now with well-worn SaaS techniques. Perhaps, but in my opinion most of these get two hurdles at best, and the Intelligent Assistants Landscape may be more attractive for machine intelligence with a significant market size.
Today's digital assistants (Siri, Google Now, Cortana, Amazon Alexa, Nuance Nina, Watson Engagement Advisor et al) seem ‘intelligent’ but they’re mostly about task completion through Q&A. They can’t explain suggestions, anticipate problems, suggest alternatives, and rarely take the initiative. To be ‘intelligent” means having conversational AI capabilities (not Q&A) with context-aware episodic memory and transitive reasoning to understand and respond with problem discovery and resolution.Episodic memory is collection of past personal experiences that can be explicitly stated at a particular time and place (events, associated emotions, and other contextual who, what, when, where, why knowledge), and reasoning by transitivity refers to ability to recognize relationships among various things with increased use of logic and inferential reasoning.
Why is this important?
Because AI is only fully useful when it knows you in multiple contexts and data + machine intelligence = artificial intuition that ‘mimics’ human intuition. And to compete and win against big players means jumping the three hurdles per above with Artificial General Intelligence plus emotional intelligence, which defines context of interaction for understanding behaviour, perception, learning. People don't change behaviour on information, they change it on emotion.Steve Ardire will be speaking at the RE•WORK Virtual Assistant Summit in San Francisco, on 28-29 January 2016. Other speakers include Dennis Mortensen, x.ai, Nick Triantos, SRI International; Deborah Harrison, Cortana, Luca Rigazio, Panasonic Silicon Valley Laboratory; Tim Tuttle, MindMeld; and Pilar Manchon, Intel.
The Virtual Assistant Summit is taking place alongside the Deep Learning Summit.
This is a guest blog and may not represent the views of RE.WORK. As a result some opinions may even go against the views of RE.WORK but are posted in order to encourage debate and well-rounded knowledge sharing, and to allow alternate views to be presented to the RE.WORK community.