What is AI all about? Trying to make computers as smart as humans
Computers work in computation cycles / second. The human brain works at approximately 10^24 cycles per second while using low power. This allows us to do complex things without even thinking about them: raising an arm, sitting down, recognizing emotions, driving a car, and more. These are incredibly difficult mathematical operations for computers (think about building a self-driving car). The recognition is still complex for simpler motions (raising an arm), especially for small devices like those found on our wrists or palms.
Fad or something revolutionary? Not a fad, more revolutionary than we think
We are moving from faster calculators to "intelligent" applications due to the exponential rise in computing power and efforts to reverse engineer as much of the brain as quickly as possible.
From learning to drive to discerning emotion, we are seeing the first waves of rapid innovation that will bring computers comparable to human intelligence. Some estimates say that we could be there by 2030.
Difference from existing software? self-improving programs
AI based software can improve itself. A typical software program once designed would focus on repeating the same pre-defined steps as quickly as possible. The software’s creator(s) would publish regular updates to the steps every so often, until the software became obsolete.
AI based software can improve itself (e.g. if you use your iPhone to reply "Got it" to every text, it pops up a clickable bubble saying “Got it” automatically). This is how apps like Google understand where Home or Work are for you, as well as how to find directions to these locations easily. The learning is like a toddler, in leaps and bounds. Before you know it, the toddler will grow into an adult and provide more coherent and contextual value.
What does it mean for me? Software is going to get a whole lot smarter
Software is already learning our meeting preferences, movie tastes, text responses, shopping habits, commute routines in good/bad weather and much more.
By analyzing our data over time, machines can learn from the pooled data to help us do a variety things like curate content, simplify chores, check bills, commute and more. AI and sensors (the largest sensor being our phone), can collect context-specific data (like location, time of day, etc.) and provide us insights from other sensors. This can be correlated with heart rates, activity levels, sleep debt, emotional states, social activity and many other items to improve overall health and improve happiness levels.
My view is that, like with mobile phones, initially more people will get access to a focused (narrow) AI based software, almost like a technological scribe than follows them around to take notes and make suggestions on well-being.
There are many questions around privacy, data ownership, job losses which each merit a post of their own.
Is AI good or bad, will it kill us? Hopefully not...
Present day applications use machine learning, which is basically a type of AI focused on a narrow problem (like AlphaGo and Go). With advances in computing power and an increase in manpower looking to solve problems in the space, we may have an AI that is ‘generally’ at the level of a human being. It will not be the smartest one but still a comparable.
However, after reaching human intelligence levels, the technology may continue to improve itself at an even faster rate, leading to powerful computers that could eventually lead to things like immortality, or annihilation of the human species and everything in between.
My perspective is that the outcome will largely depend on our own decisions, primarily around what we choose to divulge, and where we choose to offload decision making / responsibility, and uses of the data .... but things can get scary quickly.
I have attempted to summarize what is a truly exciting and deep subject, without using common terms like algorithm (way to do something), time-series (anything over a period of time) or recursion (a function that calls itself, and not the commonly used metaphor of “inception”).
kiwi.ai makes software that uses AI to recognize what people are doing, and how they are feeling. The Kiwi team is on a mission to make the lives of at least a billion people happier and healthier.