Weekly #AI Digest: Human Performance & Car Eyes

In 2016, a study was done to check with the experts in the trade of AI to find out what did they think about the evolving AI taking over the human performance. In simpler words, when will the AI driven machines becoming better performers than human beings. The results of the study are below:

ai-takeover

Image Source: The Technology Review

Some noteworthy things here.

  1. AI driven machines will be better at the game Go by 2026. Remember the graph above is from 2016 and has an approximate scale of 10 year interval on time line.
  2. Most of the periods of machines surpassing humans are with a median range of 40 to 45 years. A typical worker works for 35-40 years, so much for job security.
  3. Full automation is 125 years ahead… Really, can we think that far? Even if we can, whats the veracity of this? Did we know 125 years back that we would have see a boom of internet and smartphones 100 years back?

Going back to the first point of AI driven machines beating humans in ‘Go’ game by 2026, the ground reality remains humans are humans.  DeepMind’s AlphaGo successfully won the Go game over the world champion in 2017 itself.

Find out more here.

AlphaGo Beats Ke Jie

So much for the forecasts of the experts. Things are changing more rapidly than one could imagine. Is it good or bad? Only time will tell.

Even wondered how a self-driving cars see the world around them when they are out on the street? The next two videos will give you a glimpse in the Eyes of the Self Driving Cars:

Google’s Self Driving Car

Uber’s Self Driving Car

 

Hope you enjoyed this edition. Check out for more next week.

Milk is the secret to Sehwag’s energy and Australia’s

As the number of centuries grew for Virendra Sehwag, Australia increased its import of dairy products from India. The graph below says the rest!

Sehwag's Century

 

Source for Sehwag’s Centuries: wikipedia.com
Source for Dairy Exports to Australia: Dairy Exports

And the special why Australia did it:

Weekly #AI Digest: Persona, Personal, Official

The virtual assistants are ‘virtually’ (meaning literally) are rocking the world with the endless possibilities they are opening. From a humble beginning with Apple’s Siri that was a novelty when it was launched, there are plenty of virtual assistants today that have come up in the market. We have spoken about them in the past multiple times. Be it Apple Siri, Microsoft Cortana, Amazon Alexa, Google Assistant, Sales Force Einstein (still in works), they have featured here.

Einstein.jpg

So while we see a lot of marketing coming out of these virtual assistants and chat bots, what if we start getting answers from Donald Trump on our phones about his plans for the nuclear war, or Elon Musk to talk about his view on the next big planet to explore after Mars or Justin Beiber on his next chart buster?

MIT Media Labs is working on this to make it a reality. They are working on something called ‘Augmented Eternity’. Its based on an idea of “swappable identities” based on question prompt. With a personality lens, you can swap the identity and then you can ask the relevant question to the persona and get your answer.

This will help in streamlining the overall asymmetry we see in expertise and consultative sharing. Plus,  we will get the right answers faster and that too from the horse’s mouth.

So while this is still in works, these virtual assistants remain dumb. A social comedy experiment by Brent Rose to squeeze a flick on comedy based on stupid answers from Siri, Cortana, Echo and Assistant went horribly wrong. Check out the video here.

Now this makes one scary.

Perhaps this is the reason why Google has found that since humans suck at making good AI so they have built AutoML that is better in building AI than the company engineers.  Through a mesh of neural networks, the controller neural net defines a ‘child’  model architecture that grows as more data is fed. Catch being the child grows faster than the human child.

While world is still bracing these advances and we are still away from the trade-off of the smart, over smart and dumb version of AI, one CEO has decided to give the dirty of job of on the spot reviews about his team members to a machine over himself.

One new member that has joined the team meetings hosted by Salesforce CEO Marc Benioff is ‘Einstein Guidance’. Einstein is the company’s AI assistant that assimilates information said by all the team members in the meeting and gives on the spot feedback about a team member who may not be performing so well.

Wait till this comes to our homes and when parents will leave this dirty job to many more Einsteins.

Hope you enjoyed reading this edition.

TaaS and The 90 million Jobs Question

Most of the times simplification of things leads to such radical changes leading to far reaching impact. And typically technology has big role to play in this simplification. transportationInvention of tractors, the modern yarn, convenient information searching platforms, simplified shopping through websites and mobiles, watching TV with streaming technology, talking bots, digital payments,  etc. are all examples of simplification of an earlier complex process. But what happens when a simplified technology replaces an existing (and successful) complex technology.

It requires skills, commitment and patience.

Recently, Tony Seba, the famous Stanford university economist riding on the now very famous Seba Technology disruption model, has studied and shared his insights for the impact of Transportation as a Service (TaaS) and Electric Vehicles (EV) on the auto industry.

Seba Framework_v18

So what is TaaS?

Transportation-as-a-Service (TaaS), also known as Mobility-as-a-Service (MaaS), is a shift away from traditional vehicle owning and operating consumer habit to provide an alternative form of transport availability with on-demand services. The on-demand services include the basic like buying food to entertainment system and many more. This encompasses ride-hailing companies like Uber, Ola and Lyft providing the private transportation and bike-sharing programs like Divvy, CoGo etc and now commercial sharing options like Shuttl . When further extrapolated with the advent of electric vehicles and research being done on self-driving vehicles in Google Waymo, Apple Car, this provides a strong technology based platform for seamless, comfortable transportation experience.

  1. Transportation as a Service (TaaS) is the new normal that will hit the US in the next 7-10 years and soon after to rest of the world. There could be exceptions here too.
  2. TaaS will provide 95% of passenger miles within the next 10 years
  3. Consequently, transportation and travel costs will go down significantly
  4. The Technology disruption will grow along as Exponential S-Curve.
  5. High vehicle utilization would lead to lesser demand of new vehicles in the markets

Additionally, his key use case was the Tesloop where in the  Tesla lifetime warranty is put to test. Tesloop is a CA local startup that provides commute to people in a Tesla across CA. Tesla Lifetime Warranty ensures that the only upkeep cost that this company has to bear is the wear and tear from tires.

So how did Tesla come up with the infinite mile warranty?

Model S from Tesla had only 20 moving parts while a traditional Internal Combustion Engine (also known as ICE and is used by most cars in the Petrol and Diesel world today) have more than 10,000 moving parts. More the number of moving parts, more is the Electric Vehiclewear and tear and higher is the upkeep cost. The high price we pay for service of vehicles every 6-9 months is due to this high number of moving parts in the current ICE engines. A mere simplified design along with a clean energy source has led to a whole new set of dimensions of the auto industry. Lot more can be found here.

Impact of just this case study of Tesloop and Infinite Mile warranty:

  1. Low maintenance cost for the vehicles, Consequently, lower Total Cost of Ownership.
  2. Lesser to no consumption of oil so lower oil prices. May oil based economies will be impacted. So a certain shift in axis in the world economic powers.
  3. More vehicles in TaaS would mean lesser number of cars on the road. So, lesser auto insurance cost as there would be lesser accidents with lower density of vehicles. The insurance matrix would change.
  4. More cash in hand with people so improved consumer spending
  5. And what is not reported by Tony Seba – the 90 million jobs questions

The Indian companies do not appear to be geared up for this yet. If you go through the annual reports for Maruti Suzuki and Tata Motors, there is hardly any mention of the electric vehicles. Perhaps, it is not in a stage yet that it can be reported.

But the bigger problem appears from the famous campaign of ‘Make in India‘ initiated by the current regime of PM Narendra Modi in India.

The ‘Make in India‘ campaign for Automotive Sector has the following aspirations:

  • Automotive Industry to be “prime mover” of the Manufacturing Sector and “Make Make_In_Indiain India” initiative
  • Exports of vehicles to increase by 5 times and of components of 7.5 times
  • Add 65 million jobs on top of current 25 million commitment for the sector

When combined with Tony Seba’s research, this initiative would need a major overhaul very soon. Following are some of the impacts that are visible:

  1. Large automotive manufacturers need to start planning for appropriate interventions for electric vehicles – new technology, new assembly lines and new skills for workers
  2. Large automotive manufacturers would need to settle with lesser domestic demand and lesser exports
  3. Auto components sector will undergo a significant consolidation and shrink as ICE’s will be replaced with electric engines. Lesser complexity as lesser number of parts are required so lesser number of OEMs.
  4. Auto insurance will undergo significant overhaul. Lesser premium, pressure on margins and shrunk workforce are immediate impacts.
  5. Roadside garages will no longer be there.
  6. And of course, the 90 million additional jobs that are being planned to be added would require new skills and re-training.

Unfamiliar and unknown impact to Skill India campaign:

The astounding impact of this is surely going to benefit the Skill India sector where the current training organisations will have a ready pool of candidates to train. The question truly remains whether Skill India or other training organisations have this kind of capacity to support such a quick switch in skills. As per National Skill Development Council reports, the current capacity of their training partners for Automotive Skill Development Council is 6.5 million.

Due to several reasons these are operating at the rate of 67% capacity. So, the current number of trained and skilled workers is 4.3 million. 

ASDC Capacity

A simple math shows that it would require 20 years to skill the current commitments. A disruptive change in technology as discussed above would lead to additional number of years. Clearly, much more needs to be done. Its a huge opportunity for budding entrepreneurs.

Another big question is whether the youth is ready to brace this kind of change? No one would know answer to this till the actually impact is visible.

It is still surprising that this shrinking world has global impact of one observation and research concluded somewhere at a time when people were sleeping.

Whether the ‘Make in India’ campaign can lead to this transformation is still an open question? But these 90 million jobs are going to be impacted in some way or the other is beyond doubt! And that the current capacity is not ready to embrace this change, opens a new area for budding entrepreneurs. Surely, interesting times ahead.

Weekly #AI Digest: Fun Facts

Lets start with some fun facts in AI

  • Most the Personal Assistants and AI products are Females. Check ‘Siri‘, ‘Cortana‘ , ‘Cognea‘ to name a few. IBM Watson is the big outlier in this. But he recently got married to Cognea.

doll-1267255_960_720

  • It is predicted that by 2018, more than 3 million workers will be supervised by a ‘Robo Boss’.
  • A whooping 85% of customer interactions will not have human reps handling the requests
  • People will leave their current jobs and become Data Science Instructors
  • Policeman giving you tickets for over speeding may be a robot or a simple machine installed in one hidden corner of the road… you may not even notice.
  • Your McDonald’s burger may have no human serving you end to end
  • Most of the news, sports and markets, that you would read would be created by Natural Language Generators

ML LANDSCAPE BANNER(2)

  • -Spider-ManSpider-Man Bot shall lead you to the closest store where you can buy the comic and enjoy reading it.
  • There are plenty on-field Robots and AI Driven machines in the field and there are plenty of startups funded to control these robots.
  • More jobs? It was said that with ATMs the number of tellers would reduce leading to job cuts. Instead, banks opened more branches and more people got hired. Bots would lead to new jobs – bot controllers, AI friends to name a few.

Besides, your privacy could be at risk. Recently, Google’s DeepMind was found to have ‘legally inappropriate access‘ to a lot of private patient data from leading hospitals in UK.

ML LANDSCAPE BANNER(2)

And here is the video of the week: