Alexa for Business Solutions

I can´t find a reason to pursue my quest in using Alexa for Business Solutions.

• Paying for an Identity? There are many free identity providers out there you can use. And you won´t have yet another identity provider
• Paying for a managed device? I´d like to pay for a serious service or additional functionality. But compared with other products, some basic management capabilities are available. So why should I pay for little?
• In an well-defined and settled market for office productivity, who comes Amazon is thinking they can force customers is adding something new to the list?

There are only very few use cases:

• Having a device in the meeting room (public device)
• Having a personal device (says it all)

Conferencing capabilities are limited to Amazon Chime and Polycom, beside some others that seems to be irrelevant (especially in Europe).

There are API´s you can use to build your own functions and integration. There are currently only very few business related skills available.

How does Amazon thing that the very basics, also valid for modern workplace systems and mobile devices used in enterprises, does not affect a family of devices that are used in private and the business space?

• You will need to have your own skill store, with individually prepared skills.
• You will need to have provided skills (including some sort of usage measuring) from 3rd parties and you won configurations with regards to connected data and services to bring it to life
• There is a bunch of links to partner, like Salesforce, who offer their AI-based solutions. But that´s a separate service with APIs to be consumed by all kind of devices and other services. Not specialized on voice interaction. I see no quick and easy benefit here.

So, what the heck Amazon is offering here?

From the ServiceNow reference (a link to a whitepaper with some APIs) there is a great feedback from a developer.

He used the whitepaper to develop a bot-generic backend to deal with the IMAC and password self-service process. Alexa is just one agent who can access this service.

But Alexa in its current state is not sufficient to do reasonable business stuff:

“Alexa is not meant for conversational state, i.e. can not be used as Q&A mode. But conversational state helps to create Incidents and Changes as well including password reset or cloud integrations/orchestrations. When user providing answers to Alexa questions, its tough to map those answers to pre-defined utterances. If we can’t read everything what user says, its difficult to create an Incident or Change.”

From <https://community.servicenow.com/community/it-service-management/blog/2017/11/30/develop-amazon-alexa-skill-for-servicenow>

He described his approach and his perspective on a positive future.

If you ask me, Amazon has not much to offer that justifies the paid service.

You can grab a bunch of these devices and setup your very own application / skill backend, using Amazon cloud services for what they are useful. Doing so will also help to follow on the required security and privacy rules of your organization.

As with AWS it seems to me they haven´t really understood what is the difference between consumer, prosumer and enterprise. Same with Google. And they haven´t decided on which way to go.

As an IT Architect, I would start to develop bot based services that can be used by text, from any web site or any deployed enterprise chat application, like Jabber, Skype, Teams or whatever it is.

Once you get this sorted and ironed out, add a voice enabled device to the mix and try to solve the conversational state issue.

And if done right, you might not need to be bound to a single solution. Because there is no benefit in having Alexa on Windows and an iPhone if there is already Cortana and Siri to do the very same. Google Now for all Android devices and you are done with Android, too.

This covers already a huge area of available devices, allowing easy access to the devices. Because I had to learn, people love to talk in favor of writing (I´m just the other way around).

For anything else, e.g. voice controlled meeting rooms, removing the need to deal with the crazy dashboard to put down the blinds, turn on the TV and opening the phone conference, you might use a device like an Echo or Echo Spot.

To my approach to this quest this means to start with the following:

• Calendar integration (private and business calendar overlay)
• To Do list management (Microsoft ToDo and Wunderlist, might look into Outlook and OneNote, too. Both have great APIs for that)
• Reminder (build-in might be good enough). Need to make something that notifies me after 5 minutes if someone approaches me and I offer 5 minutes 😉
• Notifications! I want to get notifications on serious events. Like my cars alarm turned on, there is a weather warning or something like this. Maybe also so simple things like, kids come are back home, so I can welcome them and spent some time (to help to deal with my work-life balance)

 

Once these basics are done I will look more into the area of the application / skill space. There are many ways to automate stuff, Alexa might trigger this. However, it does not make much sense, if I need my PC to look up information to follow the conversation.

I´m thinking about a travel planning, where I start the process with a day and event, so the skill can check the calendar for all details.

There are travel policies and tools that must be considered.

It would be helpful I get some information on

• Gather the information for the trip. Might be there is already a hotel booked or a specific hotel has to be booked
• Make a hotel suggestion and place a reservation. Make an entry in the calendar with the required details.
• Check on whether going there by car, train, airplane or a combination. Present the suggestion, potentially create an alternative and prepare once this gets confirmed (book trains or flights, mark traveling time in the calendar with the required information)

All of this can be of use in an office or home office situation. However, I still don´t see the application of voice controlled devices in open space offices – and there are many.

As you see, I am disappointed in the offering of Amazon.

And I am still confident there is at least some help that can come with a voice controlled device. However, the number of voice enabled devices is already large. We just not started to use them properly.

The emperor’s new clothes

I recently had a meeting with some colleagues to discuss Citrix Smart Tools – an automation solution by Citrix. This fits into Citrix Cloud, an approach to make this IT veteran fit and attractive for the cloud era.

But oh my, beside Citrix Smart Tools is really something you can check out (just in case you still missed out to mature your IT to the level of automation).

All classic Citrix components and services, especially in the Citrix Cloud, are still the same.

This might be some good news to some people. However, I don´t think a Citrix session to an management console published on some server at Citrix is the right answer to deal with cloud-born applications.

It just blends the old with some new buzz.

Basically you can now:

–          Deploy a whole Citrix infrastructure and / or worker (the VDIs or SBCs)  in the cloud or on-premise on your hypervisor of choice using Citrix Smart Tools (scripted deployment)

–          You can do the same with the automation tool and suite of your choice (e.g. the ARM template of the Citrix 7.x demo environment)

–          You can order a Citrix management backend (Citrix Cloud) and host your workers wherever you want

–          Use Citrix as a service, where you don´t need much knowledge about desktop / application virtualization using Citrix XenDeskop / XenApp Essentials

Especially the last one, Citrix XD/XA Essentials comes in 25 user bunches you need to pay monthly. It is payed with the Azure subscription model. But there is no demo mode (at least not today).

Therefore you´ll be charged for a single month 25 x (12$ Citrix Fee + 6,25$ RDS CAL + Azure IaaS) is roughly 500$ per month.

Quite expensive for a demo!

And it is not that easy to set up (at least if you are unaware of the traps).

Without much knowledge, it is much cheaper to deploy either the Demo ARM template on Microsoft Azure or use Citrix Smart Tools and deploy to a cloud or hypervisor of your choice.

 I hope Citrix will invest into developing this into a true cloud service – or milk the cow until the apps and services of this world are SaaS-only.

 

Emerging work styles and your part in it

When reading IT news or talking to colleagues there is one dominant and controversial topic: the balance of work and private life.

 

This tends to hum about companies rude and anti-social behavior, lack of adoption of new technologies – or the adoption of new technologies and not protecting the employees from their evil effects.

 

Recently I came across two articles who were talking about the change in workplace technology.

 

My personal takeaway is, that we are right now not looking at the full picture. Because there is a technological, a social and a personal aspect to it.

 

We had a clear definition of computer = work in our head and many still follow this paradigm. Now we use computer (in the incarnation of smartphone and tablet) in private life.

 

This is a significant shift and we haven´t adapted to it and some just take the short cut of computers are always available therefore we always work or are supposed to work.

 

This is just too simple. As I tend to write long post, I allow myself to elaborate a bit on what happened.

If you like to save some time, it is save to skip the next chapter 😉

 

Excurse

IT began as part of our work life.

Due to cost of computer and peripherals, functional capabilities and a significant learning curve universities and large companies used to have them.

With Commodore, Apple and Microsoft (beside of others) the computer moved on into the living room or child’s bed room (if you are happy enough to have your own TV or later when you could afford a monitor).

The transition took many years and people developed pattern on how things got to be:

You have a computer at home for fun and “learning computer”. With this knowledge you can have a good job, working with computer, earning more money in a white collar job.

 

At this time, let’s say the 70´s and 80´s, there had been many “dirty” or blue collar jobs even though another wave of automation took place. I clearly recall the “disruptive” changes in print or car manufacturing. This change and peoples adoption to it took years again. The need to get along with computers increased.

 

In the late 90´s and early 2k´s the next shift happened: the Internet.

 

The only way to interact with the Internet at the time had been a computer. Many cool things happened there, so everyone wanted to rush in. It is the time of eBay, early Amazon, online games and the search engines. Information, products and fun at your fingertips – at home.

 

This offered the potential to work from home (using corporate email, VPN and extranet). Here the crisp and clear model of the work / life segregation stated to stumble. First only for a few.

Apple crashed down a long existing wall with the Telco’s to establish the eco system of mobile internet enabled devices. It started small, slow and cumbersome. We talk about January 2007 and 2G EDGE as the live wire to the internet. But things tend to develop fast and take a look where we are now.

Basically, we got more options and we can reach out without moving, faster and in better quality. Life like if you can afford it. And we are happy to transit our work from loud and dirty environments to nice and clean ones. As young people what they want to do. White collar jobs is what they want.

Re-defining the workplace – in our head and between us

Technological changes have been followed by social changes before. And we adapted and made the best out of it.

The modern workplace is not on your desk, your phone or in the cloud – it is a mix of tools and technologies to allow various work style for the benefit of the company and the employee.

Therefore my repetitive call to take the personas analysis more serious and step back from the one-size-fits-all solution. Today there are multiple supported work styles per company. But they are not available to anyone who would benefit from them, some are missing and most of the time there is no education or coaching about how to choose the best for yourself and what comes with your choice.

Even a change back must be accepted if you learn, something is not working for you. Best example is working from home.

The biggest effect is the clash of our previous values and our options.

• You can follow existing work styles, e.g. with arrival at work before anyone else and leaving later, too. No new technologies are needed to do this.
• You can be always available and work on request. New technologies help here to be available, get the information securely and get the work done.
• You can try to set up rules to balance these extremes (as proposed in many home office guides).

Just some rough and strong colored examples to express the spectrum of choices.

In any of these cases you should be aware what you have chosen to do. And you need to set up rules and make put clear expectations in place with your family and social contacts on what is to expect, what is feasible and what´s not.

This balance puts you in the center. It´s not the others or the company, it is primarily you.

Your company has done its job if it helps you in this process. This is a matter of change management if they want to be successful.

If working from home means to your partner to have full access to you, can take you out of your thoughts or have a quick ride to some shop – you will not be able to live up to you or your colleagues expectations.

You can´t start early to set up these expectations. You might lose friends and the situation in your family might get tough. If it does not work for you, you need to change or adapt. Potentially switching the role or the employer.

But we also need to be honest to ourselves and others, because the line between work and private is easily crossed. If you spend 2 hours working and 1 hour surfing the Internet, you cannot blame work for being absent for 3 hours.

Making use of workplace technology for us

And you can easily get lost. Especially as we see changes in the working habits. Asynchronous email is still present. But collaborative work, the one that I think will survive the AI introduction as it will remove all the dull work and leave us the intellectual, social and innovative tasks, asks for a more direct communication. No doubt this is part of the success story of tools like Slack. Previous technologies like SharePoint social features, Yammer and enterprise chat have been the predecessors, facets of solutions like Slack. And this will improve, as it allows to integrate bots and other office, development and analytical automation to be integrated.

Physically device form factors will be the ones we are used to, from phones to PC, extended by input sensitive TVs (e.g. whiteboard NG), AR devices, 3d printer / rapid prototyping, volumetric displays and extensible / foldable displays and devices using them.

Hopefully we will see a tighter integration of the various building blocks, removing the usability gaps we experience today.

I just read about Magic Leap Inc.. When I think the specs of the HoloLens, there is still a huge potential to increase performance and reduce cost. But it will take time to get a sub-$200 device for each one in your family so you will not buy a new TV. As outlined in the Magic Leap article, this can have an incredible impact on many industries we are used to today. And offers new chances to traditional aspects of life, like story telling, presentation and sales.

This will further remove the need of there and when we work. Potentially even how long we work (like in work smart not hart).

Side effects are already visible today. Companies need less floor space (due to work from home). Meeting rooms are booked with 3rd party professional services, based on the best way to travel to it and not based on a fixed location like your companies offices.

People will commute less, resulting in reduced costs, stress level and 1-2 hours more time to balance between work and private life. Well, some diseases might spread less fast, as you will spend less time in places with many people.

The social aspects

The perspective is quite positive I think. It also fits to my personal experience in the last couple of years in an increasing amount of time working flexible and smarter.

There is still the struggle with your family, especially if they don´t live up to the level of embracing new technologies as you and your colleague do. And if you have trouble setting up the expectations – and reasons for this are many.

If you have been traveling for 5 days a week like I did for years (even though it made no sense and there was plenty of opportunity to change it – company responsibility!) you know how it hurts when you tell your little ones you need to leave tomorrow or now. They begin to plan their lives without you.

And it is a great relief if they start to ask you, when you will be traveling again. The situation is upside down, as they plan their lives with you and just want to know when they might miss you.

I am looking forward to the day one of my kids sends me a floor plan of their flat or house, that will include dedicated home office rooms for each of the adults in addition to the obligatory bed / living / kids and bath room, kitchen and all the rest. Without a dedicated parking spot, as you will book your ride from A to B for every larger or smaller commute.

The pain I outlined above, of your kids planning their live without you and asking you why you have to leave again, will be hopefully a chapter of the past.

Friends and social contacts will be still as strong as before, but will be lived differently. There will be the risk of connected isolation – as you have it already today. But it is also a great relief to all the people who are introverts or otherwise not very happy with social interaction. The challenge is once again, to balance your distance to a social life as we know it. Or we belief it is or should be.

Quite a perspective I would think is worth working towards it. As long as it does not ends up like the fictional “real life” sketched in the book “Ready Player One” by Ernest Cline (one of my favorite stories).

Change happens elsewhere, not here

I am getting tired of reading articles suggesting to use the Personas methodology – while already having the same answer. Sometimes a bit covered you notice, all the solutions to address the needs of each group are also the same.

How can this approach be beneficial?

In my career I discovered a bunch of additional groups of people with very specific needs. And their discovery – and addressing their needs – brings much gain. High acceptance and productivity uphold or even improvements.

The Personas method can be really helpful if you work with an organization with no clear reflection on who is working how for them and where what values are generated.

Sales / Management, Knowledge Worker and Task Worker is just not good enough. How about in-house developers or task worker who need to enter large amount of information into terminal based systems?

Technology and IT services and processes have advanced and there are plenty of options to choose from to address such demand.

My advice is, to challenge anyone who is proposing a personas analysis to find out if it is made for real discovery and support on decision making or to sell more days to the benefit of the consulting company.

For myself, I must confess to have implemented too often based on higher command – and realized during design / implementation or roll-out how large that gap is. And started to mitigate.

Just like weaving IT Security into everything you do, you should take care for a thoughtful analysis of the consumers of your service. If you do this early, you might be a bit slower at start, but you end up in time and budget, instead to discuss the effects of cost of poor quality while you try to close your project.

Potentially political incorrect

This morning I waited for the daily storm, unleashed by my family preparing for school, to settle.

While I worked hard on showing equability, the latest IKEA catalog came in handy to focus my mind on it. While I turned page by page, starting in the kitchen section, I wondered why all food shown was only fruits and vegetables.

And in fact, it was all very “vegetarian / vegan” oriented. At (I guess) two occasion, fish was shown. This makes you think on

• who is the addressed group for these catalogs
• how do they want to present themself
• what messages do they transport

I am a huge fan of diversity. And my antenna reacts to inbalance. My senses started to ring here, as I think this way a very strong message is put into the pages of this IKEA catalog.

Might anyone take care of his well being based on his knowledge or believes, I don´t mind.

But I am also free to reflect on biased corporate messages send.

Not even Köttbular has been shown, or the typical kids food you cannot escape (Wiener sausage, fish finger and chicken nuggets).

Ok, feeling better now. Will return to work.

Cheers

A wild guess on how cars will look like in the future

During my quest on looking for a new family car I came across some interesting articles.

While test driving an BMW i3 I noticed the tires to be different.

Obviously there is a trend to larger cars (SUV) and larger tires. The most common car in Germany, VW Golf went up from 15″ tires to 17″. Larger tires are looking much better compared to smaller ones. With the larger tires comes a higher friction, as the contact surface is larger. As a result, the amount of fuel needed is higher. For electric vehicles this is not very favorable. Therefore the large tires of the BMW i3 are very thin. Well, electric driving requires you to break with some of your used driving habits. I realized to drive slower on a highway driveway.

Takeaway 1: People like how their cars look. Large tires are preferred

Takeaway 2: Tires might get thinner to reduce the friction

The aerodynamic is a second aspect and some car manufactures and suppliers are developing technology needed to remove the outside mirror, respectively all mirrors. Japan is leading here, as cars can drive there mirrorless with the needed alternatives. Tesla showed a modified model without mirrors a while back.

Takeaway 3: Outside mirrors (all mirrors) will go away and substituted by cameras

If mirrors go away this is just an step in-between to autonomous driving.

While outside mirrors have always been a hassle (think about broken mirrors due to parking accidents or vandalism), the coverage of the rear tires has always been an issue. Even thought this could save a lot of energy during fast driving (e.g. on the highway), people don´t want it as it ruins the look. Potentially this can be fixed with an automatic coverage while driving on higher speeds, as it is done with the aerodynamic spoilers (Porsche or the Audi TT come to my mind here).

Nevertheless, if you look closely on some previously presented prototypes or cars presented in movies presenting a potential future (for example I Robot with their Audi future show).

But the aesthetic perception can change over time.

Takeaway 4: (Rear-)Tire coverage is unlikely due to ascetics

Cars already offer many assistant systems. As usual they get introduced into the top-level cars and as prices for the technology drop, they become diversification aspects to the next lower class. Many of the business cars already offer break and distance control assistants, traffic sign reading and parking assistance. Instead of implementing a bunch of separate components (e.g. ultrasonic sensor, radar, camera etc.) they will be combined and offer surround capabilities. Like needed for the mirrorless car.

One of the components needed here is LIDAR. Some companies develop very compact and capable devices, with reduced height. Right now you notice them as an displaced and not very ascetic addition to the top of a car. To enable autonomous driving and reduce cost and complexity, these devices will be an integral part of future cars. Car designer need to make sure to integrate them nicely, without reducing their functionality. But this problem was already solved for antennas. We can assume that there will be an accepted solution.

However, the current generation of LIDAR can be tricked. Either this technology must become more robust or will be exchange or enhanced with an other technology.

Takeaway 5: Car roofs will have some kind of soft bump to house the LIDAR

The question is, how roof less cars might get autonomous drive – or if these cars will be the last ones to be driven manually.

Right now cars grow due to the safety needs of todays drivers. There are two aspects to it, I noticed myself and heard from others:

1. Easier entry and exit
2. Better overview while driving

Quite self-explanatory. But once cars become like outlined in Elon Musk´s Masterplan Part Deux, the easier entry and exit will remain (especially in the aging societies). But the overview part is not required anymore. And for the easy entry and exit, there are other alternatives we will see in the future. But this will a decade away. At least. With the autonomous car economy as outlined in the masterplan (but also in many other visions from mobility experts years ago), we might see an Uber-ish mobility. Car models will differentiate in the number of people they can take with them (e.g. 1 to 3 or many, with a higher variation of seating abilities. Ok, there will still be people who want to sit in driving direction to please their comfort. And they will differentiate in the comfort or luxury they offer.

In such an economy it will be less likely to own a car. It is more likely to see the fleet car market to become dominant. And transportation services (think flat rates) will be normal as for

Takeaway 6: Cars types will change based on technologies, ownership etc.

In this case, AI and the processing of a huge amount of data will be used to optimize the fleets and to reduce costs. Autonomous driving will on the large scale will have a tremendous social impact:

• Elderly people remain flexible, self-controlled and can travel securely (for themselves and others)
• Children can travel independently from their parents (no Mum taxi)
• Less space is needed for parking lots for customer facing companies (drive-in)
• Less space is needed for parking lots for employees
• No taxi drivers. Resulting in
  • Saved time (as they will not take the long tour)
  • No need to have strange conversations or to listen to music you don´t like
  • No robbery of taxi drivers

Cars will be easier to service, as there is now a high interest to reduce cost.

For example, the German postal service tried to work with car manufacturers for many years. Finally they approached a start-up, supported by regional universities. They´ve build a car from ground. The project had been very successful and the company is now part of Deutsche Post. Up to now 1000 of these cars that are speed limited to 80km/h and with a range of roughly 120km are in use today.

One is servicing my area, what is really sweet.

Plan is to manufacture up to 10.000 per year. The 70.000 vehicles at Deutsche post will be changed to such cars where possible (estimated 30.000 cars are needed). Right now there is already a high interest from other counties and competitors. Especially in high density areas, a clean car can make a difference.

StreetScooter spec:

• 20.4 kWh lithium-ion battery
• 50-80 km (30-50 miles) real-world range
• 80 km/h (50 mph) top speed
• 30 kW electric motor, FWD
• 650 kg payload
• 4.5-7 hour recharge

In addition, many delivery services in Germany test a robotic vehicle to transfer deliveries to your home. These little robots are small, can drive on the sidewalk and have a action radius of 6 km.

Ok, a last one. I just came across a German study on the effects of electric mobility. And this study suggests the return of a gear box into electric cars. Right now all BEV have a simple gear box, reducing the rotational speed by 3:1.

Simulation suggest to have more variety to improve the efficiency in various driving scenarios (urban, road, motorway). Sure, a more complex gear box will introduce additional weight to the car. There is an interesting analysis of existing electrified cars, arranged by car weight. And to achieve this win I would think to keep the gearing automatically, based on driving habits, trip profile (e.g. distance to destination and road profile). As online connected navigation systems are more the norm, the gearing could be changed automatically by the type of area you are driving in. And for all the people who think they can do it better than any (hopefully proper programmed) computer, can have some steering wheel paddles to switch gears on the go. However, I would charge extra for this 😉

Takeaway 7: There might be a gear box in car, but hopefully it will be fully automatic and there is no stick as in today´s cars

As said, mobility will change significantly over the next decades. Our grand- or grand-grand children might wonder why we traveled as we did today, in cars that looked like today.

 

 

 

 

Artificial Intelligence – face off

Sometimes it is hard to relax. Today I try to sit in my garden, enjoy sun, wind, flora and fauna.

All quite natural and making you think about the artificial intelligence (AI) hype lately.

The idea of intelligent machines, in human like appearance is old. We can use books and movies as reference.

When I started in the early 80´s with computers, AI research was already there. And many concepts and principles from that time are still in use. It was limited by the computer capacity. Therefore I see the current activities as a Renaissance of it.

Nevertheless, one of the oldest question has not been answered yet: What is intelligence?

So how will we know what artificial intelligence will be.

We even have problems to differentiate between the various human abilities, beside of our inability to take other beings on this planet into consideration.

For this post I will stick with the diffuse term of AI used out there while I talk about the AI Renaissance, because it comes twist.  Based on my observation, we see two types of AI emerging:

1. AI models based on classic computing hardware
2. AI models based on specific hardware

Take it like the two faces of AI.

In ancient Roman religion and myth, Janus is the god of beginnings, gates, transitions, doorway, passages and endings. He is usually depicted as having two faces, since he looks to the future and to the past

The first face of AI is a straight progression from where we started. We see a lot of benefits from it. Voice recognition, text analysis and translation jumped significant in quality and speed. Analysis of images and video to help identify objects and situations help to search for such content in a new way. For instance, I love to go to my OneDrive website and look up my uploaded images by tags. Really helpful if you search for some content.

This type of AI is very specialized to solve one problem at time. It can either convert audio into text or analyze your posts or search behavior to identify potential diseases.

Same is true for the type of AI used to make bots useful.  As I wrote in an earlier post, such bots and their AI will either need to know your identity to learn about you (something that will require trust or ignorance) or it can only use the overall base of data going though, with reduced insight into individual interests or decisions to learn from. Despite of what we sometimes do at school, we learn from context.

At the moment we gain great results from trained systems, based on conventional computing hardware, helping us to achieve things we had to do manually before. While you needed a well-trained secretary to take down a dictation, you can now record it on your phone and it is even able to identify who (by name) is talking. Some companies even base their customer identification on voice only.

Personally I found some military application really scary: Based on a simulator, human pilots and a standard PC running some code doing something that is called “deep learning” (a new entry in the dictionary for buzzword bingo), a team was able to develop an software agent with the ability to control a plane (in the simulator), reacting to the activities of opponent aircrafts and “winning” the situation. Ok, sounds like some fictional movies merged together. But the really scary part to me was, the ability to take the trained model and run it on a Raspberry Pi type of hardware. Learning is harder than to apply what you have learned. Sounds reasonable, right.

I don´t want to imaging such agent to be uploaded to a commercial drone, leaving the rest to your imagination.

A much better application is to have a trained system to validate financial transactions of a company or your private expenses, as part of you tax application or to get recommendations to rebalance your insurances. Especially when there are changes to your contract or what is available on an changing market.

The automated analysis of medical lab tests, x-ray or MRT images might help to improve to reduce misinterpretations or just errors. And having access to a large number of information might help identify quickly, if something special is going on. Like increased numbers of symptoms, the trend of influenza during the annual cycle or if there is an epidemic on the rise. We have some good measures in place. By using this technology the hope is to be quicker and more precise and can process such information 24/7 with a lower error rate as with humans.

The first face of AI is already quite amazing.

Now let´s take a look at the second face of AI. Universities and companies like IBM research on circuits able to simulate biological components like neurons and synapses. With the right amount of error and variation as we can observe it in biological systems. This seems to be important for some processes. These circuits are different form the specialized CPUs developed to speed up the first face of AI applications, using optimized GPUs and DSPs with a large amount of cores to help with this quite well scaling parallel programs. These are mathematical models, oriented on the biology, but not sharing much with it.

The new circuits of the second face are different, as they are more like the biological systems, but based on silicone and other material and “programmable” to some degree. Like with the first face you are able to define models by defining rules and structures of interconnections. It´s not like a “bunch of neurons” on an electric bus. The expectation is to build more complex and dynamic AI with them, despite of specialized functionality. Generic intelligence is their goal.

This is still quite early in the development and to see the advancements of the first face AI helps to get money for R&D of the 2nd face AI. There might be some “industrial limitations” to this approach. Of course, the hardware can be produced based on today’s industrial processes. They can be programmed to have the same initial structure. But to my knowledge, there is currently no way to make them all identical, based on a previously trained master. Especially if randomness and weakness comes into play.

Therefore we would need

• The special hardware
• The neuronal structure to start with
• A training environment with training data and tests to verify the state of training
• Some time (e.g. training cycles) to go thought the training environment

And potentially, at the end of the fabrication / training class there are some excellent, some good and some fair performing systems. When they become even more complex, we could end up with the situation in biology, when the results are quite individual and it is hard to predict if something is ideal for the task.

This degree of individualism can result in a 1:1 production scheme, as we see it in biological systems.

This is a significant differentiation to the 1st face AI. Here you have

• A performant hardware system with some specifications needed to crunch all the data during training
• A mathematical model in code (on some Git-Hub server 😉
• Some training environment and training data (can be found on Git-Hub, too)
• Test routines and test data
• Some time (e.g. training cycles) to go thought the training environment

 

But as you will end up with the same result if you feed the same data in the same sequence on this model (deterministic), you just need to do it once. The resulting model can be copied and operated on a different type of hardware, capable enough to feed in the data to process and run the model.

This will result in a 1:n type of model

At this point I find it hard to predict if we can overcome the limitations of each of the here proposed AI models.

It would require mass production of 2nd face AI like the 1st one or more general intelligence based on the 1st face approach. Therefore I don´t see any global scale threat to mankind following both approaches. And even if it comes to a first artificial conscious mind (that might have a problem to reproduce), there are some thought on whether these would be interested in us at all.

The social and economic changes resulting from this type of advancements are more imminent and need to be answered. And we will need rules to control the usage of any kind of AI for police and military services. We might need a ban on some AI use, like we have for atomic, biological and chemical usage.

I started with stating that there are many stories about AI. It seems to be an increasing interest. Here is a bunch of shows I recommend:

• Real Humans (and the U.S. remake Humans)
• Caprica (with a religious twist, explaining much from the series Battlestar Galactica)
• Almost Human
• Delete

To my bad, I missed the title of a book that had been recommended this week on 1Live radio. The plot is about how software agents and the type of AI as we know it might affect our society. Any hits or recommendations on this are welcome in the comments.

Stay natural intelligent 😉

The effects of managing mobile devices

Today I came across the following Channel 9 video I had to watch.

 

As an orange [0] veteran (or former long term Avanadee) I am always keen to see what these folks are up to. You need to know that Joseph Paradi is right on the numbers, but I missed the emphasis on the fact that Avanadees are hard core technologists and most of the time either traveling or with their customer. There are other consultancies out there with a more relaxed work model, but Avanade is that way.

Because people are working hard and busy they balance their work and private life by utilizing not only their corporate devices. Often they are even required to work with customer devices and are quite limited to reach out to the Avanade services on the orange “corporate network” called Internet as Joseph put it.

 

Therefore it is no wonder that the introduction of the mandatory usage of an MDM let to a 30% drop in connected devices. The mail system as one of the core collaboration services is a great indicate in most organizations. And communication is key in a company of regional diverse knowledge worker.

To identify this as a productivity issue and to put it to high attention is absolutely right.

 

Same as the conclusion: It is about protecting the data, not the device

 

Devices are cheap nowadays. I guess Joseph also knows the number of devices that are lost or stolen within Avanade per year.

 

Another thing you need to know about Avanade is the high standards on IT security for many years already. Not only technical measures (like the MxM approach), but also in educating the employees.

Avanadees often work with all kind of sensitive data. Not only personal data of your colleagues or candidates, but RfPs, contracts and everything you can think of. Therefore there is a long tradition of awareness and activities to keep data save.

As they are also a leader in many areas, they also lead in the area of business digitalization. Sure, in such a work environment you greatly benefit when you transfer pen and paper processes to workflows on whatever platform.

 

If you think about it, MAM and quickly moving to capable Windows 10 devices is a way to accomplish this.

 

Funny side note: These aspects had been discussed in the ASG [1] mobility group [2] years back. And not only there. Running into this issue shows once again that even with such a company, you can trouble projects by assigning the wrong people to it. But then again, please notice, Joseph is from the internal IT (no offense!). It is just a different group of people as the thousands of consultants out there serving people on that matter.

 

Beside this reflection, the discussion is worthwhile watching as it covers some core aspects like

• Offering of flexible work styles
• Dealing with BYOx
• Business digitalization
• Dealing with tech geeks
• Identifying where your productivity comes from

 

A couple of days ago I re-read a text about the benefits management brought us in the 20th century and the challenge it has to deal with in the 21th. It is about how to improve productivity of knowledge worker. Avanade is a great example.

 

The BYOx is dead discussion is based on some specifics to the European markets and offering people money instead of devices. Nevertheless it is a topic, as even though a company does not offer you money, people own private devices. And every motivated person will use the build-in ability to connect at least to the mail system to stay in the loop. Or to use a mobile phone to have access to the calendar and be aware of schedule changes. You might forget about this if you are an IT pro and used to this stuff. And this was different years ago before the percentage of people with a capable smartphone as this high. I don´t think it´s dead. It is the same discussion as years ago when it started. Only some other aspects are not so strong in Europe.

 

Joseph talked about Azure AD, but did not mention the MAM (or MxM) solution being used.

 

Microsoft is offering multiple paths

• Microsoft Intune (Cloud)
• Office 365 (Cloud)
• SCCM 1602 (On-premise, private cloud)

 

And these various options are getting aligned more and more. Sure, if you trust Gartner, Microsoft is still behind Airwatch, Mobile Iron and XenMobile, but ahead of many other.

I´ve seen customers just looking at the Quadrant and calling the Airwatch sales-rep and got a bunch of licenses. Don´t get me wrong, these are great products I love to work with.

 

Sometimes I just miss the process of developing the awareness of

• What do I want to accomplish?
• What does my landscape looks like?
• What do I want to manage?
• What is a suitable solution/product?

 

But this might be a personal thing 😉

 

The rest of the video is an interesting discussion on the silo dilemma of It organizations I talk with customers about for ages. Maybe you will not find nothing new here (good) otherwise I recommend to take this as a spark to take a look on it again.

 

Annotation

 

0 – Orange is the corporate color of Avanade. And no, this has nothing to do with the Dutch region 🙂

1 – ASG as in Austria-Switzerland-Germany

2 – At this time we had some “Solution Focus Groups”, gaining knowledge and working on offerings and solutions in particular areas. Mobility was one of them. A special challenge had been to define a boarder between the application development team and the infrastructure folks. I always voted to develop a single solid view and story. Once you get into a deeper discussion you need to show your ability to serve a customer in a holistic way, with experts from multiple technology areas.

 

Link to webcast: https://channel9.msdn.com/Series/Endpoint-Zone/Endpoint-Zone-Episode-15-Avanade-using-Intune-MAM

 

Link to Gartner Study: https://aka.ms/epzworkplacestudy