If your home has a pool, you may find some difficulties shopping for home insurance. This is because swimming pools often attract additional risks – the neighbor’s kids sneak in for a late night skinny dip or a toddler falls in and nobody is looking. Yikes! No matter who gets hurt, you can be found liable.

For this reason, most home insurance companies will add a fee if your home has a pool. They will also require a fence at least 4 ft high surrounding the pool to prevent children from wandering in.

But what if you also have a diving board or slide? You might find yourself getting rejected by nearly every home insurance company in town. I mean, the increased risk of injury is pretty obvious…

Lucky for you, Young Alfred does work with carriers that are willing to insure your home with a diving board or slide, and these include Swyfft, Hippo, and State Auto.

You can check your rates by filling out a form here.

At your service,
Young Alfred

There’s a wealth of great content on the internet when it comes to software testing. As a DevOps consultant, though, I still regularly get asked about testing best practices. So, I thought I’d collate the information into one place as part of a short series on Software Testing.

When it comes to making code changes, there are a variety of tests you can perform to ensure everything is working correctly. Testing is fundamental to building quality into software. By making sure you gain fast feedback on code to make impactful changes sooner rather than later. The rate at which you can ship or merge code depends entirely on the speed with which you can test it!

The top four types of testing produce very different outcomes, and there are certain benefits and downsides to each of them. Which testing method is most beneficial? This article covers the differences in the four primary types of tests and determines which is the most valuable to your development pipeline.

Unit Testing

Unit tests refer to the different methodologies and functions used to test the individual parts, components, and modules that make up the whole of your software. The tests are usually low level, and also relatively inexpensive to automate. Build an abundant number of Unit tests into your pipeline early. Split these into two categories; locally-run rapid sub 10-second tests and those that take more time. Move longer tests to your Continuous Integration pipeline to minimize the negative impact slower tests can have on dev productivity.

Unfortunately, Unit testing is often frequently overlooked or considered to be unimportant in favor of Integration tests. And in order to shorten development time and save financial resources, Unit testing is sometimes the first thing to be cut from projects by customers.

One of the great things about Unit tests, though, is that you can verify separate modules and functionalities prior to merging them together to be tested as a whole later. Unit testing should occur throughout the process and prior to Integration testing or System testing. This is recommended for a variety of reasons. These include bug prevention, to document trial and error throughout the code, clarify specifications with the Product Owner, and perhaps most importantly, save time and money.

Although problems can arise throughout any project, it is always simpler to fix the earlier you identify a problem. Unit tests prevent and eliminate potential issues before they even have the chance to become a problem. Even though Unit tests will not identify and solve all of the problems you may run into, they can still help prevent deadlines being pushed back.

Smoke Testing

Another popular test is the Smoke test. In terms of time frame, these should be sub 15-minute tests. as their aim is to spot leaks quickly. Commonly referred to as “Build Verification Testing,” this method verifies that critical elements and functions of the software are working properly. Therefore, create tests that identify common faults. Now, this method of software testing is not going to provide you with in-depth details about each function. However, the results will indicate as to whether or not you should proceed with further testing. If the software fails a Smoke test, it is best to go back and do a rebuild with the required fixes in place. Otherwise, you will lose time and effort running further detailed testing.

Like Unit tests, Smoke tests are most helpful in exposing issues early, and you can coordinate the two together. Smoke tests can also confirm if previously made changes are negatively affecting major areas of the program or not. Smoke tests can also be expanded as needed as well. This will allow you to automate testing throughout the application. Overall, use Smoke tests early and often. But they are not truly independent tests, as they work well with other test types.

Integration Testing

Once you have completed multiple components of your software, including Unit tests on each individual elements, you may want to run an Integration test. This type of test will allow you to verify that the varying components of your software will work well together. A downside to Integration tests is that they are more expensive to run than Unit tests. This is partly because they require multiple parts of the software application to be operational. The benefit of Integration tests is that interactions between components that do have errors will be brought to light by the test. Making their expense worthwhile. If you have not run Unit tests on each component, you may run into additional work and time fixing individual components in order to integrate them.

There are a few different ways to approach Integration testing. You can wait to implement Integration testing once the whole software is available. However, this may allow errors to continue throughout other stages of development until that point. Alternatively, some individuals prefer to use the Top Down or Bottom Up methods. Either the top-level or bottom-level units are tested first and then each level is tested step-by-step until completed. These two methods can be combined in an approach known as the Sandwich or Hybrid. Whichever method you use, verify that Unit tests are complete on each component. And try to complete Integration testing prior to System testing.

Acceptance Testing

After you have spent time running Unit tests and Integration tests, you may want to run an Acceptance test. Use these tests to verify the software meets the clients’ or business’ needs. Again, this type of testing is more expensive to run than Unit tests. On top of that, the entire software must be available and operational in order to perform Acceptance testing. Acceptance tests focus more on replicating user behavior than the other three test types, but they offer a measurable performance of the software. They also provide a way for the client/product owner to reject different changes you have made if they do not meet certain needs.

Companies tend to start out with manual acceptance tests, often in the form of peer code review. As development pipelines become more sophisticated the trend is to automate these tests and even outsource them. Execute Acceptance testing via different methods.

• Canary Deployment: Run acceptance tests through canary deployment to make sure there are no misconfigurations with the production environment.
• Alpha Testing: Which is usually completed by members of the business who are not directly involved in other aspects of development or testing, i.e., product management or customer support.
• External Acceptance testing: Which is run by individuals who are not associated with the group who has developed the software. Sometimes customers of the client or end users perform Acceptance testing, depending on the needs of the organization.

Each test type has its time and place in your development pipeline, but Unit testing can address the most crucial needs of the process and highlight many errors sooner than the others, saving time in the pipeline. Which is why Unit testing should be the foundation you start with as part of your development process. This practice is vital to DevOps implementation to prevent errors being carried further downstream and having an impact on other team members within the pipeline. Fast feedback and problem-swarming in this manner ensure quality assurance rather than quality control.

Use Unit testing in conjunction with other types of testing or independently. As such tests can be run with all or only parts of the software operating. It can be automated or maintained manually as well. Though DevOps best practices encourage automating as much of your testing as possible. Another major advantage of Unit tests is that the practice helps ensure minimal testing time, preventing lost time and effort. With clients putting more and more focus on wanting to constantly update and provide changes to their end users, Unit tests will ensure that customers are regularly kept up to date with the latest software innovations.

---

Caylent offers DevOps-as-a-Service to high growth companies looking for help with microservices, containers, cloud infrastructure, and CI/CD deployments. Our managed and consulting services are a more cost-effective option than hiring in-house and we scale as your team and company grow. Check out some of the use cases and learn how we work with clients by visiting our DevOps-as-a-Service offering.

Artificial Intelligence is all the rage in business right now, with companies clamoring to be at the front of the AI race. According to this analysis by Bloomberg, more and more organizations are mentioning AI on their earnings calls.

While the concept of AI is becoming a bigger part of business conversations, implementation is often stagnated due to misconceptions about the technology.

So,  what are the most common misgivings about AI?

AI is Going to Replace All Jobs

Understandably, some people may feel that their jobs are at risk with AI-powered solutions. A machine that is precision-driven and has no need for benefits or personal development is surely an attractive option for businesses, right?

Actually, it’s not that simple.

A report from Paysa suggests automation jobs will put 10,000 people to work, and big companies will spend $650 million on annual salaries to make it happen.

So is there truth in the belief that AI based automation may strip hard working humans from their well-earned jobs? Not likely, according to current statistics. AI takes implementation, constant improvement, and quality assurance measures that can only be provided by good old-fashioned human workers.

Which leads us to the second of the AI myths:

AI Algorithms Will Make Sense of all Data Without Continuous Optimization

One of the most common concerns about AI is that with the help of machine learning, AI bots will be able to constantly improve their level of comprehension and development without the aid of humans.

Is this a valid concern?

ScienceDirect states that there are important meta-modeling analysis and interpretation activities associated with practical decision making when it comes to AI solutions. Simply put, AI models are not “set it and forget it” solutions. An autonomous bot can become a rogue liability if it’s not guided by a provider who can assure that a certain level of quality assurance is both measured and controlled.

While automation is the name of the game with AI, any integration of the technology should be able to be measured, managed, and re-configured when the needs arise.

AI Engines are Too Expensive and Too Complex for Non-AI Specialists to Implement on Their Websites

One of the constant challenges that companies are facing is that they feel their current talent pool will have difficulties in implementing and optimizing AI solutions. Another concern is that AI technology is expensive and may not provide an acceptable ROI.

AI Specialists can make millions over a four or five year period. That’s great for them, but what about companies that want to onboard AI technology without spending tons of cash for AI-abled candidates?

The good news is that companies may already have employees in-house that have the aptitude to learn the basic skills needed to monitor and improve AI programs. In fact, Google offers a free program that trains people in navigating AI systems.

And regarding the ROI? Cylance recently announced the results of their inaugural report, Artificial Intelligence in the Enterprise: The AI Race Is On, and found that according to their survey, which included insight from over 650 IT professionals, “optimism about the value of artificial intelligence (AI)-powered solutions in the enterprise is high and plans to continue investing in the technology are widespread.” In short, the outlook for positive ROI is good.

AI Can Mimic Human Emotion (Like Empathy)

The one advantage that humans have always had over automated solutions is their ability to provide Jungian-level emotional interactions. With the advance of machine learning, worry has mounted that AI-enabled bots will eventually be able to replicate human emotions.

No matter how smart automated machines may become, is it possible that they can equal or excel with genuine human conversations? While automated bots may be able to mimic human emotions, they cannot truly replicate the experience without feeling the emotions themselves.

The definition of empathy is “the ability to understand and share the feelings of another.” Empathy requires true human experience. While AI bots may be able to create the illusion of empathy, they will never be able to completely empathize with humans. Robots do not have feelings, and it is those feelings that can convert a banal conversation into a customer experience success story.

The Human Race Will Eventually be Enslaved to Robots

Before wrapping up, let’s talk about the mechanical elephant in the room. You don’t have to scan too far down your news feed to find a think piece about how humans are doomed to be taken over by the “Robopocalypse.”

It may sound like a far-fetched conspiracy theory but just in case, here are some helpful tips on how to survive a robot apocalypse:

1. Wall Street Journal advises to “just close the door.” Apparently, in a government sponsored study, robots were stumped by a closed, unlocked door.
2. Use giant magnets. Jeff Behl, Chief Network Architect at LogicMonitor and a veteran in the world of technical operations advises that we should consider using giant magnets to disrupt their hard drives.
3. Always have a glass of water handy. As Assaf Baciu recommends in his guide on how to survive Elon Musk’s inevitable robo-domination: “electricity isn’t a fan of liquids”.

In Conclusion

While it’s been fun pontificating on how to stop an AI apocalypse, the truth remains that many of the myths surrounding Artificial Intelligence could prove to be detrimental for using the technology for what is actually is: an empowering innovation.

AI is not a plug and play be-all-end-all solution to personnel and quality assurance complications. Instead, it is an ever-developing aid. AI can be used to enhance and assist the hard-working and intelligent individuals that are necessary to deliver measure-driven results.

At the same time, when properly implemented it can save companies time, money and frustration.

It’s not often that two competing companies that are each pouring billions of dollars into similar platforms agree to work together, but that’s exactly what we saw with the just-announced partnership between Amazon Alexa and Microsoft Cortana.

Trying to predict the future is a crucial part of being a startup founder, especially when you’re working in a category as new as voice. We’re building design software for voice applications at Sayspring, so we’re constantly thinking about how the major voice platforms will unfold.

The recent news that Amazon Alexa and Microsoft Cortana will now “talk to each other” is a major development. Here are some of the reasons I think it makes sense, and what it means for the future of voice.

What Microsoft Gains

Amazon is a formidable competitor, but here’s why this agreement makes sense for Microsoft.

Voice-Powered Commerce — A product of Amazon, one of the core missions behind the Alexa platform is making it easier to purchase. Cortana doesn’t have much in the way of purchasing functionality, and by partnering with Alexa, they are bringing the ease and convenience of voice commerce to millions of Cortana users.

Improved Presence in the Home — Microsoft’s voice assistant Cortana currently has over 141M monthly active users, and roughly 25% of all Cortana queries are delivered through voice. That’s a massive user base, mostly a result of Cortana being shipped with every copy of Windows 10.

But while Microsoft has owned the desktop for decades, the company has tried hard to establish a major presence in the home with the Xbox. It seems to be a battle they are losing to Sony. It’s estimated that Microsoft has sold roughly 33M Xbox Ones since the 2013 launch, compared to Sony Playstation 4’s 66M sold. Cortana runs in almost every Xbox, but it lacks a microphone of its own and can only be spoken to through a headset or Kinect, limiting voice access to a sliver of overall Xbox owners.

With 10’s of millions devices sold, Amazon Alexa promises to become the “Operating System for the Home.” Partnering with Alexa makes not only Xbox, but the entire Microsoft ecosystem, including Outlook and Office 365, accessible through what is now the primary interface of the home: voice.

Access to More Third-Party Apps— Three of the major voice platforms (Alexa, Assistant, and Cortana) all offer support for third-party applications. Cortana just launched their offering, Skills, this past May. Alexa launched Skills over two years and already has over 17,000 third-party apps. Now Cortana can offer all of that functionality through this partnership.

Even before this partnership was announced, it’s telling that Microsoft decided to call their voice apps “Skills” the same way Alexa does, borrowing the same metaphor of these apps teaching the voice assistant something new. Google Assistant calls them Actions, and the entire third-party experience is delivered in a different voice, making the voice assistant more of an operator connecting users to different services. With a common approach between both platforms, it’s likely we’ll see the experience of using third-party apps become more and more seamless.

Insight into Alexa Usage — Whenever someone talks to Alexa through Cortana, Microsoft will know what is being discussed, giving the company a look at how users are interacting with the market-leading voice platform. That’s valuable insight into user behavior through voice interfaces. Of course, the flip-side of this is that Alexa will also see how users are talking to Cortana, but considering how far ahead Amazon here, this seems like a smart trade for Microsoft.

“You need a culture that is fundamentally not opposed to new concepts and new capabilities” -Satya Nadella

When Satya Nadella took over as CEO, he outlined his vision for a new Microsoft. The former CEO, Steve Ballmer, had viewed Microsoft as a devices and services company. Nadella’s vision is to make Microsoft a “company capable of working across any platform — even those controlled by competitors — to help people be more productive.” This partnership shows how the company is determined to leverage Alexa’s headstart in voice for their own benefit.

What Amazon Gains

Though they have 70% of the smart speaker market, here is what Amazon stands to gain from partnering with Microsoft on voice.

More Customers — As Microsoft brings purchasing to Cortana users through this partnership, Amazon gains access to those users as customers. Though it’s likely that many Cortana users already have an Amazon account, now it’s much easier for them to make purchases through Alexa when they’re in a Cortana-powered environment, like the desktop.

Windows Users, many in the Enterprise — Beyond the 141M active monthly Cortana users, Windows still powers much of the desktop world, with 400M+ Windows 10 machines out in the wild. A much more efficient method of input, voice is becoming increasingly popular on the desktop.

As just one example, Google announced that Google Analytics will soon support voice commands, letting anyone ask things like “how many visitors did we have last week” without needing to manually set date ranges and click through dashboards. That’s the kind of voice integration we’ll continue to see more software support, and Amazon wants to be a part of that shift.

Most of us sit in front of a computer all day. From controlling your smart home to starting your car, we will now have access to all of our Alexa-powered devices through Cortana without leaving our desk.

Xbox — Xbox wasn’t mentioned in this announcement, and Cortana Skills aren’t yet available on Xbox, but I can’t help but think this will become important. Gaming is a huge opportunity and Amazon knows this. They bought Twitch for $1 Billion in 2014, and recently released Pulse, a social network specifically for gamers. Every Xbox One owner has Cortana, but without a microphone, few can talk to her. Amazon Alexa-powered devices like the Echo and Dot, usually placed within speaking distance of a gaming console, can now be that microphone.

This would help Amazon connect with an important set of users, but also give them a look at how gamers are using Xbox, what apps they’re interacting with, and what type of content they are viewing. For a company that is targeting gamers, has their own gaming platform, and is in the streaming media business, this would all be good information to have.

A Rising Tide Lifts All Boats

Amazon has thousands of people working on Alexa, and they see it as being the fourth-pillar of their business (the other 3 being e-commerce, marketplace, and cloud services). Today’s voice interfaces are powered by the cloud, and the kinds of services that Amazon offers. The more people using voice, and find utility through voice assistants, the more potential customers for AWS.

The other two major players in the voice assistant market, Google and Apple, both have mobile hardware to embed their voice platforms into. Microsoft and Amazon both haven’t had successful smartphones adopted by the market. They both had issues with their mobile strategy, and with the prospect of seeing another major platform shift, they have a shared interest in making sure Google’s Assistant and Apple’s Siri don’t become the dominant voice platforms.

“It is a renaissance, it is a golden age.” -Jeff Bezos on AI

Early Days, Big Promise

There will undoubtedly be some awkward user experience issues early on that will need to be worked out (“Alexa, open Cortana” is just a starting point), but this type of partnership is a smart move to drive voice adoption forward while potentially blocking out two major competitors.

---

Mark Webster is the Founder/CEO of Sayspring, a design and prototyping platform for voice applications.

Chicory, the premier technology company that uses artificial intelligence to create digital grocery shopping experiences, today announced the launch of their Click-to-Cart technology, making their full suite of advertising solutions, shoppable. This new technology differs from similar technologies in that Chicorya��s offering sends promoted products directly to the consumera��s cart at a selected retailer, reducing touch points along the path-to-purchase.

In an effort to create more integrated digital shopping experiences, Chicorya��s Click-to-Cart technology can prompt shoppers to add one item or a full basket of pre-selected SKUs to an online shopping cart, with a single click of one of the companya��s shoppable ads. Click-to-Cart can be layered onto any of Chicorya��s advertising solutions, including social media campaigns and video, providing a truly omnichannel solution for CPG advertisers.

a�?We are seeing a change in the way consumers are shopping as they move from brick and mortar retail to online, which, in turn, changes the way brands can advertise,a�? said Yuni Sameshima, Chicory CEO and Co-Founder. a�?Grocery isna��t about buying individual items, it is about creating consumer solutions. Shoppers dona��t go to stores for one jar of peanut butter, they go to stores to stock up for their back-to-school lunches. That understanding guided our thought process in creating this digital solution for brands, to help our partners and clients pair their digital advertising with effective, actionable calls-to-action across the web.a�?

The click-to-cart ads can also be retailer specific or retailer agnostic. This unique option allows for truly tailored advertising campaigns based on selected goals. For retailer-specific ads, brands and retailers can control whether or not they would like to the shoppable ads to be sent to the consumera��s shopping cart at a specific place or location. Conversely, retailer-agnostic ads, the consumer can select which retailer or online store they prefer or is most convenient for them.

Chicorya��s Click-to-Cart is backed by the same AI and machine learning processes that powers its distributed grocery ecommerce experiences on over 1,000 recipe websites.