Reflections on “Inside Amazon’s artificial intelligence flywheel” [Levy]

Inside Amazon’s artificial intelligence flywheel — from wired.com by Steven Levy
How deep learning came to power Alexa, Amazon Web Services, and nearly every other division of the company.

Excerpt (emphasis DSC):

Amazon loves to use the word flywheel to describe how various parts of its massive business work as a single perpetual motion machine. It now has a powerful AI flywheel, where machine-learning innovations in one part of the company fuel the efforts of other teams, who in turn can build products or offer services to affect other groups, or even the company at large. Offering its machine-learning platforms to outsiders as a paid service makes the effort itself profitable—and in certain cases scoops up yet more data to level up the technology even more.

It took a lot of six-pagers to transform Amazon from a deep-learning wannabe into a formidable power. The results of this transformation can be seen throughout the company—including in a recommendations system that now runs on a totally new machine-learning infrastructure. Amazon is smarter in suggesting what you should read next, what items you should add to your shopping list, and what movie you might want to watch tonight. And this year Thirumalai started a new job, heading Amazon search, where he intends to use deep learning in every aspect of the service.

“If you asked me seven or eight years ago how big a force Amazon was in AI, I would have said, ‘They aren’t,’” says Pedro Domingos, a top computer science professor at the University of Washington. “But they have really come on aggressively. Now they are becoming a force.”

Maybe the force.

 

 

From DSC:
When will we begin to see more mainstream recommendation engines for learning-based materials? With the demand for people to reinvent themselves, such a next generation learning platform can’t come soon enough!

  • Turning over control to learners to create/enhance their own web-based learner profiles; and allowing people to say who can access their learning profiles.
  • AI-based recommendation engines to help people identify curated, effective digital playlists for what they want to learn about.
  • Voice-driven interfaces.
  • Matching employees to employers.
  • Matching one’s learning preferences (not styles) with the content being presented as one piece of a personalized learning experience.
  • From cradle to grave. Lifelong learning.
  • Multimedia-based, interactive content.
  • Asynchronously and synchronously connecting with others learning about the same content.
  • Online-based tutoring/assistance; remote assistance.
  • Reinvent. Staying relevant. Surviving.
  • Competency-based learning.

 

The Living [Class] Room -- by Daniel Christian -- July 2012 -- a second device used in conjunction with a Smart/Connected TV

 

 

 

 

 

 

 

We’re about to embark on a period in American history where career reinvention will be critical, perhaps more so than it’s ever been before. In the next decade, as many as 50 million American workers—a third of the total—will need to change careers, according to McKinsey Global Institute. Automation, in the form of AI (artificial intelligence) and RPA (robotic process automation), is the primary driver. McKinsey observes: “There are few precedents in which societies have successfully retrained such large numbers of people.”

Bill Triant and Ryan Craig

 

 

 

Also relevant/see:

Online education’s expansion continues in higher ed with a focus on tech skills — from educationdive.com by James Paterson

Dive Brief:

  • Online learning continues to expand in higher ed with the addition of several online master’s degrees and a new for-profit college that offers a hybrid of vocational training and liberal arts curriculum online.
  • Inside Higher Ed reported the nonprofit learning provider edX is offering nine master’s degrees through five U.S. universities — the Georgia Institute of Technology, the University of Texas at Austin, Indiana University, Arizona State University and the University of California, San Diego. The programs include cybersecurity, data science, analytics, computer science and marketing, and they cost from around $10,000 to $22,000. Most offer stackable certificates, helping students who change their educational trajectory.
  • Former Harvard University Dean of Social Science Stephen Kosslyn, meanwhile, will open Foundry College in January. The for-profit, two-year program targets adult learners who want to upskill, and it includes training in soft skills such as critical thinking and problem solving. Students will pay about $1,000 per course, though the college is waiving tuition for its first cohort.

 

 

 

Three shifts as big as print to digital — from gettingsmart.com by Tom Vander Ark

Excerpts (emphasis DSC):

We just lived through the biggest shift in learning since the printing press—a 25-year shift from print to digital. While it extended access and options to billions, it didn’t prove as transformational as many of us expected. It did, however, set the stage for three shifts that will change what and how people learn.

  1. Basic to broader aims.
  2. Passive to active learning.
  3. Time to demonstrated learning.

 

 

 

To higher ed: When the race track is going 180mph, you can’t walk or jog onto the track. [Christian]

From DSC:
When the race track is going 180mph, you can’t walk or jog onto the track.  What do I mean by that? 

Consider this quote from an article that Jeanne Meister wrote out at Forbes entitled, “The Future of Work: Three New HR Roles in the Age of Artificial Intelligence:”*

This emphasis on learning new skills in the age of AI is reinforced by the most recent report on the future of work from McKinsey which suggests that as many as 375 million workers around the world may need to switch occupational categories and learn new skills because approximately 60% of jobs will have least one-third of their work activities able to be automated.

Go scan the job openings and you will likely see many that have to do with technology, and increasingly, with emerging technologies such as artificial intelligence, deep learning, machine learning, virtual reality, augmented reality, mixed reality, big data, cloud-based services, robotics, automation, bots, algorithm development, blockchain, and more. 

 

From Robert Half’s 2019 Technology Salary Guide 

 

 

How many of us have those kinds of skills? Did we get that training in the community colleges, colleges, and universities that we went to? Highly unlikely — even if you graduated from one of those institutions only 5-10 years ago. And many of those institutions are often moving at the pace of a nice leisurely walk, with some moving at a jog, even fewer are sprinting. But all of them are now being asked to enter a race track that’s moving at 180mph. Higher ed — and society at large — are not used to moving at this pace. 

This is why I think that higher education and its regional accrediting organizations are going to either need to up their game hugely — and go through a paradigm shift in the required thinking/programming/curricula/level of responsiveness — or watch while alternatives to institutions of traditional higher education increasingly attract their learners away from them.

This is also, why I think we’ll see an online-based, next generation learning platform take place. It will be much more nimble — able to offer up-to-the minute, in-demand skills and competencies. 

 

 

The below graphic is from:
Jobs lost, jobs gained: What the future of work will mean for jobs, skills, and wages

 

 

 


 

* Three New HR Roles To Create Compelling Employee Experiences
These new HR roles include:

  1. IBM: Vice President, Data, AI & Offering Strategy, HR
  2. Kraft Heinz Senior Vice President Global HR, Performance and IT
  3. SunTrust Senior Vice President Employee Wellbeing & Benefits

What do these three roles have in common? All have been created in the last three years and acknowledge the growing importance of a company’s commitment to create a compelling employee experience by using data, research, and predictive analytics to better serve the needs of employees. In each case, the employee assuming the new role also brought a new set of skills and capabilities into HR. And importantly, the new roles created in HR address a common vision: create a compelling employee experience that mirrors a company’s customer experience.

 


 

An excerpt from McKinsey Global Institute | Notes from the Frontier | Modeling the Impact of AI on the World Economy 

Workers.
A widening gap may also unfold at the level of individual workers. Demand for jobs could shift away from repetitive tasks toward those that are socially and cognitively driven and others that involve activities that are hard to automate and require more digital skills.12 Job profiles characterized by repetitive tasks and activities that require low digital skills may experience the largest decline as a share of total employment, from some 40 percent to near 30 percent by 2030. The largest gain in share may be in nonrepetitive activities and those that require high digital skills, rising from some 40 percent to more than 50 percent. These shifts in employment would have an impact on wages. We simulate that around 13 percent of the total wage bill could shift to categories requiring nonrepetitive and high digital skills, where incomes could rise, while workers in the repetitive and low digital skills categories may potentially experience stagnation or even a cut in their wages. The share of the total wage bill of the latter group could decline from 33 to 20 percent.13 Direct consequences of this widening gap in employment and wages would be an intensifying war for people, particularly those skilled in developing and utilizing AI tools, and structural excess supply for a still relatively high portion of people lacking the digital and cognitive skills necessary to work with machines.

 


 

 

Per Debra Chen, Founder Rockin’ Hood Project (emphasis DSC)

It is with great pride that we announce the official launch our entrepreneurship initiative with the unveiling of our website:
www.rockinhoodproject.com

As studies have demonstrated, only one thing consistently brings children raised in poverty into the middle class: entrepreneurship education. And so it is our mission to expose students from inner city schools to successful entrepreneurs, influencers, and accomplished individuals who can inspire and educate on the principles of thinking outside the box and on believing in their own achievements.

Essentially we are building the Ted Talks for Kids.

If you’d like to get involved, know of corporate sponsors, or would like to do an interview with our organization Rockin’ Hood Project, please email us at debra@rockinhoodproject.com.

 

Also, an excerpt from:
Why Every School in America Should Teach Entrepreneurship — from time.com by Steve Mariotti (2012)

Why Aren’t We Teaching Entrepreneurship in Our Schools?
The Williams brothers’ story is one of countless examples from NFTE’s files that beg the question: If entrepreneurship education can create jobs, encourage students to stay in school, and provide economic rescue for people in our low-income communities, why aren’t we teaching it in every high school in America?

Let’s begin state and national discussions about owner-entrepreneurship education, focused on four goals:

  • Engage young people in school by teaching math, reading, writing and communication within the motivating context of starting and operating a small business.
  • Teach young people about the market economy and how ownership leads to wealth creation.
  • Encourage an entrepreneurial mindset so our youth will succeed whether they pursue higher education, enter the workforce, or become entrepreneurs.
  • Make young people financially literate so they can save and invest to achieve goals like home ownership and retirement.

 

 

 

The number of Americans working for themselves could triple by 2020 — from work.qz.com by Amy Wang

Excerpt (emphasis DSC):

Americans are as eager to work as ever. Just no longer for somebody else.

According to FreshBooks, a cloud-based accounting company that has conducted a study on self-employment for two years, the number of Americans working for themselves looks to triple—to 42 million people—by 2020.

The trend, gauged in a survey of more than 2,700 full-time US workers in traditional, independent, and small business roles about their career plans, is largely being driven by millennial workers. FreshBooks estimates that of the next 27 million independent workers, 42% will be millennials. The survey, conducted with Research Now, also finds that Americans who already work for themselves are suddenly very content to keep doing so, with 97% of independent workers (up 10% from 2016) reporting no desire to return to traditional work.

 

 

From DSC:
With the continued trend towards more freelancing and the growth of a more contingent workforce…have our students had enough practice in selling themselves and their businesses to be successful in this new, developing landscape?

We need to start offering more courses, advice, and opportunities for practicing these types of skills — and the sooner the better!  I’m serious. Our students will be far more successful with these types of skills under their belt. Conversely, they won’t be able to persuade others and sell themselves and their businesses without such skills.

 

 

Robo retail: The automated store of the future is heading closer to our doorsteps. — from jwtintelligence.com’

Excerpt:

The automated store of the future is heading closer to our doorsteps.

Self-checkout and online delivery services might soon be outmoded. Automated, cashier-less and mobile, doorstep-accessible shopping outlets are popping up globally—these offer not only a quick and seamless shopping experience, but also allow customers to handpick the items they are seeking.

Retail giant Amazon launched its Amazon Go store in Seattle in late January 2018. Amazon Go stocks everyday items, Whole Foods Market goods and Amazon-branded meal kits, but has no cashiers, no check-out lines and no barcode scanners. Shoppers enter by scanning an app, shop and leave—items purchased are automatically charged to their accounts. Dilip Kumar, vice president of technology for Amazon Go, suggests the concept is Amazon’s answer to solving “time poverty,” which he calls people’s “number one problem.”

This is just the tip of the iceberg. While Amazon Go currently only operates in Seattle, two other mobile concepts are hoping to reach a wider audience by physically bringing roving stores directly to the consumer. Robomart, based in the Bay Area, is a prototype tap-to-request grab-and-go food mart. Conventional grocery delivery services like those run by Amazon, FreshDirect or Instacart don’t let customers select products for themselves. If you’re particular about the ripeness of an avocado or conscious about bruises on tomatoes, being able to choose your own produce is essential. An autonomy-focused platform like Robomart puts consumers in the metaphorical driver’s seat, while still maintaining a high level of ease.

 

 

 

From DSC:
Speaking of cashiers, I had some comments regarding the future of cashiers towards the bottom of this posting here.  Another relevant posting is: “Tech companies should stop pretending AI won’t destroy jobs” + 6 other items re: AI, bots, algorithms, & more

 

 

 

From DSC:
With some predictions saying that the workforce is going to be composed of upwards of ~50% of us being contingent workers, (I’ve already seen figures around mid 30’s and even 40%), the question I have is:

Are we teaching students how to protect themselves, how to sell themselves, how to sell their businesses, how to plan financially, etc.? 

Consider this article:

Would our students know about these types of mistakes?

Also, it seems to me that higher education should be helping students “future proof” themselves — or at least as much as possible. One of the values higher education should be bringing to the table is to identify which jobs are going to be around for the next 5-10 years and which ones aren’t.

Along these lines, lifelong learning and learning how to learn are becoming increasingly important. Thus, I will continue to try and post articles/resources on this blog in regards to metacognition and the like.

 

 

 

Top Ten Makerspace Favorites of 2017 — from worlds-of-learning.com by Laura Fleming

Excerpt:

This year’s New Media Consortium/COSN’s 2017 K-12 Horizon Report deemed makerspaces the future of edtech and predicted full adoption in K-12 in one year or less! It is more important than ever that we plan and create GREAT makerspaces! Part of doing so is selecting the right materials, resources and supplies to support the experiences you want your students to have in your makerspace. In what has become an annual tradition, I am thrilled to highlight some of my favorite makerspace-related things for this year.  As always, I am grateful to my friend and colleague, Travis Lape, for his input on this list! It is my hope that this year’s favorites encourage you to think ‘out of the box’ as you choose products to support your makerspace themes.

I am proud to present to you, the Worlds of Making Top Ten Makerspace Favorites of 2017!

 

Buildr TAPE
You can now turn almost anything into a surface that allows you to build with legos! Buildr TAPE™ is flexible, cuttable, and wearable toy building block reusable adhesive tape.  This peel and stick TAPE™ instantly transforms virtually any surface into a bendable base for building new amazing creations with your traditional toy building blocks. Compatible with Lego®.

 

 

 

 

 

 

The next era of human|machine partnerships
From delltechnologies.com by the Institute for the Future and Dell Technologies

 


From DSC:
Though this outlook report paints a rosier picture than I think we will actually encounter, there are several interesting perspectives in this report. We need to be peering out into the future to see which trends and scenarios are most likely to occur…then plan accordingly. With that in mind, I’ve captured a few of the thoughts below.


 

At its inception, very few people anticipated the pace at which the internet would spread across the world, or the impact it would have in remaking business and culture. And yet, as journalist Oliver Burkeman wrote in 2009, “Without most of us quite noticing when it happened, the web went from being a strange new curiosity to a background condition of everyday life.”1

 

In Dell’s Digital Transformation Index study, with 4,000 senior decision makers across the world, 45% say they are concerned about becoming obsolete in just 3-5 years, nearly half don’t know what their industry will look like in just three years’ time, and 73% believe they need to be more ‘digital’ to succeed in the future.

With this in mind, we set out with 20 experts to explore how various social and technological drivers will influence the next decade and, specifically, how emerging technologies will recast our society and the way we conduct business by the year 2030. As a result, this outlook report concludes that, over the next decade, emerging technologies will underpin the formation of new human-machine partnerships that make the most of their respective complementary strengths. These partnerships will enhance daily activities around the coordination of resources and in-the-moment learning, which will reset expectations for work and require corporate structures to adapt to the expanding capabilities of human-machine teams.


For the purpose of this study, IFTF explored the impact that Robotics, Artificial Intelligence (AI) and Machine Learning, Virtual Reality (VR) and Augmented Reality (AR), and Cloud Computing, will have on society by 2030. These technologies, enabled by significant advances in software, will underpin the formation of new human-machine partnerships.

On-demand access to AR learning resources will reset expectations and practices around workplace training and retraining, and real-time decision-making will be bolstered by easy access to information flows. VR-enabled simulation will immerse people in alternative scenarios, increasing empathy for others and preparation for future situations. It will empower the internet of experience by blending physical and virtual worlds.

 

Already, the number of digital platforms that are being used to orchestrate either physical or human resources has surpassed 1,800.9 They are not only connecting people in need of a ride with drivers, or vacationers with a place to stay, but job searchers with work, and vulnerable populations with critical services. The popularity of the services they offer is introducing society to the capabilities of coordinating technologies and resetting expectations about the ownership of fixed assets.

 

Human-machine partnerships won’t spell the end of human jobs, but work will be vastly different.

The U.S. Bureau of Labor Statistics says that today’s learners will have 8 to 10 jobs by the time they are 38. Many of them will join the workforce of freelancers. Already 50 million strong, freelancers are projected to make up 50% of the workforce in the United States by 2020.12 Most freelancers will not be able to rely on traditional HR departments, onboarding processes, and many of the other affordances of institutional work.

 

By 2030, in-the-moment learning will become the modus operandi, and the ability to gain new knowledge will be valued higher than the knowledge people already have.

 

 

The Living [Class] Room -- by Daniel Christian -- July 2012 -- a second device used in conjunction with a Smart/Connected TV

 

 

 


From DSC:
From an early age, we need to help our students learn how to learn. What tips, advice, and/or questions can we help our students get into the habit of asking themselves? Along these lines, the article below,”How Metacognition Boosts Learning,” provides some excellent questions. 

Speaking of questions…I’ll add some more, but of a different sort:

  • How can all educators do a better job of helping their students learn how to learn?
  • How can Instructional Designers and Instructional Technologists help out here? Librarians? Provosts? Deans? Department Chairs? Teachers? Trainers (in the corporate L&D space)?
  • How might technologies come into play here in terms of building more effective web-based learner profiles that can be fed into various platforms and/or into teachers’ game plans?

I appreciate Bill Knapp and his perspectives very much (see here and here; Bill is GRCC’s Executive Director of Distance Learning & Instructional Technologies). The last we got together, we wondered out loud:

  • Why don’t teachers, professors, school systems, administrations within in K-20 address this need/topic more directly…? (i.e., how can we best help our students learn how to learn?)
  • Should we provide a list of potentially helpful techniques, questions, tools, courses, modules, streams of content, or other resources on how to learn?
  • Should we be weaving these sorts of things into our pedagogies?
  • Are there tools — such as smartphone related apps — that can be of great service here? For example, are there apps for sending out reminders and/or motivational messages?

As Bill asserted, we need to help our students build self-efficacy and a mindset of how to learn. Then learners can pivot into new areas with much more confidence. I agree. In an era that continues to emphasize freelancing and entrepreneurship — plus dealing with a rapidly-changing workforce — people now need to be able to learn quickly and effectively. They need to have the self confidence to be able to pivot. So how can we best prepare our students for their futures?

Also, on a relevant but slightly different note (and I suppose is of the flavor of a Universal Design for Learning approach)…I think that “tests” given to special needs children — for example that might have to do with executive functioning, and/or identifying issues, and/or providing feedback as to how a particular learner might best absorb information — would be helpful for ALL students to take. If I realize that the way my brain learns best is to have aural and visual materials presented on any given topic, that is very useful information for me to realize — and the sooner the better!

 



How Metacognition Boosts Learning — from edutopia.org by Youki Terada
Students often lack the metacognitive skills they need to succeed, but they can develop these skills by addressing some simple questions.

Excerpt (emphasis DSC):

Strategies that target students’ metacognition—the ability to think about thinking—can close a gap that some students experience between how prepared they feel for a test and how prepared they actually are. In a new study, students in an introductory college statistics class who took a short online survey before each exam asking them to think about how they would prepare for it earned higher grades in the course than their peers—a third of a letter grade higher, on average. This low-cost intervention helped students gain insight into their study strategies, boosting their metacognitive skills and giving them tools to be more independent learners.

More recently, a team of psychologists and neuroscientists published a comprehensive analysis of 10 learning techniques commonly used by students. They discovered that one of the most popular techniques—rereading material and highlighting key points—is also one of the least effective because it leads students to develop a false sense of mastery. They review a passage and move on without realizing that they haven’t thoroughly understood and absorbed the material.

Metacognition helps students recognize the gap between being familiar with a topic and understanding it deeply. But weaker students often don’t have this metacognitive recognition—which leads to disappointment and can discourage them from trying harder the next time.

To promote students’ metacognition, middle and high school teachers can implement the following strategies. Elementary teachers can model or modify these strategies with their students to provide more scaffolding.

During class, students should ask themselves:

  • What are the main ideas of today’s lesson?
  • Was anything confusing or difficult?
  • If something isn’t making sense, what question should I ask the teacher?
  • Am I taking proper notes?
  • What can I do if I get stuck on a problem?

Before a test, students should ask themselves:

  • What will be on the test?
  • What areas do I struggle with or feel confused about?
  • How much time should I set aside to prepare for an upcoming test?
  • Do I have the necessary materials (books, school supplies, a computer and online access, etc.) and a quiet place to study, with no distractions?
  • What strategies will I use to study? Is it enough to simply read and review the material, or will I take practice tests, study with a friend, or write note cards?
  • What grade would I get if I were to take the test right now?

After a test, students should ask themselves:

  • What questions did I get wrong, and why did I get them wrong?
  • Were there any surprises during the test?
  • Was I well-prepared for the test?
  • What could I have done differently?
  • Am I receiving useful, specific feedback from my teacher to help me progress?

 



From DSC:
Below are a few resources more about metacognition and learning how to learn:

 

 

 

  • Students should be taught how to study. — from Daniel Willingham
    Excerpt:
    Rereading is a terribly ineffective strategy. The best strategy–by far–is to self-test–which is the 9th most popular strategy out of 11 in this study. Self-testing leads to better memory even compared to concept mapping (Karpicke & Blunt, 2011).

 

 

 

  • The Lesson You Never Got Taught in School: How to Learn! — from bigthink.com
    Excerpt:
    Have you ever wondered whether it is best to do your studying in large chunks or divide your studying over a period of time? Research has found that the optimal level of distribution of sessions for learning is 10-20% of the length of time that something needs to be remembered. So if you want to remember something for a year you should study at least every month, if you want to remember something for five years you should space your learning every six to twelve months. If you want to remember something for a week you should space your learning 12-24 hours apart. It does seem however that the distributed-practice effect may work best when processing information deeply – so for best results you might want to try a distributed practice and self-testing combo.There is however a major catch – do you ever find that the amount of studying you do massively increases before an exam? Most students fall in to the “procrastination scallop” – we are all guilty at one point of cramming all the knowledge in right before an exam, but the evidence is pretty conclusive that this is the worst way to study, certainly when it comes to remembering for the long term. What is unclear is whether cramming is so popular because students don’t understand the benefits of distributed practice or whether testing practices are to blame – probably a combination of both. One thing is for sure, if you take it upon yourself to space your learning over time you are pretty much guaranteed to see improvements.

 

 



Addendum on 1/22/18:

Using Metacognition to Promote Learning
IDEA Paper #63 | December 2016
By Barbara J. Millis

Excerpt:

Some Definitions of Metacognition
Metacognition, simplistically defined, can be described as “cognition about cognition” or “thinking about thinking” (Flavell, Miller & Miller, 2002, p. 175; Shamir, Metvarech, & Gida, 2009, p. 47; Veeman, Van Hout-Wolters, & Afflerbach, 2006, p. 5). However, because metacognition is multifaceted and multi-layered (Dunlosky & Metcalf, 2009, p. 1; Flavell, 1976; Hall, Danielewicz, & Ware , 2013, p. 149; Lovett, 2013, p. 20), more complex definitions are called for. Basically, metacognition must be viewed as an ongoing process that involves reflection and action. Metacognitive thinkers change both their understandings and their strategies. The clearest definitions of metacognition emphasize its nature as a process or cycle.

Several authors (Nilson, 2013, p. 9; Schraw, 2001; & Zimmerman, 1998; 2000; 2002) narrow this process down to three ongoing stages. The first stage, pre-planning, emphasizes the need for reflection on both one’s own thinking and the task at hand, including reflection on past strategies that might have succeeded or failed. Following this self-reflection, during planning, metacognitive thinkers develop and implement—put into action—a plan. In the third and final stage—post-planning adjustments/revisions—subsequent analysis following implementation leads to modifications, revised decisions, and new future plans. In an excellent summary, Wirth states that “metacognition requires students both to understand how they are learning and to develop the ability to make plans, to monitor progress and to make adjustments” (as cited in Jaschik, 2011, p. 2).

 

Conclusion: As we have seen, metacognition is a complex but valuable skill that can nurture students’ learning and their self-awareness of the learning process. It is best conceived as a three-step process that can occur through deliberately designed activities. Such activities can take place before, during, and after face-to-face lessons or through online learning. They can also be built around both multiple choice and essay examinations. Immersing students in these metacognitive activities—assuming there are opportunities for practice and feedback—can result in students who are reflective learners.

 

 

 

 

Looking for something?

Use the form below to search the site:

Still not finding what you're looking for? Drop a comment on a post or contact us so we can take care of it!

© 2018 | Daniel Christian