The Low-Down: 2018-04-08

Apr 14, 2018

Silicon Valley's Astronomical Housing Prices Are Creating Problems For Startups

Tech talent can not afford to move there. And many already there no longer believe they can afford to stay. JL

Sisi Caro reports in the Observer:

House prices in San Francisco are 76% higher than five years ago. That cost translates into pressure on startup founders. For startups that rely on venture capital, rising payroll means a lower chance to survive. Software engineers in San Fran earn 20% more than in other parts of the country. “Startups need time to find a profitable business model before they burn VC money. It takes three years to infinity. When a startup compensates employees with an enormous salary, that shortens its time to find a profitable model.”

Form follows function...or vice versa? JL

Mark Wilson reports in Fast Company:

It feels like restaurants used to be made around the vision of a chef. But now more menu items, and even places, (are) designed around Instagram. Charcoal ice cream. Unicorn Frappuccinos. Rainbow bagels. Neon signs. Maybe at this point, Instagram is just such part of our culture we design with it in mind without thinking of it. For people doing new design, they consider it without making it an explicit design strategy.

Collateral damage. JL

Kevin Roose reports in the New York Times:

Recent decisions to emphasize stories shared by friends and family as well as trusted news outlets — part of the response to sensationalized clickbait and false news, and an attempt to foster “meaningful social interaction” — has hidden them from view. They argue that (social media) owes much of its growth to the kinds of entertainment they offer, and that users will spend less time on social networks if it’s not shown to them.

Shoppers' preferences appear to be more selective than has been previously understood. JL

Neil Dutta reports in Bloomberg:

The gap between the subindustries of department stores and warehouse clubs suggests that for certain goods, consumers may not value the shopping experience. Heading to a warehouse to buy staple household goods in bulk is fine, but Americans may want something more when going to a traditional department store. E-commerce has displaced many retail jobs, causing some very visible damage, such as

empty malls. But other retail categories appear to have picked up the slack.

So should consumers, who will now have to pay multiple sources for access to entertainment.

At least until there is only one entertainment conglomerate left. And Disney plans to be that one. JL

Derek Thompson reports in The Atlantic:

Disneyflix wouldn’t just be Netflix with Star Wars movies—it would be Amazon for Star Wars pillowcases and Groupon for rides on Star Wars roller coasters and Kayak for the Star Wars suite at Disney hotels. That’s could rival Netflix and create the profits Disney has enjoyed during its unprecedented century of dominance. This is a future where the movie industry as we know it ceases to exist. “We’re entering a world where theatrical movie openings will be optional. Disney (will) put its best content first on [Disneyflix].”

More to the point, Facebook has multiple ways of harvesting data from non-FB sources so the ostensible 'protections' are limited and, arguably, fraudulent.

Unless, that is, Facebook promises never to access or use such information. Which, given its history, is highly unlikely. JL

David Pierson and Tracy Lien report in the LA Times:

If Facebook were to establish a subscription service, there would be a two-tiered system where some pay for privacy and others pay with their data. If a user who pays not to be tracked interacts with a user on the free platform, the subscriber's data would likely be swept up too

Why Apple Is Facing It's Toughest Fight Since the 1980s

The battle to dominate AI not only features strong, wealthy competitors such as Google and Amazon, it plays against Apple's legacy because it requires combing large amounts of customer data. JL

Dave Gershgorn and Mike Murphy report in Quartz, photo by Benoit Tessier in Reuters:

Apple hasn’t competed in a market it hasn’t essentially defined since its early PC days. Tech companies are jockeying to usher in the next wave of personal computing. Artificial intelligence is going to be the key differentiator. AI relies on vast amounts of data to produce useful products, which means something Apple is loathe to do—gathering data from customers. This is a war Apple was supposed to win. (But Apple)’s now playing catch-up in a race where top talent is scarce and expensive.

Even in a city devoted to and dependent on tourism like New Orleans, the impact Airbnb has on business and employment, let alone housing affordability, is a concern. JL

Monica Burton reports in Eater:

“You can’t have neighborhood restaurants without neighborhoods.” Short-term rentals are pushing out residents, including restaurant and bar staff. In Marigny, 12.7% of people work in food service; in Bywater it’s 16%; if rents rise due to gentrification, workers need to relocate. Staffing has long been a problem in the restaurant industry. “Next thing you know you’re talking about less densely populated neighborhoods with less service,”

Every business is a tech business. And as carbon-based energy is increasingly challenged by alternatives, the competition for digital solutions and the people who make them more productive is becoming more intense. JL

Sarah Kent and Christopher Matthews report in the Wall Street Journal:

Big oil companies were early adapters of supercomputers, have poured hundreds of millions into upgrades, and now possess some of the most powerful computers on the planet. The efforts are part of a digital arms race among energy companies, embracing technology to produce fossil fuels more cheaply and efficiently. The companies are seeking to compete with Silicon Valley firms for top data and computer scientists.

AI is not going to fix Facebook's business model which, it is increasingly clear, is dependent on monetizing exactly the sorts of customer data analytics that have repeatedly gotten it in trouble.

And even if AI could address its core problems, no one is prepared to wait five to ten years for FB to figure it out. JL

Khari Johnson reports in Venture Beat:

Throughout five hours of testimony, Zuckerberg sprinkled magic AI pixie dust over his testimony. Hate speech? AI will take care of that. Fake news? AI. Terrorists using Facebook-owned platforms? AI. Content moderation? AI. One problem with AI: Congress (is) proposing legislation now, while Zuck talked about returns on AI in five to 10 years.

Does self-referential/circular investing suggest belief or bubble? JL

Theodore Schleifer reports in Re/code:

This is an example of a startup funding a venture capital firm, as opposed to the reverse, and it’s a venture capital firm that has previously funded it. That’s a new possibility for investors who have digital assets in their war chest, an increasingly common though still rare circumstance, and requires them and the funds they back to make decisions about how much faith they have in the particular cryptocurrency being invested.

The power of scale. The fact that on-demand or 'sharing' economy startups top the list could either be a cause for concern or a reflection of as-yet unrealized value, monetization opportunities - and continued competition for legacy companies. JL

Elissa Maercklein reports in Price Economics:

Of 157 unicorns, 139 are growing, while only 18 have negative growth. Of the top 20 fastest growing unicorns, five are transportation sharing platforms , and one (WeWork) is in shared workspace. Other industries include trading platforms, eCommerce, healthcare, and cloud data analytics. In the top 20 shrinking, eCommerce companies topped the list.

The Implication of Using Open Source Designs To Create More Specialized Chips

The returns to collaboration in tech continue to exceed those of going it alone. JL

Klint Finley reports in Wired:

Open source changed how companies build software. The embrace of open source isn't about altruism. Collaborating on freely available code enables companies and programmers to pool resources to solve common problems and avoid reinventing the wheel. The point of RISC-V isn't just to help companies save money on chip licenses. The real benefit is that it could enable companies to build chips that better meet their needs. And even if RISC-V never takes off, it shows chip makers are learning to share.

Substituting technology for trust is not a sustainable strategy. And in the post Facebook-Cambridge Analytica era, probably a non-starter. JL

Kai Stinchcombe comments in Medium:

People treat blockchain as a “futuristic integrity wand”—wave a blockchain at the problem, and your data will be valid. For anything people want to be valid, blockchain has been proposed as a solution. (But) auditing software is hard! The most-heavily scrutinized smart contract in history had a small bug that nobody noticed until someone did notice, and used it to steal fifty million dollars. If cryptocurrency enthusiasts putting together a $150m investment fund can’t audit the software, how confident are you? Projects based on the elimination of trust have failed to capture customers’ interest

Demographics is only destiny if you actually turn out and vote. And so far, Millennials have not done so. JL

Richard Fry reports in World Economic Forum, photo by Mario Anzuor in Reuters:

The Baby Boomer voting-eligible population peaked at 73 million in 2004. Since the Boomer electorate is declining in size and the Millennial electorate will continue to grow, mainly through immigration and naturalization, it is a matter of time before Millennials are the largest generation in the electorate. (But) Millennials have punched below their electoral weight in recent presidential elections. (young adults are less likely to vote than their older counterparts.) Turnout among Millennials in 2016 was 51%, lower than the 61% of the electorate who voted.

An individual's technological tracking number has become more real - and valuable - than their physical signature. JL

Stacy Cowley reports in the New York Times:

Credit card networks are finally ready to concede what has been obvious to shoppers and merchants for years: Signatures are not a useful way to prove someone’s identity. Later this month, four of the largest networks — American Express, Discover, Mastercard and Visa — will stop requiring them to complete card transactions. As forensic systems improved, signatures became a relic.

Because of the potential marketing and financial implications, medical data may the most valuable personal information. And it is also the most sensitive.

But there are trade-offs in determining how ownership, access and usage should be managed - and the solutions are neither monolithic - or satisfactory to the various competing interests. JL

Erika Fry reports in Fortune:

Personal medical data goes for five times the amount that financial or other data does on the dark web. Patients should own their data. But how that works given its value and the various demands for it, leads to more questions than answers.“From a physician standpoint, you cannot do your job unless you have access to as much of the data as is available….anything that limits that access literally puts patients at risk. We have to be very mindful of aggregated information and access to it."

Using technology to provide a relatively inexpensive service to people who fear or do not have access to traditional banks. Who are exactly the sort traditional banks disdain because they are not profitable enough under most government finance regulations.

But the tech companies realize they can hoover up lots of cash from such customers - and that all those little deposits add up globally to 'real money.' JL

Peter Rudegair reports in the Wall Street Journal:

The company’s goal is to give those excluded from the banking system access to the digital economy. Consumers with smaller balances are largely ignored by banks and have to rely on check-cashing centers and other alternative providers of financial services. Other technology firms with large user bases are looking at providing banking functions. Square Inc. gives out bank cards, and Amazon. has been in discussions to build a checking-account product for its customers.

Amazon Spent $23 Billion On R and D Last Year, More Than Any Other Company

And one might well argue that it is paying off. JL

Rani Molla reports in Re/code:

Amazon was at the top of the list, spending $22.6 billion in 2017, 41 percent more than in 2016 (when it also topped the list). R&D spending is important not only as it contributes to a company’s own innovation and dominance, but also for its contribution to national productivity, accounting for about 3 percent of the GDP. Amazon is followed in R&D spending by Alphabet, Intel, Microsoft and Apple.

As in so many other industries, the ability to analyze vast amounts of data, then have the model determine which are the most likely to be effective will, eventually, have dramatic effects on both efficiency and effectiveness. JL

Sy Mukherjee reports in Fortune:

Artificial intelligence (can) improve drug discovery at the earliest stages (when the risk of failure is also the highest). [A.I.] can help analyze large data sets from clinical trials, health records, genetic profiles, and preclinical studies; within this data, it can recognize patterns and trends and develop hypotheses at a much faster rate than researchers alone. "You don’t start out with a predetermined hypothesis—you feed the system all this data and allow the data to generate the hypotheses.”

European-style extreme data protection is unlikely in the US - but it does represent the worst case 'stick' that Congress is brandishing to force Facebook to embrace a somewhat more palatable 'carrot.' A number of the potential options from both Europe and the US are described below.

The challenge is that the regulatory options under US law are complicated and may neither satisfy critics - or solve the problem(s). That said, the gauntlet has been thrown down. Facebook had better pick it up fast or someone else will. JL

Natasha Singer reports in the New York Times, John McKinnon and Keach Hagey report in the Wall Street Journal:

Coming up with smart, effective rules to govern the huge businesses of Facebook and its rivals, from Google to Twitter , would be challenging, and sound legislation could take months, if not years, to execute. The goal would be to protect user privacy, increase transparency and give individuals greater control over their digital identities, while not stifling innovation in an industry that is the epitome of American ingenuity and entrepreneurship.

Higher costs, growing skepticism and greater scrutiny. JL

Evan Clark and Adriana Lee report in WWD:

Part of the success of tech and the companies that have adapted to tech is that they’re getting raw materials free from consumers. The fear is that’s going to end. Margins are going to shrink because there will be more rules on the use of data, less access to data and companies will have to pay more. Dense terms of service, complex privacy settings or automatic opt-ins for unwanted features can seem like hostile acts. Greater care and transparency work as powerful differentiators. Businesses capable of conveying that they’re built to serve people, not pull one over on them could stand out.

Substituting data for judgment can be a risky trade. JL

Paul Spriggs reports in Ad Week:

Over the last 10 years, the marketing industry built up an unhealthy reliance on data to lead strategy, creative ideas and media, cultivating a false sense of security. The desire for efficiency, reduced risk and immediate sales resulted in an over-reliance on technology and a focus on data-driven expertise. Marketing initiatives became shorter term, failing to build equity in enduring, emotionally resonating ideas.

It could help. But the question remains, who monitors those who do the conceptualizing and programming? JL

Eran Eyal reports in Venture Beat:

Few (businesses) recognize their customers across multiple channels and devices, relying on third-party, behavioral data that doesn’t give a complete understanding of what customers want. Most handle a small slice of each of customer’s purchases. In a blockchain ledger recording a shopper's purchases, shoppers would control access. AI engines could gain temporary access and then marry this data with a retailers'. This shift from fragmented data to comprehensive data maintained by consumers should increase the amount of data available to recommendation engines and predictive systems.

How Big Is the Gap Between Spielberg's 'Ready Player One' and Current Virtual Reality?

The gap is pretty big right now.

The question is not if some of the movie's images will become reality, but when. JL

Edd Gent reports in Singularity Hub:

Most VR tools are still out of reach of consumers. Where reality is lagging is in recreating the physical experience of VR. The big differences between current technology and the movie is how they reproduce images. There’s been slow progress on kinaesthetic feedback systems that exert force on the body to recreate sensations like weight, inertia, or resistance. They may be available at arcades and theme parks, but the more advanced technologies are aimed fields like medicine, engineering, and defense.

The use of high speed computers and data analysis has changed intellectual property protection.

There are two risks: one that an artist - currently in music, but by inference and technological development, for authors, film makers, software coders, et al that there is a legal penalty for having copied, however inadvertently.

The second is the reputational risk accrued from having been accused of copying and thus of being less creative than others or, perhaps even worse, of having lost one's gift.JL

Andy Hermann reports in Rolling Stone, illustration by Chelsea Johnston:

Forensic musicology compar(es) two [pieces of music] to figure out their objective similarity, and to attempt to infer the likelihood of copying- secondary similarities between two tracks – their "look and feel and cowbells," – rather than lyrics, melody and other elements protectable under copyright law. "If it sounds close, we can no longer rely on the fact that the notes are different, because we don't have a bright line.""The last thing you want is for somebody to accuse you, let alone publicly, of not being original."

Photos of some of the 25 people falsely accused on Twitter of being the YouTube shooter. JL

Charlie Warzel and Jane Lytvynenko report in Buzzfeed:

Almost immediately, came the garbage. 4chan hoaxers trying to trick people into thinking the shooter was a comedian named Sam Hyde; speculation the shooter was motivated by YouTube censoring political content; speculation it was religiously motivated; photos of supposed shooters in MAGA hats; unconfirmed images of potential victims and inaccurate death tolls. The incentive structures — frictionlessness, virality, scale, and anonymity — that govern Twitter help fake news travel faster than facts.

Multinational enterprises used cheap Chinese labor to wrest concessions from seemingly all powerful western governments and work forces.

They may now use automation and their influence with worried US and other western governments who perceive China as their primary threat to perform the same sort of arbitrage. JL

Reihan Salam reports in The Atlantic:

Consider that it is not the Chinese who are the predators; rather, it is the multinational enterprises that have taken advantage of China’s relative poverty, and who have leveraged its low-cost workforce and its cooperative government to extract tax and regulatory concessions from their home-country governments, and to keep domestic workers on their toes. Automation will offer U.S. multinationals an alternative to China-centric global production networks.

Humans - as traders, bankers, investors - and customers - have become increasingly superfluous in finance as computers and algorithms take over.

There are still leaders who must ultimately sign off on the decisions, but the question worth asking is whether even they are driving decisions - or are being driven by the reality of automated imperatives. JL

Hugh Son and Dakin Campbell report in Bloomberg:

The cost of computing power was collapsing. That, combined with abundant data on public exchanges, would enable automated trading. Across equities and fixed-income, apart from dwindling human traders and the human minders of the machines, algorithms will connect sellers and buyers. (But) for all the ways in which finance is a place where machines transact with other machines, the race for trading riches will be won or lost by people such as Blankfein, Gorman, and Dimon, men driven to keep the throne.

The end - as an iconic, standalone business - was signified by changes in culture, organization and, finally, by the dawning realization that it was no longer a platform that could drive the future of the enterprise. JL

Ben Thompson comments in Stratechery:

To understand that Windows is not and will not drive future growth is one thing; identifying future drivers of said growth is another. Microsoft had already lost the devices war and needed to focus on services that worked on iOS and Android. With this new reorganization, Windows is off in the corner where it belongs, leaving the Cloud and Enterprise team and Applications and Services Group free to focus on building their businesses on top of all platforms.

Foxconn and Swiss-Israeli Firm To Build World's First Blockchain-Based Smartphone

Designed to convert digital, encrypted currencies across multiple platforms and uses without having to 'obtain' the actual tokens. Awesome. Instant wealth for everyone.

What could possibly go wrong? JL

Rama Raman reports in Reuters:

Foxconn, will manufacture (a) blockchain smartphone designed to securely store and use digital currencies in fee-less transactions. The blockchain-based smartphone can automatically convert tokens for the use of decentralized applications, without having to obtain the different tokens through an exchange. The device (will be sold) through new stores located in Vietnam and Turkey (with) 25,000 units preordered.

There is an inherent dilemma in this research effort. Since prior research has demonstrated that anonymity is what fuels some of the most obnoxious online behavior, continuing to guarantee that sort of protection to those who have consistently abused the rules will not cause them to change. The conflict comes from cultural norms that, especially in Europe, demand anonymity.

And since many of the abusers appear to revel in their ability to break rules, we look forward to seeing the results - but would be betting on behavioral improvements. JL

Megan Dickey reports in Tech Crunch:

Twitter is going to start testing an idea that if it shows people its rules, behavior will improve. Researchers point to evidence of when institutions clearly publish rules, people are more likely to follow them. The privacy of Twitter users will be protected. Twitter will only provide anonymized, aggregated information.

Technology, information and declining distribution costs are challenging the economic foundation of a traditional supplier in favor of newer ones.

Sounds kinda familiar...JL

Ivan Penn reports in the New York Times:

The cost of power from utility-scale solar farms (is) on a par with natural-gas, and wind farms were less expensive still. A wind farm can literally be put on a train and brought online within a year. Power generated by natural gas declined 7.7% in 2017. A healthy grid requires consistent power, even when the sun doesn’t shine or the wind ceases to blow (but) there’s a trend across the energy sector where coal and natural gas can’t compete.“I am nervous we will build capital plant that doesn’t stand the test of time.”

No they are not morally superior.

What they have are more significant revenue streams from diverse products and services, so are not as dependent on selling customer data as are Facebook and Twitter. JL

Leonid Bershidsky reports in Bloomberg:

The guarantee that companies won't go overboard in invading privacy is that they have revenue streams which make it an unnecessary risk. Apple is a hardware company that sells content and software. Microsoft, with its cloud, software and subscription businesses is even less likely to go rogue in data collection. Monetization through ads was the choice for companies not offering products or services. That revenue stream is not as easy to hold on to as those coming directly from customers. That's why Microsoft and Apple aren't suffering much from the Cambridge Analytica-triggered tech sell-off.

Despite heart-warming stories about coastal VCs taking luxury bus trips around the industrial midwest and professing delight at the potential opportunities, the data suggest that investment not only continues to pour into San Fran and NYC, but that the trend is accelerating.

The reality appears to be that the best way of attracting venture capital is by already having secured it. Could this change? As Americans become tech savvy while businesses and research institutions become not just tech dependent, but tech-focused, possibly. JL

Richard Florida reports in CityLab, photo by Gabrielle Lurie in Reuters:

The Bay Area accounts for 45% of venture investment in the United States. Boston, New York, and Washington comprise another third. These two regions attract three-quarters of America’s venture capita. The five leading metros account for 80% of venture capital investment and 85% of its growth over the past decade. Charlotte, Nashville, Pittsburgh, Portland, Dayton, Huntsville, Tampa, San Antonio, St. Louis, Cincinnati, and Sacramento accounted for 3% of U.S. venture investment in 2017. Seattle saw no growth in venture capital over the past decade.

It's too simple to say that it's not the technology, it's genetics.

But the reality may be that like any good consumer focused device, smartphones are designed to trigger the human desire to use them as much as possible.

It's just that given mobiles' convenient size and functionality, they may be overstimulating peoples' eminently susceptible sensors. JL

Justin Dupuis reports in the World Economic Forum, photo by Stefan Wermut in Reuters:

We may be glued to our smartphones because of an evolutionary drive for socializing, rather than a technological addiction to them, new research suggests. Humans evolved to be a social species and require constant input from others to guide culturally appropriate behavior, meaning, goals, and a sense of identity. The most addictive smartphone functions all tap into the human desire to connect with other people. The pace and scale of hyper-connectivity pushes the brain’s reward system to run on overdrive.

How A Wearable Device Can Respond To Your Thoughts

Reading neuromuscular signals that transmit 'internal verbalizations.' JL

Thuy Ong reports in The Verge:

MIT researchers created a wearable device that can recognize nonverbal prompts, “reading your mind.” The system can communicate with the user via a pair of “bone-conducting headphones” by transmitting vibrations from the face to the ear. The headphones are meant to convey information to the user without interrupting their conversation or hearing. A code to analyze data found seven places on the face were able to recognize nonverbal words.

When it can intuit context, inference and meaning. JL

Will Oremus reports in Slate:

While we know Alexa is always listening, how does she think? And how much of what she hears does she “remember?" It has no sense, of words or meaning. That’s in the cloud. It’s looking for a pattern. Everything else is sound waves passing through. Context is important. Humans have no problem understanding inferences. Computers can’t do this today. Context over multiple interactions, your environment, whose present, who’s not present, where you are physically will allow a more natural conversational interface with your artificial agent. That’s the biggest challenge facing A.I

The debate may clarify what is at stake. JL

Greg Ip reports in the Wall Street Journal:

The Supreme Court long ago ruled that antitrust law is meant to protect competition, not competitors. “The integrity of antitrust require that successful firms not be attacked simply because they obtain dominant positions." The Justice Department or Federal Trade Commission could hurt by punishing the most competitive companies, (but) both productivity and innovation would be set back. Politicized regulation would add to the uncertainty that businesses regularly complain deters confidence and investment.

The impact of virtual reality does not end when the headset comes off. And no one fully understands how that will effect human behavior.

What is apparent is that neither the companies attempting to make VR commercially viable, nor the society on which it is being unleashed (nor its legal, medical and public safety professionals), nor the people who will experience it, are entirely prepared for the consequences. JL

Joshua Rothman reports in The New Yorker:

It isn’t just that we live inside a model of the external world. We also live inside models of our own bodies, minds, and selves. These “self-models” don’t always reflect reality, and they can be adjusted in illogical ways. They can portray a self that exists outside of the body—an O.B.E. The best part of V.R. comes after you remove the headset: having been immersed in a flat computer-generated world, in real life, “the most ordinary surface, cheap wood or plain dirt, is bejeweled in infinite detail for a short while.”

Forget the outrage for a minute. This is about data quality and advertising performance, which is the basis for Facebook's growth and it's valuation.

If reports prove true, that in response to its reputational crisis, the quality of Facebook's data and the efficiency that results from it is declining, the business model is in jeopardy JL .

Josh Marshall reports in TPM:

The crisis poses a threat to Facebook’s core business model, its advertising business. Facebook’s market valuation comes from advertising. And advertising comes from data and the artificial intelligence which models it into predictive efficiency. Marketers (have recently) noticed platform instability and reductions in targeting efficiency. Any breakdown in growth and consistency is a problem. It’s having to clean up the reputational mess by rejiggering its core revenue stream, something it’s not clear it knows how to do. That creates a lot of unpredictability.

Trade wars are generally initiated for emotional rather than rational business reasons.

Resentment fuels unrealistic perceptions of advantages to be gained, usually without accurately assessing a competitor's psychological and political imperatives.

The Smoot-Hawley Tariffs of 1930 were also initiated to protect domestic producers and their employees. And we know how that turned out. JL

Krishnadev Calamur reports in The Atlantic:

China has been the engine of global economic growth in recent years.“It will hurt competitive business models, global supply chains, and r&d because of the pinch on profitability. That will hurt job opportunities, productivity [and] growth.”Consumers, faced with a choice of two expensive goods,one made in China and the other in the U.S., may buy neither. Laid-off workers are consumers who will buy less, leading to a spiral that could end with recession. “The way American consumers feel pain in the short terms is in their portfolios. It’ll affect savings, and that affects discretionary spending."

Apr 14, 2018

Apr 13, 2018

Apr 12, 2018

Apr 11, 2018

Apr 10, 2018

Apr 9, 2018

Apr 8, 2018

contact

Search This Blog

Blog Archive

Labels

links