AI First

Looking to the future, the next big step will be for the very concept of the “device” to fade away. Over time, the computer itself—whatever its form factor—will be an intelligent assistant helping you through your day. We will move from mobile first to an AI first world.

Google Founder’s Letter, April 2016

My Library recently finalized a Vision Document for our virtual library presence. Happily, our vision was aligned with the long-term direction of technology as understood by movers and shakers like Google.

As I’ve written previously, the Library Website will disappear. But this is because the Internet (as we currently understand it) will also disappear.

In its place, a new mode of information retrieval and creation will move us away from the paper-based metaphor of web pages. Information will be more ubiquitous. It will be more free-form, more adaptable, more contextualized, more interactive.

Part of this is already underway. For example, people are becoming a data set. And other apps are learning about you and changing how they work based on who you are. Your personal data set contains location data, patterns in speech and movement around the world, consumer history, keywords particular to your interests, associations based on your social networks, etc.

AI Emerging

All of this information makes it possible for emerging AI systems like Siri and Cortana to better serve you. Soon, it will allow AI to control the flow of information based on your mood and other factors to help you be more productive. And like a good friend that knows you very, very well, AI will even be able to alert you to serendipitous events or inconveniences so that you can navigate life more happily.

People’s expectations are already being set for this kind of experience. Perhaps you’ve noticed yourself getting annoyed when your personal assistant just fetches a Wikipedia article when you ask it something. You’re left wanting. What we want is that kernel of gold we asked about. But what we get right now, is something too general to be useful.

But soon, that will all change. Nascent AI will soon be able to provide exactly the piece of information that you really want rather than a generalized web page. This is what Google means when they make statements like “AI First” or “the Web will die.” They’re talking about a world where information is not only presented as article-like web pages, but broken down into actual kernels of information that are both discrete and yet interconnected.

AI First in the Library

Library discussions often focus on building better web pages or navigation menus or providing responsive websites. But the conversation we need to have is about pulling our data out of siloed systems and websites and making it available to all modes like AI, apps and basic data harvesters.

You hear this conversation in bits and pieces. The ongoing linked data project is part of this long-term strategy. So too with next-gen OPACs. But on the ground, in our local strategy meetings, we need to tie every big project we do to this emerging reality where web browsers are increasingly no longer relevant.

We need to think AI First.

ProtonMail: A Survivors Tale

Beginning November 3rd, encrypted email service provider, ProtonMail, came under a DDOS attack by blackmailers. Here is my experience, as a supporter and subscriber, watching from the sidelines. It’s a survival story with many heroes that reads like a Mr. Robot script.

Why Encrypt Your Email?

ProtonMail is an encrypted email service that I just love. It overcomes the problems with email providers’ harvesting your personal data for resale, the pitfalls of these databases falling into criminal hands and just plain weirdness you feel when every word, attachment and contact is shared to whomever.

To make my point on why everyone should use encrypted email, like ProtonMail, consider this experience: I recently had to fill out an affidavit confirming my identity but did not have all the particulars with me, such as past addresses, etc. No problem, I just logged into my 12 year old Gmail account and did some searching. In no time, I had all the personal info the affidavit required to prove my identity.

It’s not that I purposely saved all this information in there. It just accumulates over the years organically.

Imagine if that data fell into the wrong hands.

ProtonMail is a crowd-funded, free email service that comes out of the CERN laboratories in Switzerland and MIT. The engineers at these research facilities were inspired by the revelations of Edward Snowdon about back doors into email servers and the general collection of data by governments, so they built ProtonMail.

The service is simple, elegant and super secure. The encryption happens through the use of a client-side password, so theoretically, nobody, not even ProtonMail, can decrypt your emails and read them.

ProtonMail Taken Down

The recent Distributed Denial of Service (DDOS) attack began on November 3rd when a group held for ransom access to ProtonMail’s email service. This was a very sophisticated attack that flooded their servers with requests, but also their ISP. The result was that ProtonMail and several other sites, including e-commerce and banking sites, were unreachable. After failing to successfully fight back, the ISP and other firms put enormous pressure on ProtonMail to pay off the cyber gang. They did so and the attack stopped…momentarily.

Less than half a day later, the attack re-commenced. This time it was even more sophisticated and destructive. And, things got even weirder. The original blackmailers actually contacted ProtonMail to let them know they were not involved in the new attack. ProtonMail is pretty certain that the second attack was likely a state entity.

You can read all the details on their blog post on the incident.

Over this past weekend, November 7-8th, ProtonMail launched a response to the ongoing attack, deploying new defensive technologies used by large Internet firms, funded through a GoFundeMe campaign. As of this writing nearly 1,500 individuals donated $50,000 in just 3 days to help in this regard.

Those would be the first, rather large, set of heroes. Thanks to you guys!

Click here to add to the fund.

Social Networks Get the Word Out

The media was really late to this story. It was not until the end of the week that the first news reports came out about the blackmail story made sexier by the fact that the ransom was paid with bitcoins.

Most of the breaking news, however, was only available on ProtonMail’s Twitter feed and their Sub-Reddit.

It was on their Twitter page that they first disclosed the moment-by-moment details of their fight to restore access and their ultimate attempt to fund new defensive technologies. It was on Reddit that the controversy and pain was aired such as reactions to their payment of the ransom and frustration of everyday users at not being able to access their email.

People really gave them a lot of credit, however. And it was heartening that, despite some rather single-minded rants, most people rallied around ProtonMail.

Lessons Learned

One thing I was surprised about were some of the complaints from business people that were using ProtonMail as their exclusive business email. They were losing money during the attack so they were often the most irate. But you have to wonder about someone using an emerging tool like ProtonMail for something so critical as company email. Obviously, new Internet services take time, especially when they are not backed by seasoned VCs who are risk adverse.

I personally had not made the switch to ProtonMail entirely. Part of this was because they don’t have an iPhone app yet, which is where I do about 50% of my emailing. But I was getting close.

So, yes, I had a few important emails get bounced back to the senders. And perhaps one or two have been lost permanently (I may never know). But it does go to show that, for the foreseeable future, ProtonMail is not a reliable sole-email solution. However, given the work they are doing in response to the latest attack, this event may be the turning point that makes them a truly stable email service.

Just this morning, they came under another attack, but unlike previous days over the past week, they were back online very quickly. Hopefully this means their new defenses are paying off.

Bottom Line

ProtonMail rocks. I really love it. The recent DDOS attack only confirms that the good team at CERN and MIT are dedicated to doing what it takes to keep this alive. I can think of other such services that have folded when they came under similar pressure. In fact, the user community around ProtonMail is as serious as ever, shelling out the money required to safeguard encrypted email just when it counted.

There will likely be further trouble ahead. The British government has suggested it might ban encrypted email services. And who knows how the US will respond long term. So, there could be more chop ahead. But for the time being, it seems that ProtonMail may have survived a very critical test of its resilience.

Stay tuned!

Is Apple Pay Really Private?

Apply Pay, the new payment system unveiled by Apple yesterday was an intriguing alternative to using Debit and Credit Cards. But how private, and how secure, is this new payment system going to really be?

Tim Cook, Apple CEO, made it very clear that Apple intends to never collect data on you or what you purchase via Apple Pay. The service, in fact, adds a few new layers of security to transactions. But you have to wonder.

A typical model for data collection business models is to promise robust privacy assurances in their service agreements and marketing even though the long-term strategy is to leverage that data for profit. Anyone who was with Facebook early on knows how quickly these terms can change.

So, when we’re assured that our purchases will remain wholly private and marketing firms will never have access to them, how can we really be confident that this will always remain the case? We can’t. So, as users, we should approach such services with skepticism.

As with anything related to personal data, we should assume that enterprising hackers or government agents can and will figure out a way to access and exploit our information. Just last week, celebrities using Apple’s iCloud had their accounts compromised and embarrassing photos were made public. And while Apple has done a pretty good job at securing Apple Pay, it’s still possible someone could figure out a way in…and then you’re not just dealing with incriminating photos, you’ve got your financial history exposed.

So ask yourself:

1. Can you think of things you buy that could prove embarrassing or might give people with malign intent a way to blackmail or do financial damage to me?
2. If my most embarrassing purchases were to become permanently public, can I live with that?
3. How would such public exposure impact my reputation, professionally and personally?
4. Does the convenience of purchasing something with my phone outweigh the risks to my financial security?

Depending on how you answer this, you may want to stick with your credit card.

Or just go the analog route and use the most anonymous medium of exchange: cash.

This Too Shall Pass – Deleting My Facebook Account

I’m killing my Facebook account.

And with it, I’m severing that company’s ability to collect data on my web habits, whereabouts, social connections (including off-Facebook connections) and financial transactions.

Apparently, I’ll also be reducing my exposure to NSA malware, as Mark Zuckerberg revealed in a public thrashing of Obama and the intelligence services that have been spreading malware through imposter Facebook sites.

This really won’t be that hard. Last year, I began experimenting with not using the social network, just to see how that was. This impulse was born from a general annoyance about FB’s murky privacy policies and the general tone of content on FB which had became increasingly irrelevant to my real social connections with people. (Remember when people started to appreciate that group emails were rude and began with the lines, “sorry for the group email!”…That’s how Facebook seems to me now, without the apologetic preface.)

BTW, if you’ve got your own suspicions about Facebook, the Electronic Frontier Foundation has put together a great timeline of Facebook’s shifting privacy policies. Reading their timeline is a great way to get your head around how free Internet services (FB, Gmail, etc.) are really about hooking you in with very clear and considered privacy policies that are planned to be revoked once they’ve got you dependent on them…or at least that’s how the timeline suggests this business model works.

Of course, deleting my Facebook account won’t be without costs.

If you’ve been a rolling stone like me, you have friends in many far-flung places. Facebook did make those connections feel stronger, so that aspect will be missed. But I’m online and quite findable, so if my pals from Japan or Europe want to find me, they only need know my name.

It turns out that completely deleting your account is a two-week process, described quite well on Digital Trends. The trick is that once you delete your account, you cannot log in for two weeks, or your account will be reactivated. That means you should first delete all apps from phones, tablets, etc. before deleting your account. You should probably delete your cookies too, just to be sure you don’t inadvertently reactivate it by triggering all those FB web beacons that mine the Interwebs.

Anyway, I’ll give my FB contacts a few days to run across my post and then I’ll zap it for good.

Better living through anonymity!

Google Book Scans and Library Catalogs

As librarians, we’ve long championed fair-use causes, and last week’s US District Court ruling in favor of Google’s book scanning project drew great interest from librarians. I think it’s exciting, but I also see a new and serious challenge to libraries ahead.

In a legal challenge brought by the US Writer’s Guild against Google, US District Judge Denny Chin ruled (PDF) that Google’s attempt to scan books was fair use. Chin’s main rationale can be boiled down to the fact that Google is not selling these books and does not make the entire book available in any practical way that might allow users to read a complete work online. Moreover, the books being scanned were often given to Google for scanning by organizations that had purchased the book and thus had every right under fair-use to scan these books and cite portions of them online. In Google’s case, this means the company can render relevant segments of a book to users in the form of in-page search results.

The final details are still to be worked out in a settlement between the two parties.

It was a good day for the public as well as Google. But this will definitely impact the library, especially the relevance of our (increasingly irrelevant) catalogs.

I can’t say this strongly enough: Libraries have missed the boat, big-time, when it comes to our approach to information retrieval over the past 15 years or so. Our anachronistic approach to marking up books, journals and other resources with metadata has been wildly inefficient for too long. Take the Internet model for retrieving knowledge, which uses algorithms augmented with elements of artificial intelligence. These retrieval functions integrate a given users social history, incorporate learnings from how users interact with results and even understand spelling, synonyms, grammar and natural language. The indexes these algorithms search, in turn, weight everything more or less democratically, based largely on the strength of interconnections between files (via hyperlinks) and keywords.

Now consider the library: where human agents try to pre-think the users intent by applying metadata with controlled vocabularies in unforgiving, and yes, arrogant fashion. We don’t use the users’ words. We use our words. Woe be to the user brought up on Google’s much more forgiving (i.e. helpful) search tool, who tries to use their own vocabularies to find relevant materials. When we do allow for keyword access, it is so weighted down the algorithmic chain (below our “authorized” terms) that users are often forced to page through results before they find what they’re looking for. If they can even find our catalog to start with (but I get ahead of myself here).

Even many of our most advanced discovery systems still don’t get this fundamental flaw in our approach. One discovery system in particular that I have experience with doesn’t account for how contemporary searchers often enter search strings with the assumption that they will be understood as phrase searches. At least, that’s the tendency they have leaned toward since Google’s algorithm reinforced this habit many years ago when it added proximity searching based on the order of words entered into their search engine. In this particular discovery system I’m referring to, if you search for a title without quotes, there is a very good chance (based on a number of factors) that your intended item won’t come up anywhere near the top ten…or even top 100 items.

So to summarize, libraries are still plodding down a dead end. Rather than forcing people to learn how we’ve organized things, we need to use machines to understand how our users (and the world) organizes things.

So how does the Google book scanning ruling relate to this dismal state of affairs?

Try an experiment: go to Google and type in the title of a currently playing movie. Chances are, Google will provide you with immediate information at the top of your results on which local cinema is showing that particular movie, and what time it is playing.

Now try searching for a recent book title. You’ll likely get Amazon.com, perhaps the publishers site, etc. What you won’t see is a link to your local library, let along library catalog.

So from this, consider the very near future (like next month) when you type in keywords into Google and begin seeing results from books. Click the book and you’re treated to alot of metadata, especially social metadata, like Goodreads reviews, links to author websites and in some cases, the ability to read the book if it’s open access.

Fortunately, there is a small link to “Find in a library” on the left-hand side once you’ve clicked into the Google Books interface from your initial search results. It’s just below the very obvious [Get Print Book] button which takes you to the big online book sellers. If you click on the library link (not the Add to your library which is something else), you’ll get the OCLC record with links to your nearest libraries. So it’s not totally a disaster, but hey, how did we get to this place where users looking instant access to any given information, don’t get a link to their local library’s request form (or directly to their local library’s e-version). But instead have to fork over mullah to learn something new. At least they still have YouTube and Wikipedia, I suppose.

We really need to get on the bandwagon here folks. A huge victory for fair-use has just been won, but until libraries start pushing our way to the front of the line (and the top of the results), we’re going to always come up last. And that’s not where our users are looking.

The Library Website Will Disappear

A new year always elicits thoughts about the future, but this month, my library has been considering our next strategic plan, which has focused those thoughts for me on my library and the profession in general.

Since I’m principally charged with managing the online aspects of my library, I come to these kinds of discussions focused on web platforms, online communication and consumer technology trends. One of the biggest trends (you may have noticed) has been the adoption of mobile, touch screen devices like iPhones and tablets. I include even Microsoft’s attempt at reinventing the PC with its Surface Tablet, despite early failures to woo consumers. And given this apparently irresistable move toward mobile, tablet-like computing, I have to ask: what does this ultimately mean for the library’s web properties: our websites, our online instruction guides, our discovery systems and our digital collections?

Palo Alto Venture Capital firm KPCB recently explored this very issue in their 2012 Internet Trends Report. Some major insights from their report include:

• Mobile traffic is now 13% of all global Internet traffic, up from just 1% in 2009. In some countries, like India, mobile Internet traffic has surpassed desktop traffic.
• Almost 1/3 of US adults own a tablet or e-reader
• Together, iOS and Android are 45% of the OS market share, vs. 35% for Windows
• The install base of Tablets + Smartphones will surpass PCs + Notebooks this year

But one trend that stands out in KPCB’s analysis is that of the Asset-Light generation. A long-form definition can be found via a very Asset-Light resource: Quib.ly. But to put it more succinctly, being asset-light means your lifestyle is one less reliant on physical commodities and personal know-how, but instead, relies on cloud and crowd-sourcing just about everything. You don’t carry around notepads, you don’t buy maps, you don’t rely on “experts” for medical advice or which movie to see next. Asset-lighters, meet their needs by streaming, connecting and sharing.

From KPCB’s perspective, this is quite important in terms of where the web is heading. Some notable examples relavent to libraries include: MOOCs (Massive Open Online Courses), Amazon Textbook Rental and of course, good-ole Wikipedia.

So, socially, we’re moving toward a very different economy and radically different means of information distribution, especially when it comes to learning.

Add to this, two major and imminent innovations and you will get a glimpse at just how different the world will be in only 5 years:

1. Wearable devices, especially eyeware will give people the ability to navigate the Internet by voice command, gesture and all through lenses fixed to their faces. Status: Google Glasses are expected to enter the market in early 2013 and Apple won’t be far behind.
2. Artificially-intelligent agents will not quite resemble HAL from 2001, but be very capable of understanding your vocal commands and then ferreting quite reliable answers to your questions or carrying out mundane tasks, like creating appointments or sending messages for you. Status: Did you see IBM Watson slaughter Jeopardy’s world champions or used SIRI on your iPhone?

So, the way we gather information and where we go to get it is already changing. And the interfaces are already being revolutionized and that pace will accelerate dramatically over the next 5 years as voice and sight overtake the the very impractical and immobile keyboard…even the touchscreen may be reduced to the point where iPads seem like a whimsical dead-end much like 8 track or Beta tapes.

All of these changes will have an immediate effect on the core of our current Internet paradigm: the Web Page.

Text and links with a smattering of images have been the key content types of of web pages since Tim Berners Lee first formulated the WWW. Mobile devices have changed all that. Not only do they steer away from typical web interfaces in favor of “apps,” they actually de-link parts of the web from each other. The result, in most cases is a much more curated and manageable Internet.

And this is important for libraries, whose pages are almost hard-wired around interconnecting pages together in rather daunting tangles of hyperlinks, portals and gateways. Unfortunately, this paradigm is increasingly less relevant to today’s devices and today’s Asset-lighters, who expect a web page to cut through the clutter and get them the answer. In fact, they want an app to do the heavy-lifting for them.

And add to this, semi-intelligent software agents and a re-conceived commercial Internet based around voice and sight and you can see how much work libraries have ahead of them.

The users 5 years from now that enter our libraries’ virtual spaces, will expect a curated, largely automated experience. Already, we see this on the ground where incoming students are completely beside themselves in the antiquated library environment. One recent Facebook post on my library’s newsfeed noted: “The Library has a website?”

Another telling anecdote: One colleague of mine defined her job as teaching people to fish. I then asked: How many people actually go fishing anymore. Fishing to them is dropping by the supermarket. Full-stop.

The world is just getting too complicated for people to be expected to take the time to find information on their own. Information will continue to be a commodity, yes. Information will continue to badger the human mind. But AI servants and wildly different means of gathering information, will mean that single individuals will never have to tackle almost any information problem alone. The crowd, the bots and the apps will do the fishing.

And the web page will be like Matrix code that few ever need to concern themselves with. Get ready…

Below the Microsoft Surface

Silverlight required to get product info: #FAIL

Coming back from Thanksgiving with my family, I came across a Microsoft Surface demo in the airport. Normally, I would have dismissed it immediately, the same way you might dismiss suggestions about trying a new dish at Chipotle after that first, stale burrito you had on your initial visit. But, I was trapped in airport limbo for 40 minutes, so the Surface seemed like a good way to pass some time.

Microsoft’s new tablet is really a new spin on the laptop, which, if you’re like me and need to create more than you consume online, is a much better solution than the typical Android or Apple tablet. That is, Microsoft kept the keyboard in mind when looking for their much required business strategy. Unfortunately, the new tablet does not come equipped with the full-blown Windows 8 OS, which seems like a big, bad burrito. That won’t come until after Xmas (did someone say missed opportunity here?).

Anyway, I was impressed, generally. Of course, we’re still talking Windoze, so the experience was fraught with so many error messages and OS-fails, that even the woman giving the demo seemed annoyed. But is still had enough new takes on what is now an old market (for tablets), that I had to give Microsoft some credit.

The most telling experience, though, came after the demo when I was asked to take a brief survey. The questions just said it all:

• What would you buy before seeing the Surface/after seeing the Surface? (Apple)
• Did you consider MSFT an innovative company before seeing the Surface? How about now? (If you have to ask, the answer is no)

But at least they woke up to the bleakness of their market position to even start asking these questions and try to change things, albeit soooooo late. On the train back from the airport, a fellow passenger and I discussed the Surface. He noted that MSFT is opening retail shops beside Apple Stores in shopping malls across the country. When I mentioned this to a colleague at work, he laughed, mentioning that he’d seen one of these shops. Apparently the Apple Store was jam-packed, while the MSFT store went ignored.

For now, at least, people will get their tacos from the place with the fresh ingredients.

Apple is the New Enemy

I was pretty shocked by the jury verdict that allowed Apple to assert proprietorship over things like pinch zoom and icon design in their proxy war with Google via Samsung. Apparently, others are too. Slate urged us to imagine the state of our world had a single car company been allowed to claim rights to the design of the steering wheel, forcing all others to create different means for directing a vehicle. Never mind if the steering wheel was the most obvious, best design. Had a single auto manufacturer been allowed to patent the steering wheel, all others would be using levers, knobs and joysticks for cars and the auto market would effectively be captured by a single company that filed its patent first.

Watching TED tonight, I saw this reality check on where patent law has brought us. Take a look:

Insight into the future of Augmented Reality

Just a good video that explores where all this augmented reality stuff is headed. If it seems far-fetched, think again. Consultants in Silicon Valley are actively engaged with the big tech firms to develop eyeware, beginning with goggles due out this year. Contact-lens AR ware, as explored in this video, is forecasted to be on store shelves in by the end of the decade.

Sight from Sight Systems on Vimeo.

We Are All Terminators Now

For years, tech analysts and insiders have speculated that augmented reality goggles were just a few years away. These products would provide a data layer superimposed over the real world similar to that depicted in the Terminator movies…albeit one less likely to place crosshairs over Sarah Conner, and more likely to allow you to friend her.

Well, last week, the NYT wrote that such products are within months of coming onto the market, and will likely be released in time for Planet Niburu’s arrival (aka Christmas 2012). That’s right, folks, all that talk about the world-as-we-know-it coming to an end in 2012 is really going to happen. Only, future historians will log 2012 not as the realization of a Mayan apocalypse, but as the beginning of the end of the Internet and the rise of something much better.

Imagine you’re walking down the street and want to find the closest biker bar. Simply input your search (I’m fascinated to see how this will get pulled off) and then start scanning the surrounding world around you for the results.

Far more than virtual reality or cyberspace, augmented reality is the real future for information seekers. My guess is that 20 years out, people will look at examples of web pages and have to laugh at how rooted in the 20th Century file cabinet mentality they were. They will scratch their heads and wonder how anyone ever found anything relevant. The very idea that you would create information that was not tied to a physical space will astonish them.

Once AR goggles come online, relevance will be tied to location…period. We’ve already seen this with our location based services via our smartphones, but these too, are merely a stepping stone. And I think if you look at Google’s product line you’ll start to see that they have been anticipating the AR future for some time.

Here’s my take: The web of the future will be something akin to Google Earth + Planet Earth. Let’s say that you’re looking for a woman named Sarah Conner. A few years ago, if you ran this search, you would be treated to all kinds of location-agnostic results that would drown you in a flood of irrelevant data. Run that search on Google today and you might get the Sarah Conners your contacts on Google Plus are associated with and perhaps the ones listed in directories in your current town. Fast-forward to an Internet that is oriented around goggles, and suddenly your top search result would be the Sarah Conner closest to you on Planet Earth at that moment (assuming she allows any random killer robot from the future to locate her online).

Or, let’s say that you really don’t want the closest Sarah, but just the ones in Los Angeles. Flip over to Google Earth in your goggles, run your search and zoom in on L.A.

Termination was never so easy!

All of this will really challenge the utility of the simple web site, suspended in the cloud, independent of the real world. It seems that at the very least, web sites will be dramatically reorganized over the next decade as we scramble to geo-locate wikipedia articles and the like. Some things may not be so helpful if they are tied to a location (where does StarWars.com go? Up in the sky? Lucas Ranch in California?), but the vast bulk of sites related to locations will move from the web to the world.

And now for the compulsory cliché: The future’s so bright, we gotta wear…