Impact and questions: An AI reading list


As part of some research into artificial intelligence (AI) and machine learning (ML) over the past few months, I’ve come across a lot of reading material.

Here’s some that stood out to me and that I can recommend looking into. Please note that this is very much on the non-technical end of the spectrum: Primers, as well as pieces focusing on societal impact, ethics, and other so-called “soft” aspects, i.e. societal, political, humanitarian, business-related ones. These are the types of impact I’m most interested in and that are most relevant to my work.

The list isn’t comprehensive by any means—if you know of something that should be included, please let me know!—but there’s a lot of insight here.


Basics, primers:

Resources, reading lists, content collections:



Reports, studies:

Presentations, talks:


For completeness’ sake (and as a blatant plug) I include three recent blog posts of my own:

Merging book stores & ebooks


filed under: to read/re-read

I love bookstores. I love them for their curation. I particularly love how smaller bookstores have a fantastic selection. However, I have a dark secret — I almost only go to bookstores when I’m traveling, and I go to great pains not to carry more stuff than I absolutely need. So I rarely buy printed books in stores, ever.

Instead — and this sounds (and is) a little sad — I jot down the books I want in the bookstore and buy them later, often online. This feels deeply unfair, almost like cheating on the bookstore owner.

So I’m wondering how both worlds can be merged for the better. In other words, how can we enjoy the curatorial services of a well-run bookstore and make sure the shop stays in business — without having to lug around books in a carry-on?

Here’s an idea: We cut bookstores in on every sale they help generate.

The mechanism could be relatively simple. A store owner signs up online with the big platforms and publishers — say, Amazon. When I go to a bookstore and find a book, I scan the book’s barcode with the store’s Kindle app to buy an electronic copy. The app checks my location, asks me to confirm the store I’m in, and registers all book sales through the app to that store. While the ebook downloads to my device of choice, the shop collects a commission. Amazon (or any other platform or publisher, for that matter) sells another book, I can keep traveling light, and the store gets its fair share. Everyone’s happy.

Where the billing takes place doesn’t really matter at this point: In this example it’s through Amazon, but it could be any number of new umbrella services or publishers’ platforms. There’s probably room in this space for half a dozen startups. But no matter how it happens, the important thing is that bookstores get a share in exchange for their curation. Because we really don’t want to have to rely on Amazon’s recommendation services alone.

Is anyone working on this yet?


Update: David D. Levine kindly pointed me to this cooperation between Powell’s Partners and Kobo. Thanks, David!

Update: I re-wrote and polished this post up a bit to re-post it to Medium.

Update: Thanks to the fine folks at Medium, notably Kate Lee, the post has been featured in Medium’s Editors’ Picks as well as the collection The Future of Publishing.

My setup: Reading, Writing, Bookmarking


As someone who reads and writes a fair amount online, I’m always looking to optimize my information flows. Particularly getting articles from one service to another can be tricky.

I found a set of tools and workflows that I’m quite happy with for the time being, so I thought I’d share them.

info flow

Feedback, tipps for improvements etc are always welcome!


I like to read primarily on Pocket, particularly Pocket for Android on my Nexus tablet. So that’s where I want my “to read” items to go.

From my browser, I send articles via the Chrome plugin that sends the links directly. Sometimes I tag them, sometimes I don’t, so far that hasn’t been super important for me.

To get links into Pocket from Twitter, I use this IFTTT recipe that sends the content linked from my Twitter faves and sends it over. In other words, I can fave a tweet, and the article linked in that tweet gets sent over straight away. Extremely comfortable particularly when skimming Twitter on the phone.

Pocket offers plugins for all major browsers and platforms.


When reading in Pocket – mostly on the tablet, even though the mobile and Mac apps are also pretty decent – I can get through a lot of info in relatively little time. If I want to save something for later, I add some tags within Pocket and fave the article there. This triggers two processes…


One, this IFTTT script takes faved articles and bookmarks them for me on Pinboard, including tags, so I have the post archived for future reference.

Two, the WordPress plugin WP Pocket WP Stacker lets me auto-generate drafts of link lists based on my Pocket faves. With some minimum editing I can get to the reading lists you’ve encountered on my blog.

Once these steps are set up, it’s very convenient and allows me to get through lots of material with a bare minimum of friction. Of course, you’ll want to adapt and tweak to match the tools your personal workflows. But thanks to more and more useful APIs and an ever-growing library of IFTTT recipes, it’s easier than ever to plug these services into one another.


Recent reading (4 links for Feb 10)



Irregularly, I post noteworthy articles I recently read. Enjoy!


Sam Jacob on 3D printing
Sam Jacob argues that instead of liberating us, 3D printing will merely “bind us even more closely to fewer and fewer corporations” and that it might not change our relationships to commercial objects the way often portrayed. (link)


MIT Initiative for the Digital Economy Warns Technology May Change Employment Opportunities : The CIO Report
MIT just launched the Initiative for the Digital Economy to research the impact of digital technologies on business, economy and society. – by Irving Wladawsky-Berger (link)


Maker Faire Rome Announced
Maker Faire comes Italy, and they’re being ambitions about it. See you there. – by Ari Honka (link)


Deutschlands Open-Data-Portal wird nur ein Data-Portal
Wenige Tage vor dem offiziellen Start des deutschen Datenportals ist das “Open” aus dem Namen verschwunden: Das einst als OpenGovData geplante Portal wird nur noch GovData heißen. – by John H. Weitzmann (link)

Too Big To Know: The Science that exists at the network level


David Weinberger has written a new book, out just now: Too Big To Know. In The Atlantic, he published an abstract. Here are two brief samples just to give you an idea:


The problem — or at least the change — is that we humans cannot understand systems even as complex as that of a simple cell. It’s not that were awaiting some elegant theory that will snap all the details into place. The theory is well established already: Cellular systems consist of a set of detailed interactions that can be thought of as signals and responses. But those interactions surpass in quantity and complexity the human brains ability to comprehend them.


We have a new form of knowing. This new knowledge requires not just giant computers but a network to connect them, to feed them, and to make their work accessible. It exists at the network level, not in the heads of individual human beings.

I tend to be very careful, bordering on wary, when it comes to US books that have even the slightest touch point with management and business. Too many one trick ponies out there that are just the ticket to the speaker circuit.

However, what David Weinberger delivered here doesn’t seem to be one of those. (In fact, he’d probably pretty appalled that I even put him in the same league with those others, and rightfully so.) I have tremendous respect for him and his thoughts, and both the Cluetrain Manifesto and Small Pieces Loosely Joined were pretty much seminal works that I’ve been working with, and revisiting, ever since they came out.

Too Big To Know seems to fall in that same category. While I’m waiting for the delivery – it won’t be out till Jan 19 in Germany – I can only recommend anyone who might have to do with Big Data or science over the next few years to consider ordering a copy.

Is Google making us stupid? (No.)


The Atlantic just ran an article asking “Is Google Making Us Stupid?” which two friends forwarded independently – always a clear indicator that it might be a good idea to actually read the article. (Thanks Puja, thanks Burkhardt!) And The Atlantic author Nicholas Carr does raise some important points.

Carr’s key argument is that the way we read skim texts in the web makes us lose the ability to immerse ourselves deeply in longer texts: We forget how deep reading works. And he’s right on when he observes that all of us information workers find it increasingly bothersome to sit down with a book and read it front-to-back while emails keep pouring in, Blackberries chirp, phones ring. We all, I suspect, know the feeling. It happened to me more than once that mere minutes after sitting down with a book I started fidgeting and was drawn almost subconsciously to my email inbox, basically by reflex.

However, that’s only half the story.

Our reading patterns are (or so I think) to a large portion context-based. Reading in the office, with my computer sitting next to me, is particularly hard because distractions are everywhere and interruptions are mostly legitimate and not just random noise. It’s an office, after all. (In case you’re wondering: I work freelance, so it actually is possible for me, in theory, to read in my office.) But I can’t really see the same mechanisms at work in different contexts. When I like to immerse myself in a story, a book, a long and complex text, I try to block out as many distractions as I can. Switching off my phone for an hour isn’t heretical, it’s a legitimate choice. It takes little time to get back from skimming mode into deep reading mode. (Of course, it may very well be that younger generations have a harder time doing so because they learned to read in a different way; this I cannot say for sure and I do not want to judge about.)

There’s another point in Carr’s article that I find somewhat disturbing. He compares the internet to a system like Taylorism, “designed for the efficient and automated collection, transmission, and manipulation of information”. Taylor’s ethic, Carr concludes, “is beginning to govern the realm of the mind as well.”

One, the facts: The internet was by no means design for efficiency. Quite to the contrary, the internet was designed for redundancy, and it’s as messy as anything: Parts are failing? No worries, it’s all decentralized. The information will find another way. The internet with it’s complex architecture is what we call an emergent system. It’s (in a way) the opposite of efficiency.

Two, the internet is not our mind. “What Taylor did for the work of the hand, Google is doing for the mind,” says Carr. Google is not trying to tell us how to think. It’s a system that was (and still is) built to find information we are looking for: Google serves our demands, not the other way round. Of course, always having all the information out there at our fingertips does have an effect on our thinking (just as it seems that we completely forgot how to make appointments with friends without cascade of cell phone calls beforehand). But that doesn’t mean changing back is impossible.

Three, Carr states that “in Google’s view, information is a kind of commodity, a utilitarian resource that can be mined and processed with industrial efficiency.” So? Google isn’t taking any workers’ rights to personal freedom here: Computers process information, in one way or another. It’s what they do, what they’re built to do. Trying to get more relevant results at the expense of nobody and nothing isn’t a bad thing to do in itself, is it? (Of course, secondary effects like low concentration span when over-using can be an issue.) Spending “days in the stacks or periodical rooms of libraries” doesn’t have an intrinsic value (which, to be fair, Carr doesn’t claim), it was (is?) rather a flaw in the system, a symbol of not having access to the information one is looking for. Efficiency is a word that comes to mind, convenience is another. Opportunity is a third, and maybe the most important one here: We now have, for the first time in human history, the chance to access a global repository of information, data, knowledge.

Is Google making us stupid? I think not. Lazy, maybe.