You can now sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
A selection of 9 links for you. They’re slinky, they’re linky. I’m charlesarthur on Twitter. Observations and links welcome.
»As an anthropologist, I wanted to interrogate AI not just as a technical agenda but as a cultural category. I wanted to look at the intellectual history of it. I found myself reading [Alan] Turing and his incredibly provocative question: “Can a machine think?” And the whole notion of the Turing test — Is there a moment where we as humans can no longer distinguish ourselves from the machines? It’s a really interesting formulation both of a technical idea but also a cultural one. It’s also where you can see the cultural ambivalences and anxieties too.
In the conversations in the press and public culture, AI is often accompanied by everything from the language around the robot apocalypse, the singularity, to the idea that they’ll replace or kill us, all depending on the narrative. I was interested in why those two stories were so tightly coupled. Why have the conversations around AI always necessitated this other conversation? Unpicking that was also a very anthropological endeavor.
link to this extract
»Almost 50m people have watched the Olympic Games on Snapchat so far, as broadcasters including NBC and the BBC use the app to reach a millennial audience.
Nearly one in three daily Snapchat users has viewed the clips in Live Stories, showing that the app could challenge other social sites such as Facebook and Twitter for dominance in live events.
The LA-based start-up partnered with seven broadcasters showing in countries including the US, the UK and Brazil, to show stories that include footage from the games and from the crowds in the last 24 hours.
In the first seven days to last Thursday, 49m unique visitors viewed Olympics content on Snapchat, almost a third of the 150m daily active users of the app.
That’s Snapchat, which was just some prototype code in September 2011, and barely known during the last Olympic games. Technological change can be fast.
link to this extract
»It was October 2013, and [DEA agent Carl] Force had spent the past couple of years working on a Baltimore-based task force investigating the darknet’s biggest drug site, Silk Road. During that time, he had also carefully cultivated several lucrative side projects all connected to Bitcoin, the digital currency Force was convinced would make him rich.
One of those schemes had been ripping off the man who ran Silk Road, “Dread Pirate Roberts.” That plan was now falling apart. As it turns out, the largest online drug market in history had been run by a 29-year-old named Ross Ulbricht, who wasn’t as safe behind his screen as he imagined he was. Ulbricht had been arrested earlier that month in the San Francisco Public Library by federal agents with their guns drawn.
Now government prosecutors were sifting through a mountain of evidence, and Force could only guess at how big it was. The FBI got around the encryption of Ulbricht’s Samsung Z700 laptop with a street-level tactic: two agents distracted him while a third grabbed the open laptop out of his hands as Ulbricht was working. The kingpin had been caught red-handed, tapping commands to his Silk Road subordinates up until the moment he was cuffed.
Force had been treating Ulbricht like his personal Bitcoin ATM for several months by this point, attempting to extort DPR one day and wrangling Bitcoin bribes for fake information the next. Now, Force didn’t want to be holding those bitcoins anymore. He opened an account with Bitstamp, a Slovenia-based Bitcoin exchange where he thought he could turn coins into cash quickly and quietly.
But when Force opened Bitstamp account #557042 on October 12, 2013, it sealed his fate.
link to this extract
»there are good arguments to be made for embracing USB audio, [Intel architect Brad] Saunders said, and the companies that make USB controller chips are very excited about the idea.
For one thing, the 3.5mm audio jack takes up precious volume inside phones, which we all want to be as slim as possible. For another, the analog circuitry of 3.5mm audio can cause interference that disrupts other electronics in a phone, Saunders said.
And digital audio opens up possibilities for lots of sound-processing options without requiring expensive headphones or earbuds. That includes audio effects to make music sound like it’s playing in a big concert hall, or signal processing to cancel noise like jet engines or rumbling trains. “All of those come into play if audio is in a digital domain,” Saunders said, which would let phone makers offer premium features without having to sign deals with premium audio companies like Dolby or Bose.
USB devices have controller chips that consume power. That’s no problem for PCs with big batteries, but it is for phones. That’s why the new USB audio standard requires power management abilities like turning off features that aren’t being used, Saunders said. As a result, with USB headphones, “the difference in battery life is negligible” compared with 3.5mm audio jacks.
»I did find the numbers quite startling. In July, NPR.org recorded nearly 33 million unique users, and 491,000 comments. But those comments came from just 19,400 commenters, [NPR social media exec Scott] Montgomery said. That’s 0.06% of users who are commenting, a number that has stayed steady through 2016.
When NPR analyzed the number of people who left at least one comment in both June and July, the numbers showed an even more interesting pattern: Just 4,300 users posted about 145 comments apiece, or 67% of all NPR.org comments for the two months. More than half of all comments in May, June and July combined came from a mere 2,600 users. The conclusion: NPR’s commenting system — which gets more expensive the more comments that are posted, and in some months has cost NPR twice what was budgeted — is serving a very, very small slice of its overall audience.
It’s not possible to tell who those commenters are; some users comment anonymously. But there are some clues that indicate those who comment are not wholly representative of the overall NPR audience: They overwhelmingly comment via the desktop (younger users tend to find NPR.org via mobile), and a Google estimate suggested that the commenters were 83% male, while overall NPR.org users were just 52% male, Montgomery said.
When viewed purely from the perspective of whether the comments were fostering constructive conversations, the change should come as no surprise. The number of complaints to NPR about the current comment system has been growing—complaints that comments were censored by the outside moderators, and that commenters were behaving inappropriately and harassing other commenters.
There’s a similar writeup by Montgomery but it doesn’t have those deep-dive numbers that Jensen offers. “The market has spoken. [Twitter and Facebook] is where people want to engage with us,” Montgomery says earlier.
News site comment software is screwed. Personally I would short Disqus, which NPR (and many others) used.
link to this extract
»At Kik, we remain bullish on bots, but we’ve noticed the same thing everyone else has: so far, there has been no killer bot. This is not yet the world that the early hype promised. But then again, we’re only four months into this. The bot ecosystem is at an earlier point now than the App Store was before Apple introduced in-app purchases. The game lasts longer than the first pitch.
Since we opened our bot platform for developers in April, more than 20,000 bots have been built for Kik. We’ve learned a lot. One of the things that has become increasingly clear is that the initial discussion about bots being powerful because of their conversational potential was somewhat misguided. It’s certainly possible to imagine a world in which we routinely carry out human-like conversations with robots to get things done or be entertained, but we don’t yet live in such a world. In fact, I believe we’ll look back on the early emphasis on “conversational commerce” as a mistake…
…It’s also important to note that we don’t think bots are going to replace apps any time soon. That’s not the point. The point is that people are increasingly spending their time in chat apps, so we’re building experiences inside chat that allow people to do more while they’re there. That’s why bots are so interesting.
Still a sceptic, personally. A visual interface does require you to launch an app, but RAM is cheap, switching apps is easy (people do it all the time) and a visual interface is quicker than a typed conversation. It’s GUI v CLI (command line interface). Unless you know the magic incantations, GUI wins every time.
link to this extract
»Online display advertising has been sold to us as superior to traditional advertising because it presumes that reaching the perfect individual is more economically advantageous than reaching a broad demographic type.
For the most part, offline advertising is sold on demographics while display advertising is sold on data-driven targeting.
While Proctor & Gamble’s experience should not be taken as conclusive proof of anything, it suggests that for big brands the demographics model is more economically efficient than the data-driven model.
Their experience with Fabreze air freshener was cited by The Wall Street Journal as an example of how highly targeted advertising failed.
For Fabreze, P&G targeted people with pets and people with large families. The presumption was that these people would have a significantly higher likelihood to purchase an air freshener than the public at large. Sales stagnated.
Then P&G targeted all adults over 18 — a very broad swath. And sales picked up.
Presumably P&G had the good sense to use the same creative so they knew what variable they were testing.
If P&G’s experience turns out to be projectible – and it has been reported that other marketers are having similar experiences – the whole model of online advertising, based on data-driven “precision targeting” and tracking – and enabled by ad tech – needs to go right down the toilet. It’s a sham.
P&G is rather large to be a canary in a coalmine, but this might be one of those moments – at least for larger advertisers.
link to this extract
»[Cyanogen co-founder and CTO Steve] Kondik said historically there have been challenges with tracking CyanogenMod users, given that its users were “privacy-focused.” (The earliest versions of the software, which started in 2009, restricted any kind of tracking.) In addition, he said, “there was a lot of uncertainty around” the user numbers at the company and, as a startup, “we don’t have the best dashboards and mechanisms” for counting them. He added that in recent weeks the company had discovered six to seven million users who had manually opted out of being tracked by the software. This group represented about two-thirds of the total user base, he said.
“We just figured this out,” he said.
Touting big numbers and growth is de rigeur for consumer tech startups, and it’s common for them to use creative definitions of a “user.”
The numbers quoted by Cyanogen executives after the company launched in 2013 made it seem like the operating system had broad consumer appeal — even bigger than Microsoft’s Windows Phone OS, they liked to say. The company has raised more than $100m in funding from investors including Andreessen Horowitz and Benchmark Capital.
So Cyanogen(Mod) had about 10m users? All of them, however, active in comment threads about how AMAZING Cyanogen(Mod) is/was. But it doesn’t look like the VC companies are going to see their money back.
link to this extract
»Bias lighting is the happy intersection of indirect lighting and light compensation. It reduces eye strain and produces a better, more comfortable overall computing display experience.
The good news is that it’s trivially easy to set up a bias lighting configuration these days due to the proliferation of inexpensive and bright LEDs. You can build yourself a bias light with a clamp and a fluorescent bulb, or with some nifty IKEA LED strips and double-sided foam tape. It really is that simple: just strap some lights to the back of your monitors.
I’m partial to the IKEA Dioder and Ledberg technique myself; I currently have an array of Ledbergs behind my monitors. But if you don’t fancy any minor DIY work, there are a wide array of inexpensive self-adhesive LED strips out there – which also have the benefit of being completely USB powered, and thus can power up and down with your monitor or TV.
Errata, corrigenda and ai no corrida: yesterday’s post included a link from Wired about a survey of Facebook users and political views, allegedly carried out by a company called Rantic. However, Rantic is a “buy followers” social media marketing company (my opinion? A parasite on the business) and I don’t think the survey is robust. Without evidence to the contrary, I doubt it even exists. I’ve emailed the Wired writer suggesting the story be withdrawn or at the very least queried for the full survey data, but haven’t heard back so far. Fingers crossed. In the meantime I’ve removed the link from the site, and urge you to ignore any findings it might have appeared to pass on.