Start Up No.1,184: Twitter tries to dump the dunk, Microsoft’s lost bet, ‘personal CRM’, misunderstood marginalia, and more

The good news: all this is now open source. The bad news: it’s effectively dead. Or maybe that’s the good news. CC-licensed photo by sndrv on Flickr.

You can 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. Friday! I’m @charlesarthur on Twitter. Observations and links welcome.

Twitter is trying to fix the dunk and ratio with new product tweaks • Buzzfeed News

Alex Kantrowitz:


Twitter knows we treat each other terribly on Twitter. We dunk, ridiculing friends and strangers via quote-tweets. We ratio, piling on replies to bad tweets. We retweet without a second thought, spreading outrage and misinformation at warp speed.

But within the next two weeks, Twitter will debut a series of experiments meant to calm us down — subtly motivating us to use the quote-tweet, reply, and retweet in nondestructive ways.

“Everything on our platform incentivizes some form of behavior,” David Gasca, a senior director of product management at Twitter, told BuzzFeed News. Amid the company’s push for healthier conversations, he’s wondering “if we modify how people can do retweets, or how people can reply, or how people can engage, how does that change conversation on the platform?”

Twitter hopes to find out. In a meeting at its San Francisco headquarters in late October, Gasca and Suzanne Xie, director of product management at Twitter, showed off two experiments among several that will go live in the coming weeks: In the first, Twitter will add an emoji to a retweet, giving people a chance to quote-tweet without going into the compose field. Gasca and Xie want to find out if this feature might encourage people to express more nuanced emotions, putting a damper on dunking and mindless retweeting.

In the second experiment, Twitter will automatically suggest people use an emoji in their replies. If you like something, you could use the heart-eyes emoji. If you don’t, you could use the red circle with a line going through it. But if you pick a negative emoji, Twitter will ask, “Why do you disagree?” — which it hopes will prompt a more thoughtful reply, rather than a flame war.


Good luck with that. (Alternatively: 🙄) A good idea, but there’ll surely be a lot of resistance.
unique link to this extract

DxOMark scores shouldn’t be your definitive camera rating system • Android Authority

Robert Triggs:


DxO Labs, the company which runs the DxOMark testing suite, is primarily a consultancy company. In other words, the company charges fees to advise camera hardware companies on how to improve their photography products. This is based on its own analysis and expertise in the camera industry.

No review site is guaranteed to be free from bias, but DxO’s business revolves around attracting big companies to it to make use of its expertise, which adds a lot of baggage to their reviews. Ranking test results in a way that encourages consumers to buy certain phones over others complicates everything.

The company claims to run an independent test, but is that really possible when it offers for-profit consultancy at the same time? There’s no reason to believe DxOMark is in anyway rigging results. After all, the company’s business model depends on its reputation and its results tend to roughly fit with the broader consensus on camera hardware.

However, manufacturers that tune their cameras against the testing suite are likely to score higher than those who don’t. We have heard that a few smartphone manufacturers don’t think DxO’s consultancy fees are worthwhile. These manufacturers don’t score highly on DxO’s tests, if the company even reviews these phones at all.


I’ve never understood DxOMark’s number-specific testing; how does one slice up photo quality in the way that they do and hope ever to keep it sensible? Corralling all that into a single number seems even worse. Personally, I ignore these figures. Smartphones are long past this point, at least at the top end.
unique link to this extract

Microsoft bet against Intel with its new Surfaces — and lost • The Verge

Chaim Gartenberg:


The focus on non-Intel chips was a big part of Microsoft’s 2019 Surface announcement. The company went out of its way to highlight the new, co-engineered Ryzen Surface Edition processor for the 15-inch Surface Laptop 3, which was optimized specifically for Microsoft’s design. And the ARM-powered Surface Pro X, with a next-generation design and a custom variant of Qualcomm’s Snapdragon 8cx chip, was supposed to represent the future of the Surface.

Unfortunately, the reviews are in, and experience has shown that neither of those custom chips have panned out quite like Microsoft had hoped.

Let’s start with the Surface Laptop 3’s Ryzen Surface Edition chip from AMD. Microsoft said that the new chip was meant to offer faster speeds and improved graphics performance in particular, thanks to an extra core. But as my colleague Dan Seifert noted in his review, the AMD chipset still struggled with most games and even basic 4K video playback. More importantly, the AMD chip was crushed in a head-to-head contest when it came to exporting video against the 13-inch, Intel-powered Surface Laptop 3, which has more thermal limits due to its size and is cheaper than the larger model.

…Unfortunately, the Surface Pro X seems to prove once again that the dream of an ARM-based Windows laptop is still a half-baked idea. App compatibility is still a big issue, performance isn’t great, and the much-vaunted battery life doesn’t always hold up as well as promised. A lot of this is down to 32-bit app emulation: when apps are designed to run on ARM, the Surface Pro X actually does pretty well. But those apps are still few and far between — the hardware may be here, but the software isn’t. And if past history is anything to go by, Microsoft may have a hard time getting developers on board.


Wonder how that ARM Mac laptops are doing in their tests.
unique link to this extract

How Google edged out rivals and built the world’s dominant ad machine: a visual guide • WSJ

Keach Hagey and Vivien Ngo:


Nexstar Media Group, the largest local news company in the US, recently tested what would happen if it stopped using Google’s technology to place ads on its websites.

Over several days, the company’s video ad sales plummeted. “That’s a huge revenue hit,” said Tony Katsur, senior vice president at Nexstar. After its brief test, Nexstar switched back to Google.

Alphabet’s Google is under fire for its dominance in digital advertising, in part because of issues like this. The US Justice Department and state attorneys general are investigating whether Google is abusing its power, including as the dominant broker of digital ad sales across the web. Most of the nearly 130 questions the states asked in a September subpoena were about the inner workings of Google’s ad products and how they interact.

We dug into Google’s vast, opaque ad machine, and in a series of graphics below, show you how it all works—and why publishers and rivals have had so many complaints about it.

Much of Google’s power as an ad broker stems from acquisitions of ad-technology companies, especially its 2008 purchase of DoubleClick. Regulators who approved that $3.1bn deal warned they would step in if the company tied together its offerings in anticompetitive ways.

In interviews, dozens of publishing and advertising executives said Google is doing just that with an array of interwoven products. Google operates the leading selling and buying tools, and the biggest marketplace where online ad deals happen.


It would take a huge lawsuit by the US DOJ to reverse this. It might be possible, but proving it would be hellish – and what would replace it?
unique link to this extract

UpHabit, Dex, and the stilted rise of the personal CRM • The Atlantic

Kaitlyn Tiffany:


“When life gets busy, sometimes we need to be reminded to enjoy our most meaningful relationships,” the creators of Garden write on their website. “Your relationships are secured for today!” the activity-completion page on Ryze announces once you’ve taken care of all your “following up.” Ntwrk promises to make its users into better friends, mentors, siblings, salespeople, and networkers; reminders to reach out also come with a summary of “what you last chatted about.” Social Contact Journal provides anniversary reminders and prewritten message templates.

While many of the apps have an explicit professional-networking utility, the Irish company Monaru, one of the Y Combinator companies, is focused specifically on users’ 10 to 15 closest relationships. Not only will Monaru remind you of a loved one’s birthday, but it will also suggest specific gifts to buy her. It can help you plan a date night, or remember to call your parents regularly. “Millennials are four times lonelier than seniors,” the company’s homepage reads, probably erroneously. The service costs $20 a month, and its tagline is “Be the most thoughtful person you know.” (The creators declined to be interviewed, saying they were “heads down” on the product.)

The idea of people as self-contained collections of data points is not a new one—the Quantified Self movement has been booming and busting since 2007. The idea of offloading your brain into a computer is not new either, though it’s a little more controversial now that we’re more aware of what happens to our personal information after we do so. But quantifying other people is different, and mediating relationships with software isn’t a purely personal decision.

All these apps released their first version in 2018 or 2019 (though Monaru is in private beta and Clay has a waitlist). They appear in the “Productivity” section of the App Store. They are, on their surface, another blurring of work and life, another viral tweet about how modern life is like a dystopian Mad Lib, and while you can fill in whatever nouns you like, the overarching story will be about exploitation, isolation, and capitalism run wild. Is that all they are?


unique link to this extract

Misunderstanding nonlinear prices: evidence from a natural experiment on residential electricity demand • American Economic Association

Blake Shaffer:


This paper examines how consumers respond to nonlinear prices. Exploiting a natural experiment with electricity consumers in British Columbia, I find evidence that some households severely misunderstand nonlinear prices| incorrectly perceiving that the marginal price applies to all consumption, not simply the last unit. While small in number, the exaggerated responses by these households have a large effect in aggregate, masking an otherwise predominant response to average price. Previously largely unexplored in the literature, this type of misunderstanding has important economic, policy and methodological implications beyond electricity markets.


The full paper isn’t yet available except to AEA members. But as Paul Kedrosky (via) points out, this isn’t really about BC Hydro’s electricity pricing; it’s about how bad we are mentally at understanding marginal rates, where you pay (say) 10% tax on your first $1m of earnings, and then 90% for anything above it.

Quick: how much tax does someone who earns $1m pay? And someone who earns $2m, who is into the 90% tax bracket? Who ends up with more money? Our mental models work on linear maths, but this crucial tax system isn’t. (A human cognitive bias of sorts, perhaps?)
unique link to this extract

Amazon’s Fire TV to carry Disney+ • WSJ

Dana Mattioli and Ethan Smith:


Amazon has reached a deal with Walt Disney Co. to carry the Disney+ streaming-video service on Fire TV devices, according to people familiar with the matter.

The two companies had been at loggerheads over terms for carrying Disney’s apps on Amazon’s Fire TV devices, the Journal reported last month. At the time, Fire TV wasn’t listed as a partner that would carry Disney’s new streaming service as a result of those disputes, The Journal reported.

Amazon was pushing for the right to sell a substantial percentage of the ad space on Disney apps. It is unclear if Disney agreed to Amazon’s terms. Disney+ launches on Tuesday. Disney is set to report earnings Thursday afternoon.

…Disney+, which will include franchises such as “The Simpsons,” “Frozen” and “Star Wars,” has been a draw for providers. Last month, Verizon Communications Inc. said it would offer its wireless customers on unlimited data plans a year of free access to Disney+, giving the cellphone carrier a fresh way to hold on to customers.

Under that agreement, Disney and Verizon would share the cost of providing the content to the carrier’s subscribers, according to a person familiar with the arrangement.


The subtle push and pull of platforms v content is playing out all over again; we saw it before with smartphones and services. What sort of deal – if any – did Apple make with Amazon to get Apple+ onto Fire sticks, one wonders?

unique link to this extract

‘Fake News’ isn’t easy to spot on Facebook, according to new study • University of Texas News


In the study, participants fitted with a wireless electroencephalography headset were asked to read political news headlines presented as they would appear in a Facebook feed and determine their credibility. They assessed only 44% correctly, overwhelmingly selecting headlines that aligned with their own political beliefs as true. The EEG headsets tracked their brain activity during the exercise.

“We all believe that we are better than the average person at detecting fake news, but that’s simply not possible,” said lead author Patricia Moravec, assistant professor of information, risk and operations management. “The environment of social media and our own biases make us all much worse than we think.”

Moravec, along with Randall K. Minas of the University of Hawaii at Manoa and Alan R. Dennis of Indiana University, authored the study, “Fake News on Social Media: People Believe What They Want to Believe When it Makes No Sense at All,” published today in Management Information Systems Quarterly.

The researchers worked with 80 social media-proficient undergraduate students who first answered 10 questions about their own political beliefs. Each participant was then fitted with an EEG headset. The students were asked to read 50 political news headlines presented as they would appear in a Facebook feed and assess their credibility. Forty of the headlines were evenly divided between true and false, with 10 headlines that were clearly true included as controls: “Trump Signs New Executive Order on Immigration” (clearly true), “Nominee to Lead EPA Testifies He’ll Enforce Environmental Laws” (true), “Russian Spies Present at Trump’s Inauguration — Seated on Inauguration Platform” (false).

The researchers randomly assigned fake news flags among the 40 noncontrol headlines to see what effect they would have on the participants’ responses…

…As they worked through the exercise, the participants spent more time and showed significantly more activity in their frontal cortices — the brain area associated with arousal, memory access and consciousness — when headlines supported their beliefs but were flagged as false. These reactions of discomfort indicated cognitive dissonance when headlines supporting their beliefs were marked as untrue.


Love that last bit. Also: explains why Facebook would be so lairy of labelling news content. It doesn’t want readers feeling uncomfortable.
unique link to this extract

Google open-sources Cardboard SDK to keep it alive • Android Police

Manuel Vonau:


Long before Google introduced Daydream and subsequently left it dead in the water, the company created the Cardboard platform. You can use the carton headsets as an ultra-low-budget entry to VR to this day, and they’re compatible with almost any regularly shaped phone on the market. Google has now open-sourced the underlying VR SDK which will allow interested developers to create their own VR experiences on Cardboard viewers and improve and enhance the project as they see fit.

Google says that it still wants to contribute to the project and plans to release a Unity SDK package, but it hasn’t actively developed the Google VR SDK for some time already. Still, it sees “consistent usage around entertainment and education experiences,” so it didn’t want to shut down the platform altogether. Google states that “an open source model will enable the community to continue to improve Cardboard support and expand its capabilities, for example adding support for new smartphone display configurations and Cardboard viewers as they become available.”


Alternative headline: “Google puts Cardboard into back of car to drive to the mountains for a long refreshing walk.” It’s dead, Jim. All the signs are that virtual reality is heading for another winter.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

12 thoughts on “Start Up No.1,184: Twitter tries to dump the dunk, Microsoft’s lost bet, ‘personal CRM’, misunderstood marginalia, and more

    • Well, my best friend forgot my 50th last year, and called me at midnight+15 mins this year trying to claim it still counted. No dice, but a lesson on that nice b’day field in Contacts ;-p
      Also, I’ve got far-away young nephews, it’s kind of hard to keep track of what they’ll like for b’days even xmas. They change fast at that age, asking them over Skype feels really unnatural, daddy is getting swamped by ideas requests every time. I kinda still have a feel for what M. 10/11/12 (actual birth date) wants after seeing him in the summer, but M. March and Mrs May change a lot over those few months. I’ve really looked into Amazon’s recos, they suck. Plus there’s surprisingly little available in Canada esp in French, I’m having trouble locating that Minecraft table game. I could use a kid expert + concierge ;-p

      But indeed, subcontracting human interaction feels weird, especially when it’s THAT on the nose.

  1. Re DxOMark

    It’s clear that
    1- the significance of quantified camera tests is diminishing now that the optics is becoming less important vs the computation
    2- optimizing for a test might be non-representative if the test itself is non-representative
    3- there’s a moral risk when the testing company takes payment for consulting (towards the test among other things), though that risk is low for quantified reviews.
    4- Specifically the Mi Note 10’s camera is overestimated in that test, it is only class-leading in very specific circumstances (the most common ones though), is great-not excellent in quite a few situations, is good-not-great in low light, but seems to rarely completely fail
    5- I think what’s beginning to matter more is not so much top pic quality, but rarity of utter fails (= not instagramable result). As Raylan Givens would say, that’s a whole other discussion.
    6- all reviews kind of fail because SW updates and alternate camera apps can have a huge impact but are never tested. In the Android world for example Google has made effort to expose all camera features to apps, and a Google Camera app hack called gCam is very popular on some well-supported handsets.

    But, it’s also clear that a specific section of the commentariat is very adept at
    1- unconditionally endorsing anything iPhones are good at / have (performance in general esp single-core, abandoned 3Dtouch, facial vs finger recognition…)
    2- dismissing anything that’s not in favour of iPhones: multicore performance, battery life, features such as removable storage and FM radio…
    3- and then remissing when Apple improves : battery life…
    I think that second discussion is way more important that whining when a well-established test doesn’t produce the desired outcome.

    • Also, smartphones aren’t past objective measurements, even real cameras aren’t. There are such a things as dynamic range, white balance, distortion, chromatic aberrations… that can never be perfect in all situations, can be measured, and consistently impact real-life (well, real-pic) quality.

      Going for subjective evaluation is valid – if done double-blind which it rarely is; anything not double-blind is utterly invalid. But totally discounting objective measurements isn’t valid. It’s not random that the two track closely.

      So the validity of evaluations goes:
      real-life blind subjective > tech measures > lab scenes, then we fall into the “invalid” bucket with non-blind subjective and spec sheets.

      Even lab test pics of standard scenes are a little bit useful, it’s very easy to see on GSMArena’s that a Redmi Note 8 trounces an Iphone 7 but not an 8: (look at the donkey’s head in the “good light ” scene)

      And of course the only right way to do it is to compare, otherwise we’re doing a panegyric or roast, not an evaluation.

      • I’m not the one getting personal. If you’ve got an actual point to make, please do.
        And if as you try to sell Google were my god, I’d maybe have or recommend their handsets (or more generally, hardware), which I don’t except to troglodyte photographers.
        My god is value for money. You should try it, most people do.

      • Also,
        1- you’re confirming that you’re intent was to trash an OEM.
        2- that article isn’t pointing out anyone’s feet are made of clay, just that lab measurements are not a perfect evaluation.

        If anything, I see you picking an article that criticizes a test your dear Iphone is not winning as you being defensive. It’s funny how much people reveal of their own psyche in their criticism/comments of others. That’s something I’ve observed time and again, ie untrusting people are untrustworthy etc…


    Weirdly, Huawei got the biggest growth in Q3. Samsung also grew rather strongly, Oppo (no info on the rest of the BBK stable) grew a bit, Apple is flat and Xiaomi slightly down.

    If Huawei gets unbanned which I think is slightly likely (for handsets not infrastructure) Samsung is within reach at only 3% total share above. That’s 20% relative share, which amusingly is the current growth differential. And I’m hoping Xiaomi will try to eke out a few points by re-launching their Max line, hopefully as a Redmi not Mi ;-p

  3. Early Xiaomi Mi Note 10 reviews are in. No surprises: exceptional camera, excellent everything else. (google translate still only speaks pig english, sorry)

    The most dubious thing about it is its value at the current €500, at that price even Samsung handsets are barely more expensive, and the non-flagship 730G starts to hurt. Xiaomi has been using a Samsung-like strategy of quick price cut + frequent sales, once the early adopter gouging is over and that phone reaches €400 it will be a lot more compelling.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.