Start Up No.986: Kurzweil on AI (yes), what 5G will and won’t do, that big old password breach, Google’s GDPR fine, Corbyn’s failure, and more

It’s a new sort of shopping. But has it screwed up, er, online shopping? CC-licensed photo by James Fleeting 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 10 links for you. Consult your doctor before installing a backstop. I’m @charlesarthur on Twitter. Observations and links welcome.

Ray Kurzweil: ‘AI is still on course to outpace human intelligence’ • Gray Scott

BJ Murphy:


Using examples of modern-day AI like AlphaGo, there are clear signs that they’re already starting to outpace human intelligence involving specific tasks. This has been a common factor of AI for the last couple decades, starting with the simple goal of defeating the world’s best (human) chess players. Today, they’re outpacing us in chess, Go, various strategic computer games, and even medical diagnostics. The question remains, however, as to whether AI will ever reach the point of superintelligence—also commonly known as the Technological Singularity.

What I found most interesting from Kurzweil’s response wasn’t so much his consistency in the belief that AI will indeed outpace human intelligence as a whole; rather our fears of a dystopian future where AI has gone astray are becoming increasingly unlikely. He makes this arguement with the understanding that there is no singular AI being controlled by singular powerful companies or people. In today’s reality, there are millions of different AIs being controlled by anyone who owns a smartphone.

One could argue that the level of power still isn’t well-balanced between centralized companies and a decentralized populace, especially as companies like Facebook, Amazon, and Google (Kurzweil’s current employer) continue making headlines as a result of their egregious negligence. However, with companies like SingularityNET working to democratize the technology, AI isn’t just moving at a pace beyond human intelligence; they’re moving at a pace that’ll empower the human species as a whole, whether that comes in the form of maintaining their longevity, increasing their cognitive capacities, or giving them access to the stars themselves.


The difference between being superhumanly good at Go and being superintelligent is like the difference between flight of stairs and climbing K2. They’re the same class of problem, separated by colossal levels of difficulty.
link to this extract

Amazon ruined online shopping • The Atlantic

Ian Bogost:


The whole affair always felt unsettling. When the buttons launched, I called the Dash experience Lovecraftian, the invisible miasma of commerce slipping its vapor all around your home. But last week, a German court went further, ruling the buttons illegal because they fail to give consumers sufficient information about the products they order when pressing them, or the price they will pay after having done so. (You set up a Dash button on Amazon’s app, selecting a product from a list; like other goods on the e-commerce giant’s website, the price can change over time.) Amazon, which is also under general antitrust investigation in Germany, disputes the ruling.

Given that Amazon controls about half of the U.S. online-retail market and takes in about 5% of the nation’s total retail spending, it’s encouraging to see pushback against the company’s hold on the market. But Dash buttons are hardly the problem. Amazon made online shopping feel safe and comfortable, at least mechanically, where once the risk of being scammed by bad actors felt huge. But now online shopping is muddy and suspicious in a different way—you never really know what you’re buying, or when it will arrive, or why it costs what it does, or even what options might be available to purchase. The problem isn’t the Dash button, but the way online shopping works in general, especially at the Everything Store.


Bogost is always worth reading, because he always finds a fresh way to come at a story.

link to this extract

773M password ‘megabreach’ is years old • Krebs on Security

Brian Krebs on the “megahack” that was big news last week:


Collection #1 offered by this seller is indeed 87GB in size. He also advertises a Telegram username where he can be reached — “Sanixer.” So, naturally, KrebsOnSecurity contacted Sanixer via Telegram to find out more about the origins of Collection #1, which he is presently selling for the bargain price of just $45.

Sanixer said Collection#1 consists of data pulled from a huge number of hacked sites, and was not exactly his “freshest” offering. Rather, he sort of steered me away from that archive, suggesting that — unlike most of his other wares — Collection #1 was at least 2-3 years old. His other password packages, which he said are not all pictured in the above screen shot and total more than 4 terabytes in size, are less than a year old, Sanixer explained.

By way of explaining the provenance of Collection #1, Sanixer said it was a mix of “dumps and leaked bases,” and then he offered an interesting screen shot of his additional collections. Click on the image below and notice the open Web browser tab behind his purloined password trove (which is apparently stored at Troy Hunt’s published research on this 773 million Collection #1.

Sanixer says Collection #1 was from a mix of sources. A description of those sources can be seen in the directory tree on the left side of this screenshot.

[CTO of Hold Security, Alex] Holden said the habit of collecting large amounts of credentials and posting it online is not new at all, and that the data is far more useful for things like phishing, blackmail and other indirect attacks — as opposed to plundering inboxes. Holden added that his company had already derived 99% of the data in Collection #1 from other sources.


So it’s basically like the fluff-covered sad-looking pick’n’mix sweet trays in Woolworths of old.
link to this extract

Google fined €50m for GDPR violation in France • The Verge

Jon Porter:


France’s data protection regulator, CNIL, has issued Google a €50m fine (around $56.8m USD) for failing to comply with its GDPR obligations. This is the biggest GDPR fine yet to be issued by a European regulator and the first time one of the tech giants has been found to fall foul of the tough new regulations that came into force in May last year.

CNIL said that the fine was issued because Google failed to provide enough information to users about its data consent policies and didn’t give them enough control over how their information is used. According to the regulator, these violations are yet to have been rectified by the search giant. Under GDPR, companies are required to gain the user’s “genuine consent” before collecting their information, which means making consent an explicitly opt-in process that’s easy for people to withdraw.

Although the €50m fine seems large, it’s small compared to the maximum limits allowed by GDPR, which allows a company to be fined a maximum of four% of its annual global turnover for more serious offenses. For Google, which made $33.74bn in the last quarter alone, that could result in a fine in the billions of dollars.


link to this extract

Jeremy Corbyn is failing the Brexit test • Financial Times

Robert Shrimsley:


It is Britain’s misfortune that at its time of need it has been blessed with two of the most inflexible, small-minded, partisan and inept figures ever to assume the mantle of leadership in the nation’s two major parties. The UK has had bad party leaders before, but until now it has been clever enough not to have them at the same time. 

Theresa May’s current failing is that after the record-breaking rejection of her Brexit deal in parliament she seems incapable of finding a plan B; Mr Corbyn’s is that he has yet to find a plan A. The Tories may own Brexit, but if Britain crashes out of the EU without a deal, the Labour leader who failed to stop it will also be widely blamed.

To be clear, Mr Corbyn’s failure does not lie in his refusal to accept Mrs May’s offer of talks in the wake of her defeat. That mistake was merely presentational. The talks were doomed because Mrs May will not budge on any of her red lines. But by refusing to see her, Mr Corbyn showed his much-vaunted belief in dialogue extends only to those who already agree with him. A more astute politician would have been filmed walking into Downing Street only to emerge sorrowfully declaring that Mrs May was deaf to compromise. But this is mere spin. Mr Corbyn’s real failure is missing the chance to step into the vacuum caused by her defeat. 


Devastatingly accurate. If you had created a book plot where the prime minister was beholden to a political group who want to Leave and represent a small but crucial slice of the UK which had voted to Remain, then you’d feel obliged to create an opposition leader who was agitating for Remain, to create dramatic tension.

Instead we just have tension. Corbyn is like Winston Churchill’s quip about Clement Attlee: “An empty taxi arrived, and he stepped out of it.”
link to this extract

President Trump posts altered photos to Facebook and Instagram that make him look thinner • Gizmodo


In recent months, Trump’s official Facebook and Instagram accounts have published photos of the president that have been manipulated to make him look thinner. If it only happened once you might be able to chalk it up as an accident. But Gizmodo has discovered at least three different retouched photos on President Trump’s social media pages that have been published since October of 2018.

The image below, published on the official Donald J. Trump Facebook page on January 17 and on his Instagram account over the weekend, has been altered to make the president look more fit.

The photo looks pretty normal at first glance. But once you compare it to the original, which is available on the White House’s Flickr page, you can see what was changed.

The original photo was taken by a White House photographer on January 14, so you know that the original hasn’t been altered by anyone in the so-called “fake news media” to make Trump look heavier. It sounds silly to even have to say that, but many Trump supporters believe that the media is involved in a coordinated conspiracy against him.


A new way to lie. But also such a pointless thing to do. He’s one of the most photographed people in the world. (Reader-submitted – thank you.)
link to this extract

In win for tech giants, EU copyright reforms stalled • Reuters

Foo Yun Chee:


EU efforts to reform copyright rules hit a roadblock on Monday when a meeting of lawmakers and officials was called off, prompting criticism of Google from publishers after it and other tech giants lobbied against the changes.

The European Commission, which launched a debate on the issue two years ago, says an overhaul is necessary to protect Europe’s cultural heritage and level the playing field between big online platforms and publishers, broadcasters and artists.

European Parliament lawmakers, representatives from EU countries and Commission officials were scheduled to meet on Monday to reconcile their positions on the reform drive. But the meeting was canceled after EU countries failed to resolve differences on Friday.

Commission digital chief Andrus Ansip expressed disappointment at the delay, saying reform was crucial and possible even at this stage.

“All involved parties have a huge responsibility: playing lightly now with a ‘No deal is better than my own maximalist position’ as I read sometimes from position statements, is dangerous and irresponsible,” he said.


Can’t think who Ansip might have been thinking of in that “No deal” quote…

link to this extract

5G: if you build it, we will fill it • Benedict Evans

Evans says that 5G will essentially be a fatter pipe, but that’s not “just”:


5G seems rather more interesting for AR. To clarify first, ‘AR’ today is used to describe three different things:

• Waving your phone at something and seeing things on the screen
• A wearable heads-up display (Google Glass) with no awareness of the world around you
• A transparent, immersive, fully 3D colour display with a sensing suite that allows it to map the room around you and recognise things and people. A bunch of companies (including Magic Leap, in which a16z is an investor) are working on this – it’s still a few years away from being a mass-market consumer product.

The third of these seems much the most interesting to me. If you could put on a pair of reading glasses that could look at the world around you and show you things in response, that could be pretty useful, in much the same way that, say, having the internet in your pocket turned out to be useful, and to enable all sorts of new and unpredictable things (imagine pitching Snapchat when our only internet experience was on a PC over dialup). This would work on 4G, but continuous low power high speed low latency connections from 5G would make it a lot better. 

At the other extreme, I also hear a fair bit about autonomy [in cars] as a 5G application. I’m not sure about this one.


link to this extract

A few days with the Luna Display • All this

Doctor Drang got a Luna Display (which plugs into a Mac and turns an iPad into a wireless second display. He didn’t like it, and took it back:


everything felt wrong when I was running Mac apps through my iPad. Buttons were too small, even when I tried tapping on them with the Pencil. Resizing windows was a chore; dragging felt off. I confess I didn’t spend time examining why the behavior just didn’t feel right, but it didn’t.

I use both my Macs and my iPad a lot, and while I don’t have any trouble switching between the two, I found it very annoying to be forced into using Mac-like actions on an iPad. This was surprising to me, as I have nearly 35 years of Mac use under my belt and only 2½ years of iPad use. But my immediate sense—a sense that didn’t change over the 4–5 days I used the Luna—was one of unease.

Would I have felt this unease had I been using the Luna Display in a more keyboard-centric manner? Maybe not. And I can see where people who are iPad-first users would find the Luna very convenient if they only occasionally need to be hands-on with their Mac mini server. But for my use, the neither-fish-nor-fowl behavior that the Luna forced me into was very inconvenient. It made me have to think about what I was doing instead of just doing it, and that got in the way of my real work.


link to this extract

Microsoft Cortana – dead or alive • Radio Free Mobile

Richard Windsor:


Microsoft held a media event where Satya Nadella extolled the virtue of Cortana, but it is clear from his commentary that there is no logical reason why anyone will ever use it again.

Microsoft will be moving it away from its position on the Windows 10 taskbar depriving it from the only place on any system where it remains default.

Microsoft will also allow compatibility with Google Assistant as well as Amazon Alexa. The unanswered question is why anyone would ask Google or Alexa to ask Cortana to do anything seeing as both of them can do everything Cortana can do and much more.

Furthermore, Microsoft has not invested in Cortana since it became resident on the Windows 10 desktop which has meant that its presence is more of an annoyance than anything else.
Microsoft claims Cortana is deeply integrated with Office 365 but asking Cortana to do anything is Office is more cumbersome and time-consuming than simply clicking with the mouse. Furthermore, most of the time Cortana has no idea what the user is talking about rendering it effectively useless.

Hence, when Cortana is removed from the desktop, I shall not be sad to see it go. I don’t think other Windows users will be either.

However, putting Cortana quietly to sleep opens the door to a much more interesting opportunity around AI licensing.


I hadn’t noticed that the Cortana chief was going (it was announced in November). Microsoft’s problem is that it doesn’t have a user-facing way to get its AI into use. But as Windsor says, there might be some potential licensing its technology to others.
link to this extract

Errata, corrigenda and ai no corrida: none notified

5 thoughts on “Start Up No.986: Kurzweil on AI (yes), what 5G will and won’t do, that big old password breach, Google’s GDPR fine, Corbyn’s failure, and more

  1. re Amazon: “online shopping is muddy and suspicious in a different way—you never really know what you’re buying, or when it will arrive, or why it costs what it does, or even what options might be available to purchase.”

    Say whaaaaaat ?

    I’m trying to support local shops. They’re making it damn hard. Except for the most fashion-y items (clothes mostly), it’s really hard to not buy from Amazon though:
    – choice of product is vastly superior
    – info on products is inferior on Amazon, but you’re at home at your PC and can look up info on other sites. That won’t beat a good and smart and honest sales assistant… when was the last time you met one of those, and a not a bad or commission-optimizing one ?
    – when it will arrive. Amazon has an almost perfect record for me (except for stuff from their Marketplace, but marketplace isn’t Amazon. A point worth making. 1,000 times over)
    – what it does. It’s a toaster: it toasts bread. If the purchase is more complicated, again, on balance, the Web will be more helpful than your average sales guy.
    – Options: same, Amazon offers rather more options than your typical brick-and-mortar shop, though i’s indeed a bit bad a listing them.
    – Prices are usually better on Amazon, not so much for the big stuff as for the small stuff. got my mom a few extension cords and doodads, local supermarket was about 2.5x more expensive for a unknown brand as Amazon for the established leader.

    I mostly wish brick and mortar shops, and other web shops, were as good as Amazon. The one single issue is the UI for verticals. IT specialist sites let you locate a 12To HDD much faster znd more reliably than Amazon: Amazon has less detailed filters by far, and *will* fuzzy-search you into buying an 8TB SAS drive instead of a 12TB SATA in you don’t double-check search results.

  2. re. Cortana “Microsoft’s problem is that it doesn’t have a user-facing way to get its AI into use.”.

    Erm. Only a billion Windows PC, probably still more than iOS+MacOS devices, though way fewer than Androids.

    The main problem is that Cortana is rather bad, It won’t even set alarms right which is like level 0 of what you want from an Assistant. When I’m at my PC I have to pick up my phone and set alarms on that. According to reviews, it goes downhill from there, and MS seem to want to invest somewhere else… there’s probably money to be made off Amazon or Google selling the Windows Assistant job.

    • The billion (but shrinking) PCs aren’t all set up with Cortana – around half will be in corporate environments where Cortana is probably disabled. Some other proportion won’t have it – the ones with Windows 7 or whatever, though there’s probably a substantial overlap there with corporate installations.
      I don’t think the desktop is a good place for a voice assistant, either. I’ve never wanted to ask Siri to do something on my laptop.
      By contrast those billion iOS + macOS devices – all of them, pretty much, are Siri-capable. Siri is the default on them. Whether that gets used is different, but it’s probably not going to have been turned off at the corporate level. And Siri on the Apple Watch for setting alarms, reminders etc is the right sort of context: you’re mobile, don’t want to interrupt yourself by stopping to write on a screen.

      • Indeed. But if MS was motivated by Cortana, they could backport it to 7/8. For example, Google Assistant was launched after Android 7.0, but in true Android “this not not iOS: no OS update needed” ran on Android 5.0 and up.

        The right place for an assistant is whatever device I’m using right now. But there’s probably value in always using the same assistant, because quirks, fortes/weaknesses and undiscoverable UI+capabilities. Plus If I set an alarm on my PC then get on my way, I want it to ring on my phone.

        I have 0 info on whether corps are turning Cortana off. Anecdotal evidence say no ?

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.