A Twitch Tale: Modern Life, a Debit Card, and Cluelessness

July 21, 2020

DarkCyber spotted an item in one of our feeds because the word “fraud” appeared in the document. The content object was “Teenager Takes $20,000 of Parents’ Money, Gives It to Twitch Streamers.” The write up explains:

the minor spent years of savings in just 17 days using a debit card. The boy paid for subscriptions, which can go as high as $24.99 per month, bought Bits—virtual goods used to Cheer in chat messages—and made uncapped donations to various streamers. Speaker to Dot Esports, the mother said that $19,870.94 was charged to a debit card between June 14 and 30.

Banks view this type of activity as a type of chargeback fraud. A consumer makes a purchase and then requests a chargeback after receiving the product or service.

One question is, “What about those parents?” Another is, “Should Twitch have a more fine grained system in place to prevent those under a certain age from spending above a threshold?”

The Twitch question could be answered with an algorithm or a simple rule based system. The gain for the Twitchers who received some financial love from a follower is good news… for them. For the parents, bad news. Perhaps the alleged adults should look into the concept of a pre-paid debit card with a hard limit? For now, it is hasta la vista $20K. For the teen? Probably back online and absorbing video streams.

And Amazon Twitch? Just another day of “good enough” safeguards for users, their parents, and talent formerly known as Dr. Disrespect, whose name has a certain je ne sais quoi.

Stephen E Arnold, July 21, 2020

A Survey of Prices from the Dark Web

July 21, 2020

The Dark Web may not be the giant repository of badness that some popularizers of sci-fi assert, but it is a challenge for some enforcement professionals.

As important as our personal and financial information is to each of us, it can come as a surprise how cheaply some hacked data can be purchased on the Dark Web. After considerable research, Privacy Affairs illustrates this point in its “Dark Web Price Index 2020.” Reporter Miguel Gomez writes:

“The privacy offered by software such as TOR creates an environment where criminals can sell their wares on the dark web without the worry of law enforcement. What’s more, many will have heard the horror stories of people’s bank accounts being cleaned out, or their identity stolen and turning up in custody in Mexico. Again, not unjustified horror. You might be asking yourself, just how easy is it to obtain someone else’s personal information, documents, account details? We certainly were. Whilst there are many marketplaces on the dark web, there are even more forum posts warning of scammers. This makes verified prices difficult to obtain without ordering the items to find out, which of course we didn’t. Our methodology was to scan dark web marketplaces, forums, and websites, to create an index of the average prices for a range of specific products. We were only interested in products and services relating to personal data, counterfeit documents, and social media.”

The researchers compiled eye-opening lists of products and going rates; interested readers should navigate there to view the entire roster. A few examples: credit card details for an account with a balance of up to $5,000 for just 20 bucks; a hacked Twitter account for $49; a 24-hour-long DDoS attack against an unprotected website, at 10-50k requests per second, for $60. Considerably more expensive, though, are passports from the US, Canada, or Europe at $1,500 or quality malware attacks at 1,000 for $1,400 – $6,000.

The article includes a few interesting details alongside the prices, like the fact that vendors usually guaranteed 8 out of 10 stolen credit cards would pay off as advertised. Also, PayPal account details were very common and cheap, but actual transfers from a hacked account were more pricy. And apparently counterfeit bills are extremely common, with the highest quality ones costing about 30% of their fake value. They even come with a “UV pen test guarantee.” See the write-up for more curious, if concerning, details.

Cynthia Murrell, July 21, 2020

Need Global Financial Data? One Somewhat Useful Site

July 21, 2020

If you need a financial number, you may not have to dig through irrelevant free Web search results or use your Bloomberg terminal to find an “aggregate” function for the category in which you have an interest. Yippy.

Navigate to “All of the World’s Money and Markets in One Visualization.” As you know, my skepticism filter blinks when I encounter the logical Taser “all.” I also like to know where, when, how, and why certain data are obtained. The mechanism for normalizing the data is important to me as well. Well, forget most of those questions.

Look at the Web page. Pick a category. Boom. You have your number.

Accurate? Timely? Verifiable?

Not on the site. But in a “good enough” era of Zoom meetings, a number is available. Just in a picture.

Stephen E Arnold, July 21, 2020

Google May Want to Spin Up Some New Jargon

July 20, 2020

Marketwatch published “Barr Blasts Apple and Google As All Too Willing to Cooperate with China.” The report states:

The criticism of U.S. companies came amid a broad speech on China, in which Barr said the Chinese Communist Party was seeking to “make the world safe for dictatorship” and accused China of waging an “economic blitzkrieg” against the U.S. in a bid for global dominance.

How has Google responded? We noted “Google’s Mission Is to Get Technology to More People: Sundar Pichai”, which is a short video. The article stated:

Google and Jio “would work together to increase internet access for millions of Indians, who do not currently have a smartphone, while improving the mobile experience for all.”

With testimony looming before Congress, Google’s alleged “fraternizing” with a country on the radar of the Attorney General and positioning investments in India as a way to improve “the mobile experience for all” does not capture several nuances about the 21st century of the Google:

  1. Google needs eyeballs to sell ads in order to keep Wall Street and stakeholders content. So “advertising.”
  2. Google appears to be keen on finding some way to generate revenue directly or on the periphery of the “China market.” So a country Google suggested change cannot kick the habit of thinking about revenue from the world’s largest market.
  3. Google seems somewhat disconnected from the increased scrutiny individuals like Mr. Barr are giving the company with the great booth give away: A Googley mouse pad.

Net net: A different PR spin may be needed. Hint: “For all” may connote Google advertising.

Stephen E Arnold, July 20, 2020

Outvoxed: The Perils of New Age Publishing in Time of Rona

July 17, 2020

CNBC which continues to delight with “real news” published “Vox Media Preparing Round of Layoffs As Business Fails to Improve Amid Coronavirus Pandemic.”

DarkCyber’s reaction was, “How can this be? So hip, so with it, so confident with its flagship podcast. So very Silicon Valley.”

The write up reports:

Vox was 40% off its revenue forecast for the second quarter and plans to miss its full-year target by 25%

Yikes.

CNBC continued:

Vox furloughed about 100 employees in April, or 9% of its staff, until July 31 as Covid-19 affected advertising budgets. Many of the furloughed workers who haven’t already taken buyouts will be laid off, according to a person familiar with the matter. These employees primarily work for parts of Vox that were especially hit hard by the Pandemic, such as SBNation, Curbed and the company’s events group. There are likely to be additional job cuts, two people said.

One possible bright spot is the over talkers program billed as Pivot. Maybe the Pivot for fee educational series will raise the Vox in exaltation.

The Rona Era may inflict further unpleasantness on informed individuals. DarkCyber particularly enjoys the management suggestions Vox experts articulate.

Well, CNBC is reporting news. Vox just makes the news.

Stephen E Arnold, July 17, 2020

Palantir Technologies and Semi Hard Numbers

July 10, 2020

Palantir Technologies is super secretive. The company plans to become publicly held. Does secret and public match up for you?

Palantir Built Itself into a $20 Billion Success with a Secretive and Controversial business. Now It’s Prepping for Life As a Transparent Public Company” offers some numbers; for instance:

  • The company is valued at $20 billion US
  • The company’s technology is 17 years young
  • The company has raised about $3 billion US in funding
  • 200 employees sent a letter to top management complaining about Palantir’s work for US Customs.

None of these numbers indicate if the company is profitable.

Important? Probably not in today’s fraught economic environment.

Stephen E Arnold, July 10, 2020

How Many Ads Can a YouTube Video Hold? Answer: Never Enough

July 10, 2020

We spotted a HackerNews post wondering if the YouTube (free version) was getting more ad love from the merrie band of Googlers.

The answer is, “Absolutely.”

The Google bean counters are well aware of the cost of the “free” video service. Thus, the free video service has to generate cash and more cash so the system can produce infinite cash. That’s logical in a Googley way I think.

In the comments to the original question on HackerNews, an entity named Operyl wrote:

If I understand correctly from a friend, the problem is YouTubers (and YouTube/Google) are currently making _much less_ money per ad. It sounds like more are getting shoved per video to make up for it (iirc, it’s up to YouTube to determine this?).

I don’t know what iirc means, but the rest of the post is clear. More money is needed.

Observations:

  • YouTube ads are more and more annoying. The fix obviously is to pay Google money. Most of the annoying ads go away. Google is discovering subscriptions. Undoubtedly Google will think subscription revenues for other services just like BMW and its heated steering wheel stroke of genius. German logic, of course. Ever read Kant? Congruent indeed.
  • The YouTube ads are increasingly irrelevant when I check out some YouTube videos. I love the tours of the Incan ruins. Ads about all sorts of things unrelated to Peruvian stone work appear. Therefore, the famous smart algorithm is just spewing ads to burn up inventory is one thought which crossed my mind.
  • The autoplay of post viewing content are interesting as well. How many of those ads are viewed BEFORE the YouTube user identifies which tab is playing the pitch to go Adobe? My hunch. Zero if these startled views are like me.

Net net: Those grousing about Google’s monetization quest have not seen anything yet. Why? The cost hole for the Google is probably close to infinite as long as there are former TikTok users looking for a home. Infinite costs can only be offset by infinite revenue. That too is logic worthy of a Google flashing logo pin.

Stephen E Arnold, July 9, 2020

Subscriptions: Spreadsheet Fever Fuels the Magazine Model

July 9, 2020

Nothing is easier. Plug in a series of four numbers, highlight the cells, and drag the little black box. Excel spits out the “projected next number.” Magic.

Think about this. Mail out 10 million snail mail pitches for a year’s subscription to a jazzy magazine, maybe Psychology Today or something similar. Fire up the spreadsheet, plug in the estimated number of sign ups, and project how much money will flow into the coffers of the magazine publisher or the third party handling the campaign from an office in Hoboken.

Subscriptions are the “next big thing” for many businesses. Here in rural Kentucky, our single car wash sells a “subscription.” The idea is that the car wash gets upfront money, and the lucky buyer can drive in one every two weeks and get the horse and buggy hosed down. Working good? Not so much.

BMW is selling subscriptions to features like heated steering wheels. Tesla, the auto company owned by Joe Rogan star, Elon Musk has subscriptions on its radar too.

Twitter, according to Bloomberg, the socially positive and continually uplifting information service, may be going to a subscription model. The DarkCyber research team has long considered Twitter a very useful tool for misinformation, disinformation, and reformation. Asking “fake personas” to pay for the service may work. On the other hand, industrious individuals may find the steady stream of innovations in encrypted messaging apps a possible complement. But look at those Excel projections. Imagine a 1,000,000 subscribers at $10 US a month. Wow, drag those tiny black squares. Count your bonus now.

The Quibi short form video service is subscription based. No one on the DarkCyber team has downloaded the app nor peered over someone’s shoulder while social distancing outside the general store in our small town. (It is near the vacant subscription car wash.)

According to a possibly specious, wildly incorrect, and statistically flawed report, Quibi’s subscription model is not selling like Rona N95 masks. The rock solid “real” news outfit Verge published “Quibi Reportedly Lost 90 Percent of Early Users after Their Free Trials Expired.”

The marketing technique implemented get six issues free and then pay only $10 US a month approach. How are magazines doing these days? Yep, stunning business.

The write up recycles data from a “research firm” named Sensor Tower and reports:

Streaming service Quibi only managed to convert a little under 10 percent of its early wave of users into paying subscribers, says mobile analytics firm Sensor Tower. According to the firm’s new report on Quibi’s early growth, the short-form video platform signed up about 910,000 users in its first few days back in April. Of those users, only about 72,000 stuck around after the three-month free trial, indicating the app had about an 8 percent conversion rate.

Short form video content is available mostly for free. Ever hear of Funimate?

Let’s step back. Advertising online is a monopoly game with two outstanding firms managing the dice, the money, and the cute little tokens. Direct mail is more expensive. With creative, list rentals, and fulfillment house fees, figure $5 to $7 per envelope delivered by snail mail. The promo can be cheaper if you go with a single “please, subscribe” flier in a ValPak envelope. Inserts in a daily newspaper. Okay, that’s a great idea. Door knob hanging? Nope. Banner ads on the Adf.ly network. Yeah, maybe?

Subscription plays are looking good when viewed through the blood shot eyes of someone with spreadsheet fever.

Reality may be different. Even National Geographic is a non profit. Hey, there’s an idea for BMW, Twitter, and Quibi. When this bout of spreadsheet fever winds down, consider the benefits of becoming a non governmental organization: Donations, fund raisers, merchandise, and more.

Stephen E Arnold, July 9, 2020

Will Insurance Companies Tie Rates to Rage?

July 7, 2020

The community-driven navigation app Waze, owned by Google, has refreshed its design. The company changed up the color scheme, logos, icons, and typeface—the sort of tweaks one would expect to keep users engaged. One particular change, however, is more intriguing. Engadget reveals, “Waze Lets Drivers Display their Moods in the App.” That could prove to be very useful information for some advertisers, individuals, and government entities. Writer Christine Fisher reports:

“Waze is also adding something called Moods, a feature that will ‘capture users’ emotions.’ ‘Celebrating the passion and authenticity of its users, Waze hopes that the update will harness the “humanness” that can often be lost within inhumane traffic conditions,’ the company wrote in a press release. It’s unclear if Moods will be shared with nearby Waze users. Letting other drivers know how you feel doesn’t necessarily sound like a great idea, but for the most part the Mood icons look too cute to induce serious road rage. ‘Hopefully our new look reminds users of the magic of our community and the way we work together for better,’ said Jake Shaw, head of creative at Waze.”

The icons are indeed very cute, we’ll give them that, and touting the “magic of community” sounds delightful. But giving away even more personal data seems like a bad idea to those of us who understand how various entities can use seemingly benign personal details. Founded in 2007, Waze is based in the San Francisco Bay area. Google bought the company for $966 million in 2013.

Cynthia Murrell, July 7, 2020

The Cost of Training Smart Software: Is It Rising or Falling?

July 6, 2020

I read “The Cost of AI Training is Improving at 50x the Speed of Moore’s Law: Why It’s Still Early Days for AI.” The article’s main point is that “training” — that is, the cost of making machine learning smart — is declining.

That seems to make sense. First, there are cloud services. Some of these are cheaper than others, but, in general, relying on cloud compute eliminates the capital costs and the “ramp up” costs for creating one’s own infrastructure to train machine learning systems.

Second, use of a machine learning “utility” like Amazon AWS Sagemaker or the similar services available from IBM and Google provides two economic benefits:

  1. Tools are available to reduce engineering lift off and launch time
  2. Components like Sagemaker’s off-the-shelf data bundles eliminate the often-tedious process of finding additional data to use for training.

Third, assumptions about smart software’s efficacy appear to support generalizations about the training, use, and deployment of smart software.

I want to =note that there are some research groups who believe that software can learn by itself. If my memory is working this morning, I think the jazzy way to state is “sui generis.” Turn the system on, let it operate, and it learns by processing. For smart software, the crude parallel is learning the way humans learn: What’s in the environment becomes the raw material for learning.

The article correctly points out that the number of training models has increased. That is indeed accurate. A model is a numerical recipe set up to produce an output that meets the modeler’s goal. Thus, training a model involves providing data to the numerical recipe, observing the outputs, and then making adjustments. These “tweaks” can be simple and easy; for example, changing a threshold governing a decision. More complex fixes include, but are not limited to, selecting a different sequence for the individual processes, concatenating models so that multiple outputs inform a decision, and substituting one mathematical component for another. To get a sense of the range of components available to a modeler, a quick look at Algorithms. This collection is what I would call “ready to run.”

The article includes a number of charts. Each of these presents data supporting the argument that it is getting less costly to training smart software.

I am not certain I agree, although the charts seem to support the argument.

I want to point out that there are some additional costs to consider. A few of these can be “deal breakers” for financial and technical reasons.

Here’s my list of smart software costs. As far as I know, none of these has been the subject of an analyst’s examination and some may be unquantified because those in the business of smart software are not set up to capture them:

  1. Retraining. Anyone with experience with models knows that retraining is required. There are numerous reasons, but retraining is often more expensive than the first set of training activities.
  2. Gathering current or more on point training data. The assumption about training data is that it is useful. We live in the era of so called big data. Unfortunately on point data relevant to the retraining task is a time consuming and can be a complicated task involving subject matter experts.
  3. Data normalization. There is a perception that if data are digital, those data can be provided “as is” to a content processing system. That is not entirely accurate. The normalization processes can easily consume as much as 60 percent of available subject matter expert and data analysts’ time.
  4. Data validation. The era of big data makes possible this generalization, “The volume of data will smooth out any anomalies.” Maybe, but in my experience, the “anomalies” — if not addressed — can easily skew one of the ingredients in the numerical recipe so that the outputs are not reliable. The output may “look” like it is accurate. In real life, the output is not what’s desired. I would refer the reader to the stories about Detroit’s facial recognition system which is incorrect 96 percent of the time. For reference, see this Ars Technica article.
  5. Downstream costs. Let’s use the Detroit police facial recognition system to illustrate this cost. Answer this question, please, “What are the fully loaded costs for the consequences of the misidentification of a US citizen?”

In my view, taking a narrow look at the costs of training smart software is not in the interests of the analyst who benefits from handling investors’ money. Nor are the companies involved in smart software eager to monitor the direct and indirect costs associated with training the models. Finally, it is in no one’s interest to consider the downstream costs of a system which may generate inaccurate outputs.

Net net: In today’s economic environment, ignoring the broader cost picture is a distortion of what it takes to train and retrain smart software.

Stephen E Arnold, July 6, 2020

« Previous PageNext Page »

  • Archives

  • Recent Posts

  • Meta