Big, irrelevant data
Delivered on 10 September 2019 by Justin Pyvis. About a 4 min read.
For those out of the loop, World Health Organisation goodwill ambassador and murderous dictator Robert Mugabe passed away last week at the ripe old age of 95. Apparently he will be granted national hero status and an official mourning. It would be funny if it weren't so sad; Lord Acton really was spot on when he opined that "Great men are almost always bad men", a description which fits Mugabe to a tee.
Moving on, this week I want to touch on Palantir Technologies, named after the "seeing stones" in The Lord of the Rings. The original "big data" company, Palantir last week announced that it may delay its IPO until 2023:
"Palantir, which has never been profitable, had reportedly been making efforts to gear up a public offering, which included hiring its first sales team and eliminating lavish employee perks, like 13-course tasting-menu lunches, that could raise eyebrows among public market investors. Employees were also reportedly no longer allowed to expense last-minute international airfare and at least one person was fired for expensing off-the-wall purchases like lingerie."
Because of Palantir’s work with the US military and government agencies, the data-mining firm has long been subject to scrutiny by outside critics. Most recently, Business Insider learned that Palantir employees themselves were becoming increasingly split over the company’s ties to government work – specifically its business dealings with Immigration and Customs Enforcement, the federal agency involved in detaining and deporting people over immigration violations."
What does Palantir do? No one really knows, but the evidence seems to suggest a mix between talent arbitrage, i.e. consulting, and software development (mostly data analysis tools). Its major clients include governments, which were always going to be the primary buyers of "big data" services. Why? The lack of market feedback - knowledge transmitting prices, profit and loss - makes it hard to know which decisions might be best and how to justify them. Bureaucrats are also easily enticed by anything that might add "rigour" to their work, regardless of whether or not it adds real value.
For that reason, I'm sceptical of Palantir's lofty promises. While not being profitable doesn't (yet) matter in this day and age: Amazon wasn't profitable for nearly 6 years; Twitter, 4; and Uber and Lyft may never make a profit, there still needs to be some plausible future upside. It was clear what the likes of Uber and Lyft did and might dominate someday, for example autonomous vehicles, but Palantir? Not so much.
For me, Palantir is up there with Google's DeepMind in terms of its real world applications, i.e. not many. Provided it doesn't run out of money first, it will continue to win government contracts and work at organisations which resemble governments, such as large financial institutions, utilities and perhaps universities. But anyone else that dips their toes into the big data Kool-Aid will quickly realise how useless most of the output actually is and promptly walk away. Unless you need a veil of opacity with which to cloak your work, big data are not the answer.
Antitrust will be the weapon used against big tech
A poor argument rarely stops a lawyer, and especially not those spending other people's money to chase a big case. This is a political, not economic, assault.
- New York attorney general investigating Facebook for antitrust violations »
- Cold War Analogies are Warping Tech Policy »
- How Microsoft escaped the Big Tech backlash »
- Facebook, Google face off against a formidable new foe: State attorneys general »
The Facebook brand is toxic and my purely anecdotal evidence suggests it's mostly used by people no longer in the dating world (i.e. older people). Releasing it under a different brand name ala Instagram may have helped but probably not; the media would be on that in a heartbeat, so it probably had to use the Facebook moniker.
- Facebook’s Dating Service Is, Unfortunately, Made by Facebook »
- Now Facebook says it may remove Like counts »
Hong Kong follow-up
Hong Kong agreed to one of the protesters' 5 demands last week but that hasn't stopped them. They're also using an app unblockable by China (bluetooth-based, so require a "chain" of people all within a very close proximity of the next person) but apparently still can't work out how to use Twitter (likely journalistic bias).
Note that the app they're using is proprietary and closed-source so essentially insecure. A shame as there is an open source mesh alternative available, although it admittedly requires a bit more technical savvy to install (it's not on the respective Apple/Android app stores). As with their use of Telegram, the Hong Kong protesters are clever but not technically adept.
- Hong Kong leader Carrie Lam announces formal withdrawal of the extradition bill »
- Hong Kong Protestors Using Mesh Messaging App China Can't Block »
- Hong Kong’s fast-learning, dexterous protesters are stumped by Twitter »
Other bits of interest
- Binance Launches Dollar-Backed Crypto Stablecoin With NYDFS Blessing »
- When Apps Get Your Medical Data, Your Privacy May Go With It Image »
- Slack plunges after posting first earnings report since going public »
- Life in an Internet Shutdown: Crossing Borders for Email and Contraband SIM Cards »
- Samsung Is Secretly Working on a Foldable Phone That Collapses Into a Square »
This week's data breaches
Facebook leaked a bunch of users' phone number, but wants people to trust it with their d**k pics. The sad part is a lot of people, blissfully unaware of the Facebook's blasé attitude towards users' privacy, will do just that.
- A huge database of Facebook users’ phone numbers found online »
- China hacked Asian telcos to spy on Uighur travelers »
- How a High-School Dropout Hacked a Million Devices »
- US City Rejects Ransom Demand and Restores Files from Backup »
- Mental-health information 'sold to advertisers' »
- Hackers Breach Forum Of Popular Webcomic 'XKCD' »
- How MuleSoft patched a critical security flaw and avoided a disaster »
Issue 49: Big, irrelevant data was compiled by Justin Pyvis and delivered on 10 September 2019.