[Glasslinger] Builds Tiny Tubes

In the early days of transistors, RCA and GE were battling against silicon with ever smaller vacuum tubes. These tubes – Nuvistors, Compactrons, and some extremely small JAN triodes were some of the tiniest tubes to ever be created. [glasslinger], YouTube’s expert on DIY valves, is pretty close to beating the tiniest tubes that were ever manufactured. He’s created a miniature diode and triode that are about 1/4″ in diameter and 1″ long.

The most difficult part of making a vacuum tube is getting a perfect glass seal around the pins. For this, [glasslinger] is using very fine tungsten wire and glass beads. A bead is placed around each wire, mounted in a stand, and melted together with a torch.

A diode is simple as far as tubes go, requiring only a filament between two pins. [glasslinger] is just stringing a fine piece of wire between two pins and welding them on with a miniature spot welder. After that, it’s just an issue of melting a 1/4″ glass tube to the base of the tube, putting it under vacuum overnight, and sealing it shut.

Filed under: classic hacks

Build Your Own Raytracing Minion

A canceled project left [Craig] with six Raspberry Pi based devices he calls “Minions”. A minion is a Raspberry Pi model A in a small enclosure with an Adafruit 2.2″ 320×240 SPI LCD. The LCD lives in a lollipop style circular housing above the base. [Craig] has found a use for one of his minions as a desktop raytracer.

The Raspberry Pi is quite capable of running Persistance Of Vision Raytracer, or POV-Ray. POV-Ray started life as an early PC based raytracer. Created as a port of an Amiga program called DKBTrace, which was itself a port of a Unix raytracer, POV-Ray first was released in 1987. For the uninitiated, raytracers like POV-Ray  literally trace rays from a light source to an image plane. As one would imagine, the Raspberry Pi’s little ARM processor would take quite a bit of time to raytrace a high resolution image. However, when targeting a 320×240 LCD, it’s not half bad.

[Craig’s] minion is running his own software which he calls ArtRays. Based upon a setup file, ArtRays can render images from several sources, including the internet via a WiFi dongle, or a local SD card. Rather than walk through the setup and software install, [Craig] has provided a link to download a full SD card image to build your own Minion. It might be worth experimenting on your own first though, rather than killing his server with a 1GB download.

We’re glad [Craig] has found use for one of his minions, now we have to see what he’s done with the other five!

Filed under: classic hacks, Raspberry Pi

Locking A Beer With A 3D Printer

Have a nice, refreshing IPA sitting in the fridge along with a ton of other beers that have ‘Light’ or ‘Ice’ in their name? Obviously one variety is for guests and the other is for hosts, but how do you make sure the drunkards at your house tell the difference? A beer bottle lock, of course.

Because all beer bottles are pretty much a standard size, [Jon-A-Tron] was able to create a small 3D printed device that fit over the bottle cap. The two pieces are held together with a 4-40 hex screw, and the actual lock comes from a six-pack of luggage padlocks found at the hardware store.

It’s a great device to keep the slackers away from the good stuff, and also adds a neat challenge to anyone that’s cool enough to know basic lock picking. Of course, anyone with a TSA master key can also open the beer lock, but if you’re hosting a party with guest who frequently carry master keys around with them, you’re probably having too good of a time to care.

Filed under: 3d Printer hacks

comScore: Ecommerce spending on Thanksgiving tops $1B for the first time, passes $1.5B on Black Friday

cash money

Reports have been coming in all weekend about how Thanksgiving and Black Friday online sales are up this year, in double-digit percentages yet again, but now the first raw number estimates are finally here. comScore has announced that Thanksgiving Day this year passed the $1 billion mark for the first time, and Black Friday managed to surpass the $1.5 billion figure.

More specifically, Thanksgiving Day saw a 32 percent gain, to $1.01 billion in spending. In addition to breaking the $1 billion record (which Black Friday first achieved in 2012), it’s also worth noting that this Thanksgiving (November 27) marked the first day of the 2014 season to reach such a level of spending. Naturally,
the record was immediately broken the following day.

Black Friday (November 28) spending was up 26 percent over the same day in 2013. This pushed online sales for the 24-hour period to $1.51 billion.


Adding the two days together, the combined spending in the U.S. was $1.96 billion in 2013. This year, that number grew to $2.51 billion.

comScore naturally suggests this is good news for the upcoming holiday season. If Thanksgiving and Black Friday broke new online sales records, there’s no reason to believe Cyber Monday (tomorrow) and other days closer to
Christmas can’t do the same.

“Thanksgiving and Black Friday both saw exceptionally strong online growth rates as each day surpassed $1 billion in desktop spending,” said comScore chairman emeritus Gian Fulgoni. “The strength we saw in the early online buying rush likely reflects a few things, including overall health in consumer spending, responsiveness to the strong deals being offered online, and perhaps some shoppers opting to stay home on Thanksgiving rather than head out to the stores that opened their doors early.”

It’s worth noting that these new records were achieved despite reports of outages and slowdowns. One firm found desktop and mobile ecommerce pages were 20 percent and 57 percent slower (respectively) this year, compared to 2013.

Why Eric Schmidt doesn’t know how Google works

Google chief Eric Schmidt.

Google is a great company — from its record $23 billion IPO in 2004 (at a time when Nasdaq was below 2000!) to its dominant mobile OS. My associates and I have such an admiration for Google that we built our latest web conferencing platform with Google’s technologies.

How Google WorksBut we also know that Eric Schmidt, together with other key executives at Google
have contributed to this success through a stable management since the early days. As a fan of his, I’ve followed most articles, interviews, and slides about his latest book How Google Works and eventually read it. It’s an interesting book that reads like a story. I highly recommend it, but not as a management book the way the authors seem to position it, because it may take technology startups in the wrong direction. In fact, I think Schmidt and Rosenberg are totally wrong about how Google works. Here’s why:

First off, the authors are confusing causation and correlation. Schmidt points out a series of characteristics of Google as a company and presents them as the reasons for Google’s success, but in my opinion, they are all consequences of Google’s success.

For example the authors write: “Their plan for creating that great search engine, and all the other great services was equally simple: Hire as many talented software
engineers as possible, and give them freedom.” Well, this worked because the search was already successful enough to fund that freedom. I would love to see one single company that isn’t dominating a market with no cash cow in-flow that can succeed without strict discipline, sharp focus, hard work, and hands-on management.

If this management style is the reason for Google’s success, then why have the majority of initiatives at Google either failed or been financially inefficient and unprofitable? If they were standalone startups, they would have most likely already been dead.

Another special characteristic of Google is its sales force. When interacting with sales people at Google, I am shocked to see how untrained and inefficient they are. If there are known companies with great sales cultures such as Oracle, or some older models such as IBM or Xerox, that could sell literally any product in their portfolio, Google is that other extreme, where sales reps have
exceptional products in their baskets but are unable to sell them. The truth about Google is, their products don’t sell, they get bought!

Schmidt and Roseberg know that 90 percent of Google’s revenue comes from advertising (thanks to Google’s de facto monopoly in search) . According to the company’s 2013 financial filings, 83 percent of Google’s revenue came from ads, about 7 percent from Motorola (which is now gone), and 10 percent from everything else. In other words, when you add up all the revenue from Google Apps (Gmail, Docs, Drive, Maps, etc.) together with the Android and other mobile businesses, and then add Chromebooks, Chromecast, Chromeboxes, and everything hardware and everything Chrome, Google Developers Network, Google+, Google cars, Google robots and drones, Google Glass and other wearables, Google Cloud, and everything else in the Google world, you get $5 billion or 10 percent of Google’s revenue. Peanuts! Google does not disclose the split of
cost among different business units, but it’s not hard to imagine their level of profitability if they were independent entities.

how does google stack up

Above: A comparison of eight tech leaders. All financial data are extracted from companies’ public statements. Employee appreciation data includes data extracted from Glassdoor public ratings. “Love Indexes” are created by the author as a way to quantify and compare numbers from different viewpoints. High margin = software, subscription, ads, everything that is neither hardware nor requires labor to bill (support, consulting, etc.).


Although Google is perceived as
heaven for its employees, the viewpoint from shareholders is a bit different. I am not saying Google’s shareholders are unhappy, but the company’s use of its profit and capital is not optimized.

Let’s compare the efficiency of Google with IBM. IBM may no longer be considered a hot company, but you may be surprised by the comparison of financial and management performance: IBM’s business is only 26 percent software (or revenue composed of high margin dollars, i.e. dollars that don’t require labor delivery or support or manufacturing); for the rest, the company has to deliver man/days for dollars. Google;s business is 83 percent high margin (90 percent in 2014). That’s why IBM’s business is by nature much less profitable: $38,000 of income per employee compared to Google’s $256,000k per employee. Google’s business is seven times more profitable. Yet the Price per Earning of Google is only three times IBM’s. Google’s P/E ratio is comparatively very low;
it’s four times less than Facebook and half of Microsoft’s in the ’90s. Not to mention Amazon, which, despite being hardly profitable, has a market cap of over $140 billion.

Google’s financial performance is not on par with the level of innovation and the “smartness of its creatives.” Worse, Google’s stock value is not on par with its financial performance. Maybe investors are not comfortable with the long-term results. Is it because of too much waste, technology explorations, and incoherent diversifications? Maybe.

Amazon and IBM use their funds more efficiently and are more loved by investors, although, granted, they’re less popular among their employees.

So what makes the difference? How does Google work?


The key is market dominance. If you have a de facto position of a monopoly in your market, money pours in, and you can afford to give your employees even more than 20 percent of their time

Google is in a situation of monopoly with its search business. In the above table, the reason Facebook and Amazon have high P/E is that investors consider them to be either in or near a monopoly position, while IBM, and to a lesser extent Oracle, are competing and have to work hard to keep their leadership positions. (Amazon’s monopoly in book publishing is already achieved, but Amazon has real challenges in other markets).

Moreover, Google’s monopoly is in a very profitable, high-margin business: fully automated ads with minimal cost of sales.

Last month, during a course at Stanford, Peter Thiel, PayPal’s founder, advised entrepreneurs to seek monopolies, stating that “competition is for losers.” As long as Google has its monopoly in search, Googlers can enjoy free food, don’t need to worry about sales skills, and can take 20 percent time off or more. But when that changes, the company needs to be ready for the shock.

In conclusion,
each company needs to find a culture, methods, and processes that work for its individual circumstances. There are rarely one-size-fits-all concepts, and those that do exist are the good old ones: Innovation & Quality, Customer Satisfaction, Employees Development & Education, and Financial Performance. Everything else is just context based.

[The opinions expressed in this story are solely my own and do not express the views or opinions of my company.]

Darius Lahoutifard is an Enterprise SaaS Executive and Enterpreneur, the founder of Business Hangouts, an Enterprise App for Google’s Hangouts on Air. You can follow him on LinkedIn here or follow his company on Twitter here.

We’re studying conversion rate optimization. Take our quick survey and we’ll share the results with you.

Uber now complies with India’s two-factor authenticatio


Uber today announced that it has complied with the Reserve Bank of India’s requirement that every transaction made with an Indian credit card use two-factor authentication (2FA). Yet the company has done so grudgingly: While 2FA is widely seen as a security measure, Uber says it has a negative impact on businesses.

For context, two-factor verification requires you to use more than one form of verification to access an account. Typically, this information includes “knowing something” such as a password and “having something” such as a mobile device. In the case of India’s 2FA requirement, a one-time code is sent to your phone, which you then need to punch into the merchant’s payment

In its announcement, Uber calls the requirement in India an “antiquated solution that is cumbersome for consumers and stifling for businesses across India” as well as “unnecessary” and “burdensome.” While the first two claims are frankly ridiculous (2FA is hardly unnecessary, and it is definitely not antiquated), the company’s last complaint is understandable.

The company unhappily notes that the rule applies to all transactions, “no matter how small the amount,” pointing out that even short cab ride in the country can’t be as quick and efficient as everywhere else Uber offers its services. Yet India’s population is well over 1.2 billion, so even small fraudulent transactions can quickly add up.

Uber also argues that the requirement is “causing a major challenge for businesses trying to offer Indian consumers a better purchasing experience.” The company further notes that
consumers prefer the old system, but “India’s one-of-a-kind 2FA requirement persists” despite the claim that it runs counter to “the face of rapidly changing business expectations.”

The good news is hidden in all these complaints. Uber does note that it has seen “a slow conversion” of riders from credit cards to a 2FA-compliant wallet, and that it is engaged in constructive discussions with the RBI. The company further adds it is talking with the Indian government “to advance regulations that support innovation and job creation.”

Uber also makes a request:

In the meantime, we would welcome an additional 45-day extension from the RBI that would give the majority of our existing riders sufficient time to transition over to a new 2FA-compliant payment system. This additional grace period not only prevents consumers from being disadvantaged but it also protects partner-drivers who rely on Uber’s
riders as their sole source of income.

This seems reasonable to us. We applaud India for pushing such a requirement to boost security, but if a company says its business will be affected less if it can delay the switch by six weeks, that seems like a valid appeal.

Uber is probably particularly annoyed because it was forced to make the change: If it didn’t, the RBI would have shut it down. Furthermore, the ride-sharing service also now has one fewer differentiator when compared to local taxi companies like Ola and Meru.

Since your credit card details are saved in Uber’s app, you don’t have to fiddle around with cash when the car pulls up at your destination, meaning you can just jump out and go. This works the same way everywhere Uber operates — except in India now, where you have to go about the additional step of putting in the one-time code.

Black Friday 2014: Ecommerce desktop pages were 20% slower than in 2013, mobile pages were 57% slower

Online shopping safety tips

Black Friday news kicked off early this weekend when the Best Buy website was hit with a massive outage, but it turns out that was only half the story. The top 50 ecommerce websites were slower overall this year compared to last, suggesting customers were frustrated even if they could get to their favorite shopping site.

Web performance monitoring company Catchpoint Systems looked at aggregate performance this weekend and compared it to the same timeframe in 2013. The results are notable: Desktop webpages were 19.85 percent slower, while mobile webpages were a whopping 57.21 percent slower.

More specifically, the firm made the following findings:

  • Median webpage response times for desktop websites for the entire group (aggregate) was 3.991 seconds, compared to 3.330 seconds in 2013.
  • Median webpage response times for mobile websites for the entire group (aggregate) was 2.954 seconds, compared to 1.879 seconds in 2013.

The data is for all of Black Friday (November 28, 2014) until 10:00 a.m. EST on Sunday, November 30. The comparison was made to the same time period in 2013: November 29 to the morning of December 1.

Catchpoint hints that the main reason for the slowdown was not simply because there were more users overloading the systems. Most webpages in the ecommerce group monitored by the firm were also bigger; there was simply more data on the page to download.

The five fastest desktop sites, according to Catchpoint, were H&M, Costco, Apple, Barnes & Noble, and Etsy. None of these five were present on the top five list for fastest mobile sites: Sears, WW Grainger, Office
Depot, Ikea, and Saks Direct.

Best Buy actually went down more than once: twice on Friday, and also experienced issues on Thursday. Neiman Marcus was down for 2.5 hours on Saturday night, with poor availability leading up to the outage. Gamestop had poor availability all weekend, and J. Crew had slower-than-normal load times as well.

Catchpoint Systems conducts its performance tests on a customized list of 50 web and mobile sites identified as leading ecommerce retailers. The tests were run from 30 backbone nodes located in various U.S. cities, which the company says allows for noise and connection variables that occur in real-user environments to be factored out of these results.

We’ve asked for full list of the 50 ecommerce sites and will update this article when we hear back.

Update: The full list is now included below.






Watch a George Lucas parody of the ‘Star Wars: Episode


On November 28, the “Star Wars: Episode VII – The Force Awakens” trailer was released on iTunes and then YouTube. On November 29, the parodies started to trickle in.

Introducing the George Lucas Special Edition:

For the sake of comparison, here is the original:

Hit play at the same time, or with a few seconds’ delay for extra laughs. The CGI additions are just perfect.

If parodies tickle your fancy, you’ll also want to watch the Lego version. If that’s still not
enough for you, check out this YouTube search for some more.

“Star Wars: Episode VII – The Force Awakens” is being directed by J.J. Abrams and is scheduled for release in December 2015.

Transferring Audio to an AVR at 12kbps

Back in the bad ‘ol days of computing, hard drives cost as much as a car, and floppy drives were incredibly expensive. The solution to this data storage problem offered by all the manufacturers was simple – an audio cassette. It’s an elegant solution to a storage problem, and something that has applications today.

[Jari] was working on a wearable message badge with an 8-pin ATTiny. To get data onto this device, he looked at his options and couldn’t find anything good; USB needs two pins and the firmware takes up 1/4 of the Flash, UART isn’t available on every computer, and Bluetooth and WiFi are expensive and complicated. This left using audio to send digital data as the simplest solution.

[Jari] went through a ton of Wikipedia articles to figure out the best modulation scheme for transferring data with audio. What he came up with is very simple: just a square wave that’s changed by turning a pin off and on. When the audio is three samples long without crossing zero, the data is 0. When it’s five samples long without crossing zero, the data is 1. There’s a 17-sample long sync pulse, and with a small circuit that acts as a zero crossing detector, [Jari] had a simple circuit that would transfer data easily and cheaply.

All the code for this extremely cheap modem is available on GitHub.

Filed under: ATtiny Hacks

Mixer for Korg Volcas

Korg, everyone’s third or fourth favorite synth company, but one of the only ones that still in business, recently put out a new line of synths, drum machines, and groove boxes. They’re called Volcas, and they’re cheap, analog, and very cool. [Jason] has a few of these Volcas, and while he enjoys the small form factor, using an off-the-shelf mixer to dump send the audio from these machines to his computer takes up too much space. He created a passive mini mixer to replace his much larger Mackie unit.

The circuit for this tiny passive mixer is an exercise in simplicity, consisting of just a few jacks, pots, and resistors. [Jason] overbuilt this; even though the Volcas only have mono out, he wired the entire mixer up for stereo.

The enclosure – something that looks to be a standard Hammond die cast aluminum enclosure – was drilled out, and a lovely laser cut acrylic laminate placed on top. It looks great, and for anyone interested in learning soldering, you couldn’t come up with a better first project.

Filed under: musical hacks