It has become common for people to post selfies, and photos of friends and family, online. Professional photographers who use models may post their photos in an online portfolio. Unfortunately, photos that include people’s faces are being used without permission by researchers who want to create facial recognition algorithms.
NBC News reported that, in January of 2019, IBM released a collection of nearly a million photos that were taken from Flickr and coded those photos to describe the subject’s appearance. According to NBC News, IBM promoted the collection to researchers as a progressive step toward reducing bias in facial recognition.
I personally feel that there are a lot of ethical problems with what IBM has done. The most obvious one is that it didn’t ask the photographers if it could use their photos.
A company as large as IBM has the money to pay photographers for the use of their photos. Stealing other people’s art is wrong. IBM is also big enough to hire a few people to get consent forms from the people who are in the photographs.
Another ethical problem is that facial recognition software is controversial. It evokes a “Big Brother is watching you” kind of feeling. Personally, I would feel disgusted if my face was used to train facial recognition software.
In July of 2018, the ACLU tested Amazon’s facial recognition tool (called “Rekognition”). It incorrectly matched 28 members of Congress, identifying them as other people who have been arrested for a crime. False matches could result in police arresting the wrong person.
NBC News reported that IBM said Flickr users can opt out of the database. However, NBC News discovered that it’s almost impossible to get photos removed.
Now would be a good time to make your Flickr and Instagram accounts private. Don’t let grabby companies steal your photos and use them in an ethically questionable algorithm.
During my recent annual Christmas vacation, I decided to finally tackle my office room closet. That closet had become stacked pretty much from floor to ceiling over the past 16 or so years with obsolete computer junk.
Most of the stuff was completely out-of-date. Lots of extra stuff that came packaged with long-gone computers. Manuals describing long-dead software procedures. Old USB and even Serial Port peripherals that do not work with modern operating systems. Lots of stuff that slowly lost any value it had as computer technology advanced over the years.
However, mixed in with what turned out to be about 15 large garbage bags of old computer junk, I found a gem that I forgot that I ever had – a classic “Model ‘M’” IBM PS/2 keyboard.
I had sort of forgotten how good classic IBM keyboards were to type on. In the intervening years of newer USB and later wireless keyboards, I began to think that my typing skills somehow weren’t what they once were. I just assumed that my typing skills had taken a sharp nosedive.
The IBM “Model M” is the holy grail of keyboards. Each Model M key uses what is called a “buckling spring” construction that gives every IBM Model M key its distinctive key travel and feel. The sculpted keys of the Model M eliminates the common mistake of hitting two keys at once on modern, style over substance keyboard designs. Modern keyboards often have very little key travel, and a very mushy feel with the keys jammed together. The Model M keyboard offers the best typing experience EVER, bar none.
So, I went on Amazon and bought an inexpensive PS/2 keyboard/mouse to USB “Y” adapter. I plugged the IBM keyboard into one of the USB ports into my Samsung DEX dock that I use with my Galaxy Note 8 smartphone along with a 24 inch curved 1080p Samsung monitor. The combination couldn’t work better. I’m typing this into Word right now using the Model M with my phone plugged into the DEX phone dock which turns the Note 8 into a real desktop experience.
I started researching the IBM Model M keyboards and discovered there is a very active market for them. IBM and later Lexmark (an IBM subsidiary) manufactured various versions of the Model M keyboard from sometime in 1984 until 1996. The employees of the IBM/Lexmark factory bought the patents and the factory equipment and continue to this day to manufacture versions of the Model M keyboard under the company name “Unicomp” using the same “buckling spring” key design. They offer both Mac and Windows versions of their keyboards, along with the original 101 classic IBM keyboard layout. Unicomp keyboards are priced from around $85 dollars to $105 dollars, depending on customizations.
No other keyboard offers the typing accuracy or sheer satisfaction of the original “buckling spring” key design. After finding out that Unicomp still sells new Model M keyboards, I ordered two of them for a couple of my computers.
Technology is improving in ways that not only allow for IOT, but also makes it easier for us to talk to our devices and appliances. IBM’s focus for the future is the “Cognitive Era” – a term they’ve coined and a market that they are going after. The overall goal is to improve the natural language understanding in IOT products.
Don and Scott spoke with Aylee H. Nielsen, who is a Social Media Execution Strategist for IBM. She talked about some of the partnerships that IBM has with different companies, including Under Armour, SoftBank, and Cognitoys. Some of the new products from those companies are IBM Watson powered devices.
Support my CES 2020 Sponsor:
$0.01 GoDaddy coupon for a New domain name cjcfs3geek
$1.99 / mo Economy Hosting with a free domain. Promo Code: cjcgeek1h
$2.99 / mo Managed WordPress Hosting with free Domain. Promo Code: cjcgeek1w
Support the show by becoming a Geek News Central Insider
Do you have video editing skills? There is cash on the table for a winning entries. Plus I have a couple of fun prizes to give away in the form of some Roku’s listen to get the details. Your really going to enjoy tonight’s show as the content flow was close to perfect.
Note: I am hiring 5 writers email me firstname.lastname@example.org
Why Would IBM want to sell their patents? And to Google, nonetheless?
Think of it this way – You’re at a garage sale and you see a box of comic books (or records if you are not a comic nut). You buy the box for $20, hoping there is a valuable comic (record) in there. You pull out the important ones and what is left is a box of comics you don’t care about.
IBM has been gobbling up different companies throughout the years and some of the patents are like the odd comic books. Nortel is a great example – IBM had over 6,000 patents that they didn’t need. Therefore in May, they decided to auction off those patents that didn’t pertain to them. Google didn’t win that bidding war.
But , according to SEO by the Sea, IBM last week did find 1,030 patents that they sold to Google for an undisclosed sum. It was a hodge-podge of patents – from fabrication to database structures. These are patents that could keep Google from going to the courts for their Android devices, new products coming out on the market and other threats to future revenue.
It’s also a case of Google picking out the ones they need, then keeping a couple in the back pocket for future need (whether for selling or future projects). A couple of those patents relate to search methods. It will be interesting how that effects other search engines like rival Bing.
We’ll have to wait and see how Google utilizes these patents.
One of the problems with watching video podcasts as an alternative to conventional television is that you have typically and deliberately watch one video at a time. On longer videos it’s not as much of a problem, but with short videos that last 5 minutes or less you have to keep manually restarting the next video after the previous one has finished.
I now have three Mac Minis – one is an old somewhat underpowered Power PC Mac Mini that I’m using as a video podcast aggregator. I have that machine’s iTunes database located on a much larger shared drive that’s available to every machine on my home network. I’m subscribed to a variety of tech podcasts, most of them in the highest resolution file sizes available.
I have two other Mac Minis that are of the latest design. I have an “Eye TV” USB HD tuner connected to one that’s connected to a substantial external antenna. Depending on atmospheric conditions I can receive up to 18 channels counting the various digital sub channels. This enables the Mac Mini to function as a DVR.
The second Intel Mac Mini is in another room and the Eye TV software also loaded on it is able to work from the other Mac Mini’s shared recordings.
Today I discovered by accident when playing around with iTunes on one of the Intel Mac Minis that the shared videos show up in the shared playlists from other iTunes databases. So, in other words, I can pick a shared iTunes list from the Power PC Mac Mini’s shared iTunes and a list of video files shows up. Since the videos are in the list just like audio would be, I am able to start a video file playing and when one file ends it will immediately start playing the next video file on the list. This is particularly useful because I can start videos playing as I do other things and it will continue to play just as if it’s a TV station. This is quite a handy capability to have. The lack of an ability to set up continuous video playback has long been one of the Apple TV’s biggest shortcomings.
Periodically I go to the Power PC Mac Mini and delete the video files that have been played, since iTunes keeps a play count, so I always have fresh material to watch.
If you aren’t in the Eastern time zone or you recorded tonight’s episode of Jeopardy! for later viewing then you may want to stop reading now. But if you have seen the show or don’t plan to watch it or don’t know what the heck I’m talking about then read on.
Day three, the final day, has just wrapped up and the final score was not particularly close, although round 2 proved to be a bit more exciting. As for final tallies, Watson racked up $77,147 to Ken Jennings’ $24,000 and Brad Rutter’s $21,600. However, during tonight’s round 2, Jennings took a brief lead over the IBM supercomputer.
We did learn a few things along the way. One was that Watson’s algorithm for wagering money is incomprehensible. It’s bets were always amounts that seemed to make little sense based on it’s total or the totals of its opponents. Another was that it did not receive information on it’s opponents answers – at one point, after Jennings buzzed in first and gave an incorrect answer, Watson buzzed in and gave the same answer. We learned that Watson thinks Toronto is a US city. We even learned that Ken Jennings has a sense of humor. In round 2, after hitting the daily double, he joked that his two choices were to either “unplug Watson or bet everything”. During Final Jeopardy of round 2 he wrote his answer as “Bram Stoker (I for one welcome our new computer overlords)”
Watson, made by IBM who also created Deep Blue, is powered by 90 32-core IBM Power 750 Express servers with a total of 16 terabytes of memory and can calculate hundreds algorithms simultaneously. Or, in short, it was an amazing display of technology and a lot of fun to watch. This was a Geek’s dream TV show.