Add Photographs To The List Of Things That Can Now Be Faked Too Convincingly To Be Trustworthy

And the race to make sure that we’ll eventually not be able to trust any goddamn thing ever again continues. Thanks, Nvidia.

The woman in the photo seems familiar.
She looks like Jennifer Aniston, the “Friends” actress, or Selena Gomez, the child star turned pop singer. But not exactly.
She appears to be a celebrity, one of the beautiful people photographed outside a movie premiere or an awards show. And yet, you cannot quite place her.
That’s because she’s not real. She was created by a machine.
The image is one of the faux celebrity photos generated by software under development at Nvidia, the big-name computer chip maker that is investing heavily in research involving artificial intelligence.
At a lab in Finland, a small team of Nvidia researchers recently built a system that can analyze thousands of (real) celebrity snapshots, recognize common patterns, and create new images that look much the same — but are still a little different. The system can also generate realistic images of horses, buses, bicycles, plants and many other common objects.
The project is part of a vast and varied effort to build technology that can automatically generate convincing images — or alter existing images in equally convincing ways. The hope is that this technology can significantly accelerate and improve the creation of computer interfaces, games, movies and other media, eventually allowing software to create realistic imagery in moments rather than the hours — if not days — it can now take human developers.

Nvidia’s images can’t match the resolution of images produced by a top-of-the-line camera, but when viewed on even the largest smartphones, they are sharp, detailed, and, in many cases, remarkably convincing.

A second team of Nvidia researchers recently built a system that can automatically alter a street photo taken on a summer’s day so that it looks like a snowy winter scene. Researchers at the University of California, Berkeley, have designed another that learns to convert horses into zebras and Monets into Van Goghs. DeepMind, a London-based A.I. lab owned by Google, is exploring technology that can generate its own videos. And Adobe is fashioning similar machine learning techniques with an eye toward pushing them into products like Photoshop, its popular image design tool.

The technology behind this is all extremely cool and I don’t doubt for a moment that it’s going to be used extensively for its intended purpose. But like I’ve said before, I also don’t doubt for a moment that it’s eventually going to fall into the wrong hands. And when that happens, we’re all doomed. Even if a majority of us somehow simultaneously develop a keen ability to think critically and put partisan agendas aside in the name of truth, what is truth going to be when it’s so easily manipulated into whatever this or that bad actor wants it to be?

I don’t have a good answer to any of this, and I’m not saying that we shouldn’t be doing any of this research. But I certainly don’t think that the people who are doing this research are giving nearly enough thought to the implications of it, at least not publicly.

Mr. Lehtinen downplays the effect his research will have on the spread of misinformation online. But he does say that, as a time goes on, we may have to rethink the very nature of imagery. “We are approaching some fundamental questions,” he said.

What a completely irresponsible, kick the can down the road statement.

We aren’t approaching anything, sir. The moment you figured out that these things are possible, those questions arrived. And before the problem has a chance to get out of anyone’s control, it’s time to start thinking about answers.

The Further Adventures of Googoo and Odessa

Here’s another video demonstrating why Grandma would not like an Amazon Echo.

It doesn’t matter how many times I watch this video, it’s just as funny. I can’t help but crack up when she yells “Ok Goo goo!”

It’s fascinating to watch her hitting the speaker thinking that will make it wake up. It totally makes sense, it just doesn’t work that way. It’s the little things like that that I forget would be super weird for an older person to get used to.

I really thought it was funny when she ran away from the speaker when it told her the weather. I wonder if she ever got it to play her song.

And just for fun, here’s a funny Amazon Echo video, but not about seniors.

That would so be happening to my dad…if he could remember to make sure it was on.

Please Do Not Spend $80 On An App Powered Coffee Mug

I have a cup or two of tea or hot chocolate pretty much every morning. And because I am an easily distracted person, sometimes I forget that it’s sitting there and by the time I want a sip, it’s freezing and I have to heat it up. So the thought of a mug that would keep my drink at just the right temperature for hours at a time is kind of an appealing one. That’s where something like the Ember Ceramic Smart Mug might come in handy, were it not for literally everything about it.

  1. Right off the bat, the damn thing costs 80 dollars. If you’re willing to spend that much money on a single mug without batting an eye, you either have too much money or you’re the reason for our country’s horrible household debt statistics.
  2. You have to control it with an app. It’s morning. I’m tired. Or maybe I’m trying to plow my way through a project on a dreary afternoon or in the dead of night. Either way, I have other things on my mind, and this is a cup. Put a temperature dial on that shit.
  3. Speaking of the app, I am going to include this passage from the review in the hopes that all of you will spend the next few minutes seething in a manner similar to that in which I am right now.

    Most of the time the Bluetooth re-connected with no problems, but on a couple of occasions it refused to pair back up until I toggled my iPhone’s wireless off and on. Ember is meant to ping you with a notification when it has reached the preset temperature, too, though often that never actually arrived.

    There is not a single thing about those two sentences that doesn’t make me want to beat someone half to death with either the iPhone or the mug. But if I decide to, I’ll probably do it with the phone, because…

  4. The mug is fragile and small.

    The matte white ceramic can scratch easily if you’re aggressive with a metal spoon, too. Ember says you should hand-wash it, and you’ll want to do that with a fairly soft cloth or sponge in order to avoid scrapes.
    My biggest issue, however, is the capacity. Fill it to the brim and you’ll get around 10 oz in there; more reasonably, figure on a usable 8 oz. I probably drink too much coffee, but even so I found I was refilling the mug more frequently than with my regular, larger cups.

  5. And apparently the battery sucks too, which leaving aside the fact that it is a cup and as such should not require a battery, kind of defeats the whole purpose, no?

    You might have rolled your eyes when your parents told you to use a coaster, but figure on using the Ember charging coaster frequently if you want to keep drinking perpetually-hot beverages through the day. Ember’s roughly hour-long battery life is at the default 130-degree temperature. If you crank it up higher, expect to see that drop: at 145-degrees, figure on around 35 minutes of use. Ambient temperature in the room and how much liquid is left will have an impact, too.

Broken record time, but seriously, who in the blue hell is this for? What’s wrong with dropping a few bucks on a decent insulated travel mug or just getting up and trucking your ass and your regular glass mug to the microwave? You’re probably sitting too much anyway.

There Is Finally A Dictation Bridge 1.0

It’s been a while since Carin or I has mentioned Dictation Bridge, which in case you’ve forgotten is a free, open source piece of software designed to help those who need it much more easily use speech recognition in conjunction with a screen reader. The project hit its funding goal in 2016, and after nearly two years of development and testing, version 1.0 has finally been officially released for use with both JAWS and NVDA.

The following are key highlights of DictationBridge that the team who created it feel are most important:

  • DictationBridge works with both NVDA and Jaws, affording the DictationBridge team the privilege of serving 78.5% of the screen reading community with a dictation solution that can be freely learned from, freely distributed, and either used as-is or modified to meet the computing needs of a single end-user or a group of end-users. Prior to DictationBridge, NVDA users had no accessible way to control either Windows Speech Recognition or the Dragon line of products from Nuance and now they have access to both.
  • Furthermore, Jaws users had no accessible way to control Windows Speech Recognition, and they had only one solution for controlling Dragon. Now they have not only a solution for controling Windows Speech Recognition, but they also have a solution for Dragon which they can freely learn from, freely distribute, and either use as-is or modify to meet their needs.
  • DictationBridge does not change your screen reader setup in any way unrelated to dictation. If you install an NVDA or JAWS update, it will not effect the version of DictationBridge designed for your screen reader and DictationBridge will in turn not change any plug-ins or scripts, including the defaults when you install it. With DictationBridge, there’s no more waiting for your dictation support to catch up with the latest version of your screen reader and there’s no potential for DictationBridge to accidentally insert bugs into the default behavior of your Screen Reader of choice.
  • DictationBridge for both NVDA and JAWS provide all of the features a user would expect in a fully featured dictation plug-in or set of scripts and configurations. DictationBridge echoes back the text you’ve dictated, it provides access to the user interface for both Windows Speech Recognition and the entire line of Dragon products and includes a full sweet of other cool features you can read about in the DictationBridge documentation.
  • DictationBridge is the first ever dictation solution for screen readers to include an extensive collection of verbal commands that users can employ to control their screen reader and do various other tasks with the Dragon line of products from Nuance, as well as with Windows Speech Recognition. The DictationBridge team designed this part of the package to make it very easy to add new commands and to modify those that we provide. It even supports navigating the entire web with only speech!
  • DictationBridge, NVDA and Windows Speech Recognition are all available at no cost to end-users, their educators, their employers, governments or anyone else.
  • Care has been taken while developing DictationBridge to ensure that it can be easily translated into languages other than English. This affords DictationBridge the possibility to be translated into any of the 35 languages supported by Windows Speech Recognition, and any of the more than 43 languages supported by NVDA which correspond with those supported by Windows Speech Recognition. Whether you’re dictating in a language other than English, or whether your entire Windows interface is in a language other than English, you’ll be able to use DictationBridge in your preferred language once there is a translation available for it, without maintaining separate software licenses for your screen reader to cover non-English interfaces. We believe all blind people, no matter where they live or what language they speak natively should have access to the software they need and DictationBridge is a part of that overall goal.
  • Because DictationBridge is free, libre open source software (FLOSS), it is not only available gratis to all who care to use it, DictationBridge affords the community the freedom to add to, modify, learn from, repurpose or do anything else they care to do with the software moving forward. People with programming skills can look at the source code as it currently stands for both the Jaws and NVDA versions, along with the DictationBridge core code, by browsing to this GitHub user profile and selecting the appropriate repository. DictationBridge provides true freedom with a lower case “f.”
  • Documentation is just as important as the software itself. You will find DictationBridge’s documentation to be among the highest quality available you’ve ever read.
  • Finally, the DictationBridge team recognizes that it is often necessary for screen reader users to use more than one screen reader throughout a day’s computer use. Consequently, DictationBridge is capable of handing off its functionality between screen readers, provided you have both the NVDA and Jaws versions installed.

There Were Car Phones In 1947

If I knew this, somewhere along the way I forgot that I did.

From the CBC Archives comes this broadcast from February 19th, 1947 demonstrating some newfangled mobile phone technology that was set to hit the market in June of that year. Like I said, I had no idea this stuff went back quite that far.

Notice how clear the call quality is, all things considered. Then stop and wonder why half the time your calls don’t sound that good *now*.

Also, there’s no reason why you couldn’t answer the phone while driving says Bill Beatty, clearly not anticipating that these car phone things are ever going to be all that popular or the trouble that might come with everyone having the same idea if he’s wrong.

In this CBC Radio report, two reporters take the new mobile phone, which is about to be introduced to the market, out for a spin. Driving along the streets of Toronto, Bill Beatty’s car is outfitted with the required antenna and receiver boxes. Calling from a telephone booth, Byng Whitteker places the call.

They’re All Gonna Laugh At You!

After reading a story about people’s Amazon Echo devices laughing at them at random, I’m really glad I didn’t get one for my grandma. That would have scared the heck out of her, understandably!

Seriously. Sometimes out of the blue, or in the middle of a conversation that has not included the word “Alexa”, or after being given a command, some Amazon Echos have decided that the last spoken phrase was a real knee-slapper. Here’s a recording of one of them laughing. Of course, the recording was created by the person triggering her to laugh, so it isn’t as freaky as if a random laugh had been captured, but still.

I can’t decide what I would find more creepy. Having it randomly laugh out of the blue would be weird and startling, and be a definite reminder of how it is listening all the time. But I can’t say I wouldn’t get the chills at the idea of it laughing in response to a command as if she has decided that the idea of her doing such a thing is hilarious. Sure, one part of my mind would say that she just misunderstood my command, but there would be another part that would have this urge to call her Hal.

It is a freaky idea that these speakers can’t do anything on their own, and all their updates are handled on some server somewhere. Nope, I still don’t want one.

Microsoft Soundscape Looks Neat, But I’ll Have To Wait To Try It

Even though this isn’t available in Canada yet, I’m still posting about it because it’s cool.

Microsoft has released another interesting app to help out us blind kids, this one is for navigation. It’s called Microsoft Soundscape. It’s kind of like Blindsquare only in a videogame-like format. As you’re walking, you will hear where points of Interest nearby are, and you will hear them in that direction. So if TD bank is to the right of you, you will hear “TD Bank” spoken in your right ear. You can create beacons on landmarks, such as the exit from the park that you take to go home, or some side door of a building, or some such and they will make knocking noises as you pass them. It made me think of the old PakMan Talks videogame when it would go “Super power. Super power…” Of course, I think you would have to wear bone-conducting headphones for this to work at all, but heck, I usually am doing that anyway. Here’s a video to show how it works.

Slightly cheesy, but hey I shouldn’t criticize, I hate making videos.

I will have to reserve judgment on how awesome it is until I can actually try it. I wonder about the accuracy of when it says something is somewhere. How does it handle when you’re in a place with a lot of things right close together? I wonder about how good the beacons are. It feels like it does a lot of the things Blindsquare does, but in a way that’s very instantaneous and close by, as if you’re just scanning your immediate area with your eyes. I definitely want to try it out when it becomes available and see how I feel.