With Google IO, Microsoft Build, and Facebook F8 all happening within a week of each other, there’s lot to talk about on this week’s show. We discuss the implications of Google’s new AI efforts, Oculus’ new prototype VR headset, and Nintendo online service plans. Plus, more details about our trip to see the NASA InSight mission rocket launch!
One Day Builds
Adam embarks on one of his most ambitious builds yet: fulfil…
Show And Tell
Adam recently completed a build of the royal St. Edwards cro…
Making
Viewers often ask to see Adam working in real-time, so this …
One Day Builds
Adam and Norm assemble a beautifully machined replica prop k…
One Day Builds
One of the ways Adam has been getting through lockdown has b…
Making
Adam unboxes and performs a quick test of this novel new hel…
Making
When Adam visited Weta Workshop early last year, he stopped …
One Day Builds
Adam tackles a shop shelf build that he's been putting off f…
Show And Tell
Time for a model kit build! This steampunk-inspired mechanic…
One Day Builds
Adam reveals his surprise Christmas present for his wife--a …
You’ve seen Infinity War multiple times in different formats but would you watch in 4DX? Not sure if it’s in the US but it just launched in some cinemas in the UK and they’re showing Infinity War in this format. https://www.cineworld.co.uk/4dx#/
2 things, while that MAY have been a Google assistant robo call, the call its self was edited , heavily edited. 2. The disturbing thing I see here is that Google thinks 2 humans talking to each other as a problem that needs fixing…..how anti social are we becoming as a society when we view communication as a problem?
I wonder if the Google assistant can help increase understanding in the middle of a call. Like real-time translation, or if you make a 911 call and you don’t know exactly where you are, the Google assistant could cut in and give the 911 evaporator your exact position.
I wonder if the Google assistant can help increase understanding in the middle of a call. Like real-time translation, or if you make a 911 call and you don’t know exactly where you are, the Google assistant could cut in and give the 911 evaporator your exact position.
As for robo-calls … I’d be concerned about the folks on the other end peppering the AI with questions that reveal credit card info, personal info, etc, whether they need it or not. I’d want to white-list whatever the AI could report; but, I might not anticipate everything. The salon wanted a first name; I might not have that “enabled”, so the process could break down right there. Maybe the AI could notify me whenever sensitive information is requested, so that I could authorize (or deny) the reply … but that will cause some “conversational latency”.
Id like to ask for skepticism regarding Uber Engineering. Detecting a bicycle is Not Hard. Here is a website that will let you download and run image-classification code in about 15 minutes of setup time https://pjreddie.com/darknet/yolo/
this classifier has been trained on stuff like people, bicycles, and fruit. you give it an image, any image, and it will draw little boxes on that image showing where the bicycles, people, and fruit are.
i downloaded some still images from the uber crash video and ran them against this program. this little classifier program was 100% certain it saw a bicycle in one frame of the uber crash, and 97% in another.
this classification took 30 seconds to run on my ancient Pentium laptop that cost me $100.00.
if you run it on a modern machine with a GPU you can do this classification instantly. in realtime.
i know next to nothing about image classification. i set this thing up and ran it, downloaded some images, and ran them as well, in about half an hour. i am not some super genius. this took next to zero knowledge and skill for me to do.
there is literally zero excuse for Uber claiming the car didn’t know there was a bicycle in front of it.
That’s a heck of a sponsor in the middle.
Norm. Dont know if you planned on talking about the Comic com center already but thanks for listening to members and making it a topic.
It’s fun to see how much of the subtle little ques that we humans use to tell when someone is listening or still there. The ums, and bits of sounds that Google has added to it’s robo calls is just some of that. I’m not sure they are trying to trick you that it’s a real person on the other end. They are added feed back so that the recipient knows that the caller is still there while the AI searches the data for the correct response. If after the salon tells you that they can’t do noon if the lines quite for a few seconds many folks would hang up figuring the caller didn’t like the news and bailed. Look at it as the engineers bending the AI to the ticks and norms of the humans instead of the humans adapting to the AI.
Remember there is no spoon.
CBB
Just curious what DJI Mavic Pro bag you were looking at using to store the Oculus Go in?
“Skin jobs”. That’s what Bryant called Replicants. In history books he’s the kind of cop who used to call black men “ni***rs”.
— So we are fast approaching the point where ethical treatment of AI may become a real topic. Currently when you encounter a voice prompted phone system you can be an snarky and rude with the “voice” on the other end because it only processes the specific response it’s looking for. When an AI on the other end of a phone call can listen to your every insult and assess the meaning, motive, and implication behind and craft responses accordinly then there will be some accountability. At what point will we judge ourselves and others on how polite we are to our AI counterparts?
“Skin jobs”. That’s what Bryant called Replicants. In history books he’s the kind of cop who used to call black men “ni***rs”.
— So we are fast approaching the point where ethical treatment of AI may become a real topic. Currently when you encounter a voice prompted phone system you can be snarky and rude with the “voice” on the other end because it only processes the specific response it’s looking for. When an AI on the other end of a phone call can listen to your every insult and assess the meaning, motive, and implication behind and craft responses accordingly then there will be some accountability. At what point will we judge ourselves and others on how polite we are to our AI counterparts?
None of you even mentioned Civil War in the top Marvel Movies?! It’s my number 1!
1. Civil War
2. Winter Soldier
3. Ragnarok
4. Iron Man 1
5. Avengers
Google is absolutely trying to trick the person on the other end, because they knew they would get the big round of applause from an unquestioning crowd at IO if they managed to do that. And that cynicism is a big part of the backlash this is getting.
I hope that by the time this tech gets into a launched product, it will tell the person opposite that it is an AI, because otherwise, it is going to be a clusterfuck of time wasting, with service minded folks desperately trying not to offend a robot that is messing up as a result of a difficult accent. This is technology that does not need to hide, and will work better if it is open about its nature. Even better if they somehow make it opt-in.
While I don’t disagree with the people who question whether we want a tool that gets us off the hook when it comes to having a normal conversation to book a hairdresser-appointment, there are a lot of cases where this will be a big help for people. A lot of people have serious problems talking on the phone, be it anxiety, language barrier, etc. and this will help them do things that would cause major strife before.
Did Norm just give away a spoiler for those of us who haven’t seen infinity wars, while discussing agents of shield?!?!?!
I have now seen the film and yes he did!