Lex Fridman
Rohit Prasad is the vice president and head scientist of Amazon Alexa and one of its original creators. This conversation is part of the Artificial Intelligence podcast.
This episode is presented by Cash App: download it & use code “LexPodcast”
This episode is also supported by ZipRecruiter. Try it: http://ziprecruiter.com/lexpod
INFO:
Podcast website:
https://lexfridman.com/ai
Apple Podcasts:
https://apple.co/2lwqZIr
Spotify:
https://spoti.fi/2nEwCF8
RSS:
https://lexfridman.com/category/ai/feed/
Full episodes playlist:
https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4
Clips playlist:
https://www.youtube.com/playlist?list=PLrAXtmErZgOeciFP3CBCIEElOJeitOr41
OUTLINE:
0:00 – Introduction
4:34 – Her
6:31 – Human-like aspects of smart assistants
8:39 – Test of intelligence
13:04 – Alexa prize
21:35 – What does it take to win the Alexa prize?
27:24 – Embodiment and the essence of Alexa
34:35 – Personality
36:23 – Personalization
38:49 – Alexa’s backstory from her perspective
40:35 – Trust in Human-AI relations
44:00 – Privacy
47:45 – Is Alexa listening?
53:51 – How Alexa started
54:51 – Solving far-field speech recognition and intent understanding
1:11:51 – Alexa main categories of skills
1:13:19 – Conversation intent modeling
1:17:47 – Alexa memory and long-term learning
1:22:50 – Making Alexa sound more natural
1:27:16 – Open problems for Alexa and conversational AI
1:29:26 – Emotion recognition from audio and video
1:30:53 – Deep learning and reasoning
1:36:26 – Future of Alexa
1:41:47 – The big picture of conversational AI
CONNECT:
– Subscribe to this YouTube channel
– Twitter: https://twitter.com/lexfridman
– LinkedIn: https://www.linkedin.com/in/lexfridman
– Facebook: https://www.facebook.com/lexfridman
– Instagram: https://www.instagram.com/lexfridman
– Medium: https://medium.com/@lexfridman
– Support on Patreon: https://www.patreon.com/lexfridman
Source
I really enjoyed this conversation with Rohit. Here's the outline:
0:00 – Introduction
4:34 – Her
6:31 – Human-like aspects of smart assistants
8:39 – Test of intelligence
13:04 – Alexa prize
21:35 – What does it take to win the Alexa prize?
27:24 – Embodiment and the essence of Alexa
34:35 – Personality
36:23 – Personalization
38:49 – Alexa's backstory from her perspective
40:35 – Trust in Human-AI relations
44:00 – Privacy
47:45 – Is Alexa listening?
53:51 – How Alexa started
54:51 – Solving far-field speech recognition and intent understanding
1:11:51 – Alexa main categories of skills
1:13:19 – Conversation intent modeling
1:17:47 – Alexa memory and long-term learning
1:22:50 – Making Alexa sound more natural
1:27:16 – Open problems for Alexa and conversational AI
1:29:26 – Emotion recognition from audio and video
1:30:53 – Deep learning and reasoning
1:36:26 – Future of Alexa
1:41:47 – The big picture of conversational AI
Great work, Lex, however, maybe because I'm German I see things more critically. I work in the same field of research, but with all the enthusiasm about new possibilities it's easy to forget that this another step to make people more dependable on technology and losing agency and touch with the world (Alexa deciding for you that you might need a table at a restaurant after going to the cinema… people are so predictable…) And after all, keep in mind, this is all targeted at doing more business and pushing consumerism.
This guy has magic skin.
I bought Echo in April 2016 and now own five devices including the Show which we just added. We enjoy music and device management most but look forward to video calling and messaging with our adult kids in the coming months as well as greater household security and convenience.
I t think Rohit hit on something that sets Alexa part from Google which is verbosity. Google responds with a explanation of your command, maybe to promote conversation, but Rohit realizes sometimes it's a human machine interaction so Alexa just says "ok" vs Google saying "turning off all lights".
I'd like to hear about how Home AI's can interpret a compound command and perhaps leveraging a command that was not very effective and weighing the follow up command as "gospel".
Alexa "its too bright", "I don't know that" "living room lights 35%" "ok"
As Elon says, all input is error. So all second input, (especially when first was impotent), represents first interpretation was error?
Apple’s AI just autocorrected “Rohit Prasad” to “Robot Prada.”
Great.
https://www.news18.com/news/tech/amazon-alexa-told-this-lady-to-kill-herself-because-humans-are-bad-for-the-planet-2434375.html weirdly Alexa
I picture the people who uses these useless devices are overweight Homer Simpsons who likes everything done for them (thinking included). Name me one "skill" this useless piece of junk can do that we humans cannot? The only thing this junk is good for is being a listening device for any dictator out there.
Excellent conversation Lex. I came away with respect for Alexa, for Rohit, and for the team. Your podcast is an excellent forum for bringing understanding to otherwise complex subjects, something missing in mainstream clickbait. Very much appreciated. Thank you to you both.
I'm surprised how little you care about data privacy. The number of ways data could get in the wrong hands and be abused is staggering. For me the potential risks with having a device like this listen 24/7 far outweigh the benefits it might have.
your sponsor spots are annoying.
just sayin.
No wonder she doesn't understand me, i just need to talk with indian accent
Alexa When were your born
Alexa : My birthdays is Nov 6, 2014.
Dude you are gong to make it big. I wouldn’t bet against that. These are important.
This is the greatest podcast/interviews I've ever heard
Delightful!
A little Henry Miller. You seem a man who would appreciate him.
“…the monstrous thing is not that men have created roses out of this dung heap, but that, for some reason or other, they should want roses. For some reason or other man looks for the miracle, and to accomplish it he will wade through blood. He will debauch himself with ideas, he will reduce himself to a shadow if for only one second of his life he can close his eyes to the hideousness of reality. Everything is endured- disgrace, humiliation, poverty, war, crime, ennui- in the belief that overnight something will occur, a miracle, which will render life tolerable. And all the while a meter is running inside and there is no hand that can reach in there and shut it off.”
― Henry Miller, Tropic of Cancer
43:50 The question should likely be – irrespective of whether one is a researcher, or a customer – Is AI meant to be a tool for humans or their overlord?
47:40 Why not delete everything by default? Who – normal customer – wants to let their daily life be annotated later by amazon or others? What are they paid back for their data being used for advancing the field of AI?
Any conversation regarding voice recognition and or facial recognition AI warrants pointed questions on privacy, ethics. I enjoy this podcast but i often feel like the difficult (and most important) questions are either glossed over or avoided entirely
I was playing this while eating dinner and the word Alexa kept triggering my echo dot in the dining room so I had to switch to the Michael Jordan episode. I'll finish this EP elsewhere… lol
Anil sabhavala from Google please
Anil sabhavala from Google please
Lex, I am addicted to you videos…keep them coming,,,
Reading the auto-captions at the same time is a lot of fun… "this is a criminal device" (communal) loll