Datasets:
willtheorangeguy
/
2018-WAN-Show-Transcripts

Dataset card Data Studio Files
xet
Community
text
stringlengths
14
502
[1924.70 --> 1925.84] They've been on the show a million times.
[1925.92 --> 1927.92] You guys kind of know what it's about, but I'm going to go over it anyways.
[1928.36 --> 1932.02] Jerky is made with the best ingredients without nitrates or preservatives.
[1932.02 --> 1933.98] Their goal is to create a snack that's full of flavor.
[1935.12 --> 1937.74] And just it isn't supposed to be bad for you either.
[1937.88 --> 1938.38] That's cool.
[1938.38 --> 1940.32] People seem to like the Sriracha bacon.
[1940.50 --> 1941.72] Apparently, that's quite popular.
[1942.26 --> 1945.02] I like pretty much everything they have.
[1945.02 --> 1950.06] It's like every one of these bags is open because they always get eaten.
[1950.24 --> 1950.58] They do.
[1950.98 --> 1952.64] People find the box and then steal them.
[1952.64 --> 1954.46] My favorite personal one is Moho.
[1954.66 --> 1957.10] But again, I like literally all of their flavors.
[1957.22 --> 1958.34] So that works for me.
[1958.78 --> 1963.16] You can use offer code LTT to save 10% off their products, which is cool.
[1963.74 --> 1966.06] And they also have one of them.
[1966.14 --> 1967.40] I'm not sure where it is.
[1967.40 --> 1970.94] Because Ivan keeps eating it, so it could be gone.
[1971.16 --> 1975.52] But one of theirs is made with the hottest peppers in the world, the Carolina Reaper.
[1975.96 --> 1977.32] They have a Reaper jerky.
[1977.66 --> 1978.30] Oh, yes.
[1978.42 --> 1979.62] That's pretty dang hot.
[1979.62 --> 1981.50] It'll probably kick your butt.
[1981.98 --> 1984.38] So if you're into that, feel free to check it out.
[1985.98 --> 1987.38] And we're back.
[1987.56 --> 1988.32] Oh, we're going back.
[1988.40 --> 1989.06] What do you want to do now?
[1989.12 --> 1990.14] Oh, let's talk about Alexa.
[1990.50 --> 1990.72] Oops.
[1990.84 --> 1991.56] Sorry, I said it.
[1991.86 --> 1992.54] Amazon Echo.
[1992.54 --> 1993.14] Oh, no.
[1994.66 --> 1998.62] So there is a family in, I believe, Oregon, Portland.
[1999.62 --> 2000.78] And they were hanging out.
[2000.78 --> 2004.28] And suddenly a phone call came in to the husband.
[2004.78 --> 2008.18] And it was one of his employees who lives in, I believe, Seattle.
[2008.94 --> 2010.32] Many, many miles away.
[2010.76 --> 2016.20] And the guy was like, quick, unplug all your A-L-E-X-A devices.
[2016.74 --> 2018.08] You're getting hacked.
[2018.82 --> 2022.06] And what had happened was, for whatever reason,
[2022.54 --> 2027.26] the Echo started recording at a time that the inhabitants of the house
[2027.26 --> 2031.04] did not know it was recording and recorded a big conversation they had.
[2031.34 --> 2037.30] Then it proceeded to send the audio files to a random person in their contact list.
[2038.34 --> 2039.40] And so this guy called.
[2039.68 --> 2044.04] And the husband's like, what are you talking about?
[2044.26 --> 2044.98] You're just joking.
[2045.20 --> 2045.76] Like, yeah, right.
[2045.88 --> 2046.16] Ha ha.
[2046.24 --> 2046.58] Good one.
[2046.62 --> 2048.56] And the guy's like, no, you were talking about your hardwood floors.
[2049.12 --> 2049.92] Like, I heard it.
[2049.96 --> 2050.76] And they're like, oh, God.
[2050.76 --> 2053.16] And so then they contacted Amazon.
[2053.42 --> 2056.24] And Amazon put an engineer to investigate right away.
[2056.80 --> 2058.46] And eventually a statement came out.
[2058.58 --> 2061.60] First, the engineer said, like, there will be a fix.
[2062.32 --> 2063.46] It kind of made it seem.
[2063.60 --> 2065.14] He didn't just brush it off as user error.
[2065.24 --> 2066.40] He was like, we're going to fix something.
[2066.68 --> 2068.08] But then the statement comes out.
[2068.48 --> 2071.78] And it's kind of, it's not user error.
[2071.78 --> 2075.42] But the system acted kind of as intended.
[2075.80 --> 2078.44] So here's what the, where the heck are we here?
[2079.32 --> 2080.90] Here's the, here's the, the events.
[2081.00 --> 2081.38] What happened?
[2082.38 --> 2087.40] Echo woke up due to a word in a background conversation sounding like the hot word.
[2087.40 --> 2093.54] Then the subsequent conversation was heard as send message request.
[2093.66 --> 2097.20] At which point the echo said out loud to whom?
[2097.88 --> 2101.74] Now, if the volume was turned down and they were ranked far away, then they wouldn't have heard that.
[2101.80 --> 2101.96] Right.
[2103.06 --> 2108.30] At which point the background conversation was interpreted as a name in the customer's contact list.
[2108.30 --> 2113.36] So they kept having this conversation and the echo was just kind of picking out where it's like, I think it said send message.
[2113.48 --> 2115.10] I think it said to Craig.
[2115.90 --> 2120.38] And then, and then it carried out that, that request.
[2120.46 --> 2127.14] And so as unlikely as a string of events like that is, Amazon is evaluating the options to make the case even less likely.
[2127.64 --> 2133.50] At the same time, like, you have, you have a smart home device.
[2133.72 --> 2133.80] Yeah.
[2133.80 --> 2136.80] How often have you had it spring up and do something weird?
[2136.80 --> 2138.74] Lately, it seems like more than before.
[2138.98 --> 2139.42] Quite a bit.
[2139.42 --> 2139.58] Yeah, I have too.
[2139.76 --> 2140.22] Quite a bit.
[2140.46 --> 2143.78] So I've got my, my Google home mini.
[2144.44 --> 2150.74] I've went into the settings and, and turned on the settings so that when you hail it, it goes ba-ding, which is off by default.
[2151.28 --> 2152.42] Normally it just would light up.
[2152.46 --> 2157.30] And I had people across the house, they'd be trying to hail it and maybe they're not very experienced with it.
[2157.32 --> 2159.40] And they wouldn't know whether or not it had heard them.
[2159.54 --> 2161.28] So I put on the feedback noise.
[2161.28 --> 2167.78] But that's also super helpful too, because then if I'm watching a movie and the movie says something like, no, you go.
[2168.60 --> 2173.20] And that sounds like, hey, then I'll hear it.
[2173.46 --> 2174.82] And I can just say, never mind.
[2174.94 --> 2176.52] And it'll, it'll turn off again.
[2176.52 --> 2193.42] But I guess what's happening here is the way that these, that this, that machine learning works basically with this kind of thing is there's a probability that when I hear something, it'll say, okay, that's a, that's a 0.8 out of one.
[2193.54 --> 2196.62] Let's say there's an 80% chance that they just said what I think they said.
[2196.76 --> 2198.42] And then there's a decision threshold.
[2198.42 --> 2203.54] So the engineers will say anything greater than a 0.75 is good.
[2203.86 --> 2206.78] And so that counts as, we're going to call that a positive.
[2207.04 --> 2212.94] So you can shift that decision threshold based on the stakes of whatever you're talking about.
[2213.00 --> 2223.14] So if you're talking about a spam email, for example, calling something, letting in spam when it's not spam, that's not that high stakes.