WEBVTT - Computers Know If You're Sarcastic? Yeah, Right. 0:00:00.160 --> 0:00:07.160 Brought to you by Toyota. Let's go places. Welcome to 0:00:07.360 --> 0:00:14.160 Forward Thinking. Hey there, and welcome to Forward Thinking, the 0:00:14.240 --> 0:00:17.200 podcast that looks at the future and says no dark 0:00:17.239 --> 0:00:23.280 sarcasm in the classroom. I'm Jonathan Strickland and I'm Joe McCormick. So, guys, 0:00:23.640 --> 0:00:26.520 I thought today we could talk a little bit more 0:00:26.560 --> 0:00:29.720 about how, you know, computers just don't get me, man, 0:00:30.280 --> 0:00:34.920 they just don't understand. They're like parents. You know, parents 0:00:35.000 --> 0:00:39.680 never understand the importance of the flavor of apple Jack's. 0:00:39.720 --> 0:00:43.400 That's true, among many other things. True. In fact, they 0:00:43.440 --> 0:00:46.760 also don't understand our music, right, they don't. They don't 0:00:46.800 --> 0:00:51.200 get my clothes, my style. Uh. Now, we're we're talking 0:00:51.320 --> 0:00:54.639 about how computers, you know, trying to teach computers how 0:00:54.680 --> 0:00:58.120 to understand what we mean when we say things. And 0:00:58.160 --> 0:01:00.640 we've talked about this before. We've talked about computers in 0:01:00.760 --> 0:01:04.759 natural language and how that's a really tough problem getting 0:01:04.840 --> 0:01:09.040 computers to understand the way we humans communicating naturally yea, 0:01:09.120 --> 0:01:11.720 and especially trying to get them to mimic it or 0:01:11.880 --> 0:01:14.360 or feed it back to us and participate in a 0:01:14.400 --> 0:01:19.240 conversation that's convincingly human. That's the standard problem of the 0:01:19.240 --> 0:01:22.160 touring test. Yeah, even if it's not convincingly human, at 0:01:22.240 --> 0:01:26.120 least useful, right like we you know, I would settle 0:01:26.200 --> 0:01:28.840 for a computer that would understand what I mean when 0:01:28.880 --> 0:01:31.160 I say things a certain way and give me the 0:01:31.160 --> 0:01:33.720 stuff that I'm requesting, even if it can't hold a 0:01:33.720 --> 0:01:37.280 conversation with me. Right right, We don't always need Siri 0:01:37.400 --> 0:01:39.760 to to respond back if if she could just find 0:01:39.800 --> 0:01:43.920 what we're bloody looking for. Cortana would be another great 0:01:43.920 --> 0:01:46.440 example when those ten has just been released the day 0:01:46.480 --> 0:01:48.720 we're recording this, and I'm seeing a lot of reports 0:01:48.760 --> 0:01:53.880 about how Cortana is almost but not quite awesome. You know. 0:01:53.960 --> 0:01:56.120 One of the things that's interesting to me about Siri 0:01:56.280 --> 0:02:01.160 is that Siri is playful Siri. Siri will sometimes be 0:02:01.320 --> 0:02:04.480 koy and sometimes like play along when you're when you're 0:02:04.480 --> 0:02:08.720 being I don't know, suggestive or weird at her. I 0:02:08.720 --> 0:02:11.440 don't say this from personal experience. I've read about this, 0:02:12.360 --> 0:02:16.280 but I get the impression that that is not original 0:02:17.360 --> 0:02:21.359 like synthetic behavior, that it has been assembled by series 0:02:21.440 --> 0:02:25.560 logical engines. That is more of a hard coded behavior. 0:02:25.800 --> 0:02:27.880 I think that's more of their sort of jokes put 0:02:27.919 --> 0:02:31.720 in directly by the programmer. The programmers anticipate the kind 0:02:31.720 --> 0:02:34.079 of like they just say, Well, if I found out 0:02:34.120 --> 0:02:38.320 that this audio personal assistant could answer questions that I 0:02:38.320 --> 0:02:40.920 ask of it, what are some of the ridiculous questions 0:02:40.960 --> 0:02:44.000 I might ask? And they probably made a list and 0:02:44.000 --> 0:02:46.560 and that's probably the basis for a lot of those 0:02:46.600 --> 0:02:49.519 funny responses, things like where can I hide a body? Yeah? Yeah, 0:02:49.560 --> 0:02:51.840 but but that's all coded by your word choice. It's 0:02:51.880 --> 0:02:54.240 not like it's not like, you know, Siri is gonna 0:02:54.280 --> 0:02:57.120 listen to you say something like, h Siri, where can 0:02:57.160 --> 0:03:00.520 I find another pizza place? Because I'm so interested in 0:03:00.639 --> 0:03:04.000 pizza today? Like she She's not going to come back 0:03:04.080 --> 0:03:06.840 with with with like, well, maybe you should try some 0:03:06.919 --> 0:03:09.600 chicken nuggets this time, fatty, Like I mean, like I 0:03:09.680 --> 0:03:13.799 was thinking, it's Taco Tuesday, chicken nuggets. The cure to 0:03:14.000 --> 0:03:19.960 facts that would be a revolutionary episode forward thinking. However, 0:03:20.320 --> 0:03:24.079 we wanted to talk today about I learned it from Laura. 0:03:24.840 --> 0:03:28.720 We wanted to talk today more about how computers might 0:03:28.840 --> 0:03:33.160 soon or at least down the road, learned to recognize 0:03:33.160 --> 0:03:35.960 not just what we say, but what we mean when 0:03:35.960 --> 0:03:40.040 we say it. Right, So could Siri participate in a 0:03:40.280 --> 0:03:45.120 sarcastic sort of caustic back and forth humorous exchange with 0:03:45.160 --> 0:03:48.560 you without that being hard coded by the programmers, could 0:03:48.600 --> 0:03:53.120 could Siri detected and even and even synthesize it herself. Right, 0:03:53.240 --> 0:03:55.920 So to start this, we thought it might be useful 0:03:55.960 --> 0:03:58.880 to kind of give a quick overview of why there's 0:03:58.960 --> 0:04:03.200 this us connect between natural language and what computers understand 0:04:03.560 --> 0:04:07.600 and basically computers machines work in uh in machine code 0:04:07.680 --> 0:04:10.600 or assembly language. Assembly language is generally defined as a 0:04:10.680 --> 0:04:14.200 very low level programming language, sometimes a one to one 0:04:14.280 --> 0:04:18.160 correspondence with the architecture of the actual machine you are 0:04:18.279 --> 0:04:22.320 using it on. So the thing about this, the defining characteristic, 0:04:22.400 --> 0:04:26.359 is that computers understand it and humans have real trouble 0:04:26.400 --> 0:04:29.640 with it because it's it's so far removed from any 0:04:29.640 --> 0:04:32.839 sort of language that we use. It would be like 0:04:32.880 --> 0:04:36.360 looking at a just enormous page, just a block of 0:04:36.440 --> 0:04:38.680 texts that are all zeros and ones, and trying to 0:04:38.760 --> 0:04:42.440 make any sense of it. Um. I assume that most 0:04:42.520 --> 0:04:46.480 human beings, the vast majority, would be unable to do so. 0:04:46.480 --> 0:04:50.520 So we humans communicate natural language. That's what I'm doing 0:04:50.640 --> 0:04:53.920 right now. And natural language isn't just agree and agreed 0:04:54.000 --> 0:04:56.839 upon syntax and grammar. It also has all these other 0:04:57.279 --> 0:05:00.440 elements to it, implications where you know, you don't say 0:05:00.480 --> 0:05:03.640 something outright, but it is implied by the way you 0:05:03.680 --> 0:05:06.359 say it, or the tone of how you deliver something, 0:05:06.400 --> 0:05:10.240 including sarcasms. Sure, even the gestures that you're making or 0:05:10.279 --> 0:05:13.120 the facial expression. Yeah. I think we talked pretty recently 0:05:13.120 --> 0:05:15.520 on this program about how one major difference is that 0:05:15.640 --> 0:05:20.960 humans learn language by induction rather than by you know, 0:05:21.040 --> 0:05:24.599 we don't have an explicit list of rules and uh 0:05:24.640 --> 0:05:27.479 and definitions, Like we don't learn how to speak our 0:05:27.600 --> 0:05:30.359 language by reading a dictionary in a grammar book. You 0:05:30.480 --> 0:05:33.440 just sort of gradually into it the rules based on 0:05:33.520 --> 0:05:36.240 experiencing them. Yeah. Yeah, and we free associate a lot 0:05:36.240 --> 0:05:38.920 of stuff with a lot of other stuff. That's true. Yeah. 0:05:38.960 --> 0:05:40.479 So one thing that we have to do is find 0:05:40.480 --> 0:05:42.640 a way to bridge that gap between the way we 0:05:42.720 --> 0:05:46.960 humans communicate and the way machines understand commands. And we 0:05:47.040 --> 0:05:51.240 do that through programming languages and compilers. A programming language, 0:05:51.279 --> 0:05:53.960 depending upon whether you call it low level or high level, 0:05:54.400 --> 0:05:57.360 is something that may be easier for a human to grasp. 0:05:57.440 --> 0:05:59.880 In general, if we say it's a high level programming language. 0:05:59.880 --> 0:06:02.480 That means it's more like the kind of languages we use, 0:06:02.520 --> 0:06:06.040 although if you're unfamiliar with programming language, it still looks 0:06:06.080 --> 0:06:09.800 like gibberish exactly. Yeah, this is a funny fact about 0:06:09.800 --> 0:06:13.520 programming languages. People often think of them as something that 0:06:13.600 --> 0:06:17.039 helps the computer understand. That's not what it is. It's 0:06:17.080 --> 0:06:21.080 something that helps the human understanding its programming language is 0:06:21.080 --> 0:06:23.719 a tool for you, right, it's so that the human 0:06:23.839 --> 0:06:26.920 can say, uh, this is this is the input I'm 0:06:26.960 --> 0:06:29.760 going to give you, this is the desired output, make 0:06:29.839 --> 0:06:33.080 it go. And so on the machine side they have 0:06:33.440 --> 0:06:36.440 there are compilers and compilers. The job of a compiler 0:06:36.520 --> 0:06:40.640 is to essentially translate, to switch from that programming language 0:06:40.640 --> 0:06:44.200 into machine language. So it's this combination that allows machines 0:06:44.240 --> 0:06:48.360 to understand what we want them to do. But more 0:06:48.400 --> 0:06:51.360 recent work, really this has been going on for decades, 0:06:51.400 --> 0:06:53.800 but truly in the last decade we've seen a lot 0:06:53.880 --> 0:06:56.839 of work done in trying to come up with ways 0:06:56.880 --> 0:07:00.080 where machines can deal with natural language and the and 0:07:00.080 --> 0:07:03.839 if it is obvious, because it removes a barrier between use, 0:07:04.000 --> 0:07:05.960 you know, a human trying to use a machine, and 0:07:06.000 --> 0:07:09.200 the machine doing what what the human wants it to do. So, 0:07:09.720 --> 0:07:12.720 for a very simple example, if you go on your 0:07:12.760 --> 0:07:16.520 computer and you wanted to find a particular piece of information, 0:07:16.560 --> 0:07:18.760 you could just type something in and it would understand 0:07:18.760 --> 0:07:21.960 what you meant by that, so it wouldn't return a 0:07:22.160 --> 0:07:25.239 page of search results that are in some way related 0:07:25.280 --> 0:07:28.920 to what you wanted. It would return exactly what you wanted. Uh. 0:07:29.000 --> 0:07:32.080 So that that's a simple example. There are lots of 0:07:32.080 --> 0:07:36.320 other examples of this, but we're still just at the 0:07:36.400 --> 0:07:41.080 very early stages of getting a good grasp on natural language. 0:07:41.080 --> 0:07:44.920 We're getting better at computers parsing sentences to at least 0:07:44.920 --> 0:07:48.200 make an educated guess as to what we want, but 0:07:48.720 --> 0:07:52.120 we're not all the way there yet. But we wanted 0:07:52.120 --> 0:07:55.520 to talk about some interesting tools that are getting us closer, 0:07:55.920 --> 0:07:58.880 and one was one that was reported on in July 0:07:59.320 --> 0:08:01.840 two thousands of team that I really wanted to talk about, 0:08:01.840 --> 0:08:05.280 which was IBMS tone Analyzer. Yeah. I had some fun 0:08:05.360 --> 0:08:08.400 playing with this earlier. Today. We can talk about the 0:08:08.480 --> 0:08:11.440 results I came up with later, but I think a 0:08:11.560 --> 0:08:14.320 common theme in our analysis of this will be that 0:08:14.400 --> 0:08:17.400 it's interesting, but maybe more for the reasons that it 0:08:17.520 --> 0:08:21.160 fails than for the reasons it succeeds. Yeah, I would 0:08:21.200 --> 0:08:23.760 agree with that. It's kind of a very advanced version 0:08:23.760 --> 0:08:27.120 of Clippy almost in that way I'm like, oh, it's 0:08:27.120 --> 0:08:29.400 really cute. The way that you have no idea what's 0:08:29.400 --> 0:08:31.480 going on. I almost think of it as as like 0:08:31.800 --> 0:08:35.200 and of evolution of just word count. Yeah, like it's 0:08:35.240 --> 0:08:38.120 still counting words, it's just now classifying what word is, 0:08:38.160 --> 0:08:41.400 what like or what general categories words fall into. But 0:08:41.840 --> 0:08:45.440 it's part of the Watson development cloud. Uh. And that's 0:08:45.559 --> 0:08:49.040 that if you are familiar with IBM S Watson, you know, 0:08:49.200 --> 0:08:51.960 you know that that refers to the machine that played 0:08:52.000 --> 0:08:55.960 on Jeopardy and beat to Jeopardy Champions. Yeah, which makes 0:08:55.960 --> 0:08:59.319 it hilarious that we're using words like cute to describe 0:08:59.360 --> 0:09:02.200 something is so ceated with Watson yea one of the 0:09:02.240 --> 0:09:06.720 most amazing computers ever put together. It's really a function 0:09:06.760 --> 0:09:09.800 of how difficult the task is. Yeah, more so than 0:09:09.880 --> 0:09:13.600 than how dumb this that's not that's not what I'm saying, 0:09:13.640 --> 0:09:16.520 No offense, wats no, no, totally that I'm on the 0:09:16.559 --> 0:09:20.600 same page. Because Watson was obviously amazing. It was able 0:09:20.679 --> 0:09:23.520 to take Jeopardy clues which are in the form of 0:09:23.520 --> 0:09:25.840 an answer and come up with the appropriate question more 0:09:25.880 --> 0:09:28.600 frequently than not. And we've talked a little bit about 0:09:28.640 --> 0:09:30.920 how it did this, and that it was able to 0:09:31.559 --> 0:09:35.400 come up with potential answers or rather questions if you prefer, 0:09:35.840 --> 0:09:40.360 and then judge how probable that particular response was to 0:09:40.440 --> 0:09:43.079 being the correct one, and if it met a certain threshold, 0:09:43.240 --> 0:09:46.200 Watson would buzz in and offer that up as the 0:09:46.320 --> 0:09:50.400 response to the clue. Yeah. And so what made Watson 0:09:50.640 --> 0:09:54.880 interesting was different than what makes most Jeopardy contestants interesting. 0:09:55.800 --> 0:09:59.280 It was because most Jeopardy contestants are competing based on 0:09:59.400 --> 0:10:04.120 their essentially their volume of trivia that they have contained 0:10:04.160 --> 0:10:08.000 in their brain. They're not having trouble discerning what the 0:10:08.080 --> 0:10:11.440 clue is trying to say, right, yeah, yeah, like like 0:10:11.480 --> 0:10:13.920 they get the puns in it, and and they get 0:10:13.960 --> 0:10:17.319 the sarcasm or whatever it is. I guess Alex Trebeck 0:10:17.400 --> 0:10:20.640 isn't sarcastic very often. But you know, references as well, 0:10:21.160 --> 0:10:25.120 they cultural references contest. Yeah, right, So for them what's 0:10:25.160 --> 0:10:28.080 impressive is, Wow, I can't believe that lady knew all 0:10:28.120 --> 0:10:32.560 that stuff about ancient Rome and about you know, records 0:10:32.640 --> 0:10:37.400 and about video games whatever they talk about. Yeah, but 0:10:37.720 --> 0:10:40.760 you wouldn't be impressed at all that she, you know, 0:10:40.880 --> 0:10:43.360 makes sense of what the clue is saying. And that's 0:10:43.400 --> 0:10:46.320 the deal with Watson. What it's not interesting that a 0:10:46.360 --> 0:10:49.080 computer has all that knowledge because it can just you know, 0:10:49.160 --> 0:10:53.000 have terabytes of memory that it runs through. What's interesting 0:10:53.120 --> 0:10:57.079 is that it knew what the question was referring to, right, 0:10:57.280 --> 0:11:00.000 and that that was one of those things that really 0:11:00.200 --> 0:11:01.840 opened up a lot of eyes and said, wow, this 0:11:01.920 --> 0:11:05.319 is exciting. A computer is, at least on some level 0:11:05.679 --> 0:11:10.000 understanding what it's supposed to be looking for. Whether that 0:11:10.120 --> 0:11:12.080 you know, that understanding doesn't go on to the same 0:11:12.160 --> 0:11:14.640 level as it would with a human, but it's still 0:11:14.640 --> 0:11:17.080 really impressive because generally, you would have a computer and 0:11:17.120 --> 0:11:20.720 you'd give it some sort of word play response, and 0:11:21.240 --> 0:11:23.920 chances are it would not come up with an appropriate answer. 0:11:24.280 --> 0:11:26.720 It would either not answer at all because it wouldn't 0:11:26.800 --> 0:11:29.840 hit that threshold of certainty, or it would give them 0:11:29.800 --> 0:11:31.959 an incorrect answer. And you know, once and once or 0:11:32.000 --> 0:11:36.040 twice Watson gave funny wrong answers. It's not like it 0:11:36.080 --> 0:11:38.960 was infallible, but it was. It worked more than it 0:11:39.040 --> 0:11:42.320 didn't work and in fact beat the two champions, so 0:11:42.480 --> 0:11:47.640 pretty exciting. Well, the Watson Developer Cloud is UH sort 0:11:47.640 --> 0:11:49.520 of it's it's kind of an umbrella. It's a it's 0:11:49.559 --> 0:11:53.040 an application program interface also known as an a p I, 0:11:53.559 --> 0:11:58.360 and developers can use it to leverage this cognitive computing approach, 0:11:58.480 --> 0:12:03.240 this language wrecking mission approach natural language recognition UH, and 0:12:03.280 --> 0:12:07.960 then leverage that into other applications. So Watson itself has 0:12:08.000 --> 0:12:10.600 gone on to do lots of work in other fields, 0:12:10.600 --> 0:12:16.240 and most notably medicine, being able to help doctors end 0:12:16.320 --> 0:12:21.319 up looking at various means of treatment, even personalizing treatment 0:12:21.360 --> 0:12:25.520 for patients with certain illnesses or conditions. But this was 0:12:25.600 --> 0:12:28.080 something where you know, if you were a developer and 0:12:28.160 --> 0:12:30.960 you wanted to try and leverage this technology that had 0:12:31.000 --> 0:12:33.920 already been built, you could then build your own thing 0:12:34.360 --> 0:12:36.600 and make something really special. And one of those things 0:12:36.679 --> 0:12:42.640 was the Tone Analyzer UM. So it it searches for social, 0:12:42.720 --> 0:12:46.200 written and emotional cues and a body of text in 0:12:46.280 --> 0:12:49.520 order to analyze the tone of the overall text and 0:12:49.520 --> 0:12:53.120 to tell you how is that coming across in an 0:12:53.120 --> 0:12:56.640 analytical way, Like you know, I I have tested the 0:12:56.880 --> 0:12:59.320 block that you have written and it comes across as 0:12:59.400 --> 0:13:03.680 seventy three percent cheerful, which sounds really weird, but that's 0:13:03.800 --> 0:13:06.560 essentially what you're getting. Yeah, yeah, And it's important to 0:13:06.600 --> 0:13:09.800 note that that this is harder in written text hypothetically 0:13:09.800 --> 0:13:11.640 than it is in spoken text because of all of 0:13:11.679 --> 0:13:14.400 those clues that we get from each other when we're 0:13:14.400 --> 0:13:16.720 talking out loud. Sure, I mean, it's much easier to 0:13:16.800 --> 0:13:20.679 detect sarcasm in person. I mean, you've probably had this experience, 0:13:20.679 --> 0:13:23.200 even though you're not a computer. You are a real 0:13:23.360 --> 0:13:26.800 human I assume, Yeah, we're guessing at least half of 0:13:26.840 --> 0:13:30.360 you are. Assuming you are a real human, you have 0:13:30.440 --> 0:13:34.280 probably had instances before where you can't read sarcasm in 0:13:34.280 --> 0:13:39.440 an email Why why is this person so angry at me? 0:13:39.640 --> 0:13:42.280 And then you need to include a winky face right there, 0:13:42.320 --> 0:13:44.640 like I was just joking, and then you feel foolish. 0:13:44.679 --> 0:13:47.520 But I mean, the lack of all this context that 0:13:47.559 --> 0:13:50.120 we get from body language and tone and stuff can 0:13:50.160 --> 0:13:54.800 be difficult, and in fact, that's part of the application 0:13:54.960 --> 0:13:56.680 of this, but I'll get into that in a second. 0:13:56.760 --> 0:14:00.960 So essentially, what this tone and lizer is doing is 0:14:01.000 --> 0:14:04.160 it's it's examining all the words within that block of 0:14:04.240 --> 0:14:10.280 text and tagging those words based upon how a predetermined 0:14:10.400 --> 0:14:14.240 categorization has been set up for the tone analyzer. So 0:14:14.440 --> 0:14:18.440 certain words fall under the general category of cheerful. That's 0:14:18.440 --> 0:14:22.440 a great example because it pops up a lot and uh. 0:14:22.560 --> 0:14:26.200 Then it ends up giving you a percentage of the 0:14:26.280 --> 0:14:29.720 overall tone of the piece. You can actually do it 0:14:29.760 --> 0:14:31.400 in two ways. You can get a word count or 0:14:31.440 --> 0:14:34.160 you can get a percentage approach. The word count just 0:14:34.240 --> 0:14:36.760 tells you how many of the words within a given 0:14:36.800 --> 0:14:40.440 block of text fall into each category, So whether it's social, 0:14:40.560 --> 0:14:42.480 or it's a writing que or it's an emotional cue. 0:14:42.680 --> 0:14:45.760 And then the percentage gives you kind of a what 0:14:45.760 --> 0:14:49.000 what is the thrust? Like? What was the impact on 0:14:49.120 --> 0:14:53.320 the reader? So that suggests that there are waitings to 0:14:53.440 --> 0:14:56.640 these words, right, that certain words are weighted as being 0:14:56.640 --> 0:14:59.840 more powerful. I found the percentage results to be more 0:15:00.000 --> 0:15:03.040 interesting because the word count results just tell you most 0:15:03.040 --> 0:15:06.640 of your words or social orientation, which could be simple 0:15:06.680 --> 0:15:12.000 things like the and but or like that doesn't have 0:15:12.280 --> 0:15:15.120 that doesn't impact a lot unless you're using them really well. 0:15:15.440 --> 0:15:20.880 But the words that were or butts can be extremely 0:15:21.200 --> 0:15:29.760 extremely uh powerful. Yes, yes, but but Ideally, what this 0:15:29.800 --> 0:15:33.480 would allow you to do is tell how your work 0:15:33.640 --> 0:15:37.640 might impact a reader before the reader has a chance 0:15:37.680 --> 0:15:42.560 to to see this written work, and you might be 0:15:42.600 --> 0:15:45.720 able to then revise that work if it doesn't seem 0:15:45.760 --> 0:15:49.080 like it's getting across the message you had intended. So 0:15:49.840 --> 0:15:53.720 it's it's kind of like a hey, listen, uh, I 0:15:53.760 --> 0:15:56.960 know you think you might be coming across this way, 0:15:56.960 --> 0:15:59.600 but this is analytically how you are coming across. Do 0:15:59.640 --> 0:16:02.760 you want to option to change up some of this stuff? Yeah, 0:16:02.760 --> 0:16:06.000 maybe you could change some of these aggressive words to 0:16:06.040 --> 0:16:08.880 be more friendly, for example, yeah yeah. Or it may 0:16:08.880 --> 0:16:10.840 be that if it's in a business letter, maybe the 0:16:10.840 --> 0:16:13.320 aggressive tone is exactly what you wanted. So it gives 0:16:13.320 --> 0:16:16.120 you the opportunity to go the other way. And in fact, 0:16:16.160 --> 0:16:19.200 what will happen is when you use the the app, 0:16:19.360 --> 0:16:22.720 it highlights each word or most of the words, almost 0:16:22.720 --> 0:16:25.480 all of them in a block of text, and then 0:16:25.720 --> 0:16:28.760 by clicking on it, you can have options to change 0:16:28.800 --> 0:16:32.280 that word to to a synonym, and it even categorizes 0:16:32.320 --> 0:16:36.880 the synonym according to the various social writing or emotional cues. 0:16:37.120 --> 0:16:39.720 So if you wanted to make it a more emotional plea, 0:16:40.040 --> 0:16:43.440 you could do that by going under the emotional category 0:16:43.800 --> 0:16:47.720 and choosing a synonym under there, and presumably this would 0:16:47.760 --> 0:16:52.280 make your presentation or letter or whatever it is more 0:16:52.520 --> 0:16:55.040 have more of an emotional impact. I didn't I didn't 0:16:55.120 --> 0:16:57.880 understand some of the options it was giving me with 0:16:57.960 --> 0:17:01.240 the synonyms, because at one point it's suggested to make 0:17:01.320 --> 0:17:04.560 my thing more agreeable, I changed the word bad to 0:17:04.640 --> 0:17:09.680 the word lousy. Yeah. Yeah, I mean, well it's more 0:17:10.200 --> 0:17:12.600 jokey like, kind of like like it's got I think 0:17:12.640 --> 0:17:19.359 that lousy has a slightly softer connotation. Lousy just bad, 0:17:19.440 --> 0:17:22.400 I would say, comes across as a little more playful, right, 0:17:22.520 --> 0:17:24.600 Like if I were to say Lauren is a bad 0:17:24.680 --> 0:17:28.840 person versus Lauren is a lousy person. Yeah. Yeah, you know, 0:17:29.280 --> 0:17:31.600 if you're just a lousy person, then you know you 0:17:31.600 --> 0:17:36.760 can pick those louses off of that shampoo. Yeah, much 0:17:36.800 --> 0:17:40.119 better than being bad. Yeah. It reminds me of that 0:17:40.200 --> 0:17:44.880 amazing Michael Jackson song Lousy video for that was amazing. 0:17:45.800 --> 0:17:49.000 So the idea of being with a tool like this 0:17:49.080 --> 0:17:52.920 in the future, you could actually build this functionality directly 0:17:52.960 --> 0:17:55.439 into something else like a word processor or an email 0:17:55.480 --> 0:17:59.439 program or ta most useful. Right, the idea being that, 0:18:00.119 --> 0:18:03.159 or you save a document before you send an email, 0:18:03.200 --> 0:18:05.760 before you send off that text message, or update your 0:18:05.800 --> 0:18:09.159 Facebook post. The analyzer could tell you, oh, hey, by 0:18:09.160 --> 0:18:11.040 the way, you're coming across as a told jerk face, 0:18:11.160 --> 0:18:15.159 this should be beside the tweet button. Yes, it should 0:18:15.200 --> 0:18:18.200 be well and I think, you know, making it something 0:18:18.240 --> 0:18:22.040 that is um like word count or spell check, spell 0:18:22.160 --> 0:18:24.399 check or even grammar check these days. Yeah, I think 0:18:24.440 --> 0:18:25.879 as long as it didn't get in the way of 0:18:25.920 --> 0:18:27.920 what you were doing, if it were an optional thing 0:18:27.960 --> 0:18:30.240 for you to check, or it was you know, unobtrusive 0:18:30.480 --> 0:18:33.800 and be great, underlined it in in fuchia or something 0:18:33.880 --> 0:18:36.280 like that, you know, whatever color is left over, right, 0:18:36.320 --> 0:18:39.600 so exactly something that's not going to easily be confused 0:18:39.640 --> 0:18:42.199 with the colors were already using for everything else with 0:18:42.280 --> 0:18:46.480 grammar and spelling errors. But it would be helpful, and uh, 0:18:46.560 --> 0:18:49.200 it might mean that you would avoid some of those 0:18:49.200 --> 0:18:51.960 situations where you dash off a message to someone thinking 0:18:52.000 --> 0:18:55.320 it's perfectly fine and they receive it and are offended 0:18:55.440 --> 0:18:58.520 or confused or or hurt or whatever that may be. Well, 0:18:58.640 --> 0:19:01.439 let's talk about some examples. Yeah, so there is a 0:19:01.480 --> 0:19:04.680 demo available online. Uh, you can go to the tone 0:19:04.720 --> 0:19:07.439 analyzer website and actually try the live demo. It allows 0:19:07.440 --> 0:19:10.240 you to paste in or type in a block of 0:19:10.320 --> 0:19:14.000 text and then you can analyze it um And so 0:19:14.400 --> 0:19:17.000 I decided I would throw in. It took me a 0:19:17.040 --> 0:19:20.600 long time to decide which opening line I would use 0:19:20.640 --> 0:19:22.800 from all the different novels that I love, but I 0:19:22.880 --> 0:19:25.240 chose this one. See if you recognize it. It is 0:19:25.240 --> 0:19:28.880 a truth universally acknowledged that a single man in possession 0:19:28.920 --> 0:19:33.400 of a good fortune must be in want of a wife. Yes, 0:19:33.520 --> 0:19:39.480 is from Black Books. The series actually is used in 0:19:39.480 --> 0:19:42.320 Black Books, but Black but that's not obviously where it's from. 0:19:42.320 --> 0:19:45.879 At any rate, it's Jane Austen quote obviously. So the 0:19:45.880 --> 0:19:50.240 analyzer said that the emotional tone was cheerful, the social 0:19:50.240 --> 0:19:57.679 tone was open and agreeable, and the writing tone was analytical. Uh. 0:19:57.880 --> 0:20:00.399 Looking at a word count like Joe was saying, not 0:20:00.520 --> 0:20:02.879 as interesting. So I just went with the percentiles. But 0:20:03.600 --> 0:20:05.720 if you did look at the word count, only five 0:20:05.720 --> 0:20:09.320 percent the sentence is in an emotional tone, in social 0:20:09.359 --> 0:20:12.040 tone and five percent in writing tone just by words. 0:20:12.280 --> 0:20:15.320 But when it looks at the impact, the emotional impact 0:20:15.359 --> 0:20:18.640 and the UH and the social um or the writing 0:20:18.800 --> 0:20:22.679 impact rather is much greater than the social impact. So 0:20:22.720 --> 0:20:24.800 in that case it was almost like a third or 0:20:24.840 --> 0:20:27.240 third and a third. So even though there were fewer 0:20:27.520 --> 0:20:31.639 writing cues and emotional cues in the sentence I picked, 0:20:32.119 --> 0:20:34.560 they had a greater impact than the social cues did, 0:20:35.160 --> 0:20:39.040 which I thought was kind of interesting. But it also 0:20:39.119 --> 0:20:41.040 shows that when you when you do this, you see 0:20:41.080 --> 0:20:44.399 how how the analyzer is picking out each word and 0:20:44.440 --> 0:20:47.680 classifying it. Yeah, it breaks down by color. Yeah. Yeah, 0:20:47.680 --> 0:20:50.120 So if you are looking at the emotional words, those 0:20:50.200 --> 0:20:53.920 are highlighted in shades of red or pink. Uh. And 0:20:53.960 --> 0:20:56.440 you know, the the emotional one is divided up a 0:20:56.520 --> 0:20:58.800 little bit too, so that way, like cheerful is one 0:20:58.800 --> 0:21:00.840 of them, and I think that's a very pink color 0:21:00.920 --> 0:21:02.800 that they used for cheerful, And so all the cheerful 0:21:02.840 --> 0:21:06.800 words were bright pink um. Then social words were all 0:21:06.800 --> 0:21:09.520 in shades of blue, and writing tone words well in 0:21:09.520 --> 0:21:12.360 shades of green. And if you clicked on them, that's 0:21:12.359 --> 0:21:15.040 where it would give you the option to switch those out, 0:21:15.119 --> 0:21:16.919 so that you can you know, if you did you 0:21:19.640 --> 0:21:22.840 did you really mean in want of a wife? Uh? 0:21:23.160 --> 0:21:26.240 But also I should point out that the results we 0:21:26.320 --> 0:21:28.040 get are you know, one thing you have to keep 0:21:28.040 --> 0:21:30.200 in mind is what is the basis of comparison, because 0:21:30.200 --> 0:21:35.680 it's not just a universal you know, text analyzer, it's 0:21:35.680 --> 0:21:39.399 actually analyzing it against a standard. And in this case, 0:21:39.720 --> 0:21:42.200 the standard they were using was the standard you would 0:21:42.240 --> 0:21:45.240 use for business letters. So according to you know, compared 0:21:45.280 --> 0:21:49.439 to your business letters standard, Jane Austen, you know, it's cheerful, 0:21:49.680 --> 0:21:53.600 so I'll buy that. But yeah, it's one of the 0:21:53.600 --> 0:21:56.160 other things is that I think a full tone analyzer 0:21:56.200 --> 0:21:59.119 would have different comparisons you could use, not just the 0:21:59.119 --> 0:22:02.439 business letter roach, it would also like compare you to 0:22:02.600 --> 0:22:06.720 threatening letters from creditors. Yeah, and and maybe maybe we 0:22:06.760 --> 0:22:10.280 can finally get that direct computer comparison of Jane Austen 0:22:10.359 --> 0:22:13.159 to like Hemingway or something like that. It would be interesting. So, 0:22:13.600 --> 0:22:16.480 you know, I know there was an article in which 0:22:16.520 --> 0:22:20.160 an author, uh or the writer of the articles said 0:22:20.240 --> 0:22:24.040 that they had compared themselves to Mark Twain. Oh yeah, 0:22:24.080 --> 0:22:26.639 I read that, yeah, which was entertaining, and said that 0:22:26.760 --> 0:22:29.760 according to the analysis, the two wrote in a very 0:22:29.800 --> 0:22:33.280 similar way, like the percentages came out in a similar way, 0:22:33.640 --> 0:22:36.399 and actually used that as a means of talking about 0:22:36.520 --> 0:22:39.160 the limitations of that but Joe, you used an example 0:22:39.200 --> 0:22:42.840 that also kind of showed some limitations of the tone analyzer. Yeah. 0:22:42.920 --> 0:22:45.480 I was like, well, what's it gonna make of something 0:22:45.520 --> 0:22:49.919 really philosophical? So I put in a quote from Dustevsky's 0:22:50.000 --> 0:22:55.159 Notes from the Underground. Okay, cheerful stuff that the underground 0:22:55.160 --> 0:22:58.399 man is talking in his long sort of diary section 0:22:58.480 --> 0:23:02.159 in the first half, and he says I could not 0:23:02.280 --> 0:23:06.239