XChrisUnknownX/SpeechTimer.py

## SpeechTimer.py
import time
import random
article = "I would like to take the time to demonstrate a little something that I discussed with others on Facebook. It pertains to AI and stenography. The writer wrote that they had heard that an artificial intelligence or algorithm at a recent AAERT conference did very well transcribing words in real time. Of course, I don't deny that I believe one day technology will be able to do what we do, but I also am a bit of a skeptic when it comes down to this sort of thing because of some observations I have about human nature, which we'll get to later. The first part of this conversation that has to happen is that the environment under which artificial intelligence or automatic voice recognition succeeds is very important. You can throw a recording system just about anywhere in the same way you can throw a stenographer in there, but if the conditions are not ideal, the stenographer will outperform the voice recognition system. We can see this in two ways. One is to observe what we see around us. News articles have come out since as early as 2016 saying voice recognition is 98 percent accurate. The catch is that the accuracy is in their lab. Also consider that Alexa, Siri, and just about any speech recognition software I've seen have their brilliant moments, but then also have moments where they pause, stutter, do nothing, or fail to comprehend a single sentence. Secondly, we can test these things for ourselves. I recently tried out a voice recognition software called Live Transcribe while a colleague of mine was dictating in a closed room and it got him almost perfectly. I tried that same program in a bar, or recording a YouTube video from a speaker, or outside on a nice day, and it got precious little of what we had to say. Environment matters, but less so to an adaptable stenographic reporter. The second part of this conversation is that people have biases and beliefs. Let's go back to when I used the Live Transcribe and my colleague dictating. I just said it got him perfectly, but that's not actually true. It missed some punctuation and grammar. Perfectly is really just a word I was using for 'good enough.' Similarly, when people talk about AI and how good it is, expect them to use these exciting terms like perfect. People like to share excitement. They aren't ever going to give you a play by play that explains exactly what was wrong or right unless they make that their conscious objective. Third, let's not forget that computers can be rigged or trained to do specific tasks very easily. I have enough amateur programming skill that I can make words appear on a screen. If I really cared to, I could make them appear at a particular time or on a specific time scale. As a matter of fact, a good programmer figures out the simplest way to do something for a given task. As an example, there are video games where you can wander in a large open world and explore virtually forever. The trick is that the game only has to draw or render what you are looking at, and the rest of the world doesn't have to exist. It does this because to actually simulate an entire planet would be too much, even for modern hardware. With regard to speech recognition, you can test this out yourself using YouTube. If you upload a dot TXT transcript that is 100 percent accurate, it will actually be better than YouTube's speech recognition software. All of this is to say that when you think of computers solving problems, what you see with your eyes may not be what is actually happening. For a final example, in computer programming, if I want to put the word hello on the screen, I can simply tell it to put what they call a string that says hello on the screen. I can also have the computer search a list of letters and randomly choose letters until it matches the word hello. These things would be identical if you looked at the end product, but one takes a lot more time and work. Again, what you see with your eyes may not be what is actually happening. My final observation is a simple one. People love to sell things. People love to brag. Check out my new watch. It doesn't matter if the timing is off as long as you don't ask for the time. Check out my new shoes. It doesn't matter if the soles are worn, as long as you don't want to walk in them. How many times in life has someone sold you something you didn't need because they made you feel good? Some of the largest proponents of automated speech recognition are doing just that. They are trying to sell the future. The future is here. Buy our revolutionary new product and all your problems will go down the drain. While I'm happy to have people selling, I just want to warn every consumer and competitor, don't buy hype. Google Glass hasn't replaced our smart phones and virtual reality hasn't ended the traditional video game market. Voice recognition, as it is, doesn't equate to the death of stenography. So, while I am a huge technology lover, when people talk about AI and stenography, I am skeptical. The two major ways that I can see AI doing what we do, unless quantum computing becomes a reality, is machine learning or modeling the human brain. I have said before that modeling the human brain seems an arduous task on modern hardware. So then we turn to machine learning. Machine learning is basically providing the program with training data and an algorithm or set of instructions for how it should interpret the data and utilize what it has learned. Ultimately, the quality of the data and algorithm matter. This is why there have reports of so- called racist or discriminatory AI that choose a specific type of resume or mislabel pictures. Many in popular culture assume that these issues are the product of an inherently biased computer science team, but that's not likely the case. It's far more likely that the simple truth is when you're feeding a computer thousands or millions of data points, it's easy to overlook something trivial. In my own amateur coding, I have overlooked a number or misplaced a colon, and it has caused an entire program to crash or continue doing infinite calculations when I really only needed it to do one hundred thousand. So if you are someone interested in the field of stenography or stenographic court reporting, I say that you should jump in and give it a try. It is likely that someday technology will be able to transcribe speech perfectly, and there are actually some cases in which it is good enough today, as is the case with many automated customer service apps. That acknowledged, we do not know if technology will reach that point tomorrow or one hundred years from now, and it seems foolish to give up on an illustrious career based on fear or the say so of people whose business revolves around peddling VR to people."
test = True
if test == True:
	# logic
	article = "Hello everyone. I am making a demonstration video because I won't have the time to do out a full video until later in the week or possibly even next week. I just wanted to show everyone exactly what I was talking about with regard to AI and stenography. Basically, computers can be told to do things and solutions to problems can get very creative. In this particular example, I have simply timed this text to show up at about 60 WPM, and I am dictating it at about 60 WPM, and there is no voice recognition at all in this entire program. I will even release the computer code publicly so that you can examine it if that's your thing. So this just goes to the whole discussion about whether or not AI can do what we do -- and my point is not that it cannot, but my point is that what we are told or even what we see with our eyes may not be what's occurring. It's sad but true. People can hype something up for the purpose of selling it off to someone and making back their investment. It's the way the world works."
else:
	pass
alist = article.split()
wpm = input("Input WPM: ")
try:
	wpm = int(wpm)
except:
	print("ERROR WPM DEFAULT TO 225")
	wpm = 225
wps = wpm / 60
spw = 60 / wpm
for i in alist:
	time.sleep(spw)
	print(i,end=" ",flush = True)
	import time
	import random
	article = "I would like to take the time to demonstrate a little something that I discussed with others on Facebook. It pertains to AI and stenography. The writer wrote that they had heard that an artificial intelligence or algorithm at a recent AAERT conference did very well transcribing words in real time. Of course, I don't deny that I believe one day technology will be able to do what we do, but I also am a bit of a skeptic when it comes down to this sort of thing because of some observations I have about human nature, which we'll get to later. The first part of this conversation that has to happen is that the environment under which artificial intelligence or automatic voice recognition succeeds is very important. You can throw a recording system just about anywhere in the same way you can throw a stenographer in there, but if the conditions are not ideal, the stenographer will outperform the voice recognition system. We can see this in two ways. One is to observe what we see around us. News articles have come out since as early as 2016 saying voice recognition is 98 percent accurate. The catch is that the accuracy is in their lab. Also consider that Alexa, Siri, and just about any speech recognition software I've seen have their brilliant moments, but then also have moments where they pause, stutter, do nothing, or fail to comprehend a single sentence. Secondly, we can test these things for ourselves. I recently tried out a voice recognition software called Live Transcribe while a colleague of mine was dictating in a closed room and it got him almost perfectly. I tried that same program in a bar, or recording a YouTube video from a speaker, or outside on a nice day, and it got precious little of what we had to say. Environment matters, but less so to an adaptable stenographic reporter. The second part of this conversation is that people have biases and beliefs. Let's go back to when I used the Live Transcribe and my colleague dictating. I just said it got him perfectly, but that's not actually true. It missed some punctuation and grammar. Perfectly is really just a word I was using for 'good enough.' Similarly, when people talk about AI and how good it is, expect them to use these exciting terms like perfect. People like to share excitement. They aren't ever going to give you a play by play that explains exactly what was wrong or right unless they make that their conscious objective. Third, let's not forget that computers can be rigged or trained to do specific tasks very easily. I have enough amateur programming skill that I can make words appear on a screen. If I really cared to, I could make them appear at a particular time or on a specific time scale. As a matter of fact, a good programmer figures out the simplest way to do something for a given task. As an example, there are video games where you can wander in a large open world and explore virtually forever. The trick is that the game only has to draw or render what you are looking at, and the rest of the world doesn't have to exist. It does this because to actually simulate an entire planet would be too much, even for modern hardware. With regard to speech recognition, you can test this out yourself using YouTube. If you upload a dot TXT transcript that is 100 percent accurate, it will actually be better than YouTube's speech recognition software. All of this is to say that when you think of computers solving problems, what you see with your eyes may not be what is actually happening. For a final example, in computer programming, if I want to put the word hello on the screen, I can simply tell it to put what they call a string that says hello on the screen. I can also have the computer search a list of letters and randomly choose letters until it matches the word hello. These things would be identical if you looked at the end product, but one takes a lot more time and work. Again, what you see with your eyes may not be what is actually happening. My final observation is a simple one. People love to sell things. People love to brag. Check out my new watch. It doesn't matter if the timing is off as long as you don't ask for the time. Check out my new shoes. It doesn't matter if the soles are worn, as long as you don't want to walk in them. How many times in life has someone sold you something you didn't need because they made you feel good? Some of the largest proponents of automated speech recognition are doing just that. They are trying to sell the future. The future is here. Buy our revolutionary new product and all your problems will go down the drain. While I'm happy to have people selling, I just want to warn every consumer and competitor, don't buy hype. Google Glass hasn't replaced our smart phones and virtual reality hasn't ended the traditional video game market. Voice recognition, as it is, doesn't equate to the death of stenography. So, while I am a huge technology lover, when people talk about AI and stenography, I am skeptical. The two major ways that I can see AI doing what we do, unless quantum computing becomes a reality, is machine learning or modeling the human brain. I have said before that modeling the human brain seems an arduous task on modern hardware. So then we turn to machine learning. Machine learning is basically providing the program with training data and an algorithm or set of instructions for how it should interpret the data and utilize what it has learned. Ultimately, the quality of the data and algorithm matter. This is why there have reports of so- called racist or discriminatory AI that choose a specific type of resume or mislabel pictures. Many in popular culture assume that these issues are the product of an inherently biased computer science team, but that's not likely the case. It's far more likely that the simple truth is when you're feeding a computer thousands or millions of data points, it's easy to overlook something trivial. In my own amateur coding, I have overlooked a number or misplaced a colon, and it has caused an entire program to crash or continue doing infinite calculations when I really only needed it to do one hundred thousand. So if you are someone interested in the field of stenography or stenographic court reporting, I say that you should jump in and give it a try. It is likely that someday technology will be able to transcribe speech perfectly, and there are actually some cases in which it is good enough today, as is the case with many automated customer service apps. That acknowledged, we do not know if technology will reach that point tomorrow or one hundred years from now, and it seems foolish to give up on an illustrious career based on fear or the say so of people whose business revolves around peddling VR to people."
	test = True
	if test == True:
	# logic
	article = "Hello everyone. I am making a demonstration video because I won't have the time to do out a full video until later in the week or possibly even next week. I just wanted to show everyone exactly what I was talking about with regard to AI and stenography. Basically, computers can be told to do things and solutions to problems can get very creative. In this particular example, I have simply timed this text to show up at about 60 WPM, and I am dictating it at about 60 WPM, and there is no voice recognition at all in this entire program. I will even release the computer code publicly so that you can examine it if that's your thing. So this just goes to the whole discussion about whether or not AI can do what we do -- and my point is not that it cannot, but my point is that what we are told or even what we see with our eyes may not be what's occurring. It's sad but true. People can hype something up for the purpose of selling it off to someone and making back their investment. It's the way the world works."
	else:
	pass
	alist = article.split()
	wpm = input("Input WPM: ")
	try:
	wpm = int(wpm)
	except:
	print("ERROR WPM DEFAULT TO 225")
	wpm = 225
	wps = wpm / 60
	spw = 60 / wpm
	for i in alist:
	time.sleep(spw)
	print(i,end=" ",flush = True)