0.000000 2.700000 Tech It From Me is an independent and solo-produced podcast. 2.700000 5.880000 Welcome to TechEd for me podcast, I'm Mike Madole. 5.880000 9.720000 Artificial intelligence didn't just appear with chat GPT. 9.720000 15.080000 It's the result of nearly a century of breakthroughs in logic, neuroscience, and computing, 15.080000 20.580000 from Alan Turing's thought experiments to today's generative models reshaping how we work. 20.580000 24.680000 In this episode, we're taking a journey through the real history of AI, 24.680000 30.960000 no myths, or science fiction, just the key milestones that got us from punch cards to transformers. 30.960000 35.160000 And along the way, I'll share how some of what I've learned through certifications in 35.160000 41.440000 generative AI, ethics, and machine learning connects back to those foundational ideas. 41.440000 44.840000 This is Tech It From Me, let's go. 44.840000 47.080000 This is the Tech It From Me podcast. 47.080000 49.720000 All right, so let's set the stage here. 49.720000 55.240000 We're in the early 20th century, the world is rebuilding after two world wars, computing 55.240000 60.380000 as we know it doesn't really exist yet, but the idea of machines doing complex tasks 60.380000 62.480000 is starting to take root. 62.480000 65.640000 The first major leap comes from Alan Turing. 65.640000 71.800000 In 1936, years before he would help crack Nazi codes, he publishes a paper introducing 71.800000 74.480000 the concept of a universal machine. 74.480000 79.440000 This abstract machine could simulate any other machine's logic, and it becomes the foundation 79.440000 81.080000 of modern computing. 81.080000 84.240000 Today, we call it the Turing machine. 84.240000 86.820000 Fast forward to 1950. 86.820000 92.360000 Turing poses a question in his paper titled computer machinery and intelligence. 92.360000 94.320000 Can machines think? 94.320000 98.880000 To answer it, he proposes a test, what we now call the Turing test. 98.880000 102.640000 Imagine having a text-based conversation with someone. 102.640000 107.000000 If you couldn't tell whether it was a person or machine, the machine would be considered 107.000000 108.240000 intelligent. 108.240000 110.380000 This is more than just a thought experiment. 110.380000 115.960000 It was the first serious challenge to the idea of artificial intelligence that machines 115.960000 119.600000 could eventually simulate human reasoning. 119.600000 124.840000 Around the same time, you've got people like Claude Shannon, known as the father of information 124.840000 126.000000 theory. 126.000000 132.280000 He laid the groundwork for how computers encode, store, and transmit information, critical 132.280000 134.640000 for any intelligent system. 134.640000 138.460000 He even tinkered with early chess-playing programs. 138.460000 143.720000 Then there's John Vaughan Newman, who gave us the modern architecture of computing. 143.720000 147.960000 The idea that data instructions could be stored in the same memory. 147.960000 154.040000 Without this, there would be no chat GBT, no iPhones, no anything. 154.040000 158.880000 So by the mid-1950s, you had the math, the theory, and the hardware. 158.880000 161.000000 Or at least the blueprints. 161.000000 166.180000 Then came the summer of 1956, the Dartmouth conference. 166.180000 169.440000 This was the moment AI became its own field. 169.440000 176.080000 John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon gathered a group of researchers 176.080000 181.860000 to explore whether every aspect of human intelligence could be described precisely enough to be 181.860000 184.380000 simulated by a machine. 184.380000 189.920000 They wrote in their proposal, it is conjectured that every aspect of learning or any other 189.920000 195.600000 feature of intelligence can, in principle, be so precisely described that a machine can 195.600000 198.100000 be made to simulate it. 198.100000 203.780000 That quote became the rallying cry for a new scientific discipline, artificial intelligence. 203.780000 209.560000 Now, I want to pause here for a second because during my studies, this era was described 209.560000 213.360000 as being equal parts genius and hubris. 213.360000 218.560000 The researchers were incredibly smart, but also incredibly optimistic. 218.560000 224.360000 They thought human level AI was, 20 years away, spoiler alert, it wasn't. 224.360000 229.740000 Still, what they set in motion at Dartmouth would echo through every generation of computer 229.740000 231.820000 science that followed. 231.820000 235.640000 And that's where this story really begins on AI. 235.640000 239.200000 So after the Dartmouth conference, researchers got to work. 239.200000 244.840000 The late 1950s and 60s were full of promise, and honestly, some pretty impressive early 244.840000 249.200000 wins, given the limited computing power of the time. 249.200000 255.720000 Let's start with the logic theorist, developed in 1956 by Alan Newell and Herbert Simon at 255.720000 258.000000 Rand Corporation. 258.000000 264.360000 This program could prove mathematical theorems from Principia Mathematica, and even found 264.360000 268.240000 a more elegant proof than the original authors. 268.240000 269.680000 That was a big deal. 269.680000 275.800000 It showed that machines could reason, at least in narrow rule-based domains. 275.800000 284.080000 Then, there was Eliza, built by Joseph Weisenbaum at MIT in 1966. 284.080000 292.200000 Eliza was an early natural language processing program that simulated a Rogerian psychotherapist. 292.200000 298.160000 It used pattern matching and substitution to respond to user's input in a way that felt 298.160000 299.160000 conversational. 299.160000 304.760000 So, here's an example, a user, "I'm feeling anxious about work." 304.760000 308.760000 Eliza would respond, "Why do you say you're feeling anxious about work?" 308.760000 314.720000 It was simple, almost mechanical, but it created the illusion of understanding. 314.720000 317.560000 People became emotionally attached to it. 317.560000 320.720000 Weisenbaum was actually disturbed by this. 320.720000 327.600000 He intended Eliza as a parody, not a therapist, but people poured their hearts out to it. 327.600000 334.000000 This is one of the earliest ethical red flags in AI, a reminder that just because something 334.000000 338.480000 sounds intelligent doesn't mean it necessarily is. 338.480000 342.640000 Same could be said about some humans today, but I digress. 342.640000 349.040000 In this theme, user perception versus actual intelligence is still relevant today, especially 349.040000 352.080000 with tools like ChatGPT. 352.080000 358.600000 Despite these early examples, the reality behind the scenes was clunky to say the least. 358.600000 364.440000 These systems worked well in controlled environments, but they couldn't deal with ambiguity, nuance, 364.440000 367.560000 or real world variability. 367.560000 370.440000 Take machine translation, for example. 370.440000 377.360000 In the 1960s, the US government heavily funded programs to translate Russian into English. 377.360000 381.600000 They hoped to decode Cold War intelligence faster. 381.600000 384.360000 But the results were laughably bad. 384.360000 391.720000 There's a famous example where "the spirit is willing, but the flesh is weak." 391.720000 398.360000 This was translated to Russian, and then back again, and it came out as "the vodka is good, 398.360000 401.840000 but the meat is rotten." 401.840000 406.960000 Expectations were far ahead of what the reality actually was. 406.960000 411.840000 By the early 1970s, the honeymoon was really over. 411.840000 417.280000 Governments realized that AI wasn't delivering on its promises, so funding was cut and projects 417.280000 419.080000 were shelved. 419.080000 426.120000 And it brings us to what's now called the first AI winter, a period of deep skepticism, 426.120000 429.680000 limited progress, and reduced investment. 429.680000 435.240000 From a business and policy standpoint, AI was seen as overhyped. 435.240000 439.560000 From a technical standpoint, the tools weren't just ready yet. 439.560000 446.120000 Computers were still too slow, storage was limited, and models simply couldn't scale. 446.120000 451.120000 The biggest mistake of this era wasn't technical, it was narrative. 451.120000 453.480000 The field over-promised and under-delivered. 453.480000 460.640000 And when you lose trust, whether with funders, users, or the public, you don't just lose momentum, 460.640000 462.600000 you lose credibility. 462.600000 466.320000 And that's a lesson we're still learning today. 466.320000 468.600000 But here's the thing about winters. 468.600000 470.480000 They don't last forever. 470.480000 475.680000 And AI, as a field, it wasn't dead, it was simply regrouping. 475.680000 481.480000 Waiting for that next spark and a spark came in the form of a very different approach. 481.480000 487.640000 Coming out of that first AI winter, the 1980s brought a new approach to artificial intelligence, 487.640000 492.480000 one that looked less like mimicking the human brain and more like mimicking human 492.480000 494.280000 expertise. 494.280000 497.920000 This was the rise of expert systems. 497.920000 503.920000 Unlike earlier attempts to recreate general intelligence, expert systems focused on narrowly 503.920000 506.040000 defined problems. 506.040000 508.040000 The concept was straightforward. 508.040000 513.680000 Take the knowledge of a domain expert, such as a doctor or a mechanic, and encode it as 513.680000 516.720000 a set of rules the computer could follow. 516.720000 519.960000 And for a while, this approach worked. 519.960000 525.360000 One of the most famous systems was Mycin, developed at Stanford. 525.360000 532.760000 Mycin, spelled M-Y-C-I-N, diagnosed bacterial infections and recommended treatments. 532.760000 537.280000 In control tests, it even outperformed junior doctors. 537.280000 542.960000 But despite the accuracy, Mycin was never deployed in real hospitals. 542.960000 544.760000 You might ask why. 544.760000 550.040000 When the answer is quite simple, it couldn't explain its reasoning in plain language. 550.040000 555.400000 And when it comes to life and death decisions, people need to understand why the machine made 555.400000 557.680000 its recommendation. 557.680000 560.680000 That lack of transparency created a trust gap. 560.680000 565.960000 And that's something we still deal with today in AI systems, particularly when decisions 565.960000 571.400000 affect people's health, finances, or legal outcomes. 571.400000 575.560000 Take this, expert systems caught on in the corporate world. 575.560000 581.280000 Companies like American Express and Digital Equipment Corporation invested in internal tools 581.280000 585.480000 for credit scoring, troubleshooting, and support. 585.480000 587.120000 The logic was appealing. 587.120000 592.400000 Capture what your best people know and use it to scale decision making. 592.400000 595.160000 But expert systems had a critical flaw. 595.160000 597.000000 They didn't learn. 597.000000 599.920000 Every rule had to be programmed by hand. 599.920000 606.760000 If the world changed, say, fraud tactics evolved or new regulations came in, someone had to update 606.760000 608.280000 the rules. 608.280000 613.160000 They were only as smart as the last time someone told them what to do. 613.160000 617.840000 As complexity grew, maintaining these rule-based systems became a burden. 617.840000 618.840000 They couldn't adapt. 618.840000 623.520000 They couldn't generalize, and they couldn't improve on their own. 623.520000 627.560000 By the late 1980s, companies started abandoning them. 627.560000 630.840000 The second AI winter had begun. 630.840000 633.260000 This one hit harder, though. 633.260000 637.560000 Businesses had poured real money into expert systems. 637.560000 642.120000 Expectations were higher, and when those systems couldn't keep up, budgets were slashed, 642.120000 645.640000 and reputations were ultimately damaged. 645.640000 647.320000 But there was a silver lining. 647.320000 653.600000 Expert systems taught us a lot about how humans reason under constraints, and about the importance 653.600000 657.760000 of explainability, maintenance, and scale. 657.760000 659.560000 They also made one thing clear. 659.560000 665.400000 If AI was going to survive and grow, it had to move beyond fixed rules. 665.400000 668.320000 It had to start learning. 668.320000 673.160000 Coming out at a second AI winter, the field of artificial intelligence found its footing 673.160000 674.240000 again. 674.240000 676.960000 But this time with a different mindset. 676.960000 682.400000 Instead of trying to teach machines how to think like humans using hard-coded rules, 682.400000 687.960000 researchers started asking, "What if we let machines learn from examples?" 687.960000 693.120000 This was the rise of machine learning, and it marked a major pivot. 693.120000 696.600000 The difference is simple, but profound. 696.600000 702.680000 Expert systems are programmed, meaning someone tells them exactly what to do. 702.680000 704.880000 Machine learning models are trained. 704.880000 708.160000 They learn patterns from data. 708.160000 714.160000 Other than defining what a cat looks like in code, you feed the system thousands of cat 714.160000 720.520000 and non-cat images, and it figures out the patterns that distinguish one from the other. 720.520000 721.920000 This idea wasn't new. 721.920000 727.800000 It dated back to early neural networks in the 1950s, but back then, we didn't have the 727.800000 731.880000 data or the computing power to make it practical. 731.880000 738.760000 By the 1990s and 2000s that had changed, three things came together. 738.760000 744.640000 One, more data, thanks to the internet and digital sensors, we had access to massive data 744.640000 745.640000 sets. 745.640000 748.520000 Number two, better algorithms. 748.520000 754.400000 New techniques like support vector machines and random forests improved accuracy. 754.400000 757.280000 And number three, faster computing. 757.280000 762.920000 More powerful processors made training models faster and more affordable. 762.920000 766.880000 Machine learning started quietly transforming industries. 766.880000 773.040000 Email providers used it for spam detection, banks applied it to detect credit card fraud, 773.040000 777.520000 and retailers optimized recommendation engines. 777.520000 780.680000 So one great example, spam filters. 780.680000 786.880000 In the early days, filters looked for specific keywords, but spammers adapted quickly. 786.880000 792.240000 Machine learning systems, however, could adapt just as fast, spotting subtle patterns that 792.240000 794.600000 weren't obvious to humans. 794.600000 797.320000 The same thing happened in finance. 797.320000 801.560000 Traditional fraud rules flagged fixed dollar amounts. 801.560000 807.560000 Machine learning models flagged behavioral anomalies, like a sudden transaction in a country 807.560000 809.680000 you've never visited. 809.680000 815.080000 This wasn't just automation, it was adaptation, and it was quite powerful. 815.080000 820.560000 For most of us, this was the era where AI started quietly working in the background, shaping 820.560000 825.480000 search engines, social media feeds, and even how we shop. 825.480000 831.560000 And while machine learning made AI useful again, the real revolution was still ahead. 831.560000 837.200000 Because what if machines could not just learn from data, but learn in layers, and extract 837.200000 843.000000 meaning from complex inputs like images, speech, and language. 843.000000 847.080000 That's what brought us into the next phase, deep learning. 847.080000 854.120000 By the early 2010s, machine learning had made AI practical again, but it still had limits. 854.120000 858.920000 Many models needed human assistance to identify which features mattered. 858.920000 864.040000 If you were building a model to identify cats and images, for example, a data scientist 864.040000 868.680000 still had to define things like pointy ears or whiskers. 868.680000 874.960000 It's known as feature engineering, and it was time consuming and inflexible. 874.960000 878.880000 But what if a system could learn those features on its own? 878.880000 881.880000 That's where deep learning changed the game. 881.880000 886.920000 Deep learning is a subset of machine learning that uses neural networks with many layers, 886.920000 889.400000 hence the word deep. 889.400000 894.360000 Each layer processes the data at a different level of abstraction. 894.360000 901.480000 The first might detect edges in a photo, the next shapes, then eyes, then entire faces, 901.480000 904.920000 it builds understanding as it moves deeper. 904.920000 909.560000 These architectures were inspired by how we think the human brain works. 909.560000 912.680000 But for years they didn't work well in practice. 912.680000 917.320000 The training took too long, the results weren't better than other methods. 917.320000 919.600000 Then came 2012. 919.600000 925.520000 A deep learning model called AlexNet entered the ImageNet competition, a benchmark for 925.520000 927.800000 image classification. 927.800000 934.520000 It dramatically outperformed everything else, cutting the error rate by nearly half. 934.520000 937.280000 Why was AlexNet such a big deal? 937.280000 943.480000 Because it showed deep learning could outperformed traditional models if you had the right ingredients. 943.480000 950.960000 Those ingredients were large data sets, GPU acceleration, and improved training techniques, 950.960000 955.040000 like relu, activations, and dropout. 955.040000 957.440000 The ImageNet win was a turning point. 957.440000 960.440000 Suddenly, deep learning wasn't just theoretical. 960.440000 964.320000 It was stated the art, and companies noticed. 964.320000 970.280000 Google, Facebook, Amazon, Microsoft, they all began building deep learning into their 970.280000 971.280000 products. 971.280000 979.480000 AI could now identify faces and objects in photos, transcribe and understand speech, translate 979.480000 987.680000 languages in real time, drive cars, detect tumors, recommend content, in marketing, deep learning 987.680000 994.720000 models started to write personalized ad copy, predict campaign performance, generate branded 994.720000 1000.840000 visuals based on product descriptions, and for the first time AI could handle unstructured 1000.840000 1010.000000 data, images, video, audio, and natural language just as easily as structured numbers in a spreadsheet. 1010.000000 1015.000000 But deep learning came with a cost, and that was explainability. 1015.000000 1018.680000 These models were accurate, but they were also opaque. 1018.680000 1022.880000 You couldn't always trace why they made a certain decision. 1022.880000 1026.560000 Even today, we often refer to them as black boxes. 1026.560000 1032.640000 It's a big problem in high stakes domains like health care, finance, and law. 1032.640000 1037.120000 It's also part of a broader conversation we're still having today. 1037.120000 1040.680000 How do we balance capability with control? 1040.680000 1046.920000 Still, the success of deep learning set the stage for the next and current phase of AI's 1046.920000 1048.400000 evolution. 1048.400000 1053.240000 Because if deep learning gave AI the power to see and hear what came next gave it the 1053.240000 1060.000000 power to understand and communicate that being transformers. 1060.000000 1067.840000 In 2017, a paper titled Attention is All You Need was published by researchers at Google. 1067.840000 1074.120000 With it came a brand new architecture called the Transformer, and it fundamentally changed 1074.120000 1077.120000 the trajectory of AI. 1077.120000 1085.440000 Up until then, most models dealing with language relied on recurrent neural networks, or RNNs, 1085.440000 1089.800000 which processed language sequentially one word at a time. 1089.800000 1097.360000 That meant training was slow and these models struggled to retain context over long passages. 1097.360000 1100.320000 Transformers flipped that approach on its head. 1100.320000 1106.480000 Instead of processing words one after another, they used a mechanism called self-attention, 1106.480000 1111.800000 which allowed the model to look at all the words in a sentence or even an entire paragraph 1111.800000 1116.840000 simultaneously and understand the relationships between them. 1116.840000 1119.920000 So let's look at an example. 1119.920000 1123.920000 The cat sat on the mat because it was tired. 1123.920000 1131.080000 A Transformer-based model can figure out that the word it refers to the cat and not the 1131.080000 1136.240000 mat because it considers the full sentence at once. 1136.240000 1142.240000 This ability to model relationships across long text makes Transformers incredibly good 1142.240000 1145.720000 at understanding meaning and context. 1145.720000 1150.560000 This architecture became the foundation for modern AI language models. 1150.560000 1158.160000 Burts developed by Google, GPT developed by OpenAI, and dozens of others from meta 1158.160000 1169.880000 anthropic, mistrol, and more. 1169.880000 1176.880000 A model capable of writing essays, summarizing content, translating languages, and even generating 1176.880000 1178.360000 code. 1178.360000 1185.960000 But it wasn't until late 2022, when chat GPT made GPT 3.5 publicly accessible that the 1185.960000 1188.240000 world truly took notice. 1188.240000 1194.880000 Suddenly, anyone, not just developers or researchers, could chat with an AI that sounded 1194.880000 1197.120000 almost human. 1197.120000 1205.880000 We entered the era of generative AI, and with it came a tidal wave of real-world applications, 1205.880000 1212.720000 automated writing for blogs, ads, and emails, content summarization, and meeting notes. 1212.720000 1219.000000 My assistance embedded into search engines and productivity software, image generation 1219.000000 1225.880000 from text prompts, voice cloning, video generation, and more. 1225.880000 1232.240000 Generative AI quickly became a part of the everyday toolkit, not just for tech professionals, 1232.240000 1236.520000 but for marketers, students, writers, and small businesses. 1236.520000 1239.880000 But with all this capability came new questions. 1239.880000 1244.360000 How do we know when an AI is hallucinating, or making things up? 1244.360000 1248.680000 What happens when biases from training data show up in outputs? 1248.680000 1251.880000 Who owns AI-generated content? 1251.880000 1256.320000 These questions aren't just philosophical, they're practical, and they're urgent. 1256.320000 1262.760000 And while the promise of these tools is massive, so are the risks, that's why the conversation 1262.760000 1267.600000 is shifting from what can AI do to what should it do? 1267.600000 1270.640000 And how do we build in accountability? 1270.640000 1276.040000 Because whether it's writing a press release, designing a logo, or giving business advice, 1276.040000 1281.680000 AI is no longer just in the background, it's sitting in the room with us. 1281.680000 1285.240000 So over the last 20 minutes, we've covered how we got here. 1285.240000 1288.120000 Now let's talk about what's next. 1288.120000 1293.480000 We're living in a moment where AI is evolving rapidly, and the road ahead could go in a few 1293.480000 1300.840000 different directions, but most of the conversation today centers around two major trajectories. 1300.840000 1305.880000 Path 1 is a smarter, broader AI. 1305.880000 1312.560000 This path we're already on, building smarter, more capable, and more integrated AI tools. 1312.560000 1317.560000 We're seeing language models that don't just generate text, they can interpret images, 1317.560000 1323.080000 they can write code, analyze video, and connect across different types of input. 1323.080000 1327.520000 These represent the next leap in general usability. 1327.520000 1333.520000 Companies are embedding these models into everyday software, Microsoft's co-pilot in Word, Excel, 1333.520000 1341.920000 and Teams, Google's Gemini in Docs and Gmail, Salesforce's Einstein GPT for customer service, 1341.920000 1347.400000 AI is becoming a layer, something that enhances almost every digital experience. 1347.400000 1354.880000 There's also a growing ecosystem of open source models, like Meta's Lama or Mistral, 1354.880000 1362.200000 which are smaller, more customizable, and available to developers and researchers around the world. 1362.200000 1367.200000 This is really powerful, but it also makes regulation harder, because now anyone can 1367.200000 1373.200000 build powerful AI tools with fewer restrictions or oversight. 1373.200000 1380.080000 The second path is toward artificial general intelligence, or AGI. 1380.080000 1384.280000 This one's more speculative, but also more ambitious. 1384.280000 1391.720000 It's the pursuit of artificial general intelligence, or AGI, an AI that can reason, learn, and 1391.720000 1396.560000 apply knowledge across any task the way a human can. 1396.560000 1404.720000 Some leading AI labs like OpenAI and DeepMind believe AGI is possible within the next decade. 1404.720000 1410.360000 Others are more cautious arguing that while models like GPT are impressive, they still rely 1410.360000 1415.680000 on pattern prediction and not true understanding. 1415.680000 1421.960000 And there's a third group, Ephesus, researchers, and policy advocates focused on making sure 1421.960000 1427.640000 that regardless of the outcome, we don't lose control of the systems we're building. 1427.640000 1434.720000 These folks talk about alignment, interpretability, and safety, and their concern is this. 1434.720000 1441.360000 Even narrow AI, if deployed at scale and without oversight, can have unintended or catastrophic 1441.360000 1443.600000 consequences. 1443.600000 1449.920000 From job displacement to misinformation to deepfakes, the risks aren't just theoretical 1449.920000 1450.920000 anymore. 1450.920000 1452.720000 We're already here. 1452.720000 1456.480000 That's why governments around the world are now stepping in. 1456.480000 1460.760000 The European Union has passed the world's first AI law. 1460.760000 1466.760000 Canada, the US, and others are drafting policy frameworks, but regulation alone won't 1466.760000 1468.680000 solve everything. 1468.680000 1475.080000 What we really need are more people in business, in education, and in public life, who understand 1475.080000 1479.000000 both what AI is and what it isn't. 1479.000000 1485.000000 Because AI isn't just a technical challenge, it's an economic one, it's a social one, it's 1485.000000 1486.600000 a human one. 1486.600000 1491.400000 In the more voices we have at the table, the better chance we have of guiding it towards 1491.400000 1495.840000 something that truly benefits everyone. 1495.840000 1497.960000 So here's my take. 1497.960000 1504.640000 We've gone from rule-based systems to machine learning, to generative models that can write, 1504.640000 1505.840000 draw, and interact. 1505.840000 1508.440000 It's an incredible arc. 1508.440000 1512.760000 But every leap in capability comes with new blind spots. 1512.760000 1518.800000 So we look ahead, let's not just chase what's possible, let's focus on what's responsible. 1518.800000 1524.680000 The future of AI isn't just about what we build, it's about what we choose to build and 1524.680000 1525.680000 why. 1525.680000 1530.280000 Thanks for listening, and coming along on this journey through the history of artificial 1530.280000 1532.000000 intelligence. 1532.000000 1537.160000 If this episode sparked a new insight, helped you connect some dots, or gave you something 1537.160000 1541.600000 to think about, consider sharing it with someone else who might enjoy it. 1541.600000 1547.280000 You can follow Tech It From Me on Spotify, Apple, or wherever you get your podcasts from. 1547.280000 1550.800000 And if you got some feedback, or simply just want to say hello, I'd love to hear from 1550.800000 1551.800000 you. 1551.800000 1556.480000 Drop me a line at Inspire@TechItFromMe.com. 1556.480000 1562.600000 Until next time, thanks for listening, I'm Mike Madole.