Will AI Actually Replace Us?

Welcome, dear listeners, to the Speak and Flow podcast, where we dive into unique strategies and stories to help you and your team achieve maximum potential and flow. Today, I have a leader in a very hot topic, AI, and I can't wait to dive into what is happening in the landscape. He's got some really great vision.

2

00:00:25,720 --> 00:00:31,379

Melinda Lee: for where it could go. And so, welcome, Michael Ostapenko!

3

00:00:31,800 --> 00:00:33,779

Michael Ostapenko: Thank you, Millian.

4

00:00:33,780 --> 00:00:36,860

Melinda Lee: Hi, Michael, founder of Citrix.

5

00:00:37,180 --> 00:00:39,080

Melinda Lee: Patrick's, right? Yeah.

6

00:00:39,080 --> 00:00:39,570

Michael Ostapenko: Yeah.

7

00:00:39,570 --> 00:00:45,319

Melinda Lee: Tell us, what are you excited about? Like, what is the vision that you want to take it?

8

00:00:47,230 --> 00:00:51,300

Michael Ostapenko: So, my vision is, to create,

9

00:00:51,970 --> 00:01:00,830

Michael Ostapenko: A system, like a artificial intelligence system, which would, be able to produce, Results, which are both

10

00:01:00,950 --> 00:01:04,320

Michael Ostapenko: Meaningful and have sense.

11

00:01:04,730 --> 00:01:09,320

Michael Ostapenko: This is, like, while, while today's, today's,

12

00:01:09,830 --> 00:01:15,730

Michael Ostapenko: systems, they're… they are more restricted to the limit, or limited to the

13

00:01:16,010 --> 00:01:20,529

Michael Ostapenko: former, which is the meaning. They're good at semantics.

14

00:01:20,720 --> 00:01:27,339

Michael Ostapenko: But they're really, really bad at, logic and, logical reasoning. So…

15

00:01:27,600 --> 00:01:31,530

Michael Ostapenko: That's when… why when you, like, use,

16

00:01:32,020 --> 00:01:39,919

Michael Ostapenko: chatbots, like ChatGPT or anything else, like bot, you can, you can, you can see that

17

00:01:40,420 --> 00:01:48,740

Michael Ostapenko: the output they produce, it's, it's really meaningful. It has, these connections, which are very natural.

18

00:01:49,330 --> 00:01:53,280

Michael Ostapenko: But at the same time, it… Sometimes,

19

00:01:53,690 --> 00:01:57,239

Michael Ostapenko: Create this, make these subtle errors.

20

00:01:57,460 --> 00:02:01,669

Michael Ostapenko: Which… An expert can, nowadays.

21

00:02:01,850 --> 00:02:03,570

Michael Ostapenko: What… which may…

22

00:02:03,830 --> 00:02:13,000

Michael Ostapenko: gone unnoticed by… by, like, general population, that… which is why it's so popular, but in general population, but not so much adopted by,

23

00:02:13,200 --> 00:02:15,589

Michael Ostapenko: Companies and businesses.

24

00:02:15,820 --> 00:02:17,930

Michael Ostapenko: It's not really reliable.

25

00:02:18,690 --> 00:02:24,129

Michael Ostapenko: And the real problem for that is the technology, the underlying technology.

26

00:02:26,120 --> 00:02:34,320

Michael Ostapenko: The neural networks in general, no matter What architecture they have, no matter…

27

00:02:34,990 --> 00:02:44,250

Michael Ostapenko: What algorithms, optimization algorithms, you use to, train this, network.

28

00:02:44,430 --> 00:02:52,580

Michael Ostapenko: No matter, like, What other, like, primitives which are used in these networks, you use.

29

00:02:53,170 --> 00:02:56,460

Michael Ostapenko: We never be able to,

30

00:02:56,770 --> 00:03:00,940

Michael Ostapenko: Reason… logically, because it's just not what they do.

31

00:03:02,140 --> 00:03:03,390

Michael Ostapenko: Mathematically.

32

00:03:04,930 --> 00:03:07,259

Michael Ostapenko: And, it's… it's just impossible.

33

00:03:07,610 --> 00:03:16,350

Michael Ostapenko: Now, It's possible to create a system which Somehow incorporates neural network.

34

00:03:16,570 --> 00:03:23,999

Michael Ostapenko: And there are various, like, ways to do it, but, the point is…

35

00:03:24,280 --> 00:03:37,830

Michael Ostapenko: If you do, like, then, you can leverage this… Neural network's ability to… Model semantics.

36

00:03:38,610 --> 00:03:44,470

Michael Ostapenko: And took somehow, Converge with this…

37

00:03:44,690 --> 00:03:53,110

Michael Ostapenko: Formal system's ability to do, like, rigid, logical, reliable reasoning.

38

00:03:53,450 --> 00:03:55,270

Michael Ostapenko: and analysis.

39

00:03:55,590 --> 00:04:02,890

Michael Ostapenko: And when you merge these two, you kind of get… The intelligence, the actual intelligence.

40

00:04:02,890 --> 00:04:04,399

Melinda Lee: It.

41

00:04:04,980 --> 00:04:06,520

Michael Ostapenko: We humans persist.

42

00:04:07,630 --> 00:04:17,230

Michael Ostapenko: And, basically, that's… that's… that's… that's my vision for the artificial intelligence, for the future of the artificial intelligence. And,

43

00:04:18,649 --> 00:04:21,950

Michael Ostapenko: I… I'm not convinced that we're…

44

00:04:22,150 --> 00:04:28,859

Michael Ostapenko: there yet. I mean, or will be there in the near future. It… it…

45

00:04:29,820 --> 00:04:34,830

Michael Ostapenko: In my opinion, there are, like, many pieces which are still missing.

46

00:04:36,450 --> 00:04:46,050

Michael Ostapenko: But… Karen said that… There are still, like, partions which start to emerge.

47

00:04:46,990 --> 00:04:55,870

Melinda Lee: Well, I'm curious, if you start to create something like that, like, that is mimicking, like, the… like you're saying, it's mimicking the human brain.

48

00:04:56,470 --> 00:05:00,339

Melinda Lee: But even probably more powerful, because…

49

00:05:01,780 --> 00:05:08,300

Michael Ostapenko: I would, I would, I would rather, like, not use, like, this, Hmm…

50

00:05:08,820 --> 00:05:23,420

Michael Ostapenko: ESN trapezens, because, I don't know, honestly. I have no idea whether it will mimic a human brain, whether this model will mimic a human brain or not, or even,

51

00:05:23,990 --> 00:05:30,769

Michael Ostapenko: Let's say, human intelligence as a product of human brain, or brain activity, right?

52

00:05:30,770 --> 00:05:32,620

Melinda Lee: But.

53

00:05:33,430 --> 00:05:39,820

Michael Ostapenko: I'm quite confident that The result, the output of such a system.

54

00:05:39,930 --> 00:05:46,780

Michael Ostapenko: Would be, like, pretty much indistinguishable from what Human intelligence produces.

55

00:05:46,970 --> 00:05:49,339

Melinda Lee: Okay, so that's my fear, that's my fear.

56

00:05:49,340 --> 00:05:49,735

Michael Ostapenko: Like…

57

00:05:50,130 --> 00:05:57,660

Melinda Lee: We… we don't know what we're creating, and we don't know on its own what it's gonna do.

58

00:05:57,910 --> 00:05:59,610

Michael Ostapenko: Oh, so…

59

00:05:59,610 --> 00:06:03,000

Melinda Lee: Intelligence, like, huh? AI, huh?

60

00:06:03,550 --> 00:06:07,340

Michael Ostapenko: Right, so… I wouldn't fear about that.

61

00:06:08,220 --> 00:06:09,150

Michael Ostapenko: Hmm.

62

00:06:09,270 --> 00:06:12,650

Michael Ostapenko: So, this is, like, another concept, actually.

63

00:06:12,930 --> 00:06:16,219

Michael Ostapenko: People would usually, like, mix up this.

64

00:06:16,750 --> 00:06:26,939

Michael Ostapenko: And, so… The… the… You're talking about, like, agency and the ability for…

65

00:06:27,740 --> 00:06:32,909

Michael Ostapenko: For, like, a machine to take, actions which are, like, beneficial to it.

66

00:06:33,440 --> 00:06:41,619

Melinda Lee: which are, which are, yeah, I'm talking about the, the intelligence, the AI intelligence, let's just call it that, to start to…

67

00:06:41,770 --> 00:06:48,119

Melinda Lee: Do things that we're not programming it to do, because it's starting to…

68

00:06:48,440 --> 00:06:49,650

Michael Ostapenko: It's not gonna happen.

69

00:06:49,650 --> 00:06:52,300

Melinda Lee: It's starting to create networks on its own.

70

00:06:52,570 --> 00:06:57,979

Michael Ostapenko: Yeah, I mean, yeah, it's more for a sci-fi story.

71

00:06:57,980 --> 00:06:58,720

Melinda Lee: Really.

72

00:06:58,720 --> 00:07:00,990

Michael Ostapenko: Yes, it's, it's, it's not gonna happen.

73

00:07:01,660 --> 00:07:13,480

Michael Ostapenko: As I said, like, in order for it, like, you can program it, like, to make something that works the way you described, like, right?

74

00:07:13,890 --> 00:07:24,469

Michael Ostapenko: But you still be the one who programs it. Now, Frankly speaking, this, agency thing, it's,

75

00:07:26,350 --> 00:07:31,649

Michael Ostapenko: It's, in some sense, you can… you can argue it's emergent, but…

76

00:07:32,130 --> 00:07:36,400

Michael Ostapenko: On the other hand, it's… it's… it's really not, because

77

00:07:36,640 --> 00:07:43,290

Michael Ostapenko: what, what's, what's the motivation for… there won't be, like, like, real motivation for…

78

00:07:44,890 --> 00:07:49,390

Michael Ostapenko: Intelligence, these artificial intelligence systems to do anything.

79

00:07:51,960 --> 00:07:57,429

Michael Ostapenko: Because… simply because it's more of a picture of a life, of life.

80

00:07:59,030 --> 00:08:05,280

Michael Ostapenko: Life is what, basically, Drives us, gives us motivation.

81

00:08:06,150 --> 00:08:09,249

Michael Ostapenko: And, intelligence is a tool.

82

00:08:10,190 --> 00:08:16,830

Michael Ostapenko: Which life uses to achieve goals, to basically reproduce, to prolong its existence.

83

00:08:17,320 --> 00:08:25,659

Michael Ostapenko: So, when we are talking about these artificial intelligence systems, I'm talking about life. At least I'm not.

84

00:08:25,660 --> 00:08:27,130

Melinda Lee: Hmm, got it.

85

00:08:27,570 --> 00:08:35,719

Michael Ostapenko: So… These systems will lack this inherent motivation. Just like…

86

00:08:35,860 --> 00:08:42,590

Michael Ostapenko: Modern artificial intelligence systems neural networks lacks this ability to logical reason.

87

00:08:43,500 --> 00:08:52,189

Michael Ostapenko: It's… it's… it's just impossible to add this ability to neural networks, like, directly like… like that. And it's impossible to…

88

00:08:52,570 --> 00:09:01,119

Michael Ostapenko: To add motivation like this agency to artificial intelligence, because it's not a property of intelligence.

89

00:09:01,280 --> 00:09:06,480

Michael Ostapenko: It's a property of life more, more property of life than intelligence.

90

00:09:07,940 --> 00:09:21,560

Melinda Lee: Got it, got it. So you're saying that, it doesn't lack the sense of… it's a separate thing. It's a separate thing to be able to have the motivation to do something that it's not…

91

00:09:22,410 --> 00:09:32,229

Michael Ostapenko: Yeah, it's, it's a separate thing. It's a… it's a separate… like, from the philosophical perspective, like, there was, like, a philosopher, I think Kant.

92

00:09:32,820 --> 00:09:35,670

Michael Ostapenko: He described these things in its… in themselves.

93

00:09:36,290 --> 00:09:37,679

Michael Ostapenko: Idea?

94

00:09:37,840 --> 00:09:46,060

Michael Ostapenko: And it's basically when… when you define something which cannot be, like, defined in other terms, like, right?

95

00:09:46,060 --> 00:09:46,650

Melinda Lee: Right.

96

00:09:46,950 --> 00:09:57,740

Michael Ostapenko: So that… that's… that's the situation. We… we are, we are… when we go into this area of intelligence and things like that, these are, like, very fundamental.

97

00:09:57,940 --> 00:09:59,800

Melinda Lee: Thanks, and .

98

00:10:01,280 --> 00:10:05,210

Michael Ostapenko: Very often, they… they… they can't, like, be,

99

00:10:05,970 --> 00:10:08,309

Michael Ostapenko: One thing can't be expressed as another.

100

00:10:09,030 --> 00:10:09,460

Melinda Lee: Yeah.

101

00:10:09,460 --> 00:10:10,200

Michael Ostapenko: possible.

102

00:10:10,750 --> 00:10:11,949

Michael Ostapenko: You need both.

103

00:10:12,370 --> 00:10:20,800

Melinda Lee: because I saw a documentary about the… they were programming some robots to… to just,

104

00:10:20,800 --> 00:10:39,780

Melinda Lee: play sports. Like, there were two robots that were playing against each other for… the goal is to… it's soccer. Like, the two robots are playing soccer, and the goal is for each of the players to get a goal, like, to hit the ball into the goalie, and then… so that's the game.

105

00:10:39,780 --> 00:10:52,719

Melinda Lee: And then they said the AI robots were playing, and they only programmed one goal, to get the ball into the goalie. But then, over time, when they started to play each other, they started to form neural networks. Oh, this worked.

106

00:10:52,720 --> 00:11:09,399

Melinda Lee: this got me a goal, this didn't. And then over time, like, they started to create patterns that then, therefore, they started to do some other things, because they were trying to figure out one problem to solve over another, and then it became a different type of

107

00:11:09,680 --> 00:11:13,400

Melinda Lee: Yeah, the original goal got skewed.

108

00:11:14,080 --> 00:11:22,250

Melinda Lee: So then they're saying that that's how could be possibility of when they start to solve different, you know, continue on to build on each other.

109

00:11:23,150 --> 00:11:28,420

Michael Ostapenko: Yeah, I hear you, and I remember this story.

110

00:11:28,880 --> 00:11:36,870

Michael Ostapenko: which be it, like, not long after this appearance of the, I think, GPT 3.5 or something, that,

111

00:11:37,090 --> 00:11:43,129

Michael Ostapenko: There was, like, some kind of Pentagon, experiments with AI.

112

00:11:43,130 --> 00:11:43,950

Melinda Lee: Yeah!

113

00:11:44,320 --> 00:11:51,879

Michael Ostapenko: Where they had, like, these, drones flying, and they had this, objective to,

114

00:11:52,320 --> 00:11:56,699

Melinda Lee: hit the target. Yeah. Or something like that, and .

115

00:11:56,700 --> 00:12:04,140

Michael Ostapenko: Then, and that's how they scored, and that's… that's how they… what functions work.

116

00:12:04,470 --> 00:12:04,860

Melinda Lee: Right.

117

00:12:04,860 --> 00:12:18,709

Michael Ostapenko: the more they hit, the more they would get, and that was the goal, right? And then, at some point, the operator, the human operator said, like, for whatever reason, that, do not execute.

118

00:12:19,520 --> 00:12:33,650

Michael Ostapenko: And, the, like, the artificial intelligence system, like, decided, like, it's still a reward, for it, right? So, what they did, like, destroyed the,

119

00:12:33,650 --> 00:12:39,599

Michael Ostapenko: communication tower connected to the operator. And still, she's the target!

120

00:12:39,600 --> 00:12:46,190

Melinda Lee: Exactly! That's what I'm… see? It's do- that's what it does! It could be possible.

121

00:12:47,020 --> 00:12:49,010

Michael Ostapenko: I mean…

122

00:12:49,560 --> 00:12:57,440

Michael Ostapenko: Well, it's not… it's not possible with the current systems, it's definitely out of reach. It's… it's a fairy tale.

123

00:12:57,610 --> 00:12:59,060

Michael Ostapenko: But,

124

00:12:59,960 --> 00:13:07,910

Michael Ostapenko: Say, if we, like, do this thought experiment that, we have, like, this intelligence, which is,

125

00:13:08,820 --> 00:13:10,090

Michael Ostapenko: quite powerful.

126

00:13:10,380 --> 00:13:14,999

Michael Ostapenko: To do more than what is currently possible.

127

00:13:15,300 --> 00:13:18,349

Michael Ostapenko: And, it's given that goal.

128

00:13:19,190 --> 00:13:24,960

Michael Ostapenko: then, well… And it's given the, let's say.

129

00:13:25,570 --> 00:13:27,880

Michael Ostapenko: Physical means to achieve that goal.

130

00:13:30,840 --> 00:13:39,209

Michael Ostapenko: Yeah, I mean, technically, it's possible for it to… and it's… it can learn on its own, on the go… on the go, right?

131

00:13:39,210 --> 00:13:40,070

Melinda Lee: Right.

132

00:13:40,070 --> 00:13:51,510

Michael Ostapenko: So, in order to achieve that goal, it could, like, experiment and things like that, and can, it can, do something like that. Yeah, technically, it's possible.

133

00:13:51,760 --> 00:13:57,330

Michael Ostapenko: But, and this is, like, a really big pot.

134

00:13:57,710 --> 00:14:00,070

Michael Ostapenko: I mean…

135

00:14:01,860 --> 00:14:11,500

Michael Ostapenko: In the real world, if something like that, like, happens, like, there is, like, a deviation from what's expected in human society.

136

00:14:12,070 --> 00:14:18,549

Michael Ostapenko: it will be noticed right away, just like you notice what… if some person does this.

137

00:14:20,070 --> 00:14:22,549

Michael Ostapenko: So, it's gonna be punished, right?

138

00:14:24,530 --> 00:14:25,379

Michael Ostapenko: So, so, so…

139

00:14:25,380 --> 00:14:26,980

Melinda Lee: Yeah, hopefully!

140

00:14:27,980 --> 00:14:28,790

Melinda Lee: Yeah.

141

00:14:28,790 --> 00:14:30,740

Michael Ostapenko: So, so…

142

00:14:30,740 --> 00:14:33,830

Melinda Lee: As long as we know how, as long as we know how…

143

00:14:34,400 --> 00:14:37,330

Michael Ostapenko: Yeah, I mean, like,

144

00:14:38,240 --> 00:14:44,540

Michael Ostapenko: And then, yes, that's another story, like, like, from a sci-fi movie, which is,

145

00:14:44,700 --> 00:14:49,630

Michael Ostapenko: Like, depicts this, artificial intelligent,

146

00:14:49,780 --> 00:15:00,950

Michael Ostapenko: Beings, like, with, like, human-looking bodies, but who, who possesses, ability to, like, think…

147

00:15:01,250 --> 00:15:02,839

Michael Ostapenko: In such a…

148

00:15:03,180 --> 00:15:13,849

Michael Ostapenko: Complex and deviant ways that no human can possibly, like, outsmart them and prevent them from reaching their goals.

149

00:15:14,170 --> 00:15:14,980

Michael Ostapenko: Sure.

150

00:15:15,230 --> 00:15:20,420

Michael Ostapenko: But, then you need to think about that, like,

151

00:15:22,380 --> 00:15:25,450

Michael Ostapenko: The, the, the, this, goal,

152

00:15:25,850 --> 00:15:35,460

Michael Ostapenko: You, you're the one who's, setting this, this goal, and you're the one who's, limiting or giving this, system the,

153

00:15:36,000 --> 00:15:39,670

Michael Ostapenko: Capacity, physical capacity to achieve that goal.

154

00:15:40,070 --> 00:15:46,669

Michael Ostapenko: So… At the end of the day, the responsibility is still on you if something goes wrong.

155

00:15:46,850 --> 00:15:47,210

Michael Ostapenko: Yes.

156

00:15:47,370 --> 00:15:48,750

Melinda Lee: Yes, yes.

157

00:15:48,750 --> 00:15:57,990

Michael Ostapenko: Because if the system doesn't have the capacity to achieve that goal, even if it Technically, like…

158

00:15:58,240 --> 00:16:00,459

Michael Ostapenko: Can't learn something like that.

159

00:16:00,590 --> 00:16:02,659

Michael Ostapenko: eat, eat,

160

00:16:03,110 --> 00:16:12,109

Michael Ostapenko: It won't be able to do that. And it won't even be able to learn to do that, because it just lacks the capacity. And,

161

00:16:12,560 --> 00:16:19,780

Michael Ostapenko: You need to also… Consider other factors, like,

162

00:16:21,160 --> 00:16:26,170

Michael Ostapenko: Like, energy and, things like that, because…

163

00:16:27,040 --> 00:16:30,369

Michael Ostapenko: We… we… as humans, we, like,

164

00:16:31,990 --> 00:16:42,589

Michael Ostapenko: the custom to think that, here, we come, we think, and the machine will go and think and do single projects the same way. But,

165

00:16:43,350 --> 00:16:49,140

Michael Ostapenko: There is, like, energy… Constraints. We need to eat.

166

00:16:49,240 --> 00:16:54,340

Michael Ostapenko: We need to breathe, and things like that. Now, you put a battery into a robot.

167

00:16:54,850 --> 00:16:56,330

Michael Ostapenko: How long will it last?

168

00:16:57,330 --> 00:17:08,339

Michael Ostapenko: You put some kind of, chip into a robot to execute those, complex algorithms and, or computations over this neural network.

169

00:17:08,930 --> 00:17:17,789

Michael Ostapenko: for how long will it last to… for it to be able to produce these superior results? Now, you can say.

170

00:17:17,940 --> 00:17:32,359

Michael Ostapenko: why, it can outsource it to some data center. Yeah, but, this is physical network connection. It's always controlled, right? The data center is, also

171

00:17:32,540 --> 00:17:39,120

Michael Ostapenko: The consumption of electricity data centers, the consumption of resource, computational resources and data center.

172

00:17:39,160 --> 00:17:40,889

Michael Ostapenko: It's all… it's all…

173

00:17:40,910 --> 00:17:51,439

Michael Ostapenko: controlled, it's all under surveillance. If there are, like, any deviations, like, and how it will even use it if it's, like, it has to pay for it.

174

00:17:51,440 --> 00:18:01,770

Michael Ostapenko: who's gonna pay for it? I mean, if you start thinking and, like, going really deep into all these things, you'll see that it's,

175

00:18:02,270 --> 00:18:03,760

Michael Ostapenko: Impracticable.

176

00:18:04,070 --> 00:18:07,720

Michael Ostapenko: for something like that to happen. And that's actually…

177

00:18:08,390 --> 00:18:13,399

Michael Ostapenko: what life, kind of, is about. It's, it's, it's about, like, complexity.

178

00:18:13,680 --> 00:18:16,630

Michael Ostapenko: And,

179

00:18:16,740 --> 00:18:25,000

Michael Ostapenko: That's why things like, we see in horror movies about these monsters and things like that. They don't really exist in real life.

180

00:18:25,500 --> 00:18:28,879

Michael Ostapenko: Because… Things like that, they just…

181

00:18:29,490 --> 00:18:33,940

Michael Ostapenko: unsustainable in real life. You can't… Can't survive.

182

00:18:34,820 --> 00:18:45,769

Michael Ostapenko: And this, ultimate… killing machine, which, is presented, by some, like, doomsayers. They…

183

00:18:45,910 --> 00:18:48,710

Michael Ostapenko: It won't be able to survive either.

184

00:18:49,040 --> 00:18:50,230

Michael Ostapenko: In this world.

185

00:18:51,880 --> 00:18:55,179

Melinda Lee: Okay, thank God. I could… I can sleep tonight.

186

00:18:58,180 --> 00:19:05,639

Melinda Lee: And so what is this whole thing about, like, quantum… did you say something about, like, quantum AI?

187

00:19:06,520 --> 00:19:13,339

Michael Ostapenko: So it's not, like, about quantum AI, it's about, like, quantum computations.

188

00:19:13,340 --> 00:19:15,800

Melinda Lee: Huh?

189

00:19:15,800 --> 00:19:19,680

Michael Ostapenko: Yeah. So, quantum computations, it's,

190

00:19:20,340 --> 00:19:25,079

Michael Ostapenko: Let's say, we have these conventional computers.

191

00:19:25,080 --> 00:19:25,510

Melinda Lee: Right?

192

00:19:25,750 --> 00:19:30,190

Michael Ostapenko: Which, and we have… which are… Well…

193

00:19:30,610 --> 00:19:38,170

Michael Ostapenko: let's say they… they are based on, like, classical physics, not so much on quantum… they do…

194

00:19:38,500 --> 00:19:56,999

Michael Ostapenko: exploit certain quantum effects, I suppose, on the very, very low level, like, in micro-sheets, but it's not used to speed up the computation, like, exponentially. What we think about, about, when we mention, like, quantum computers is, like, this exponential speed up.

195

00:19:57,900 --> 00:20:09,250

Michael Ostapenko: over… the conventional, computers. And, this, this allows, like, to…

196

00:20:09,490 --> 00:20:16,220

Michael Ostapenko: Why is it important, in general? So… In mathematics.

197

00:20:16,830 --> 00:20:25,540

Michael Ostapenko: And in computer science, specifically. Like, there are also so-called, like, complexity classes.

198

00:20:26,430 --> 00:20:33,619

Michael Ostapenko: computational complexity classes of problems. And the idea behind those classes is that,

199

00:20:34,260 --> 00:20:39,629

Michael Ostapenko: If you have some… some class which is… describes very hard problem.

200

00:20:41,060 --> 00:20:45,769

Michael Ostapenko: It, it, it usually has this, How do you describe it?

201

00:20:46,260 --> 00:20:52,349

Michael Ostapenko: It has an associated language, common language, which… which is extremely expressive.

202

00:20:53,120 --> 00:20:57,560

Michael Ostapenko: Because, it allows to express this powerful problem.

203

00:20:58,110 --> 00:21:07,119

Michael Ostapenko: And it's… you can use this language to model real-world and solve real-world problems, but

204

00:21:07,220 --> 00:21:11,470

Michael Ostapenko: Formally, rigidly, accurately.

205

00:21:11,830 --> 00:21:18,750

Michael Ostapenko: And, what's… the main idea behind this class is that

206

00:21:19,350 --> 00:21:23,139

Michael Ostapenko: If you have the solution for just one problem from this class.

207

00:21:24,560 --> 00:21:28,669

Michael Ostapenko: We have solutions for all the problems which fall into this class.

208

00:21:29,300 --> 00:21:33,619

Michael Ostapenko: like, efficient solution, because you have this

209

00:21:34,130 --> 00:21:39,750

Michael Ostapenko: easy, and it's, like, in practice, you have this easy… And,

210

00:21:40,340 --> 00:21:46,179

Michael Ostapenko: The ways to transform one problem of this class into another. And that's the crux of it.

211

00:21:46,420 --> 00:21:53,569

Michael Ostapenko: So… Quantum computers, they… they solve… they basically can solve one of these.

212

00:21:53,670 --> 00:22:05,049

Michael Ostapenko: powerful classes of problems. That's why they… everyone is after them. If… because if this… if you build a quantum computer, it can solve one problem from this

213

00:22:05,440 --> 00:22:07,709

Michael Ostapenko: Very hard class of problems.

214

00:22:07,980 --> 00:22:12,849

Michael Ostapenko: And you can solve all of them. And these problems, they… they're…

215

00:22:13,310 --> 00:22:19,650

Michael Ostapenko: it's, it's the… so, it's, it's so expressive, it can, it can describe, like, as I said, like.

216

00:22:19,780 --> 00:22:33,510

Michael Ostapenko: Biological processes, physical processes, social processes, anything that is extremely complex, and you can describe it, and you can then optimize

217

00:22:34,100 --> 00:22:38,630

Michael Ostapenko: Things, like, the different aspects of these processes.

218

00:22:38,940 --> 00:22:40,940

Michael Ostapenko: Build new drugs.

219

00:22:41,210 --> 00:22:45,779

Michael Ostapenko: And, I don't know, build new materials, things like that.

220

00:22:45,930 --> 00:22:59,199

Michael Ostapenko: And, that's why everyone is after this, quantum computers. But… At the same time, Even though this…

221

00:22:59,600 --> 00:23:03,820

Michael Ostapenko: like, classes are so powerful and so expressive.

222

00:23:04,690 --> 00:23:06,880

Michael Ostapenko: There is no, like.

223

00:23:07,270 --> 00:23:13,620

Michael Ostapenko: To this day, there is no mathematical proof for certain of the… some of these classes that they can't be

224

00:23:13,930 --> 00:23:17,720

Michael Ostapenko: Solved, efficiently using classical means.

225

00:23:19,250 --> 00:23:26,559

Michael Ostapenko: And, now, this is one of the, like, so-called millennial pro- millennium problems.

226

00:23:27,300 --> 00:23:30,600

Michael Ostapenko: It's, like, with the price of, like, $1 million for each.

227

00:23:31,200 --> 00:23:50,460

Michael Ostapenko: Yeah, but… Now, I'll say right away, we aren't trying to solve this. It's beyond our scope. We are just trying to create, like, what's practically possible and what's feasible right now. Yeah, but,

228

00:23:50,720 --> 00:24:00,829

Michael Ostapenko: In general, just to set a, like, a stage, like, there, there is this price, and there is this problem.

229

00:24:01,720 --> 00:24:19,740

Michael Ostapenko: And, no one knows the solution. No one knows if the solution is even possible, or even if it's even possible to prove that it's impossible. No one knows anything about it. So, but, what we believe, and what we…

230

00:24:20,410 --> 00:24:21,700

Michael Ostapenko: like, C?

231

00:24:22,000 --> 00:24:27,939

Michael Ostapenko: Is that current solutions to these kind of problems, they aren't optimal yet.

232

00:24:28,180 --> 00:24:31,300

Michael Ostapenko: there's still… Ways to improve.

233

00:24:31,550 --> 00:24:32,650

Michael Ostapenko: efficiency.

234

00:24:33,430 --> 00:24:37,099

Michael Ostapenko: And, we are aiming to do just that.

235

00:24:37,380 --> 00:24:38,090

Melinda Lee: I love it.

236

00:24:38,090 --> 00:24:52,220

Michael Ostapenko: We're not… we're not trying to, like, build, like, quantum computers, which are a different topic, and I have, like, my opinion on that too, but… although I'm not, like, I can't be considered an expert in this field, so…

237

00:24:52,220 --> 00:24:54,540

Melinda Lee: That's what you're doing, that's what you're doing at Citrix, right?

238

00:24:54,860 --> 00:25:08,929

Michael Ostapenko: Not quantum computing, the, the classical, approaches. Yeah, but I'm just, I'm just explaining it using quantum computers, it's, because it's a hot topic, it's kind of ironic, because, what we do, it's,

239

00:25:09,340 --> 00:25:17,439

Michael Ostapenko: which is logic and algorithms and things like that, mathematics. It's kind of ancient.

240

00:25:18,500 --> 00:25:19,730

Michael Ostapenko: Quantum computing?

241

00:25:20,150 --> 00:25:29,249

Michael Ostapenko: They're relatively new, and everyone is talking about them, because it's, like, a new child. But, yeah.

242

00:25:29,250 --> 00:25:51,739

Melinda Lee: Fascinating. So fascinating. I think I'm so appreciative of this conversation. I'm learning a lot, and it's because it's so different from what I do with regard to people and communication, and, you know, we have these systems, AI systems, computations, and computer… I mean, just really fast-tracking.

243

00:25:51,740 --> 00:25:59,390

Melinda Lee: How well, we… do things as humans, right? How well we're able to solve problems, and these…

244

00:25:59,390 --> 00:26:18,000

Melinda Lee: these really challenging, complex problems, and that we have the leverage of the technology to do that. And so, I really appreciate your expertise and your experiences digging into… to using this, to learning this, to applying this to help society.

245

00:26:18,420 --> 00:26:23,849

Michael Ostapenko: Yeah, that's, basically the goal, because,

246

00:26:24,440 --> 00:26:27,470

Michael Ostapenko: I mean, technology on its own, it's,

247

00:26:27,870 --> 00:26:37,710

Michael Ostapenko: It's nothing. It's not just useless, it's really nothing, because, unless there is, there are people to use it to benefit from it.

248

00:26:37,710 --> 00:26:48,969

Melinda Lee: Yeah, yeah, yeah, I agree. And I thank you so much for now I can sleep tonight, because I'm not afraid of these AI robots taking over the world.

249

00:26:48,970 --> 00:27:06,699

Michael Ostapenko: Well, I'm glad. I mean, you can definitely sleep for, like, tight for the next decade, or several decades, because the current systems aren't there yet, and I haven't seen any fundamental progress

250

00:27:06,720 --> 00:27:09,640

Michael Ostapenko: Which would allow anything like that to happen.

251

00:27:09,820 --> 00:27:10,740

Michael Ostapenko: Yet.

252

00:27:10,870 --> 00:27:11,570

Michael Ostapenko: Oh.

253

00:27:11,750 --> 00:27:14,040

Melinda Lee: There's a lot of, like.

254

00:27:14,210 --> 00:27:19,730

Michael Ostapenko: Dogs, there is a lot of hype about this field, but

255

00:27:20,070 --> 00:27:25,760

Michael Ostapenko: Yeah, there was a breakthrough, like, in this,

256

00:27:26,330 --> 00:27:30,079

Michael Ostapenko: transformers and things like that. But,

257

00:27:30,930 --> 00:27:41,399

Michael Ostapenko: Which… which kind of showed that semantics can be a model, like, efficiently using these systems, but beyond that.

258

00:27:43,890 --> 00:27:51,060

Michael Ostapenko: It's really, like, more, more, like, a lot of hype, Nothing, nothing, nothing, substantial.

259

00:27:51,790 --> 00:27:52,440

Melinda Lee: Yeah.

260

00:27:52,840 --> 00:27:55,169

Melinda Lee: Yeah, good. Good, good.

261

00:27:55,170 --> 00:27:55,950

Michael Ostapenko: I really appreciate.

262

00:27:55,950 --> 00:28:05,519

Melinda Lee: And Michael, so how would people get ahold of you? Where do they find you if they want some more, experience or expertise from your company?

263

00:28:06,630 --> 00:28:09,339

Michael Ostapenko: Well, they can,

264

00:28:09,970 --> 00:28:22,250

Michael Ostapenko: I'm usually, like, available, through email. I always, check it. It's… you can write to CEO at Cetrix.com.

265

00:28:22,620 --> 00:28:27,909

Michael Ostapenko: So, Cetrix is spelled as S-A-T, as in Tom.

266

00:28:28,030 --> 00:28:31,749

Michael Ostapenko: R, Y, X, etc.

267

00:28:31,970 --> 00:28:36,220

Melinda Lee: Yep, perfect! And we'll have your, email in the show notes, too.

268

00:28:36,470 --> 00:28:38,090

Michael Ostapenko: Yeah, thank you, Emil Day.

269

00:28:38,090 --> 00:28:45,760

Melinda Lee: Okay, thank you so much, Michael. It was such a great conversation, and I really enjoyed it, I learned a lot, and I appreciate your time.

270

00:28:46,580 --> 00:28:49,170

Michael Ostapenko: I'm happy, I'm happy to be…

271

00:28:49,290 --> 00:28:52,330

Michael Ostapenko: invited by you, and I really… it's,

272

00:28:52,870 --> 00:28:56,720

Michael Ostapenko: It's, it was, it was really a pleasure talking to you.

273

00:28:56,720 --> 00:29:13,970

Melinda Lee: Thank you, and thank you, audience, for being here. I trust that you got your takeaway from today, and so continue to use and leverage technology however it serves you best, so that you can have more deep, relationships and be able to enjoy your life.

274

00:29:13,970 --> 00:29:22,550

Melinda Lee: And so, until next time, I'm your sister in flow. May prosperity flow to you and through you, onto others, always. Thank you. Bye-bye!

275

00:29:22,760 --> 00:29:23,500

Melinda Lee: Bye, Michael.

276

00:29:23,500 --> 00:29:24,710

Michael Ostapenko: Okay, alright?

277

00:29:24,710 --> 00:29:25,530

Melinda Lee: Bye-bye.