Latest / The Rogue Recap with Lynda / AI - The Good, the Bad, and the Energy Crunch

Transcript

AI - The Good, the Bad, and the Energy Crunch · The Rogue Recap with Lynda

0:07
Speaker 1: Welcome to the Rogue Recap.
0:09
Speaker 2: Hot takes, cold facts, and zero respect for the official narrative.
0:15
Speaker 1: Sit back, roll your eyes, and let's recap rogue style.
0:20
Speaker 1: What's up, everybody, and welcome to the Roague Recap. I
0:22
Speaker 1: am Lynda McLaughlin, your host. Here, I am again at
0:25
Speaker 1: Linda Mick at Rogue Recap roguerecap dot com. You know
0:28
Speaker 1: what to do, YadA, YadA, YadA. So there's a million
0:31
Speaker 1: things going on, like I say every single day because
0:33
Speaker 1: it's true, so I think it bears repeating. But one
0:36
Speaker 1: of the things that is not getting discussed is AI.
0:39
Speaker 1: So we're talking about all the cool things we can
0:42
Speaker 1: do with it, all the cool things it can do
0:45
Speaker 1: for us to help us with time management and expediting projects,
0:48
Speaker 1: and that's all wonderful, But what else is it doing?
0:52
Speaker 1: I think that's my question. And I am certainly not
0:54
Speaker 1: an expert. I use it every day because I work
0:56
Speaker 1: in media, as most of you know. But I have
0:58
Speaker 1: a dear friend I call doctor Dave. He's telling me
1:01
Speaker 1: he's not a doctor. I think it's fake news, but
1:03
Speaker 1: he's telling me he's not. So I'm going to leave
1:05
Speaker 1: that there. But his name is Dave. Laboue. He's wicked smart.
1:07
Speaker 1: He's got about seventy eight master's degrees in artificial intelligence
1:12
Speaker 1: and language algorithms and all sorts of sciences that regular
1:16
Speaker 1: people just go, Hey, I need a guy for that,
1:18
Speaker 1: and he's the guy. So Dave, welcome to the show.
1:21
Speaker 2: Thank you, happy to be here.
1:22
Speaker 1: Linda, good Glad for those of you who you know
1:25
Speaker 1: haven't listened to my show. I like to bring my
1:28
Speaker 1: friends on. Dave is a great friend, and he just
1:31
Speaker 1: happens to be super wicked smart. So I said to him,
1:35
Speaker 1: can you come on and talk to me about AI
1:38
Speaker 1: because I'm seeing all this news come out and one
1:42
Speaker 1: of the biggest things that people are asking is how
1:45
Speaker 1: the hell are we going to run this? We don't
1:47
Speaker 1: have the energy. And so I told Dave, I said,
1:49
Speaker 1: I want to talk about what the president is saying.
1:51
Speaker 1: I want to talk about what Elon Musk is saying,
1:55
Speaker 1: but I also want to talk about the functionality of
1:58
Speaker 1: how this really happens. So I'm going to Da've talked
2:00
Speaker 1: to us for a hot second about his background so
2:02
Speaker 1: that we all know why he's worthy of talking about
2:04
Speaker 1: these things, and then we're going to get into what
2:07
Speaker 1: happens next here, you know, for the regular American citizen
2:11
Speaker 1: on their electric bill, to how they use AI in
2:13
Speaker 1: everyday life. So Dave, tell us a little bit about
2:16
Speaker 1: your background and why you love to nerd out on
2:19
Speaker 1: this stuff.
2:20
Speaker 2: Yeah, sure thing. And for my background, it comes from
2:23
Speaker 2: a combination of reason, philosophy, mathematics, data science, and artificial intelligence.
2:30
Speaker 2: So I've seen the evolution of language processing as it's
2:34
Speaker 2: come through different algorithms, and the world we're living in
2:38
Speaker 2: today is really a quantum leap from what we saw
2:41
Speaker 2: ten to fifteen years ago. And from a neuroscience standpoint,
2:45
Speaker 2: from a language and linguistic understanding standpoint, from a compute standpoint,
2:52
Speaker 2: all of these areas have evolved very rapidly in the
2:55
Speaker 2: last five years, and we're seeing major increases on what
3:01
Speaker 2: these AI algorithms can do the large language models, who
3:05
Speaker 2: owns them, who's developed them, who has the control over them,
3:08
Speaker 2: how they train them, and what that means for us
3:11
Speaker 2: as users to interact with them. So there's a lot
3:14
Speaker 2: of open questions that we have to go through and
3:16
Speaker 2: think about at a deeper level, both in terms of
3:21
Speaker 2: what kind of dependency and interaction do we want to
3:24
Speaker 2: have with these ais.
3:27
Speaker 1: So can I interrupt you on that point? Can you
3:29
Speaker 1: explain to us in layman's terms? What is language processing?
3:35
Speaker 3: Sure?
3:35
Speaker 2: So, language processing? The quick background is you're taking text
3:40
Speaker 2: as it is typed in a computer, or word documents
3:43
Speaker 2: or anything that you use digitally. And about ten to
3:47
Speaker 2: fifteen years ago, we were basically looking at collections of words,
3:51
Speaker 2: strings of words, how do they relate to each other?
3:54
Speaker 2: What combinations words are used together? To identify likeness with
4:00
Speaker 2: the something that came out of actually Google Brain, which
4:03
Speaker 2: was the paper written called the Attention is All you
4:06
Speaker 2: need to develop the GPT revolution that we know about.
4:09
Speaker 2: Open Ai now created that and evolved it. It's sort
4:15
Speaker 2: of a major breakthrough that allows language processing to understand context.
4:21
Speaker 2: So what that means is, as I'm speaking right now,
4:25
Speaker 2: I'm thinking about the next word or set of words
4:27
Speaker 2: that I'm going to deliver, and that's based on the
4:30
Speaker 2: previous words that I've used and sort of my background
4:32
Speaker 2: and knowledge, and language models now have the ability to
4:36
Speaker 2: look backwards in history that they never had the ability
4:39
Speaker 2: to do before. They were only looking at sort of
4:41
Speaker 2: certain strings of text place side by side, or certain documents.
4:46
Speaker 2: And this relates to the energy usage. So we're looking
4:48
Speaker 2: now not just in a single, very narrow sense across
4:52
Speaker 2: you know, hundreds or thousands of documents. We're looking across
4:55
Speaker 2: documents that spanned all digital information potentially, so anything we
4:59
Speaker 2: have online that's and scanned and catalog and organized that
5:02
Speaker 2: is now available memory to generate the next word that's
5:06
Speaker 2: being predicted as you talk to an LLF. So it's
5:10
Speaker 2: been quite a change in how we can use these models,
5:16
Speaker 2: what we can expect from them, and what they're able
5:18
Speaker 2: to do and to evolve from that. Language models are
5:22
Speaker 2: no longer just about language. They have multimodal aspects. As
5:26
Speaker 2: we know, we can use image generation, voice generation, you
5:30
Speaker 2: can do image to text, text image, image to image,
5:34
Speaker 2: and they understand all these different modalities in a deep
5:39
Speaker 2: way that humans would. So we're not quite at general
5:43
Speaker 2: intelligence for AI's still a long shot to go there,
5:46
Speaker 2: but we have a lot of advanced understanding for these
5:50
Speaker 2: algorithms to be able to understand an image, understand a video,
5:54
Speaker 2: modify it or edit it and give us back something
5:57
Speaker 2: based on what we're requesting of it. It's gone from
6:00
Speaker 2: just language to basically anything digital now can be used
6:04
Speaker 2: through a language model.
6:05
Speaker 1: Okay, So two questions to that. The first thing is
6:08
Speaker 1: when you say it's collecting information, just how you and
6:12
Speaker 1: I are having a conversation and we're picking and choosing
6:15
Speaker 1: the words we're going to say next based upon the
6:17
Speaker 1: word said previously. No, we're saying hmm, we're going to
6:21
Speaker 1: go here, We're going to go there. I want to
6:22
Speaker 1: ask this question, are these models, these AI models, are
6:27
Speaker 1: they pulling from data sets that it's getting from inputs
6:31
Speaker 1: all over the world, or is it just pulling from
6:34
Speaker 1: the data set given to it in a particular user moment,
6:39
Speaker 1: so a combination.
6:40
Speaker 2: Up until recently, that was how they were trained, quote unquote,
6:46
Speaker 2: So you had large open web documents, basically anything that
6:51
Speaker 2: was open source to the web or digitally for that matter. Actually,
6:55
Speaker 2: very private data is very valuable to anything that is
6:58
Speaker 2: not broadly available in a very specific niche area is
7:02
Speaker 2: a valuable data asset for language models because those are
7:05
Speaker 2: areas that they haven't been trained on. But otherwise more generally,
7:10
Speaker 2: you know, after about the GPT from open ai came
7:13
Speaker 2: out in twenty twenty, twenty twenty one, Yeah, a lot
7:16
Speaker 2: of that energy was used in just training of all
7:18
Speaker 2: of the data that was available on the web. People
7:21
Speaker 2: were scraping Twitter at the time, bringing those conversations in
7:25
Speaker 2: using different social media, Reddit post all that was brought in.
7:28
Speaker 2: That was training the model that was basically building its
7:30
Speaker 2: brain of how to think, not to reference when it's
7:33
Speaker 2: constructing its sentences about what it knows. So obviously there's
7:36
Speaker 2: a bias there if you scrape through information from a
7:39
Speaker 2: certain leaning area, that's all it's going to know. So
7:44
Speaker 2: balanced understanding and sort of training of these models is
7:48
Speaker 2: very important. But to get back to your point now
7:52
Speaker 2: that we've sort of evolved beyond the base training of
7:54
Speaker 2: these language models or sort of called the fun foundational models,
7:58
Speaker 2: a lot of them do integrate real time data, especially
8:00
Speaker 2: groc that's a great segue from x XAI into groc AI.
8:07
Speaker 2: You have the ability for that LLM, that algorithm to
8:11
Speaker 2: basically feed real time all conversations, news that's posted back
8:16
Speaker 2: and forth, direct messages, anything that's happening in x can
8:19
Speaker 2: now be integrated real time into that language model and
8:21
Speaker 2: used to generate responses. And this goes for other models too,
8:25
Speaker 2: But it really is about the data they have access
8:27
Speaker 2: to and how quickly they can reference it and bring
8:30
Speaker 2: it back to the user.
8:32
Speaker 1: So it's interesting and this goes directly to GROC and
8:36
Speaker 1: I'm looking this up on x as we speak, because
8:39
Speaker 1: I want to make sure that I quote this accurately
8:42
Speaker 1: to that exact point of creating, for lack of a
8:46
Speaker 1: better example or analogy, partisan robots, if you will, right, So,
8:50
Speaker 1: we're creating the libs and the conservatives in the AI world,
8:54
Speaker 1: which is so creepy because we've got plenty of them
8:57
Speaker 1: in the human world. We don't need anymore. But I'll
8:59
Speaker 1: leave that there. So Mark Benioff of Salesforce, he's the CEO.
9:04
Speaker 1: I saw this interview yesterday and he says that he
9:09
Speaker 1: used rock to make the viral image for his dream
9:15
Speaker 1: Force Salesforce, you know, positioning, because chat gpt refused to
9:23
Speaker 1: create it. So chat gpt said nope, I won't do it.
9:27
Speaker 1: And he was like, this is the difference between open
9:30
Speaker 1: creativity and corporate censorship. And I was like, that's wild.
9:36
Speaker 1: Like this this interface literally said to you, no, I
9:40
Speaker 1: won't do this, no, and I will tell you interestingly
9:44
Speaker 1: enough to your point. We were talking about this the
9:48
Speaker 1: other day. The ability to go image to image, text
9:52
Speaker 1: to image, image to video, video to video, or image
9:55
Speaker 1: to image to create a video, and the prompts right
9:59
Speaker 1: the language that you're using. So the other day, I
10:03
Speaker 1: was trying to create an announcement for my daughter and
10:06
Speaker 1: where she's going to college. So I had a picture
10:08
Speaker 1: of my daughter with her lacrosetick, and then I had
10:12
Speaker 1: where she was going to college, and I made the
10:13
Speaker 1: super cool graphic. But then I took the images that
10:17
Speaker 1: I made and I had to put them into a
10:19
Speaker 1: real editing software because I know how to edit, to
10:22
Speaker 1: piece it all together and do it the way that
10:24
Speaker 1: I really wanted. What's interesting is I couldn't make it
10:27
Speaker 1: in AI. I had to do it if you will
10:30
Speaker 1: buy hand because that particular interface told me that she
10:34
Speaker 1: was a miner and it wouldn't let me use the
10:37
Speaker 1: image of her. She was in a lacrosse uniform with
10:41
Speaker 1: a lacrossetick. There was nothing inappropriate about this image whatsoever.
10:45
Speaker 1: I did not supply her age. She is a minor,
10:48
Speaker 1: but I she could have been eighteen right and still
10:51
Speaker 1: be in high school. But it wouldn't let me do it.
10:54
Speaker 1: And I thought, wow, that's interesting that it's not allowing me.
10:58
Speaker 1: It's telling me, no, you're not allowed to make this
11:01
Speaker 1: so it was like a direct example for me on
11:04
Speaker 1: a much lower level obviously, because he's using it for Salesforce,
11:08
Speaker 1: which is taking over so many of the in house
11:10
Speaker 1: communicat for many many corporations. I don't actually care for it,
11:14
Speaker 1: but that's a different conversation. But anyways, it's not user friendly,
11:17
Speaker 1: it's not easy to use. But I thought that was
11:21
Speaker 1: so interesting that this computer told me.
11:23
Speaker 2: No, yeah, And that's a good example of a guardrail
11:29
Speaker 2: that's on. And they do have that capability. Obviously, the
11:33
Speaker 2: owners that have built these foundation models, only a few
11:36
Speaker 2: of them big tech. Basically you've got yeah, Google, Gemini,
11:38
Speaker 2: You've got now Rock from XAI. You have a few
11:41
Speaker 2: open source models. But the other one is really Facebook
11:44
Speaker 2: has their meta Lama models. So there's really a few
11:48
Speaker 2: paths you can funnel into when you're interacting with modern AI.
11:53
Speaker 2: Typically they're referencing one of these larger models through a
11:56
Speaker 2: back end connection, and they control of that. I mean,
12:00
Speaker 2: you can place things as sort of an intermediary business
12:06
Speaker 2: lying AI software, but if it's from the source determined
12:11
Speaker 2: to have this guardrail or that guardrail, they have the
12:13
Speaker 2: control over what is able to be spoken about. Describe
12:18
Speaker 2: utilize and there are good use cases, you know, relating
12:21
Speaker 2: to what you just said about having identified mind children,
12:25
Speaker 2: right and mental health. That's another one where interactions are
12:28
Speaker 2: sensitive among children. You want to make sure that the
12:30
Speaker 2: right conversations are being had, it's not going off the
12:33
Speaker 2: rails in a negative way. And then the other side
12:35
Speaker 2: of that is again the balance of what freedoms do
12:38
Speaker 2: we allow them to maintain, because that's part of you know,
12:43
Speaker 2: our rights and AI as it becomes a bigger and
12:46
Speaker 2: broader part of our society, we won't know if we're
12:49
Speaker 2: interacting with a human behind the screen or but it's
12:51
Speaker 2: an AI bought or agents, because they will become so
12:54
Speaker 2: realistic that that agent that bought that AI, if it's
12:59
Speaker 2: pre built to act a certain way, that bias will
13:03
Speaker 2: sort of suddenly condition us to adapt. So it's a
13:08
Speaker 2: form of censorship in a way, yeah, right, a control
13:15
Speaker 2: and it definitely needs to be monitored, yeah, because it
13:18
Speaker 2: can happen subtly and gradually and then altogether at some point.
13:24
Speaker 1: It's a little scary, you know. And there's a few
13:27
Speaker 1: other things I want to get to, but one more
13:29
Speaker 1: one more aspect of this part is I don't know
13:32
Speaker 1: if you've seen on all of the social media platforms,
13:35
Speaker 1: whether you're on Insta or you're on TikTok, or you're
13:37
Speaker 1: on x or whatever. But people showing two examples, right,
13:42
Speaker 1: they'll show I put this question into GROCK and I
13:45
Speaker 1: put this question into chat GPT, and the answers are astounding.
13:50
Speaker 1: It's truly shocking. And I'm like, oh my god. I'm like,
13:54
Speaker 1: it's like I have It's like I have Trump and
13:58
Speaker 1: Biden doing my AI. I'm like so creepy that it's
14:02
Speaker 1: not just like I just don't know where to go
14:04
Speaker 1: to get the lack of bias, and I don't know
14:08
Speaker 1: that we can.
14:11
Speaker 2: Yeah, and that's a good example. I think some of
14:14
Speaker 2: the earlier models. I say earlier, you know, two years
14:17
Speaker 2: ago when Gemini came out, I think they got some
14:19
Speaker 2: bad pr based on exactly that is, it really went
14:23
Speaker 2: too far in the direction and it was obvious. Yeah,
14:26
Speaker 2: you and even generate factual images of history without it
14:29
Speaker 2: being latly injected with some kind of bias. You will, right, right,
14:35
Speaker 2: And there are ways too well, there are limits to
14:39
Speaker 2: how you can work around that. But again, if the
14:42
Speaker 2: base model, and if we route basically back to three
14:45
Speaker 2: or four foundation models owned by three or four companies.
14:48
Speaker 2: They ultimately control the gates of what can and cannot
14:52
Speaker 2: be done. And it's hard. I mean you can use,
14:56
Speaker 2: like you said, the prompt engineering, the prompt sort of
14:59
Speaker 2: adjustments you train or adjust how it responds. Yeah, there
15:03
Speaker 2: are limits. There are limits if it's guarded from the
15:06
Speaker 2: back end.
15:07
Speaker 1: And to that point, what's interesting is if you're on
15:10
Speaker 1: groc versus Chat GBT, so on grock, you know, after like,
15:14
Speaker 1: I'll say something like, I feel like I could write
15:16
Speaker 1: this better. Right, it's I have it in twelve sentences.
15:19
Speaker 1: It could probably be for I'll run it through. I
15:22
Speaker 1: never let I never let anything go out for my
15:27
Speaker 1: personal round, my business life that's written by a robot.
15:30
Speaker 1: But I have many times run it through for grammar
15:34
Speaker 1: or said I want to do this plan. How could
15:36
Speaker 1: I make this more user friendly? Because you know, to
15:39
Speaker 1: me it's very obvious, but maybe to somebody else it's
15:41
Speaker 1: not right. But after you run it through in grock,
15:44
Speaker 1: it'll say to you, make more concise, make more persuasive,
15:49
Speaker 1: make more professional. It gives you options so that you
15:53
Speaker 1: can take that tone somewhere else. With Chat GBT, you're
15:55
Speaker 1: not getting that It's not that kind of thing at all.
15:59
Speaker 1: And I will say the conjecture that it puts in
16:02
Speaker 1: like I'll say, you know, Obama sent you know, one
16:05
Speaker 1: point seven billion dollars to a run in the final
16:08
Speaker 1: months of his presidency. This is a fact. There is
16:11
Speaker 1: no I'm not putting any like context around that. And
16:15
Speaker 1: it'll come back and it'll give me this filler in
16:18
Speaker 1: chat GPT of Barack Obama felt that the sanctions needed
16:22
Speaker 1: to be left, and as a result, he felt like he, what,
16:26
Speaker 1: what's the word? You're using a lot of feel here.
16:28
Speaker 1: There's a lot of felt and maybes, and I'm not
16:30
Speaker 1: interested in that. I'm just asking for like the date,
16:33
Speaker 1: the time, did it happen, and it can't do it,
16:35
Speaker 1: whereas Grock is much more like this happened there and
16:38
Speaker 1: it doesn't matter if it's right or left, which I
16:41
Speaker 1: think is very interesting.
16:43
Speaker 2: Yeah, And I think that's part of what Elon has
16:45
Speaker 2: said about the development of his model is he wants
16:48
Speaker 2: it maximally truth seeking, which I mean, what more can
16:52
Speaker 2: you ask for in an AI than why would seek
16:56
Speaker 2: the objective truth whether humans can see it or not? Yes,
17:00
Speaker 2: because implicit bias from human perspective makes its way into
17:03
Speaker 2: how we think about things, that we write about, things
17:06
Speaker 2: that we talk about things. So to have sort of
17:09
Speaker 2: a pre set standard that focuses on truth of above
17:14
Speaker 2: all else is the ideal approach to AI in the future.
17:19
Speaker 2: I mean, I think otherwise, subtly and gradually it will
17:23
Speaker 2: go off the rails and come back to harm us
17:25
Speaker 2: a society unless it's seeking truth.
17:28
Speaker 1: So to that point, I'm going to play a quick
17:29
Speaker 1: cut that I have of ELIN and I thought was
17:32
Speaker 1: very apropo for this conversation. We'll talk on the other
17:34
Speaker 1: side of this.
17:36
Speaker 3: If you just press the rock icon on any x post,
17:40
Speaker 3: analyze THEPLOYD, and research it as much as you want
17:43
Speaker 3: so you can you can basically have just by tapping
17:46
Speaker 3: the rock icon, you can assess whether that that post
17:50
Speaker 3: is the truth, the whole truth, or nothing but the truth,
17:52
Speaker 3: or whether there's something supplemental you need to be explained.
17:55
Speaker 3: So I think I think it's actually we've made a
17:58
Speaker 3: lot of progress towards, yeah, freedom of speech and people
18:04
Speaker 3: being able to tell whether something is false or not. Propaganda.
18:07
Speaker 3: The recent update to GROWK is actually I think very
18:09
Speaker 3: good at piercing through propaganda, and then we used that
18:14
Speaker 3: latest person of groc to create Crokipedia. It's not just
18:17
Speaker 3: I think more neutral that then and more accurate than
18:22
Speaker 3: than Wikipedia, but actually it's a lot more information than
18:24
Speaker 3: a Wikipedia page.
18:26
Speaker 1: So to that point about Wikipedia, which is a very
18:28
Speaker 1: interesting point that he makes. As somebody that manages talent
18:33
Speaker 1: and works in media, a lot of my talent has
18:37
Speaker 1: Wikipedia pages. And one of my clients had said to me,
18:41
Speaker 1: there's some really weird stuff on my Wikipedia. Can you
18:44
Speaker 1: go and change it? And I was like, oh yeah, sure,
18:46
Speaker 1: something I had never done, never even crossed my mind,
18:49
Speaker 1: and looked, okay, fine, So I go to Wikipedia. It's like,
18:53
Speaker 1: I mean, literally almost everything in it was not true
18:56
Speaker 1: about my client, who he was married to, how many
18:59
Speaker 1: children he had, were he lived, had he'd been arrested
19:02
Speaker 1: thirty eight times. I'm like, what I mean, this guy
19:04
Speaker 1: is like a stalwart lawyer married for four years. It
19:08
Speaker 1: was just and I thought, oh my god, if somebody
19:12
Speaker 1: googles him and that comes up, they're and be like,
19:15
Speaker 1: holy shit, you got to be kidding me. But none
19:18
Speaker 1: of it was true. But because it's open to public edit,
19:24
Speaker 1: it wasn't protected from any of that, which I thought
19:27
Speaker 1: was super scary and I didn't know it before that experience.
19:32
Speaker 2: Yeah, and that's that is a very concerning aspect of
19:37
Speaker 2: how these models have been trained historically. Is right, if
19:41
Speaker 2: there are you know, public figures, there's enough out there
19:44
Speaker 2: that someone's going to see it or catch it. But
19:45
Speaker 2: sure if someone has a Wikipedia change or not checking
19:48
Speaker 2: it every single day or you know, really once a year,
19:51
Speaker 2: and that false information goes into that page. If an
19:55
Speaker 2: AI has, as I mentioned, ear you've been trained where
19:58
Speaker 2: it's reading through documents, and Wikipedia was one of the
20:00
Speaker 2: earlier knowledge corpuses that it was trained and learned on. Yeah,
20:05
Speaker 2: that's what it knows. So it doesn't know and I'll
20:07
Speaker 2: turn to the point of view if that's the only
20:08
Speaker 2: thing it knows about person A or person B. So
20:11
Speaker 2: when you ask about that, it'll respond with confidence that
20:15
Speaker 2: this is exactly what it is, because as far as
20:18
Speaker 2: the AI is concerned, it is confident it did find
20:21
Speaker 2: that information, it is returning it, it is retreating it
20:23
Speaker 2: as is reported. So as according to the AI, that
20:25
Speaker 2: is definitively factually true. So it's yeah, and this is
20:31
Speaker 2: this goes into the bias of how do we manage
20:34
Speaker 2: these ll ms, because if they're not trained on objective
20:40
Speaker 2: sources like Wrokipedia. If that is objective, I'm sure there's
20:43
Speaker 2: going to be errors everywhere, unintentional, but of course at
20:46
Speaker 2: least closer to the truth. Sure that will be the
20:49
Speaker 2: reference that the language model uses, as opposed to Wikipedia,
20:52
Speaker 2: where anything could be in there, anything true or false.
20:55
Speaker 1: And edited by a stranger.
20:56
Speaker 2: Who stranger. Yeah, and if someone is you talking to
21:00
Speaker 2: that AI, and you know, referencing that, the AI will say, yes,
21:04
Speaker 2: I am certain that this is true and put that down.
21:09
Speaker 2: There's no guarantee it's wrong. And then again, if you
21:12
Speaker 2: don't know that, you're not the wiser, you'll take that, Hey,
21:14
Speaker 2: this Alb's right about most things, it must also be
21:17
Speaker 2: right about this.
21:18
Speaker 1: It's so crazy, I have to say, I really anyways,
21:23
Speaker 1: it's something I think we'll need to follow. I just
21:25
Speaker 1: don't know as this grows exponentially and quickly, I mean literally,
21:32
Speaker 1: you know, as you spoke to a moment ago, the
21:36
Speaker 1: way that you are looking at AI as an engineer,
21:39
Speaker 1: as a scientist, as somebody who has studied this, you
21:43
Speaker 1: know extensively the amount of changes that are happening and
21:47
Speaker 1: at the speed at which they're happening. To me as
21:49
Speaker 1: a user, I'm like, oh my, I couldn't do this
21:51
Speaker 1: a week ago. Yeah, that to me is a little scary,
21:56
Speaker 1: right as a user.
22:00
Speaker 2: Yeah yeah, And if you lay out the timeline, you
22:04
Speaker 2: know that there is a little bit of a frighteningly
22:07
Speaker 2: concerning fast pace to it. But yes, also if we
22:10
Speaker 2: look back, you know when cell phones rolled out against
22:13
Speaker 2: the iPhone and all that, it happened like a blinkin
22:15
Speaker 2: eye and then we just normalized it. I guess the
22:18
Speaker 2: same type of pattern is happening here where you know,
22:21
Speaker 2: four years ago, you had a language model that could
22:24
Speaker 2: create a paragraph and it could start to pass as
22:28
Speaker 2: you know, pleading college level tests, you know, twenty two
22:31
Speaker 2: to twenty three. Right then it evolved into okay, now
22:34
Speaker 2: it's best a PhD.
22:36
Speaker 1: It can it can write code, it can do JavaScript,
22:39
Speaker 1: it's in Python.
22:40
Speaker 2: I'm like, what, Yeah, Well that's where we're at now,
22:44
Speaker 2: where we have we're sort of in the agent aspect
22:46
Speaker 2: of this evolution. And what agents are is you can
22:50
Speaker 2: basically take specialized AI and they have the intelligence to
22:56
Speaker 2: do anything digitally that you can do. So if you
22:58
Speaker 2: move your mouse on your screen, it can mimic that.
23:01
Speaker 2: If you go and log into a website. If you
23:03
Speaker 2: go in and you know, start doing what you were
23:05
Speaker 2: saying earlier, editing a photo document. It can do all
23:09
Speaker 2: of that. If you it will ask for your credentials,
23:12
Speaker 2: It will, It can reason, and it can actually deliver
23:15
Speaker 2: anything you do on a computer. It can now do
23:17
Speaker 2: that independently, which has a lot of other security risks associated.
23:24
Speaker 1: As it goes into your bank account.
23:26
Speaker 2: It can absolutely do that. Yeah, if you give it control,
23:28
Speaker 2: and I mean a lot of times it does property. Hey,
23:29
Speaker 2: would you like me to use this and do that?
23:31
Speaker 2: But it is at the level now where we're no
23:33
Speaker 2: longer interacting with something that can chat back to us
23:37
Speaker 2: and refer information and it can tell us things, It
23:39
Speaker 2: can do things, It can do things. And you think
23:41
Speaker 2: about how much we can do digitally through our phones,
23:43
Speaker 2: through our computers that we do and interact with every day.
23:47
Speaker 2: Now an AI has the ability to do those same
23:50
Speaker 2: types of tasks, right, and which leads us to, you know,
23:54
Speaker 2: what's the next evolution of the role We know right
23:57
Speaker 2: who we're talking to on the other end of a
23:59
Speaker 2: phone or a digital computer. Because voice generation is getting better, Yes,
24:06
Speaker 2: before we know it, regular voices voice, your voice will
24:11
Speaker 2: be available to be used.
24:12
Speaker 1: Yes, it's I have to say, I there are aspects
24:16
Speaker 1: of that, you know. I was talking to somebody other
24:17
Speaker 1: day and they were saying, you know, there are people
24:20
Speaker 1: in the trades, whether it's HVAC or electrical or plumbing
24:24
Speaker 1: or what have you, and they are just marching down
24:25
Speaker 1: the street. They're like, we're breaking it in because you
24:27
Speaker 1: people are gonna be out of work. And Mark Cuban
24:31
Speaker 1: was saying, I saw two things in the last two
24:34
Speaker 1: days that really made me think. Kevin O'Leary, who obviously
24:38
Speaker 1: is a real estate mogul among just a serial entrepreneur,
24:41
Speaker 1: and then Mark Cuban, Who's Mark Cuban. And we'll leave
24:43
Speaker 1: that there, but they both spoke in the last two
24:47
Speaker 1: days about AI, the advance of AI and what it
24:50
Speaker 1: does to the workforce. Kevin O'Leary said it would make
24:54
Speaker 1: sense for us to look into helping to build the
24:58
Speaker 1: data centers and and that that would be the next
25:01
Speaker 1: big business, Like if you can help to facilitate building
25:04
Speaker 1: out these data centers, managing the data centers, helping to
25:08
Speaker 1: create the space and the functionality without overtaking, I don't know,
25:12
Speaker 1: sort of like the same idea of like when people
25:14
Speaker 1: are like, yeah, we need wind turbines, you're like, you're
25:16
Speaker 1: a moron. That shit doesn't work and you're killing the
25:18
Speaker 1: birds and the farms, so knock it off, right, So
25:21
Speaker 1: we want to get around that. We don't want to
25:23
Speaker 1: have the data center wind turbine moment. But Mark Cuba
25:26
Speaker 1: made the point which I thought was really interesting, and
25:30
Speaker 1: he said, if I were starting out today and I
25:34
Speaker 1: wasn't handy or in a trade or the kind of
25:37
Speaker 1: person that was more tactical, right, he said, I would
25:41
Speaker 1: be a person that would be helping people who are
25:45
Speaker 1: in single owner operated businesses to learn how to use
25:50
Speaker 1: AI because a lot of them could use the help
25:54
Speaker 1: and they just don't know how to use it. And
25:56
Speaker 1: I thought that's really interesting. He made a comment that
25:59
Speaker 1: about thirty three out of thirty three million businesses that
26:03
Speaker 1: we have in the United States, thirty million of them
26:06
Speaker 1: are single owner operated. They don't have staff, they don't
26:10
Speaker 1: have anything. It's just them their company and they're doing it.
26:12
Speaker 1: And I thought, huh, so if they did have the
26:15
Speaker 1: help of an AI agent, to your point, they could
26:18
Speaker 1: reduce their workflow right and probably expedite the process of
26:23
Speaker 1: whatever business they have and I thought, huh, that's interesting,
26:28
Speaker 1: But then it comes into how are we facilitating the energy?
26:33
Speaker 1: So to your point, you and I were speaking before
26:35
Speaker 1: we got started that this is going to take an
26:38
Speaker 1: exorbitant amount of energy.
26:42
Speaker 2: Yeah, absolutely in that scene.
26:44
Speaker 1: Yeah yeah.
26:47
Speaker 2: Where the sort of AI economy is going is write
26:52
Speaker 2: down those paths two of what I'll use it to agents.
26:56
Speaker 2: That's the big buzzword right now. How do you generate
26:59
Speaker 2: an HVA management agent? For example? Obviously they're not going
27:03
Speaker 2: to go and install hvacs and do repairs and all that.
27:06
Speaker 2: All the business operations that are involved with that, or
27:09
Speaker 2: you know a family owned business or you know someone
27:11
Speaker 2: who's only up one or two employees, whether it be
27:15
Speaker 2: like I was saying, answering calls in the future, with
27:17
Speaker 2: you have voice capabilities, managing your invoices, managing your calendar,
27:23
Speaker 2: all those ordering parts, all of that can be moved
27:28
Speaker 2: to an agentic AI that can solve some of those tasks.
27:31
Speaker 2: I mean, not all at this exact moment, but the
27:35
Speaker 2: speed we're going at, I mean it's going to be
27:36
Speaker 2: here very shortly here And those are some examples where
27:42
Speaker 2: that's a new business. That's if someone creates that business
27:44
Speaker 2: that the agent that now solves or helps with you know,
27:48
Speaker 2: marketing plans or outreach and things like that. You scale
27:51
Speaker 2: that across that's not now energy usage. So you're creating
27:55
Speaker 2: not only the original lllms being used by meta by
27:59
Speaker 2: right by open AI, but now the intermediary businesses who
28:05
Speaker 2: are using those ais to solve a practical human problem.
28:10
Speaker 2: And that's all energy demand. So as more and more
28:13
Speaker 2: of these AI businesses, you know, become up, they become invented.
28:19
Speaker 2: It's more demands, more strain, more power resources, and the
28:25
Speaker 2: grid that's already at its capacity may not be able
28:29
Speaker 2: to handle where it's headed.
28:32
Speaker 1: So to that point, obviously, people like us that deal
28:36
Speaker 1: in this type of work. You know, we've talked many
28:38
Speaker 1: times about e MP issues as well as the Texas
28:43
Speaker 1: grid being the main supplier of electric for the entire country.
28:47
Speaker 1: Under Joe Biden, he released the contract for the chips
28:50
Speaker 1: that go into that grid to come from China. Trump
28:55
Speaker 1: had gotten rid of that. He's like, I only want
28:57
Speaker 1: American made chips in the electric grid for the simple
29:00
Speaker 1: fact that China is not a friend, They are a foe.
29:04
Speaker 1: Who the hell knows what's in those chips, and now
29:07
Speaker 1: they're in charge of the electric that feeds the nation.
29:10
Speaker 1: I mean, have we all lost our mind? Right? So
29:13
Speaker 1: Trump is now telling AI, we're not going to use
29:17
Speaker 1: that grid. First of all, it cannot handle it. Second
29:21
Speaker 1: of all, it would jack up the cost of the
29:23
Speaker 1: electric costs for every average American citizen, many of whom
29:27
Speaker 1: are not using AI right, or they're using it without
29:30
Speaker 1: knowing it right. They're talking to SERI, they're using their iPhone.
29:33
Speaker 1: They're not actively using AI like you and I aren.
29:36
Speaker 1: But I read this post which I've shared with you,
29:38
Speaker 1: and I'm going to share it with the audience now.
29:41
Speaker 1: And President Trump yesterday had a roundtable with the very
29:44
Speaker 1: people you're talking about, the CEOs from Amazon and Meta
29:47
Speaker 1: and Microsoft and open Ai and Oracle and XA and
29:50
Speaker 1: all these people that are sort of like the they're
29:54
Speaker 1: the lifting and the starting spot for all things AI.
29:57
Speaker 1: And they had the roundtable which one of the main
30:00
Speaker 1: points that he was talking about was the ratepayer protection Pledge,
30:03
Speaker 1: which he's saying Americans will not pay higher electric bills
30:07
Speaker 1: do to power usage from AI data centers. Well, how
30:10
Speaker 1: do we get there? Oh, I know, We'll make all
30:13
Speaker 1: of these big CEOs figure out how to fund, create,
30:17
Speaker 1: and provide their own utilities. Okay, so now they don't
30:22
Speaker 1: only control AI all of these quote unquote as you're
30:26
Speaker 1: saying AI agents, but they own all the power from
30:30
Speaker 1: nuclear and they're their own utility. So they run all
30:34
Speaker 1: the power and utility in the United States, because that's
30:36
Speaker 1: going to be some massive utility. And he made a
30:38
Speaker 1: great example. This guy's post as Ricardo T was his
30:42
Speaker 1: name full disclosure. I have no idea who this dude is,
30:45
Speaker 1: but he made an excellent analogy which says, by the
30:50
Speaker 1: end of this year twenty twenty six, there will be
30:52
Speaker 1: five US data centers which will each consume over one
30:57
Speaker 1: gigawatt of continuous power. To give you perspective, one gigawatt
31:02
Speaker 1: powers eight hundred and fifty thousand homes. So five of
31:07
Speaker 1: these facilities will use more electricity than some entire countries,
31:12
Speaker 1: which obviously the grid cannot handle. So we're talking about
31:16
Speaker 1: now they're going to own all this power. They're going
31:20
Speaker 1: to be buying nuclear reactors. We've got Meta signing twenty
31:24
Speaker 1: year nuclear deals. Chevron's getting a two and a half
31:27
Speaker 1: gigawat natural gas plant in West Texas. So what exactly
31:32
Speaker 1: are the tech companies controlling. So they're controlling computer computering, data,
31:37
Speaker 1: energy infrastructure. Like we're exactly talking about tech companies anymore.
31:43
Speaker 1: These are becoming like superpowers, they're becoming their own I
31:48
Speaker 1: hate to say it, but like their own countries. Yeah,
31:53
Speaker 1: your thoughts, big setup.
31:56
Speaker 2: Right, it's another one. And I feel like this one
31:58
Speaker 2: goes two ways because you know, most people they get
32:04
Speaker 2: their energy bill, they see the total, right, maybe they
32:06
Speaker 2: look at their kill a lot usage for the last month,
32:09
Speaker 2: but they don't really go into the intricacies of how
32:11
Speaker 2: that energy bill works. And it's actually somewhat complex from
32:15
Speaker 2: the energy standpoint, because you have peak pricing, you have
32:18
Speaker 2: time of use, You have different categories that a user
32:22
Speaker 2: can be used or placed into based on their location,
32:26
Speaker 2: their region obviously of the three grids. But then for example,
32:29
Speaker 2: the let's take peak pricing certain hours of the day.
32:33
Speaker 2: You know they recommend don't run your dishwasher or whatever
32:36
Speaker 2: it might be, right, because that's when the general societal
32:39
Speaker 2: demand is at its highest, so you get built at
32:41
Speaker 2: a higher rate you use sometimes you I mean, that's
32:45
Speaker 2: obviouslyp because's when everybody's active and doing things. And also
32:50
Speaker 2: most likely when people are going to be using AI
32:53
Speaker 2: as well, because they're up indirectly, right, So you think
32:57
Speaker 2: about it, how it's like you said, it's it's going
32:59
Speaker 2: to be in so people will be using it indirectly
33:01
Speaker 2: without knowing it. So they'll be using it through their phones,
33:04
Speaker 2: talking to you know, various agents on the other side
33:07
Speaker 2: of a website. But that's all going to be additional
33:08
Speaker 2: AI calls to the data center and back and forth.
33:13
Speaker 2: So to that kind of point there is that now
33:18
Speaker 2: will increase all of our energy bills because we're directly
33:21
Speaker 2: and indirectly using AI, either deliberately or just by living life.
33:26
Speaker 2: So as that energy consult goes up, the pricing goes
33:30
Speaker 2: up because the demand goes up. The energy companies have
33:33
Speaker 2: to sort of buy from third party providers, which goes
33:35
Speaker 2: into that whole equation. So moving them into their own
33:39
Speaker 2: independent system does separate that usage from what we use
33:45
Speaker 2: as regular consumers. Okay, so we're no longer sort of
33:50
Speaker 2: using them and then paying the energy companies to use
33:54
Speaker 2: them as well, you know, in this sort of cyclical
33:56
Speaker 2: loop of just energy demand and consumption. So right, it
34:01
Speaker 2: does create the message of their own independent modes of
34:06
Speaker 2: energy companies, data companies, information companies, knowledge companies, really knowledge companies.
34:12
Speaker 1: Well, it's also that idea, Dave, if you think about it, right,
34:15
Speaker 1: we're talking about and it's a double on tone in
34:18
Speaker 1: every way. It's a power company. It's company. It's a
34:22
Speaker 1: power company. It has power over you, and it has
34:25
Speaker 1: the power literally to source and feed its own demands.
34:31
Speaker 1: So at what point does the nation start saying, hey,
34:35
Speaker 1: we would like to lease power from you that you're
34:38
Speaker 1: not using at xyz rate and then we're going to
34:43
Speaker 1: send that back to the American public Because you probably
34:47
Speaker 1: know better than me. How long have we been saying
34:50
Speaker 1: the Texas grid is completely and totally overweighted. It cannot
34:54
Speaker 1: withstand the amount that we are pushing, know on it,
35:00
Speaker 1: like the demand is so much greater than its ability
35:04
Speaker 1: to provide. Am I speaking out of turn? Or is
35:06
Speaker 1: that accurate? No?
35:08
Speaker 2: I mean you see it sometimes in those peak summer days.
35:12
Speaker 1: Where blackouts, Sure, the brownouts in California and then the
35:15
Speaker 1: blackouts and other places.
35:16
Speaker 2: Sure, but it is it is already here because I mean,
35:19
Speaker 2: then you think about the other aspects. We have increased
35:22
Speaker 2: ev usage, you know, more and more ED cars are
35:24
Speaker 2: being bought, more of people are moving to the electric society.
35:29
Speaker 1: Rights, not just the ballots, which is still, by the way,
35:32
Speaker 1: completely and totally supported by fossil fuels. But we'll leave
35:35
Speaker 1: that there for the environmentalists.
35:37
Speaker 2: Yeah, exactly, Like, don't let the.
35:39
Speaker 1: Facts get in the way of your green efforts. Okay, please,
35:43
Speaker 1: good luck.
35:45
Speaker 2: Back to the power plants.
35:46
Speaker 1: I just can't. I'm like, some things just are I'm like,
35:49
Speaker 1: that's cool if you want to plug in, but by
35:51
Speaker 1: the way, that plug doesn't work unless there's an energy
35:54
Speaker 1: grid that is supported by fossil fuels to give you
35:57
Speaker 1: that energy. But keep going. You're doing great, no car right,
36:01
Speaker 1: good talk.
36:02
Speaker 2: Yeah, yeah, I mean, well, hey, if they control the
36:04
Speaker 2: power and the knowledge, then they can convince you of anything.
36:06
Speaker 1: So that's right, and the message.
36:08
Speaker 2: To AI, right, And you know that's a good point
36:11
Speaker 2: too about the power control, because you know, right now
36:15
Speaker 2: we're kind of on the transition to AI as our
36:17
Speaker 2: knowledge repository. For years, we had Google, where we'd go
36:21
Speaker 2: and Google search something and whatever. One of the top hits,
36:23
Speaker 2: top pages are what we could find and access. So
36:27
Speaker 2: if you were a piece of information that was factual,
36:29
Speaker 2: but you were not in Google's top two three pages,
36:33
Speaker 2: you wouldn't be discoverable really because very people are clicking
36:36
Speaker 2: that far into so the knowledge you only get shown
36:40
Speaker 2: You either have to say, all right, do I take
36:41
Speaker 2: this because it's being shown to me through Google. They're
36:44
Speaker 2: not so many other alternatives, so I guess this must
36:47
Speaker 2: be the truth, you know, if you're deductively questioning it.
36:50
Speaker 2: Same thing with AI, same thing with power control, and
36:53
Speaker 2: sort of the concern of where do we draw the
36:56
Speaker 2: lines of autonomy what these companies can do, because if
37:02
Speaker 2: they're you know, going back to the bias of their
37:05
Speaker 2: not only independent energy wise, you know what they can
37:09
Speaker 2: share with us knowledge wise, and then they become integrated
37:12
Speaker 2: with all of our digital applications, we have no alternative
37:15
Speaker 2: in some prices, and then we're sort of behold into
37:18
Speaker 2: what they tell us or don't tell us, or what
37:20
Speaker 2: truth they want to convey, or how they want to
37:23
Speaker 2: subtly massage it. Like you were saying about the four
37:26
Speaker 2: million dollars in palates. They'll kind of maybe they'll give
37:29
Speaker 2: you the answer, but they'll position it in a way
37:31
Speaker 2: that convinces you that, oh, you should think about it
37:34
Speaker 2: this way as opposed to that way.
37:36
Speaker 1: Well, I'll give I'll give you a little insight into
37:39
Speaker 1: how I research, which is really funny. Obviously, if you're
37:45
Speaker 1: in your thirties or forties, you used Britannica, you used encyclopedias,
37:49
Speaker 1: you had pencils, you know, you did things a little differently.
37:52
Speaker 1: I personally think we need to go back to that
37:54
Speaker 1: for a few reasons. One, we have nine year olds
37:56
Speaker 1: wearing glasses because of screen time and the strain on
37:59
Speaker 1: your right. And then they're playing video games. And you know,
38:03
Speaker 1: my older kids are on Snapchat. I mean, and you
38:06
Speaker 1: know where we used to like text each other and
38:09
Speaker 1: we thought that was so cool, or we had instant
38:11
Speaker 1: Messenger we thought that was so cool, or Google Chat.
38:15
Speaker 1: These kids aren't even using words. They send pictures of
38:18
Speaker 1: their face. If they don't want to talk, they send
38:21
Speaker 1: pictures of the ceiling. I'm like, why are you sending
38:25
Speaker 1: what is happening? Why are you sending a picture of
38:26
Speaker 1: our ceiling? I don't understand. Oh, that's just slip them
38:28
Speaker 1: and I don't want to talk. Okay, you could just
38:30
Speaker 1: not answer them. That's also a good way. I've tried
38:33
Speaker 1: that a couple of It's the old school way of
38:35
Speaker 1: not answering. I'm not interested, Thanks for reaching out. I
38:37
Speaker 1: mean just this idea of always being connected, always doing
38:41
Speaker 1: a I always having all this integration. But when I search,
38:46
Speaker 1: I go to the twelfth or the thirteenth page and
38:49
Speaker 1: I work backwards because I want to see what's actually
38:54
Speaker 1: what is the truth of the story, right, Because I
38:56
Speaker 1: know the SEO, which for those of you listening is
38:59
Speaker 1: search engine optimization, and you can pay for that will
39:02
Speaker 1: put you in that top page or the very top
39:05
Speaker 1: of the second page. Well probably, you know, if you're
39:08
Speaker 1: paying to get to number one, I probably don't want
39:11
Speaker 1: to know who you are. I want to know who's
39:12
Speaker 1: on page four because they're the guy that's grinding it
39:14
Speaker 1: out and there is some truth to that.
39:16
Speaker 3: You know.
39:17
Speaker 1: I'll get back to page one and two. That's not
39:18
Speaker 1: a problem. But chat GPT, while it was first and
39:23
Speaker 1: Opening Eye was first, I will say I think GROCK
39:26
Speaker 1: has given it a run for its money, which really
39:29
Speaker 1: speaks to the change, because I don't feel like anybody
39:32
Speaker 1: ever beat Google, Like I like Safari, I like Brave,
39:35
Speaker 1: I like Duck, dup Goo. I feel like all of
39:37
Speaker 1: them are less biased than Google, but Google is still
39:41
Speaker 1: the king.
39:44
Speaker 2: Yeah, yeah, I mean Google to be fair. They are
39:47
Speaker 2: the Google Brain, the Google labs, all of their sort
39:50
Speaker 2: of research. Are the scientists and academics who fraught us
39:55
Speaker 2: the algorithms initially. I mean obviously a lot of people
39:58
Speaker 2: who are involved. But you look at of the fundamational
40:00
Speaker 2: fundamental papers and research work, a lot of it was
40:03
Speaker 2: out of Google one hundred percent. Other companies sort of
40:05
Speaker 2: took it and evolved in advanced and commercialized it. But
40:08
Speaker 2: the fundamental kind of ideas of how do you mathematically
40:12
Speaker 2: create an AI word generating thing was based out of
40:18
Speaker 2: Google apps. So yeah, yeah, they are. They know they
40:24
Speaker 2: know too much, they.
40:25
Speaker 1: Know too much. Yeah scares me. Now they're going to
40:28
Speaker 1: know too much and they're going to be in charge
40:29
Speaker 1: of how much we know about what they know. Like,
40:32
Speaker 1: I'm out, I need I don't know, I need a
40:34
Speaker 1: farm far away with dogs and other things, chickens. I'll
40:38
Speaker 1: take anything for with four legs over this crap. But Dave,
40:41
Speaker 1: you're the best. Thank you for joining Roague Recap today. Guys,
40:44
Speaker 1: if you're listening, thank you for being with us. You
40:46
Speaker 1: know what to do at Roague Recap at Lindamickrogue recap
40:49
Speaker 1: dot com. We will be back tomorrow with more stacyfe
40:52
Speaker 1: out there and pray for our troops.
40:53
Speaker 2: Night.
40:53
Speaker 1: Everybody,