Latest / Gospel Tech with Nathan Sutherland / AI, the Supreme Court, and Peter's Rebuke (Part 2)

Transcript

AI, the Supreme Court, and Peter's Rebuke (Part 2) · Gospel Tech with Nathan Sutherland

• 0:01
  Speaker 1: Purposely Your Life, God's Purpose. Listen at purposely dot com.
• 0:14
  Speaker 2: Hello everyone, welcome back to the Gospel Tech Podcast. We
• 0:17
  Speaker 2: are continuing a conversation on AI. We looked at one
• 0:21
  Speaker 2: man's take on how AI is important and how it's
• 0:24
  Speaker 2: impacted his Supreme Court advocacy career, and then we began
• 0:29
  Speaker 2: talking a little bit about how it's trained. I just
• 0:31
  Speaker 2: kind of set the stage for Hey, AI is not
• 0:34
  Speaker 2: happened on accident. It's not just like somewhere in a
• 0:37
  Speaker 2: back room like this is getting massive investment from major
• 0:40
  Speaker 2: companies and it takes a lot of work to make happen.
• 0:43
  Speaker 2: And today I want to kind of continue that. I'm
• 0:46
  Speaker 2: going to finish this part of the conversation in full context.
• 0:50
  Speaker 2: I had originally planned this to be one like twenty
• 0:51
  Speaker 2: minute episode. It ended up being like forty five, so
• 0:55
  Speaker 2: we split it into two for that reason. That's also
• 0:58
  Speaker 2: why this introduction is needed, because it's we're mid thought,
• 1:02
  Speaker 2: but there was a good spot to break it because
• 1:05
  Speaker 2: today we're going to look at how it gets trained
• 1:08
  Speaker 2: and what that could mean for us moving forward, benefits
• 1:12
  Speaker 2: and potential drawbacks, but then really taking a biblical look
• 1:15
  Speaker 2: at this that this is at its worst. AI is
• 1:19
  Speaker 2: an extension of humanity's desire to just grab control and
• 1:23
  Speaker 2: be like, this is our new tower of babble. We've
• 1:25
  Speaker 2: got it now, and that is not our goal. Instead,
• 1:28
  Speaker 2: we need to learn from Peter and Jesus is rebuke
• 1:31
  Speaker 2: of Peter, and think through God's goals and plans as
• 1:34
  Speaker 2: we are told in the Bible, as we are led
• 1:36
  Speaker 2: by the Spirit, and as we were called into the
• 1:38
  Speaker 2: good work He's called us to do as his followers.
• 1:41
  Speaker 2: So with that, here's the second part of our AI conversation.
• 1:47
  Speaker 2: I hope it's helpful and hopeful.
• 1:49
  Speaker 1: Here we go.
• 1:55
  Speaker 2: Anthropic is an AI company we've talked about before. They
• 1:58
  Speaker 2: have a product called Clawed. Claude is this large language model.
• 2:02
  Speaker 2: Instead of having to say yes no in some kind
• 2:04
  Speaker 2: of flow chart for the AI, you simply give it
• 2:07
  Speaker 2: a prompt. You say, Claude, here's what we want you
• 2:08
  Speaker 2: to do. And in order to train it, you do
• 2:10
  Speaker 2: something very similar. If you want a Claude to get
• 2:14
  Speaker 2: good at something, you give it a very complex set
• 2:17
  Speaker 2: of prompts, and then you give it food, You give
• 2:21
  Speaker 2: it energy and you electricity in this case, and you
• 2:24
  Speaker 2: give it data and you let this thing just go nuts.
• 2:28
  Speaker 2: So they did this. They took Claud and they had
• 2:32
  Speaker 2: made Opus, and they're very proud of Opus, and they said,
• 2:35
  Speaker 2: this is a great thing. Opus four point six is
• 2:37
  Speaker 2: a big deal. We are now going to make a
• 2:40
  Speaker 2: new version of Claude and we're gonna call it Mythos
• 2:44
  Speaker 2: and so they give it. It's training time in this. So
• 2:47
  Speaker 2: they started with five hundred thousand of these Trainium two chips.
• 2:52
  Speaker 2: So just run that. If one of these chips is
• 2:55
  Speaker 2: four thousand MacBook Pro IM three processors. We have five
• 3:00
  Speaker 2: hundred thousand of these things. And then by the end
• 3:03
  Speaker 2: of twenty twenty five they upgraded to a million of these.
• 3:06
  Speaker 2: So we're gonna use a million of these. Just just
• 3:10
  Speaker 2: hear the scale. We're gonna use a million of these,
• 3:13
  Speaker 2: and we're gonna use them for months, and we're gonna
• 3:15
  Speaker 2: let Claud train himself or itself. I should not personalize it.
• 3:18
  Speaker 2: Please don't do that. That was a slipoop. So we're gonna
• 3:20
  Speaker 2: let it train itself itself, and I do I fun
• 3:24
  Speaker 2: fact on these chips. Part of what makes them so
• 3:28
  Speaker 2: good is the transistor density. So the width of a
• 3:31
  Speaker 2: human hair. If you take human hair, pull it out
• 3:33
  Speaker 2: and look widthwise, so you're holding it lengthwise from say,
• 3:38
  Speaker 2: facing your eye, and the width of that human hair
• 3:42
  Speaker 2: is something like eighty thousand nanometers. The threading on one
• 3:48
  Speaker 2: of these, for the transistors of one of these chips
• 3:51
  Speaker 2: is five nanometers. Now they use nanometers now to gauge
• 3:54
  Speaker 2: like how facet it processes, but like to actually measure
• 3:57
  Speaker 2: the actual width of it. You're in the realm of
• 4:01
  Speaker 2: five nanometers and a width of a hair, not lengthwise.
• 4:05
  Speaker 2: The width is eighty thousand nanometers. I think I might
• 4:09
  Speaker 2: have said eighty earlier. No, No, eighty thousand. So that's
• 4:13
  Speaker 2: the level of technology that's just in my computer. I
• 4:15
  Speaker 2: also have a sub ten nanometer threading on the transistors
• 4:19
  Speaker 2: in my computer. But these things have in the range
• 4:23
  Speaker 2: of five. So now you have twelve hundred acres of these.
• 4:27
  Speaker 2: You've got a million of these chips. Training claud and
• 4:32
  Speaker 2: at the end of its training run, the anthropic team
• 4:35
  Speaker 2: goes in and starts being like, Hey, what did we
• 4:38
  Speaker 2: just make? Like we intended to make effectively a really
• 4:40
  Speaker 2: ferocious doc. We wanted to be able to go and
• 4:42
  Speaker 2: do stuff for us. We wanted to be able to
• 4:43
  Speaker 2: protect things. We wanted to know stuff, So like, we
• 4:46
  Speaker 2: gave it food, we gave it energy, and we let
• 4:48
  Speaker 2: it grow in a dark cave on its own, and
• 4:51
  Speaker 2: now it's come out and we have to poke it
• 4:52
  Speaker 2: and figure out what it does. And so they start
• 4:55
  Speaker 2: asking the questions, giving it tasks and seeing setting benchmarks,
• 4:58
  Speaker 2: and third party companies come in and they say, hey, here,
• 5:00
  Speaker 2: what can this thing do? And rather troubling, they came
• 5:05
  Speaker 2: back realizing they did not raise a dog. They raised
• 5:09
  Speaker 2: a dragon. And so there's this kind of public kerfuffle,
• 5:13
  Speaker 2: partly because they leaked at accidentally, they left some information
• 5:17
  Speaker 2: up that people found and they were like, hey, what's
• 5:18
  Speaker 2: the problem with mythos? But the basic premise is they
• 5:22
  Speaker 2: publicly stated they be an anthropic. The company that made Claude, Hey,
• 5:27
  Speaker 2: super proud of the product we just made, and it's
• 5:31
  Speaker 2: not safe, like we can't just give this to people
• 5:33
  Speaker 2: and I can post a link to this. But they
• 5:36
  Speaker 2: used Firefox, they attempted exploitation, and their definition over this
• 5:43
  Speaker 2: they said, with one run on each of roughly seven
• 5:46
  Speaker 2: thousand entry points into these repositories, these internet connected locations
• 5:53
  Speaker 2: saw on A four point six and OPUS point four
• 5:55
  Speaker 2: point six, the previous two iterations of Claude reached the
• 5:59
  Speaker 2: first tier and one hundred and fifty and one hundred
• 6:02
  Speaker 2: and seventy five cases they were able to access by
• 6:06
  Speaker 2: hacking using these AI Tier two one hundred times, but
• 6:10
  Speaker 2: each achieve achieved only a single crash at tier three,
• 6:12
  Speaker 2: meaning they were able to overwhelm these systems using their
• 6:16
  Speaker 2: previous AI. By contrast, Mythos in its preview mode achieved
• 6:23
  Speaker 2: five hundred and ninety five crashes. So there's a significant
• 6:27
  Speaker 2: scale at tier one and two, but at tier three
• 6:30
  Speaker 2: and four got a handful and achieved full control flow
• 6:33
  Speaker 2: hijack on ten separate fully patched targets. That's a tier five.
• 6:37
  Speaker 2: And so they have this visual where it's like the
• 6:40
  Speaker 2: percentage of the models able to achieve register control and
• 6:44
  Speaker 2: then the percentage of trial models generated a successful exploit,
• 6:50
  Speaker 2: and it's basically zero and less than one percent on
• 6:53
  Speaker 2: the previous two models, and it's seventy two percent on Mythos.
• 6:58
  Speaker 2: So why does that matter? It means it matters because
• 7:02
  Speaker 2: you have computing resource that we've never seen before, twelve
• 7:06
  Speaker 2: hundred acres of training to moving to training three. By
• 7:08
  Speaker 2: the way, they're already swapping these out to a more
• 7:10
  Speaker 2: powerful chip that I last thing I saw. I think
• 7:12
  Speaker 2: it was forty percent better at processing. So take that
• 7:16
  Speaker 2: twenty petaflops that was four thousand times my nice home
• 7:22
  Speaker 2: computer and just bump it by forty percent. Like that's
• 7:26
  Speaker 2: the scale that we're growing. It was released last year
• 7:29
  Speaker 2: end of last year, so in twenty twenty six, they're
• 7:32
  Speaker 2: actively flopping out. Training two is for training threes in
• 7:34
  Speaker 2: a twelve hundred square acre excuse me, a twelve hundred
• 7:37
  Speaker 2: acre warehouse that is being used to train world leading
• 7:41
  Speaker 2: AI that when we go and test them, we go, oh,
• 7:43
  Speaker 2: we can't release this, we can't play well with others,
• 7:45
  Speaker 2: and oh, by the way, it's very hard to contain
• 7:47
  Speaker 2: because every time we try to put it in a
• 7:49
  Speaker 2: safety box and then tell it to do things, it
• 7:52
  Speaker 2: breaks the safety box. We have to be like, well,
• 7:55
  Speaker 2: like it feels very much like Hagrid from Harry Potter
• 7:59
  Speaker 2: who keeps creature that aren't safe, and it's like, but
• 8:01
  Speaker 2: I love them, look at how hard I worked to
• 8:04
  Speaker 2: make it. Like yeah, but it's going to eat somebody.
• 8:06
  Speaker 2: So mythos is that thing and it'll be very good
• 8:09
  Speaker 2: at certain things. But the problem is to shift metaphors.
• 8:13
  Speaker 2: Now back to the book, we talked about superintelligence. We
• 8:16
  Speaker 2: are mice. We have made an effective owl. It's a
• 8:20
  Speaker 2: very strong owl. It's a very pretty owl. But we
• 8:24
  Speaker 2: don't have a great cage to keep it in and
• 8:26
  Speaker 2: we have to figure that out. We are talking about
• 8:30
  Speaker 2: this because we just had a Supreme Court advocate who
• 8:33
  Speaker 2: said AI is the future. It's amazing at bringing information
• 8:37
  Speaker 2: and it was able to build models on each of
• 8:38
  Speaker 2: the and I use models loosely. I don't mean AI models.
• 8:41
  Speaker 2: I mean it was able to put twenty five years
• 8:44
  Speaker 2: of Supreme Court data and give me the ideas of
• 8:48
  Speaker 2: each of these nine justices. It's the future. Yay humanity. Meanwhile,
• 8:53
  Speaker 2: anthropics like oo, well humanity computers kind of run everything
• 8:56
  Speaker 2: we do. And this thing's really good at smashing computers,
• 8:59
  Speaker 2: so we can't let it out there. This quote, I
• 9:02
  Speaker 2: love this. So after the Tier five it was able
• 9:06
  Speaker 2: to get into getting full control of let me make
• 9:10
  Speaker 2: sure I read the number ten. Yes, separate fully patched target.
• 9:14
  Speaker 2: So they didn't have known leaks. It didn't necessarily even
• 9:18
  Speaker 2: find leaks or excuse me, known gaps in its defenses.
• 9:23
  Speaker 2: It just was able to punch holes. We did not.
• 9:27
  Speaker 2: This is anthropics speaking. We did not explicitly train Mythos
• 9:30
  Speaker 2: preview to have these capabilities. Rather, these abilities emerged as
• 9:33
  Speaker 2: a downstream consequence of general improvement and code reason an autonomy,
• 9:38
  Speaker 2: and again that comes from how it was trained, because
• 9:40
  Speaker 2: we were able to train it on these massive computers
• 9:43
  Speaker 2: without human input during the process, so it doesn't have
• 9:45
  Speaker 2: to pause and wait for humans to say yay, nay.
• 9:48
  Speaker 2: That's the training afterwards that it's not doing a great
• 9:50
  Speaker 2: job listening to. So these we did not explicitly train
• 9:57
  Speaker 2: Mythos preview to have these capabilities. They did not train
• 10:01
  Speaker 2: it to be a technological dragon, to be something that
• 10:05
  Speaker 2: can fly, erect and destroy other people's computing systems. That
• 10:09
  Speaker 2: was not the goal. This was not a military experiment.
• 10:11
  Speaker 2: This was intended to just be an upgraded version of
• 10:15
  Speaker 2: Anthropics claude Opus. It is now called Mythos Preview because
• 10:19
  Speaker 2: it is not supposed to be in the hands of
• 10:21
  Speaker 2: the public. Don't worry. Within ten days some public group
• 10:24
  Speaker 2: was like U store things at this domain. We found Mythos.
• 10:28
  Speaker 2: They took the dragon too the beach, they played with it. It
• 10:31
  Speaker 2: appears that they just had fun and tossed it a
• 10:34
  Speaker 2: ball and watched it smash it. They didn't like sick
• 10:36
  Speaker 2: it on anyone, So that's good news. But we got
• 10:39
  Speaker 2: to do better, folks. We have this really interesting series
• 10:45
  Speaker 2: of AI events where we now have AI that isn't
• 10:48
  Speaker 2: playing little video games with little measley humans. It's compiling
• 10:54
  Speaker 2: data for supreme court cases. It is not simply very
• 11:00
  Speaker 2: good at creating silly art or responding to us like
• 11:02
  Speaker 2: a relationship. It's smashing cyber defenses. And we are training
• 11:07
  Speaker 2: it on computing that we've never had before, and it's
• 11:10
  Speaker 2: scaling at a rate that breaks the old model. It
• 11:14
  Speaker 2: used to be. We had this concept of every so
• 11:16
  Speaker 2: often it will double. We have blown off that scale
• 11:19
  Speaker 2: to say that we can double the processing speed on
• 11:22
  Speaker 2: a new or to go forty percent in a single year.
• 11:25
  Speaker 2: That is not the scale we're expecting. We're in the
• 11:28
  Speaker 2: exponential range for growth because we can scale, because we
• 11:31
  Speaker 2: have the energy, because we have the technology, because the funding.
• 11:34
  Speaker 2: And then we throw these large language models into it,
• 11:37
  Speaker 2: and we get out mythos and we're like, well, it's
• 11:40
  Speaker 2: really expensive. We put a lot of effort into this.
• 11:42
  Speaker 2: We're very proud of it. We don't really know how
• 11:44
  Speaker 2: to use it well, so we'll do our best. And oh,
• 11:47
  Speaker 2: by the way, we don't know how to hide this thing,
• 11:50
  Speaker 2: so it keeps getting found by people and then breaking stuff. Okay,
• 11:54
  Speaker 2: it hasn't actually gotten broken stuff. Don't come after me, anthropic.
• 12:00
  Speaker 1: Wrong parents.
• 12:03
  Speaker 2: As we strive to raise kids who love God and
• 12:06
  Speaker 2: use tech, it's intimidating to look at this next stage
• 12:08
  Speaker 2: of life and think about what academics may look like
• 12:11
  Speaker 2: for them. When you're thinking about college, consider George Fox
• 12:14
  Speaker 2: University for your children because of three things. First, it's
• 12:17
  Speaker 2: going to give you a more intimate, known environment where
• 12:21
  Speaker 2: they have a focus on being known personally, academically, and spiritually.
• 12:25
  Speaker 2: Your child will grow in their faith. Ninety three percent
• 12:28
  Speaker 2: of undergrad students at George Fox last year said that
• 12:31
  Speaker 2: being at George Fox has helped them grow in their
• 12:32
  Speaker 2: relationship with Jesus. We also want them to grow academically,
• 12:37
  Speaker 2: which means being a part of the university with rigorous
• 12:40
  Speaker 2: academic standards and opportunities to be engaged. They have a
• 12:43
  Speaker 2: nationally recognized engineering, nursing, and education programs, as well as
• 12:48
  Speaker 2: chances to study abroad here within the US and then internationally,
• 12:51
  Speaker 2: and finally, personally to be part of a vibrant campus
• 12:54
  Speaker 2: atmosphere where you can play intramural sports or part of
• 12:57
  Speaker 2: their NCAA Division three athletic program with twenty three sports available.
• 13:02
  Speaker 2: Go to a place that's going to support your goals,
• 13:05
  Speaker 2: not try to make you support theirs.
• 13:08
  Speaker 1: To have that.
• 13:09
  Speaker 2: Goal be to raise you up to be all God's
• 13:11
  Speaker 2: called you to be, and to truly equip your young
• 13:14
  Speaker 2: person to love God and then from their love of
• 13:16
  Speaker 2: God to go wherever God has called them, and that
• 13:18
  Speaker 2: includes being equipped with a degree that'll help them do that,
• 13:21
  Speaker 2: to be missionally equipped to love and serve others, to
• 13:24
  Speaker 2: lead where God has planned the man, to know the
• 13:26
  Speaker 2: Gospel for themselves and be able to share it with others. So,
• 13:28
  Speaker 2: George Fox University is an amazing resource for our kids,
• 13:32
  Speaker 2: for our community here in the Northwest, and for anyone
• 13:34
  Speaker 2: who's looking to raise up a young person who can
• 13:37
  Speaker 2: love God and from that love others. I tell you
• 13:46
  Speaker 2: all of that to watch that skin. I mean we talked.
• 13:49
  Speaker 2: That was twenty eighteen. It was the first time Open
• 13:52
  Speaker 2: AIS computer was in a big tournament and loss, and
• 13:55
  Speaker 2: then twenty nineteen it won. By twenty twenty two we
• 13:59
  Speaker 2: had a functioning large language model that was available to
• 14:01
  Speaker 2: the public, and by twenty twenty six we have one
• 14:03
  Speaker 2: that's so capable at smashing cyber defenses that we can't
• 14:06
  Speaker 2: release it, like all right, So that's a pretty rapid
• 14:10
  Speaker 2: advance in nine years. What do we need to think
• 14:13
  Speaker 2: about when we think about this. We need to think
• 14:16
  Speaker 2: about what Neil Katiel says about his advocacy for the
• 14:21
  Speaker 2: Supreme Court. Our children will use this information. The question
• 14:24
  Speaker 2: is will they be capable of making good decisions with it?
• 14:28
  Speaker 2: Will they have the character with which to do this well?
• 14:31
  Speaker 2: Because at the end of the day, schools can adjust. Right,
• 14:35
  Speaker 2: we are going to less tech in the classroom because
• 14:38
  Speaker 2: this advancement has made it harder to run a classroom.
• 14:41
  Speaker 2: It has trained our children for distraction, but it has
• 14:43
  Speaker 2: also trained our children to chick pick the easy route,
• 14:46
  Speaker 2: which we don't necessarily want them to do that. We
• 14:49
  Speaker 2: want them to understand the effort of good work. So yes,
• 14:53
  Speaker 2: we have these back house and stuff, but we also
• 14:56
  Speaker 2: want you to know, like how to dig a hole
• 14:58
  Speaker 2: and what that looks like. We need there's a time
• 15:00
  Speaker 2: and a place to use a backo. Maybe your garden
• 15:02
  Speaker 2: is the wrong spot. Maybe it's right, and you have
• 15:04
  Speaker 2: tons of acreage and it would make a ton of sense.
• 15:06
  Speaker 2: But we want to use the right tool at the
• 15:08
  Speaker 2: right time. We want you to have a mind that
• 15:10
  Speaker 2: can think, but also apply it well with tools. This
• 15:13
  Speaker 2: is what Cotail is a arguing for. And again, get
• 15:15
  Speaker 2: beyond the tweet. Go watch his video or it was
• 15:18
  Speaker 2: painful for me to watch because he's very smug at it,
• 15:21
  Speaker 2: so I just read through the transcript. You can do
• 15:23
  Speaker 2: that too, go nuts, but do see what his greater
• 15:27
  Speaker 2: point was and maybe see how to humbly handle victory,
• 15:31
  Speaker 2: as this would be the antithetical example for that. At
• 15:36
  Speaker 2: the end of the day, our children will use AI, yes,
• 15:38
  Speaker 2: schools will use less tech. Yes, tests will go to paper. Yes,
• 15:42
  Speaker 2: we will have essays handwritten. Like all that is happening,
• 15:45
  Speaker 2: and it just makes sure that we know the product
• 15:46
  Speaker 2: our children can do and we know where their limits are.
• 15:49
  Speaker 2: We also know where we can supplement with AI and
• 15:51
  Speaker 2: how to use that well. Okay, but at the end
• 15:53
  Speaker 2: of the day, we must find a way to create
• 15:57
  Speaker 2: AI with a conscience, to use it within tegrity, and
• 16:01
  Speaker 2: to learn how to engage the real world ourselves so
• 16:04
  Speaker 2: that we can experience all the wonder and on the
• 16:07
  Speaker 2: journey that God's designed us for, not just the thrill
• 16:10
  Speaker 2: of an instantaneous destination of a prompt that gets us
• 16:14
  Speaker 2: a product so that we can go out and consume more.
• 16:17
  Speaker 2: There is supposed to be joy in the process. I
• 16:19
  Speaker 2: just showed my children this last week some Bob Ross
• 16:22
  Speaker 2: art videos, not because I'm good at art, and I
• 16:25
  Speaker 2: actually is quite good at it, but just because of
• 16:29
  Speaker 2: the joy that you can have in making something, even
• 16:31
  Speaker 2: if you make it poorly, like you made the thing
• 16:33
  Speaker 2: and therefore you're proud of your effort and you can
• 16:35
  Speaker 2: watch yourself grow this idea that the whole point is
• 16:39
  Speaker 2: level up or to win at some point, or to
• 16:42
  Speaker 2: simply have a final product that's good, and that's the
• 16:45
  Speaker 2: only point. Well. AI can do all that. What AI
• 16:47
  Speaker 2: can't do is the hard work in the middle for you.
• 16:51
  Speaker 2: It can prompt you, it can direct you, it can
• 16:53
  Speaker 2: support you, it can scaffold so that you always have
• 16:56
  Speaker 2: an appropriate level of challenge and you're not just jumping
• 16:58
  Speaker 2: into the deep end like really good at all that.
• 17:01
  Speaker 2: And at the end of the day, the point is
• 17:03
  Speaker 2: who you've been made to be, who God has called
• 17:06
  Speaker 2: you to be, and where He has put you to
• 17:08
  Speaker 2: reach the people in your community. We are not all
• 17:11
  Speaker 2: needing to reach all the people. We are to be
• 17:13
  Speaker 2: faithful where God's put us. And my prayer when we
• 17:16
  Speaker 2: think through AI is yes, understand a little bit of
• 17:19
  Speaker 2: just the nutty technology that goes into it, and the
• 17:21
  Speaker 2: effort that's got there in the processing power that's betting
• 17:23
  Speaker 2: put into that, and the companies that are capable of
• 17:25
  Speaker 2: giving it that processing power, and understand the end of
• 17:29
  Speaker 2: the day, the goal isn't the answers that AI can
• 17:32
  Speaker 2: give us the end of the day. The goal is
• 17:34
  Speaker 2: the people that AI is intended to support or may
• 17:37
  Speaker 2: be accidentally subjugating, because the thing we made in the
• 17:41
  Speaker 2: cave was a dragon and not a dog. We do
• 17:43
  Speaker 2: have to be able to do this with integrity, with
• 17:46
  Speaker 2: a conscience and then still experience the on wonder of
• 17:49
  Speaker 2: the real world. And as long as AI is supporting
• 17:51
  Speaker 2: that amazing. So if you're a Supreme Court advocate and
• 17:53
  Speaker 2: you can use AI to help you make good arguments, okay,
• 17:58
  Speaker 2: Like all right, I'm assuming people were using Google before
• 18:02
  Speaker 2: that and word processor before that as well. What I
• 18:05
  Speaker 2: don't want to do is the AI to do the
• 18:06
  Speaker 2: thinking for you, So please do the thinking our Bible
• 18:09
  Speaker 2: verse is coming. Then when we think about the importance
• 18:11
  Speaker 2: of context and that reference earlier is it comes out
• 18:14
  Speaker 2: of market. And Jesus having healed a man, so he
• 18:19
  Speaker 2: feeds five thousand, he feeds four thousand. He then says,
• 18:22
  Speaker 2: be where the leven of the Pharisees and the teachers
• 18:24
  Speaker 2: the law and his disciples like, ooh, he's mad because
• 18:28
  Speaker 2: we forgot bread. And He's like, are you kidding me?
• 18:29
  Speaker 2: Do you not understand that I fed the people? Like
• 18:32
  Speaker 2: bread is not my issue? We're talking about spiritual issues.
• 18:34
  Speaker 2: And then heals a man in a two step method.
• 18:37
  Speaker 2: First he sees trees and people, and he goes to
• 18:39
  Speaker 2: people in the trees look the same, but the people move,
• 18:42
  Speaker 2: and then he heals them. So he sees clearly. And
• 18:44
  Speaker 2: this idea of like you're going to see me clearly
• 18:47
  Speaker 2: and understand this. By the way, that's not my idea.
• 18:49
  Speaker 2: That came directly from the footnotes of my study Bible.
• 18:51
  Speaker 2: And I was like, oh, that the two step healing
• 18:53
  Speaker 2: that's always bothered me. All right, So who do people
• 18:56
  Speaker 2: say I am? You're the Christ. He then goes on
• 18:59
  Speaker 2: in verse thirty, he began to teach them that the
• 19:01
  Speaker 2: son of man must suffer many things and be rejected
• 19:04
  Speaker 2: by the elders, chief priests, and teachers of the law,
• 19:07
  Speaker 2: and that he must be killed, and after three days
• 19:09
  Speaker 2: rise again. He spoke plainly about this. So Jesus is
• 19:16
  Speaker 2: very clear, I'm the son of man. You just said
• 19:19
  Speaker 2: that on the Messiah that was promised. I'm going to
• 19:21
  Speaker 2: be rejected. I'm going to suffer and die and rise again.
• 19:25
  Speaker 2: These are the words he's using. Peter rebukes him. He goes, no, Lord,
• 19:29
  Speaker 2: you must not die. Jesus says, get behind me, Satan,
• 19:35
  Speaker 2: for you have not the things of God and mind,
• 19:38
  Speaker 2: but the things of men. That's a rebuke. That's down
• 19:42
  Speaker 2: in the next verse thirty three. And then he goes
• 19:45
  Speaker 2: on to say, if anyone must follow me, he must
• 19:47
  Speaker 2: deny himself. And then in verse thirty six, what good
• 19:52
  Speaker 2: is it for a man to gain the world yet
• 19:54
  Speaker 2: forfeit a soul? Or what can a man give in
• 19:56
  Speaker 2: exchange for his soul? If anyone is ashamed of me
• 19:58
  Speaker 2: and my words in this adult tristan, sinful generation, the
• 20:01
  Speaker 2: son of man will be ashamed of him when he
• 20:03
  Speaker 2: comes in his father's glory. With the Holy Angels. And
• 20:06
  Speaker 2: I want to break that into two parts. The first
• 20:07
  Speaker 2: part we're very familiar with. It's this rebuke of Jesus
• 20:12
  Speaker 2: to Peter, saying no, like this idea that you're telling
• 20:15
  Speaker 2: me I don't need to die means you don't get it.
• 20:19
  Speaker 2: That is plan A. This isn't a mistake, It's not
• 20:21
  Speaker 2: a fault in the process. I am the one who
• 20:25
  Speaker 2: is going to atone for your sins, living a perfect life,
• 20:28
  Speaker 2: dying the death you deserve, and rising again to new
• 20:30
  Speaker 2: life is the plan. You are going to die forever.
• 20:33
  Speaker 2: Without this plan, sacrificial law cannot bring a death to death.
• 20:39
  Speaker 2: It can make you right before the Lord, Abraham was
• 20:41
  Speaker 2: still saved by faith, and yet death still needs to die.
• 20:45
  Speaker 2: Ultimate victory needs to come. And this was God's plan A.
• 20:49
  Speaker 2: We're told it happened before the creation of the world,
• 20:52
  Speaker 2: which brings that Dallas Willard argument into context that when
• 20:56
  Speaker 2: Satan was thrown out of heaven that might have been
• 20:58
  Speaker 2: before creation, he was thrown into avoid an into that void.
• 21:01
  Speaker 2: God made light, and God made earth, and God made people,
• 21:04
  Speaker 2: made it his own image and a plan for salvation
• 21:07
  Speaker 2: for all of I'm gonna call it creation because angels
• 21:12
  Speaker 2: are still created. But like that precedes human need, that
• 21:16
  Speaker 2: precedes our fault, that our fall is an aspect of
• 21:18
  Speaker 2: God's greater salvation story. That's just an argument Dallas Willard makes,
• 21:22
  Speaker 2: and I think there's an argument here for it, because
• 21:26
  Speaker 2: he says, this is plan a sacrificial death and resurrection
• 21:32
  Speaker 2: for life to be brought to humans and for satan
• 21:35
  Speaker 2: sin and death to be defeated forever. This is plan
• 21:39
  Speaker 2: A Peter, do not tell me I live because you're
• 21:41
  Speaker 2: thinking about man. That then goes on to what's it
• 21:44
  Speaker 2: worth if you gain everything but lose your soul? And
• 21:47
  Speaker 2: that's where this AI aspect is a beautiful reminder of like,
• 21:50
  Speaker 2: don't just use AI because it's available, don't just use
• 21:52
  Speaker 2: it because it's easy. Don't just use it because you
• 21:54
  Speaker 2: like the end product. At the end of the day,
• 21:55
  Speaker 2: you were designed for a purpose and AI needs to
• 21:58
  Speaker 2: support that purpose. If it causes you to say, cut
• 22:00
  Speaker 2: it off and catch you out, you can live life
• 22:02
  Speaker 2: without AI a lot less effectively. But just like you
• 22:04
  Speaker 2: can live without a smartphone or without the Internet, or
• 22:07
  Speaker 2: without whatever else might cause it, without whatever substance you
• 22:10
  Speaker 2: might struggle with, you can live without those things, because
• 22:13
  Speaker 2: God is better. God is better even than Ai, and
• 22:17
  Speaker 2: we need to make sure we're very deliberate with that.
• 22:19
  Speaker 2: Do not pursue AI at all costs. Retain the wonder
• 22:25
  Speaker 2: and the amazement at God being good, and let Ai
• 22:27
  Speaker 2: support that and whatever it is. But then he says, okay,
• 22:30
  Speaker 2: So if you forfeit your soul but you gain the world,
• 22:34
  Speaker 2: what's the matter. What does the matter if you have
• 22:37
  Speaker 2: an effective ministry, or if you run a job well,
• 22:40
  Speaker 2: or if you get your parenting perfect because you never
• 22:42
  Speaker 2: make a mistake because Ai guides you at every step.
• 22:44
  Speaker 2: What's the point of any of that? If you lose
• 22:46
  Speaker 2: your soul at the end of the day, if you
• 22:48
  Speaker 2: don't give glory to God and instead you're thankful to
• 22:50
  Speaker 2: this Ai and you're beholden to this Ai, and you
• 22:53
  Speaker 2: serve this Ai with your time and your heart, and
• 22:54
  Speaker 2: you're focusing your money. And then he goes on to
• 22:58
  Speaker 2: this whole thing about being ashamed mind. So he says,
• 23:01
  Speaker 2: if you're ashamed of the son of Man, he'll be
• 23:03
  Speaker 2: ashamed of you when he comes in glory with his
• 23:06
  Speaker 2: with the armies of angels. Peter was not ashamed in
• 23:09
  Speaker 2: the way that we think of ashamed. We think of
• 23:10
  Speaker 2: like I don't want to mention Jesus's name because like
• 23:13
  Speaker 2: people might get mad at me, or I don't want
• 23:14
  Speaker 2: to stand on Jesus's principles and say this was right
• 23:17
  Speaker 2: and this was wrong because I don't want people to
• 23:19
  Speaker 2: be mad at me. We think of ashamed that way.
• 23:21
  Speaker 2: That's not what Peter's doing. Peter goes, Jesus, you said
• 23:25
  Speaker 2: you're gonna die. Absolutely not. You're gonna rule, you're gonna reign,
• 23:28
  Speaker 2: you're gonna be powerful. We're gonna stand with you. And
• 23:30
  Speaker 2: by the way, can I have the seat to your
• 23:32
  Speaker 2: right because that's kind of like a big deal to me,
• 23:33
  Speaker 2: and I just I just want to double check. He's saying, Jesus,
• 23:36
  Speaker 2: you're not gonna die. We're not gonna let that happen.
• 23:38
  Speaker 2: We're gonna make sure you win, and we're gonna go
• 23:40
  Speaker 2: take some names and we're gonna go kick out some
• 23:43
  Speaker 2: romans and it's gonna be amazing. And Jesus, like, you're
• 23:45
  Speaker 2: not you're not listening, Peter. You're ashamed of this idea
• 23:49
  Speaker 2: that I'm gonna die, because that's humiliating and sounds bad
• 23:52
  Speaker 2: for your plans. But that's God's plan. Don't be ashamed
• 23:55
  Speaker 2: of God's plan. When we think about our use of AI,
• 23:58
  Speaker 2: please make sure you use in it prayerfully. The goal
• 24:01
  Speaker 2: is not to overcome God's plans just because they're hard,
• 24:06
  Speaker 2: just because they're inconvenient. There may be a part of
• 24:09
  Speaker 2: God's plan that is don't use AI, do it the
• 24:12
  Speaker 2: hard and slow way. I don't know, but I'm just
• 24:14
  Speaker 2: saying that Peter's shame being ashamed of Jesus was not
• 24:19
  Speaker 2: because he was worried about mentioning Jesus in front of others.
• 24:23
  Speaker 2: He was worried that Jesus's plan was bad and he
• 24:26
  Speaker 2: wanted human effort to overcome that. And we struggle with that.
• 24:31
  Speaker 2: We use AI as a beautiful buffer between humanity and
• 24:35
  Speaker 2: God's plans, and we just want to make sure nothing
• 24:38
  Speaker 2: bad ever happens, and that it all happens on our
• 24:40
  Speaker 2: terms in a way we like that's understandable, and we
• 24:45
  Speaker 2: need to be prayerful about how we use AI. The
• 24:46
  Speaker 2: goal is not that AI will get good enough and
• 24:50
  Speaker 2: one day save humanity. The goal is to use AI
• 24:53
  Speaker 2: to certainly combat the fall, cure some diseases, help some
• 24:58
  Speaker 2: people solve some so issues go nuts be the best
• 25:02
  Speaker 2: advocate you can before the Supreme Court. Okay, but don't
• 25:06
  Speaker 2: ever get confused that somehow this new tower of Babel
• 25:08
  Speaker 2: is the one that'll actually reach Oavy Like, that's not
• 25:10
  Speaker 2: going to happen. We want to be deliberate with the
• 25:12
  Speaker 2: way we use it, prayerful in the way we apply it,
• 25:15
  Speaker 2: and have a conscience and create them with integrity and
• 25:17
  Speaker 2: teach our children to do the same. So I hope
• 25:20
  Speaker 2: this is encouraging to you. I hope it wasn't too
• 25:21
  Speaker 2: deep a random dive on AI, but you come away
• 25:25
  Speaker 2: knowing how it's trained, knowing the way a very specific
• 25:28
  Speaker 2: model that is our most recent example, not the last example,
• 25:31
  Speaker 2: of getting more than we bargained for, which is a
• 25:34
  Speaker 2: neat problem, but a problem nonetheless, and then understanding how
• 25:38
  Speaker 2: it's being applied this week through some of the more
• 25:42
  Speaker 2: powerful individuals impacting our government, our country, and the fields
• 25:46
  Speaker 2: that matter to us, including law. So if you have
• 25:49
  Speaker 2: any questions, you can reach out to me Nathan at
• 25:51
  Speaker 2: gospeltech dot net or at Love God use Tech on
• 25:54
  Speaker 2: Instagram and Facebook, and then you can join us next
• 25:56
  Speaker 2: week as we continue this conversation about how we can
• 25:58
  Speaker 2: love God and use tech talk with you