[English] Connor McDonald - Cheap Headphones for Connor - What happened in between?

00:00:13: Kai: Hello and welcome to another episode of

00:00:15: Kai: Devs on Tape.

00:00:16: Kai: Today we have the first time.

00:00:18: Kai: I guest for the second time, year after

00:00:21: Kai: year, maybe next year the third time we

00:00:23: Kai: would see.

00:00:24: Kai: Welcome, connor McDonald.

00:00:26: Connor: It's great to be back.

00:00:27: Connor: That sounds like I've come back to some

00:00:29: Connor: sort of rock console, but it's very

00:00:31: Connor: privileged to be invited back for a second

00:00:33: Connor: time.

00:00:33: Connor: I obviously didn't offend you enough in the

00:00:35: Connor: first time.

00:00:36: Connor: You have another try.

00:00:37: Caro: Yeah.

00:00:38: Kai: I'll do my best.

00:00:39: Connor: I'll do my best.

00:00:40: Kai: Australian.

00:00:40: Kai: Yeah, all right.

00:00:41: Kai: So in I call it preparation for the podcast,

00:00:44: Kai: we thought about what could be possibly

00:00:47: Kai: interesting for the people to hear again,

00:00:50: Kai: or what are the topics we can talk about.

00:00:52: Kai: We didn't talk earlier, so one question is

00:00:54: Kai: definitely what happened in between?

00:00:57: Connor: Well, the last time it was here wasn't it a

00:01:00: Connor: DOAG that we did a podcast yeah different

00:01:01: Connor: room, but yeah different room.

00:01:04: Connor: Well, what's happened is well.

00:01:05: Connor: I'll start with the simple things first.

00:01:07: Connor: Last time I was here, I had a really nice

00:01:09: Connor: pair of headphones to wear as part of the

00:01:10: Connor: podcast and for the listeners tuning in.

00:01:13: Connor: I want to let them know that Caroline and

00:01:15: Connor: Kai have a wonderful pair of headphones

00:01:17: Connor: each, and I have some cheap consumer

00:01:20: Connor: supermarket headphones, making my voice

00:01:23: Connor: sound tinny and lame, because I want to

00:01:25: Connor: assure all the listeners that my voice is

00:01:26: Connor: normally a beautiful, deep baritone.

00:01:29: Connor: But it's the headphones, it's purely the

00:01:31: Connor: headphones.

00:01:31: Connor: So I do feel like I've been cheaped out on.

00:01:34: Connor: But moving on from there, no, it's been a.

00:01:36: Connor: It's been a year of interesting change.

00:01:38: Connor: Don't you give us a chance to even clarify

00:01:40: Connor: this situation?

00:01:41: Kai: No, no great.

00:01:42: Connor: We're moving on.

00:01:43: Connor: We're moving on, go ahead, go ahead.

00:01:45: Connor: Just in terms of if I'll start with the

00:01:48: Connor: tech part of the year, the thing that blew

00:01:49: Connor: me away is a year ago.

00:01:51: Connor: If I said to you hey, have you heard about

00:01:53: Connor: this AI stuff?

00:01:55: Connor: Everyone would say, well, maybe a little

00:01:57: Connor: bit, you know, maybe it's something you

00:01:59: Connor: know.

00:02:00: Connor: People would say I'm sick of digital chat

00:02:02: Connor: bots and assistants.

00:02:03: Connor: I hate that nonsense on a website where it

00:02:05: Connor: says hi, I'm Sally, your automated

00:02:07: Connor: assistant.

00:02:08: Connor: Everyone's view of AI was pretty much it's

00:02:11: Connor: some junk.

00:02:13: Connor: It's the modern version of Clippy from.

00:02:15: Connor: Microsoft Office.

00:02:16: Connor: You fast forward what nine months.

00:02:18: Connor: And now everyone's going this is the future

00:02:21: Connor: of IT, that everything will be AI and, to

00:02:24: Connor: the extent that we're all like, it's going

00:02:25: Connor: to take over the world.

00:02:26: Connor: It's going to launch robots against us.

00:02:29: Connor: Everyone's, you know, quoting terminator

00:02:31: Connor: scenarios and all this kind of stuff.

00:02:32: Connor: So that changes wild and I put my oracle

00:02:38: Connor: head on.

00:02:38: Connor: Oracle, like every other big corporation,

00:02:40: Connor: is just going wow, we need to scrap it, you

00:02:43: Connor: know to come up to speed.

00:02:44: Connor: So I've been doing some AI demos and every

00:02:46: Connor: other company's been doing some AI demos,

00:02:48: Connor: but that's, I imagine it's similar to you.

00:02:50: Connor: It's like it's, you know it's blown.

00:02:52: Connor: It's just blown up so fast.

00:02:53: Kai: Yeah, exactly.

00:02:54: Kai: I saw a talk yesterday from Sasha Lobo and

00:02:58: Kai: he was also just summarizing not in deep

00:03:01: Kai: tech, it wasn't a deep tech talk, it was

00:03:03: Kai: just about one of the major AI researcher

00:03:06: Kai: said, like in January 2022, like last year,

00:03:10: Kai: he said that GPT 3000 in a couple of years

00:03:13: Kai: would not be able to clarify the situation.

00:03:16: Kai: When you put your phone on a table and move

00:03:18: Kai: the table, what happens to the phone?

00:03:19: Kai: Because no one is telling GPT what happens

00:03:22: Kai: with different objects on.

00:03:24: Kai: And then the demo stopped the video

00:03:26: Kai: interview.

00:03:27: Kai: He asked chat GPT, what happens when you

00:03:29: Kai: put a phone on the table and move the table?

00:03:31: Kai: And then you get like a big essay about

00:03:33: Kai: forces between objects that are close to

00:03:35: Kai: each other and the move and how it's been

00:03:38: Kai: calculated and what happens.

00:03:39: Kai: And they said all right, not even one of

00:03:41: Kai: the most advanced research in the AI space

00:03:44: Kai: was able to do a forecast what's happening

00:03:47: Kai: in the next 11 months?

00:03:49: Kai: Right, so there's happening so much stuff

00:03:51: Kai: right now.

00:03:52: Kai: We cannot see what's happening in the next

00:03:54: Kai: month, so we don't know what to expect,

00:03:56: Kai: right, but I'm sure we are very mind blown

00:03:59: Kai: by now and like 10, 10 weeks, for example.

00:04:02: Connor: I saw an interesting article, which it was

00:04:04: Connor: both interesting and scary in the sense

00:04:07: Connor: that someone they said you think about how

00:04:09: Connor: we treat dogs, right?

00:04:11: Connor: Dogs believe that they're sort of almost on

00:04:14: Connor: an even a par with us because they provide

00:04:17: Connor: us companionship and we feed them, and the

00:04:19: Connor: dog thinks that somehow the owner has

00:04:22: Connor: prepared and generated this food.

00:04:24: Connor: It is unaware of the complexity.

00:04:26: Connor: It didn't know that we went to the shop and

00:04:28: Connor: we had a credit card and we drove there All

00:04:30: Connor: that stuff.

00:04:30: Connor: The dog just thinks my life is complete

00:04:33: Connor: because this person, I look after him, they

00:04:35: Connor: give me food and that's all they understand.

00:04:37: Connor: They think that's the complete equation.

00:04:39: Connor: This person then wanted to say this is how

00:04:41: Connor: AI will treat us eventually.

00:04:43: Kai: Yeah.

00:04:43: Connor: In the sense that we will think that we're

00:04:45: Connor: dealing with AI, we'll think that we're in

00:04:46: Connor: control, but there'll be so much going on

00:04:49: Connor: behind the scenes that we simply cannot

00:04:51: Connor: comprehend, and so we will become dogs to

00:04:54: Connor: the AI.

00:04:54: Connor: And I was like okay, that sounds.

00:04:55: Connor: It's an interesting metaphor and it scares

00:04:57: Connor: the hell out of it.

00:04:58: Kai: Definitely so.

00:04:59: Kai: Do you think?

00:05:00: Kai: There's a big, big AI network by now?

00:05:02: Kai: And it's just saying, oh, give them GPT 3.5

00:05:05: Kai: and show them how to tweet.

00:05:09: Kai: Oh, I cannot process files for now.

00:05:12: Kai: And behind that, the big model is just

00:05:13: Kai: laughing and saying you'll see next year.

00:05:16: Connor: Who knows?

00:05:16: Connor: I mean, even though I'm a little bit

00:05:18: Connor: skeptical.

00:05:19: Connor: My partner, genevieve, works in business

00:05:21: Connor: intelligence and I used to be her AI for

00:05:24: Connor: lack of a better term in the sense that

00:05:25: Connor: when she had to write a complex SQL to do

00:05:27: Connor: some data analysis, my phone would go bing.

00:05:30: Connor: And like hey, I'm trying to work out this

00:05:33: Connor: running total, or a month to date, or

00:05:35: Connor: whatever.

00:05:36: Connor: Here's the tables.

00:05:37: Connor: What do I do?

00:05:37: Connor: And I said, well, why don't you try just

00:05:39: Connor: throwing that text into chat GPT?

00:05:42: Connor: And don't get me wrong, it doesn't give the

00:05:44: Connor: right SQL, but it gives her 90% and that's

00:05:47: Connor: pretty much what I used to give her.

00:05:48: Connor: I said, look, this is roughly what you want

00:05:49: Connor: to do.

00:05:50: Connor: And she goes okay, I can take it from here

00:05:52: Connor: and I can correct the column names, fill in,

00:05:54: Connor: correct the syntax a little bit, but yeah,

00:05:56: Connor: so in terms of a tool, she thinks it's

00:05:58: Connor: awesome.

00:05:59: Kai: Is it the modern way of telling like, let

00:06:02: Kai: me Google that for you, because you say,

00:06:04: Kai: okay, you can write it in the iMessage

00:06:06: Kai: field or not?

00:06:07: Kai: Your Android guy, you can write it into

00:06:09: Kai: WhatsApp and you get the answer, or you can

00:06:11: Kai: even just ask chat, gpt, exactly.

00:06:14: Connor: So I you know, I think the benefits are

00:06:16: Connor: there, but yeah, like, and of course, as a

00:06:18: Connor: result, everyone's saying that you know,

00:06:19: Connor: we'll all be out of a job as developers or

00:06:22: Connor: whatever.

00:06:22: Connor: Who?

00:06:23: Kai: knows.

00:06:23: Kai: That's another interesting point of the

00:06:25: Kai: talk yesterday that developers should be

00:06:27: Kai: open-minded for new technology and they

00:06:29: Kai: should say that we were, say, we are going

00:06:32: Kai: to say the next years like, oh, back in the

00:06:35: Kai: days, we did everything by hand and we have

00:06:37: Kai: written, written the programs all by

00:06:39: Kai: ourselves, like craftsmanship and so on,

00:06:42: Kai: and we have to say goodbye to this, this

00:06:44: Kai: handmade applications.

00:06:46: Kai: This is not the point where we are working

00:06:47: Kai: in the next years.

00:06:48: Kai: It is just focusing on how to train models

00:06:51: Kai: and to find ways to apply the AI stuff to

00:06:53: Kai: our things.

00:06:54: Kai: But I mean, we have a long career ahead of

00:06:57: Kai: us, so we should think about that.

00:06:59: Kai: What's happening in the next years?

00:07:01: Kai: Many people don't think about that.

00:07:02: Kai: What, what will be changed?

00:07:04: Kai: Right?

00:07:04: Connor: So I just want to point out to listeners

00:07:07: Connor: that you pointed at.

00:07:08: Connor: Caroline said you and I have.

00:07:10: Connor: You know, you two have long careers ahead

00:07:12: Connor: of you by inference, looking at me saying

00:07:14: Connor: I'm a dinosaur and don't have many years

00:07:16: Connor: ahead of me.

00:07:17: Kai: Exactly kind of that's what I meant,

00:07:18: Kai: because you got the cheap headphones and I

00:07:21: Kai: have to bring something back right so this

00:07:24: Kai: is exactly so.

00:07:25: Connor: I'll tell you a funny story about that.

00:07:27: Connor: Actually, in Poland, I've just come from

00:07:28: Connor: Poland, from the Poland Oracle User Group

00:07:30: Connor: fantastic event Shout out to Louisa and

00:07:32: Connor: Camila.

00:07:32: Connor: It was a really well-run event.

00:07:35: Connor: A guy came up to me between sessions and

00:07:37: Connor: and very, I was very appreciative.

00:07:38: Connor: He said oh yeah, hi, connor, I just want to

00:07:40: Connor: say I like your YouTube videos and I've had

00:07:42: Connor: a lot of value out of them and that made me

00:07:43: Connor: feel really good.

00:07:44: Connor: It's nice for people to make that effort.

00:07:45: Connor: And then he said I'm also doing some video.

00:07:48: Connor: I'm starting to do some little tech videos

00:07:49: Connor: from my office and stuff like that.

00:07:51: Connor: And he said you've been an inspiration.

00:07:52: Connor: I was like oh wow.

00:07:53: Connor: And he goes.

00:07:54: Connor: It's just so good to see someone just so

00:07:57: Connor: much older than me doing video.

00:08:01: Caro: And I was going.

00:08:02: Connor: I think that's a compliment.

00:08:05: Connor: It was like thank you and not just old up

00:08:08: Connor: so much old.

00:08:10: Connor: Okay, thank you very much, but but either

00:08:13: Connor: way I still, even with those kind of slaps

00:08:15: Connor: in the face, I thoroughly enjoy doing my

00:08:18: Connor: YouTube videos.

00:08:19: Kai: So this is one highlight of the year for

00:08:20: Kai: you that somebody is telling you that

00:08:22: Kai: you're so much older than right.

00:08:23: Connor: So much older.

00:08:24: Connor: Well, I have to.

00:08:24: Connor: My children do that every single day, so

00:08:27: Connor: I'm very much used to it.

00:08:28: Connor: So I've gone from being.

00:08:30: Connor: When they were young and they learned how

00:08:31: Connor: to do things like video, they thought I was

00:08:33: Connor: cool dad.

00:08:34: Connor: And then, as they've got older and matured,

00:08:36: Connor: they've reverted to normal child speak,

00:08:39: Connor: which is I'm uncool dad.

00:08:40: Kai: I mean, they saw your content right.

00:08:42: Connor: So exactly.

00:08:43: Connor: I have to admit, though, one of the things

00:08:44: Connor: that they get instructed to do is, every

00:08:46: Connor: time I publish a video, I said you boys

00:08:48: Connor: better get on there and you better press

00:08:49: Connor: that like.

00:08:51: Connor: It's all about the likes.

00:08:53: Connor: What are the young YouTubers say Smash that

00:08:55: Connor: like button.

00:08:56: Kai: Yeah, hit the bell right, or what is it in

00:08:58: Kai: English?

00:08:58: Kai: That's right, the notification thing Hit

00:09:00: Kai: the bell smash that like button so they get

00:09:04: Kai: instructed.

00:09:04: Connor: So at least I get one like on each video.

00:09:07: Kai: So this is also what happens in between,

00:09:09: Kai: because last time you had something like I

00:09:12: Kai: don't know if I get it completely right,

00:09:13: Kai: but it's not the count of followers what is

00:09:17: Kai: important.

00:09:18: Kai: It is the followers that count.

00:09:20: Caro: It's not the followers.

00:09:21: Kai: So we were completely prepared, like

00:09:24: Kai: completely.

00:09:24: Kai: Yeah, you prepared your statement that it's

00:09:27: Kai: not necessary to have many followers.

00:09:29: Kai: It should be the right followers.

00:09:30: Kai: You want to help and now you're just.

00:09:31: Kai: Yeah, I need the followers, I want likes, I

00:09:34: Kai: want the diamond play button.

00:09:35: Connor: No, my children to love me, that's, that's

00:09:37: Connor: the most important.

00:09:38: Connor: And you know, if that, if forcing them to

00:09:40: Connor: hit like button is the way to do it, then

00:09:42: Connor: that's the way I do it and, yeah, that

00:09:44: Connor: might be also the way you the reason why

00:09:48: Connor: you started TikTok, right?

00:09:49: Kai: So that's another question of what happened

00:09:50: Kai: to your TikTok.

00:09:51: Connor: So you did test, you told us about how did

00:09:55: Connor: this this that's been an interesting

00:09:57: Connor: experience and the motivation came I can't

00:10:00: Connor: remember this was before or after I spoke

00:10:02: Connor: to you last year, which was I had a podcast

00:10:06: Connor: interview with an intern at Oracle who's

00:10:08: Connor: now an employee of Oracle, so that's been a

00:10:10: Connor: success story there and she was saying that

00:10:13: Connor: she learns a lot of her tech Ideally, she

00:10:15: Connor: said if she can find a 30 second sound bite,

00:10:18: Connor: it doesn't matter whether it's.

00:10:18: Connor: TikTok, instagram, youtube.

00:10:20: Connor: But she said, yeah, when you've got a

00:10:22: Connor: problem, if I can find a 30 second sound

00:10:24: Connor: bite to solve that problem and let me move

00:10:26: Connor: forward, then that's the way you can learn.

00:10:28: Connor: Yeah, and that was pretty confronting for

00:10:30: Connor: me who's used to blog posts and you know

00:10:32: Connor: what we now call a long YouTube video,

00:10:35: Connor: which is like five minutes, and so, as a

00:10:38: Connor: result, I thought well, let's go to TikTok

00:10:39: Connor: ago, and I have to admit it is surprising

00:10:43: Connor: that if I publish a YouTube video after a

00:10:47: Connor: day, it might have been watched by a few

00:10:49: Connor: hundred people.

00:10:50: Connor: If I publish a TikTok video, it's been

00:10:52: Connor: watched by a few thousand, literally within

00:10:54: Connor: a few days.

00:10:55: Connor: Now, don't get me wrong, do I honestly

00:10:57: Connor: think 3,000 people spending a quality

00:10:59: Connor: amount of time looking at the content?

00:11:01: Connor: Because we all know, with TikTok, your

00:11:04: Connor: index finger is just flicking through next

00:11:06: Connor: video, next video.

00:11:07: Connor: I don't know how the TikTok analytics work,

00:11:09: Connor: but certainly there seems to be a tech

00:11:11: Connor: audience there, intermingled with cat

00:11:13: Connor: videos and dog videos and people running

00:11:16: Connor: into water fans and all sorts of ridiculous

00:11:18: Connor: accidents, which seems to be most of TikTok.

00:11:20: Connor: I still struggle to be, enthusiastic about

00:11:23: Connor: TikTok because it just I think it just

00:11:25: Connor: creates that doom scrolling mentality of

00:11:27: Connor: just scroll, scroll, scroll.

00:11:28: Connor: But the reality is I don't get to make that

00:11:30: Connor: decision.

00:11:31: Connor: If the next generation of programmers is

00:11:33: Connor: getting their tech from TikTok, then my job

00:11:36: Connor: description is get content on TikToks.

00:11:39: Connor: I'll keep doing it and we'll see how it

00:11:41: Connor: goes.

00:11:42: Kai: I mean, this is something that is very

00:11:45: Kai: specific to you, that you are always like,

00:11:47: Kai: yeah, very happy, and looking forward and

00:11:50: Kai: checking everything out.

00:11:52: Kai: So this is not.

00:11:53: Kai: No, I don't get the trend of TikTok, I

00:11:56: Kai: don't want to do it anymore and no, I'm not

00:11:58: Kai: following the trends.

00:11:59: Kai: So this is one question we prepared.

00:12:03: Kai: The one question how do you get this?

00:12:06: Kai: How are you still happy every single day

00:12:08: Kai: and every single moment?

00:12:10: Kai: So or ask the question the other way around

00:12:13: Kai: what drives you nuts?

00:12:14: Kai: I didn't see that in any video or in any

00:12:17: Kai: discussion on the conference.

00:12:19: Connor: Well, plenty of things do.

00:12:20: Connor: I have to admit, though, one of the things

00:12:22: Connor: I learned a long time ago is I worked for a

00:12:25: Connor: company this is all a mining company.

00:12:28: Connor: This is well sort of in the late 90s, late

00:12:31: Connor: 80s, early 90s, worked the FEN for about 12

00:12:34: Connor: years, and I found myself at work getting

00:12:37: Connor: more and more bitter, yeah, and more and

00:12:39: Connor: more things just wound me up and I'd come

00:12:41: Connor: home and I wasn't really enjoying it, and

00:12:44: Connor: at the time I blamed the company.

00:12:46: Connor: I thought you know, they're not, you know,

00:12:48: Connor: doing intelligent things in IT, they're not

00:12:50: Connor: treating their employees well, et cetera.

00:12:52: Connor: I had all sorts of bitterness against them

00:12:54: Connor: and it's sort of I'm not sure what the sort

00:12:55: Connor: of moment was, but I sort of realized that

00:12:57: Connor: in reality, it's actually probably I won't

00:12:59: Connor: say my fault, but it's me.

00:13:01: Connor: If I'm going to harbour the bitterness,

00:13:03: Connor: then nothing in the company is probably

00:13:05: Connor: fixed there.

00:13:06: Connor: And don't get me wrong wherever I've worked

00:13:07: Connor: since then, including my time at Oracle,

00:13:10: Connor: sometimes things upset you, and what

00:13:12: Connor: happens is I found it's much better now to

00:13:14: Connor: come home and just expunge it.

00:13:16: Connor: So sometimes I'll just come and I'll just

00:13:18: Connor: let rip at the monitor, whatever you know

00:13:22: Connor: just you know, and just unleash all the

00:13:23: Connor: frustrations and they're gone.

00:13:26: Connor: And it's much better to expunge them than

00:13:28: Connor: to harbour them.

00:13:29: Connor: And one of the things I do find every day

00:13:31: Connor: is you do bump into people, even at

00:13:33: Connor: conferences, people at conferences here,

00:13:34: Connor: and you say you know how things going and

00:13:36: Connor: they're just pissed off at the world.

00:13:38: Connor: And I think that's just so exhausting.

00:13:40: Connor: But you just you know there's not enough

00:13:42: Connor: time for I don't have enough years left to

00:13:44: Connor: spare to spare time, so you said it this

00:13:46: Connor: time, right.

00:13:48: Connor: I'm so much older than other people doing

00:13:52: Connor: videos, so, yeah, so I find that it's

00:13:56: Connor: life's too short to basically be sitting

00:13:58: Connor: there and being pissed off about your work

00:14:01: Connor: life, because you, as we all know, you're

00:14:03: Connor: going to spend a good amount of time of

00:14:04: Connor: your life working, so you may as well get

00:14:07: Connor: some reward out of it.

00:14:08: Connor: And don't get me wrong I'm very fortunate

00:14:10: Connor: because I have the best job in the world.

00:14:12: Connor: You know, I fully can see that there are

00:14:14: Connor: plenty of people who have working

00:14:16: Connor: environments that it would be incredibly

00:14:18: Connor: difficult to find that enjoyable or

00:14:19: Connor: rewarding, but I'm just very lucky.

00:14:23: Caro: I've always heard a sentence you love it,

00:14:26: Caro: you leave it or you change it.

00:14:28: Caro: In this case, that's cool so yeah, you

00:14:31: Caro: should change it or love it or just say

00:14:34: Caro: okay, it's fine.

00:14:35: Connor: Exactly, and you know, all of us in the IT

00:14:38: Connor: industry are in a privileged position

00:14:40: Connor: because, I kind of remember a time in the

00:14:42: Connor: IT industry where it hasn't been possible

00:14:45: Connor: to find work.

00:14:46: Connor: Maybe not your dream job, but if one job

00:14:49: Connor: changes, companies do have layoffs from

00:14:51: Connor: time to time and stuff.

00:14:52: Connor: But the reality is very few IT people can't

00:14:54: Connor: find work relatively quickly, Whereas other

00:14:58: Connor: people they lose their jobs and it's a life

00:15:00: Connor: altering event.

00:15:02: Kai: I mean, if you're focused and do some

00:15:04: Kai: craftsmanship or work with your bare hands

00:15:06: Kai: right.

00:15:07: Kai: And if you learn to work with wood, for

00:15:09: Kai: example, and then at some point you're just

00:15:11: Kai: pissed off or upset from wood, what should

00:15:13: Kai: you do differently?

00:15:14: Kai: Right.

00:15:14: Kai: But in IT, if you are in the company and

00:15:16: Kai: you see over the years that your interest

00:15:19: Kai: changed and you want to do something else,

00:15:21: Kai: there are ways to do it easily, right.

00:15:24: Kai: So you can say I'm not interested in being

00:15:26: Kai: a DBA, for example, but I want to develop

00:15:28: Kai: something because I did it in private time.

00:15:31: Kai: So now I'm looking into Apex, for example,

00:15:34: Kai: or whatever database environment, and this

00:15:37: Kai: is more easy.

00:15:38: Kai: So I think that everyone is able in IT,

00:15:41: Kai: able to refocus after a couple of years

00:15:44: Kai: instead of being bitter because I think

00:15:46: Kai: it's right.

00:15:47: Kai: So, at every conference, when I asking

00:15:49: Kai: people, how is it going, how do we feel?

00:15:51: Kai: Most of the time the first thing they say

00:15:53: Kai: is what's not okay, what's not working If

00:15:56: Kai: it's the political system in Germany or a

00:15:58: Kai: different country, or if it's the company

00:16:00: Kai: they are working for, or what are the

00:16:02: Kai: crisis or whatever.

00:16:04: Kai: But they're not saying hey, I had a pretty

00:16:06: Kai: nice week.

00:16:07: Kai: Nothing bad happened.

00:16:08: Kai: I had a great time working, met great

00:16:11: Kai: friends and most of the time they're just

00:16:13: Kai: focusing on the negative stuff, and this

00:16:15: Kai: has to change, because when I'm talking

00:16:17: Kai: about stuff that is not good, then I'm not

00:16:19: Kai: happy to.

00:16:21: Connor: Let it be known that 19 minutes went by

00:16:23: Connor: before Kai managed to swing Apex into the

00:16:25: Connor: conversation, so that's probably a record

00:16:28: Connor: for you.

00:16:28: Kai: I'm trying to.

00:16:29: Kai: I'm trying hard.

00:16:33: Connor: But in terms of things in last year,

00:16:35: Connor: obviously we've had a couple of Apex

00:16:36: Connor: releases.

00:16:37: Connor: I work as part of the Apex team.

00:16:40: Connor: I report to my kitchen but obviously I'm

00:16:43: Connor: not actively involved in the building of

00:16:45: Connor: the product, but certainly because I look

00:16:48: Connor: after the Aston system one of the nice

00:16:51: Connor: things I've been enjoying is the latest

00:16:53: Connor: version of Apex with the work and copy and

00:16:55: Connor: stuff.

00:16:55: Connor: So that's some cool stuff.

00:16:57: Connor: I try Keeper Breast of Apex features, but

00:16:59: Connor: the days of Apex being a little tiny app

00:17:01: Connor: where you could be fully cognizant of all

00:17:03: Connor: the bits and pieces are sort of gone Now.

00:17:05: Connor: It's just such a bigger, more impressive

00:17:07: Connor: product.

00:17:07: Connor: But yeah, that's the other thing that I've

00:17:09: Connor: been enjoying seeing some new Apex features

00:17:10: Connor: this last year.

00:17:11: Kai: This was our first moment in DevSontape

00:17:14: Kai: where we got some information no one else

00:17:16: Kai: outside of the Apex team had.

00:17:18: Kai: Max Hifts was announcing working copies in

00:17:21: Kai: DevSontape first, like in July.

00:17:23: Kai: We recorded that casecode and, yeah, we

00:17:26: Kai: were able to release the the podcast

00:17:29: Kai: episode where he was teasing this

00:17:30: Kai: functionality before any blog or lab stuff

00:17:35: Kai: Say that.

00:17:35: Kai: So be prepared, this question will come to

00:17:37: Kai: you to tell us some secrets from Oracle

00:17:40: Kai: right after the last words in this podcast.

00:17:43: Caro: I just wanted to say Kai pushed him to a

00:17:48: Caro: secret.

00:17:50: Kai: Not only him.

00:17:51: Kai: So this is the Doak where we squeeze the

00:17:53: Kai: secrets out of Oracle developers, not

00:17:55: Kai: developers Oracle people.

00:17:58: Connor: Okay, I'll just make something up and at

00:18:00: Connor: the end I'll say but of course say father,

00:18:03: Connor: but that way you've got a big announcement,

00:18:05: Connor: You've got something exciting.

00:18:06: Connor: It's probably fiction, but that's okay.

00:18:09: Kai: I mean clicks likes.

00:18:12: Kai: We'll see what happens.

00:18:14: Caro: So you talked about your TikTok career and

00:18:18: Caro: you also post a lot of videos and helpful

00:18:21: Caro: things as an old dinosaur on YouTube and

00:18:26: Caro: I've seen this year you reached a special

00:18:29: Caro: number of videos.

00:18:30: Connor: I didn't date, I was, I have to admit it's.

00:18:32: Connor: It did feel a bit self-indulgent to do a

00:18:34: Connor: little sort of GIF celebrating the fact

00:18:36: Connor: that I got to 700.

00:18:38: Connor: So I did a pun on the 007 of James Bond.

00:18:43: Connor: Oh yeah, no, I got it.

00:18:45: Connor: It's one of those things where I like

00:18:48: Connor: celebrating those little milestones,

00:18:49: Connor: because it comes back to the thing about,

00:18:50: Connor: you know, enjoying the job you do.

00:18:53: Connor: And I figured if you are happy to celebrate

00:18:56: Connor: the things you achieve yourself, if you're

00:18:58: Connor: proud of them, then you don't have to be

00:19:00: Connor: constantly thinking, oh, to someone else

00:19:01: Connor: approve you know why not be personally

00:19:04: Connor: satisfied with what you're achieving, and

00:19:06: Connor: also the reason I picked video as something

00:19:08: Connor: to celebrate is the reality is, oracle in

00:19:11: Connor: particular, is often viewed as an old tech

00:19:13: Connor: company.

00:19:14: Connor: All the big companies, you know your

00:19:15: Connor: Microsofts and your IBMs are often viewed

00:19:17: Connor: as old tech, and even people at conference

00:19:20: Connor: attendees occasionally say to me things

00:19:22: Connor: like oh well, you know Oracle, you know, we

00:19:24: Connor: know them for 55 page technical white

00:19:27: Connor: papers and things like that, or massive

00:19:31: Connor: documentation sets, etc.

00:19:33: Connor: Which are perhaps not as popular as they

00:19:35: Connor: used to be.

00:19:36: Connor: So I'm, you know, very keen on doing things

00:19:38: Connor: like video, because yeah, it's a, it's the

00:19:40: Connor: modern form and that's you and I.

00:19:44: Connor: I would love to say I was, you know, sort

00:19:45: Connor: of a ground breaker and I worked all this

00:19:48: Connor: out by myself, but it was purely self

00:19:51: Connor: fulfilling in the sense that the amount of

00:19:53: Connor: times I'm at home and I've got some sort of

00:19:54: Connor: obstacle the bath in the kitchen sink is

00:19:57: Connor: blocked, or I'm trying to erect a set of

00:19:59: Connor: shelves, and I've got seven IKEA pieces

00:20:01: Connor: left over and I'm trying to work out what

00:20:02: Connor: the hell went wrong.

00:20:03: Connor: You Google for it, and what I found myself

00:20:05: Connor: doing more and more of is not clicking on

00:20:07: Connor: the text solution but clicking on the

00:20:10: Connor: YouTube solution.

00:20:11: Connor: Here's how I unblock my sink, here's how

00:20:13: Connor: you do it without flooding.

00:20:14: Connor: You know the bathroom, etc.

00:20:15: Connor: And I thought well, if I'm sitting here

00:20:17: Connor: writing blog posts but at the same time

00:20:20: Connor: going, I'm getting better solutions by

00:20:22: Connor: clicking on a video.

00:20:23: Connor: Yeah, maybe I need to rethink what I'm

00:20:24: Connor: doing.

00:20:25: Kai: I mean the risk.

00:20:26: Kai: The risk is very big that you sit there for

00:20:28: Kai: like two hours because the recommendation

00:20:30: Kai: videos are right up Also interesting, right.

00:20:32: Kai: So this is the difference between that and

00:20:34: Kai: in text.

00:20:34: Kai: Yeah, so in the generation of my father

00:20:37: Kai: he's printing everything out.

00:20:38: Kai: So if he has a new product he's interested

00:20:40: Kai: in, then he's printing out the white paper

00:20:42: Kai: for that, everything, the technical paper,

00:20:44: Kai: everything.

00:20:45: Kai: I said, dad, you just look to YouTube and

00:20:48: Kai: see a review of that like four minutes.

00:20:50: Kai: If you're interested, you can go to the

00:20:51: Kai: website and see it and even try it at a

00:20:54: Kai: store.

00:20:54: Kai: Right, he's printing it out, never read it,

00:20:57: Kai: throw it away.

00:20:58: Kai: But this is a different time, right?

00:21:01: Caro: Yeah, I just remembered.

00:21:03: Caro: Last time or last year, we recorded our

00:21:06: Caro: podcast session.

00:21:07: Caro: You put.

00:21:08: Caro: You took your phone and looked at it and

00:21:11: Caro: said oh, there are 10,000 downloads now for

00:21:14: Caro: my podcast and you're not impressed.

00:21:19: Kai: Happy, but not impressed.

00:21:20: Caro: Yeah, yeah, it was kind of like, oh cool

00:21:23: Caro: Wow.

00:21:24: Connor: I have to admit I feel guilty doing a

00:21:26: Connor: podcast because one of the things I have

00:21:28: Connor: probably lapsed a bit this year is my own

00:21:30: Connor: podcast.

00:21:31: Connor: Yeah, I saw that Simply because I just

00:21:33: Connor: struggled to find the time.

00:21:34: Connor: But I have to admit I really like doing

00:21:36: Connor: them, even when I'm not doing them with

00:21:38: Connor: guests.

00:21:38: Connor: I'm just sitting there just blah, blah,

00:21:41: Connor: blah into the microphone.

00:21:42: Connor: My problem is I go edit my podcast

00:21:45: Connor: afterwards, unlike you guys.

00:21:47: Connor: You know, do it nice and raw and unfiltered,

00:21:49: Connor: which is something I should have done from

00:21:50: Connor: day one.

00:21:52: Connor: But now when I have podcast guests, podcast

00:21:54: Connor: guests I promise them that I will clean up

00:21:57: Connor: the audio at the end.

00:21:58: Connor: So at the end there's no ums, there's no

00:22:01: Connor: urs, there's no breaks, etc.

00:22:03: Connor: Because I promised my guests I said you'll

00:22:05: Connor: sound great, you'll sound like you were

00:22:06: Connor: perfectly literate, that you'd never had to

00:22:08: Connor: pause for thought, whatever.

00:22:10: Connor: That was a rod for my own back, because now

00:22:12: Connor: I can record a podcast in an hour and then

00:22:14: Connor: it's two days of going through it and

00:22:16: Connor: cleaning it up.

00:22:16: Kai: That was a stupid decision To give you a

00:22:19: Kai: recommendation as a podcast which is not as

00:22:22: Kai: successful as your own.

00:22:24: Kai: You could just do like an uncut version of

00:22:27: Kai: your podcast same name, just uncut or raw

00:22:30: Kai: Raw so say we're doing it raw, so every

00:22:33: Kai: single ur is there.

00:22:35: Kai: And the second recommendation is the tool

00:22:36: Kai: we are using which is filtering out a

00:22:39: Kai: couple of noises and sounds.

00:22:42: Kai: So we are trying all the different kind of

00:22:44: Kai: software right now and we talked about AI.

00:22:46: Kai: Those tools are coming up right now.

00:22:48: Kai: In like once a week or something new is

00:22:50: Kai: there and there are already tools that are

00:22:53: Kai: doing this stuff for you.

00:22:54: Caro: Yeah, but actually we are in the meanwhile

00:22:57: Caro: we also cutting out a lot of amps and urs

00:23:00: Caro: because there were some complaints about it.

00:23:03: Kai: Yeah, there were some.

00:23:05: Kai: Most of the time we were the amps and urs

00:23:08: Kai: not, I guess but yeah, there was a tool

00:23:10: Kai: that is called DeCipher, I think, and this

00:23:13: Kai: is exactly cutting out.

00:23:14: Kai: So it's transcribing a podcast and it's

00:23:17: Kai: showing the urs and amps underlying them

00:23:19: Kai: and you can just say remove all or remove a

00:23:22: Kai: couple of them, but most of the time it

00:23:24: Kai: doesn't sound natural right after that,

00:23:25: Kai: because the pause is not so long, but who

00:23:27: Kai: are we to tell you how to make your content

00:23:30: Kai: right?

00:23:30: Kai: I?

00:23:30: Connor: mean it's got to the stage now where, when

00:23:32: Connor: I bring up the recording and I use Audacity

00:23:36: Connor: to go through the recording, I know what

00:23:38: Connor: the waveform of an arm looks like.

00:23:40: Connor: So, I'm just looking at, I'm going that's

00:23:41: Connor: probably an arm, and the other problem I

00:23:43: Connor: have is, occasionally I'll fill in a space

00:23:46: Connor: with so, so blah, blah, blah, blah, blah,

00:23:48: Connor: blah, blah, and then that's how we build

00:23:50: Connor: the database.

00:23:51: Connor: So now we can be out of user.

00:23:53: Connor: So yeah, and I'm sitting there going why?

00:23:56: Caro: Why do I do it?

00:23:56: Connor: Because it just makes work for me.

00:23:58: Connor: But it's funny, you talk about doing a raw

00:24:00: Connor: or uncut version.

00:24:01: Connor: The problem there is my podcast would then

00:24:02: Connor: go from PG-13 to over 18s only.

00:24:07: Connor: Oh yeah, too much swearing, too much

00:24:08: Connor: profanity, too much frustration.

00:24:10: Connor: So.

00:24:11: Kai: I didn't know that you're cutting out so

00:24:13: Kai: much.

00:24:15: Kai: We try to behave at this podcast besides

00:24:16: Kai: guests that are complaining about the

00:24:18: Kai: headphones they get.

00:24:20: Connor: But funny enough.

00:24:20: Connor: In terms of podcasting, I have to admit

00:24:22: Connor: we're going on a podcast right here which

00:24:24: Connor: I'm sure interests us and doesn't interest

00:24:26: Connor: the listeners, but they can just bear with

00:24:27: Connor: it.

00:24:28: Connor: One of the things I think would be an

00:24:29: Connor: opportunity for both of us is in terms of

00:24:31: Connor: bringing it back to AI is obviously, you've

00:24:33: Connor: got a number of episodes in German.

00:24:36: Connor: All my episodes are obviously in English

00:24:37: Connor: because I'm a spectacularly monolingual.

00:24:41: Connor: I did work with an AI recently on a video,

00:24:44: Connor: where I took a video, uploaded it to an AI

00:24:46: Connor: service and they transcribed the video to

00:24:49: Connor: Spanish, including all the mouth movements

00:24:52: Connor: and everything.

00:24:53: Kai: So literally it looks like I'm speaking.

00:24:55: Connor: Spanish with my voice in Spanish and I was

00:24:58: Connor: like, surely this can't be any good.

00:25:01: Connor: But it blew me away.

00:25:02: Connor: It was amazing.

00:25:03: Connor: In fact, the only thing that got wrong was

00:25:05: Connor: because I was talking about 23C.

00:25:08: Connor: It didn't understand tech, so every time I

00:25:10: Connor: said 23C, it said 23 degrees Celsius.

00:25:13: Connor: But other than that, I obviously didn't

00:25:16: Connor: know if it was any good because I don't

00:25:17: Connor: speak Spanish, but I sent it to a few

00:25:19: Connor: Spanish speakers inside Oracle.

00:25:21: Connor: You probably know Monica from the APN team

00:25:22: Connor: and she came back and she said yeah, she

00:25:24: Connor: said the conversational Spanish is great,

00:25:26: Connor: it's just the tech stuff it struggled with.

00:25:28: Connor: But I think, especially because podcast

00:25:30: Connor: doesn't involve video, I reckon it would be

00:25:33: Connor: good for, for example, you could take your

00:25:34: Connor: German episodes and use an AI to publish

00:25:37: Connor: them in English and I could take my English

00:25:39: Connor: and publish them in any language.

00:25:41: Kai: This is also an announcement for our

00:25:44: Kai: listeners.

00:25:44: Kai: Now in the English episode, no, carol is

00:25:47: Kai: very scared.

00:25:47: Kai: Now what we are doing.

00:25:49: Kai: Like half a year in the middle of the year,

00:25:51: Kai: I started to transcribe each and every

00:25:53: Kai: episode with I think it's Mac Whisper, the

00:25:57: Kai: framework under the hood, and this is

00:25:59: Kai: transcribing everything we say and

00:26:00: Kai: translate it into English, spanish and

00:26:02: Kai: whatever, and now we are including that

00:26:05: Kai: into the podcast stream so each and every

00:26:07: Kai: podcast player that supports it can do some

00:26:10: Kai: subtitles for your talk and you can read

00:26:12: Kai: with it, or it's just during the talk.

00:26:14: Kai: It's going on a screen so you can read it.

00:26:17: Kai: And also for YouTube, so you can choose

00:26:19: Kai: from each and every language and it's

00:26:21: Kai: transcribing from the German translation

00:26:23: Kai: automatically.

00:26:24: Kai: So for every listener who doesn't speak

00:26:26: Kai: German, you can follow the episodes we

00:26:28: Kai: already did with the subtitles, which is

00:26:30: Kai: basically just reading book, because it's

00:26:32: Kai: text you can follow and you can read an

00:26:34: Kai: interview.

00:26:34: Kai: This is something I had in the beginning.

00:26:36: Kai: I tried to do that.

00:26:38: Kai: It was everything fine, Besides the name

00:26:40: Kai: Devs on tape.

00:26:41: Kai: It was death on tape, Death on tape.

00:26:43: Kai: And I said welcome to death on tape and I

00:26:45: Kai: felt like we were some rock band or

00:26:48: Kai: something I don't know.

00:26:49: Kai: And then I did like search and replace and

00:26:52: Kai: I found the next word and said search and

00:26:54: Kai: replace, search and replace.

00:26:56: Kai: Come on, you can't even get the name of the

00:26:58: Kai: guest right.

00:26:59: Kai: This would be embarrassing if we published

00:27:01: Kai: that.

00:27:01: Kai: We did some tuning of that and at some

00:27:03: Kai: place.

00:27:04: Kai: I didn't care.

00:27:05: Kai: I said okay, everyone knows that it's not

00:27:07: Kai: death on tape.

00:27:08: Kai: Besides, we have a very old guest.

00:27:10: Connor: But yeah, Right, so much older so much

00:27:14: Connor: older than we.

00:27:15: Connor: But another cool thing with AI in that

00:27:17: Connor: space was I was hoping to go on the Latin

00:27:20: Connor: America tour earlier this year and just

00:27:22: Connor: timetables and scheduling didn't work it

00:27:24: Connor: out.

00:27:24: Connor: But I saw that they were happy to embrace

00:27:27: Connor: English speakers because even things like

00:27:29: Connor: PowerPoint now can do you start presenting

00:27:32: Connor: and it could do dynamic translation and

00:27:34: Connor: subtitles as you speak.

00:27:35: Caro: Yeah.

00:27:36: Connor: And so the English speakers there were

00:27:37: Connor: having translation to Spanish and

00:27:39: Connor: Portuguese, as long as the Wi-Fi was good.

00:27:41: Caro: Yeah.

00:27:41: Connor: They were.

00:27:42: Connor: The audience was much happier because I

00:27:44: Connor: could actually follow along, even though

00:27:45: Connor: the speakers in English, so I'm hoping to

00:27:47: Connor: use that to justify a trip to non-English

00:27:52: Connor: speaking countries next year.

00:27:54: Kai: This is what is happening on Max Hifts each

00:27:56: Kai: and every presentation.

00:27:57: Kai: When he puts a new USB key in the laptop.

00:27:59: Kai: It's doing some key like pretending a

00:28:02: Kai: keyboard or something which is doing in

00:28:04: Kai: PowerPoint, a kind of a key combination

00:28:06: Kai: which enables the live transcription on the

00:28:08: Kai: screen.

00:28:09: Kai: And then if he presents on German and it's

00:28:11: Kai: not configured well, then you have like

00:28:13: Kai: English words that sounds like German words

00:28:16: Kai: and you have no one even can follow what's

00:28:18: Kai: on the screen right now, so it doesn't have

00:28:20: Kai: anything to do with what he's talking about.

00:28:22: Kai: It said all right, this is happening to me

00:28:24: Kai: each and every presentation.

00:28:26: Kai: I put out my USB key, put it in my pocket,

00:28:28: Kai: disable it on the screen and then go

00:28:30: Kai: forward.

00:28:31: Kai: So yeah, that's my experience with this

00:28:33: Kai: service thing.

00:28:34: Connor: My worry is always that, because I don't

00:28:36: Connor: speak the other language, for all I know,

00:28:39: Connor: I'm saying you should be using 23C because

00:28:42: Connor: of this and the subtitles are going.

00:28:44: Connor: This speaker is a loser because you don't

00:28:47: Connor: know it's great if you do your presentation.

00:28:50: Kai: Someone comes to you after and says oh, the

00:28:52: Kai: thing where you talk about yachting right

00:28:55: Kai: or a golfing was pretty nice and you said

00:28:57: Kai: what the?

00:28:58: Kai: hell happened, so I was not doing next

00:29:00: Kai: course to golf or whatever.

00:29:02: Kai: Yeah, we will see where the future leads us.

00:29:04: Kai: So the office suite is bringing so much AI

00:29:06: Kai: in future.

00:29:06: Kai: I saw that.

00:29:07: Kai: So you have like write a text into office

00:29:10: Kai: and say I get the earnings report in Excel,

00:29:13: Kai: put it into this PowerPoint presentation.

00:29:15: Kai: I got via email and outlook yesterday and

00:29:18: Kai: please do the annotations from Frank, who

00:29:20: Kai: did it like yesterday send an email, put

00:29:22: Kai: everything together to a PowerPoint

00:29:24: Kai: presentation and this is also from Sasha

00:29:26: Kai: Lover, by the way.

00:29:28: Kai: We try to be like greenwashing the company,

00:29:30: Kai: so put some palm leaves on the slides and

00:29:33: Kai: what you get is like all the information

00:29:35: Kai: from all the different office products and

00:29:37: Kai: put them into a PowerPoint presentation.

00:29:39: Kai: Right, this is something I can.

00:29:41: Kai: Yeah, I'm looking forward to see.

00:29:42: Kai: So they said that 60 companies are

00:29:45: Kai: currently trying that in Germany and they

00:29:47: Kai: tried to roll it out in the next year.

00:29:49: Connor: I think, I worry that I'll give my slides

00:29:52: Connor: to an AI and it'll say you have 400 slides.

00:29:55: Connor: This is a bad idea.

00:29:57: Connor: And it'll shrink it down to 10 slides and

00:29:59: Connor: I'll go no, I don't want 10 slides, I want

00:30:00: Connor: my original 400.

00:30:01: Kai: I want my 500 megabyte power, power which

00:30:05: Kai: struggles to open on every presentation

00:30:07: Kai: laptop.

00:30:08: Connor: I have to know I do have some issues with

00:30:09: Connor: PowerPoint in terms of performance, but

00:30:11: Connor: that's another one, another one for the.

00:30:12: Connor: AI.

00:30:13: Connor: But speaking of AI and dragging in data

00:30:14: Connor: from other sources, it would be.

00:30:16: Connor: It behooves me to at least do a little bit

00:30:19: Connor: of a promo for the fact that.

00:30:20: Connor: Have you seen our select AI stuff that's

00:30:21: Connor: coming in Oracle?

00:30:23: Connor: Here's something that you know.

00:30:23: Connor: This is I won't call this an announcement,

00:30:25: Connor: because this was already demoed at

00:30:27: Connor: Cloudworld, but not a lot of people got to

00:30:29: Connor: go to Cloudworld from Europe, which is

00:30:31: Connor: facilities where typically, if you're

00:30:34: Connor: running a SQL statement, you would just

00:30:35: Connor: select and you would have the columns and

00:30:37: Connor: all the complexities of SQL.

00:30:39: Connor: One of the things that we demoed at

00:30:40: Connor: Cloudworld is you do select, then the

00:30:43: Connor: keyword AI and then your chat GPT style of

00:30:45: Connor: question Show me the sales for the last

00:30:49: Connor: year.

00:30:49: Connor: But unlike chat GPT, which this says well,

00:30:52: Connor: I don't know what sales are in your company

00:30:54: Connor: I can give you a generic sort of well, this

00:30:57: Connor: is how I would write the SQL.

00:30:58: Connor: This is what I would roughly do, but of

00:31:00: Connor: course it bears no relationship to whatever

00:31:02: Connor: I can't the acronyms called RAG, retrieval,

00:31:06: Connor: augmented generation, I don't know, rag.

00:31:09: Connor: Yeah, and that is one of the facilities we

00:31:12: Connor: have now is you can pass information about

00:31:15: Connor: what you have in your database in terms of

00:31:18: Connor: what are the tables, what are the columns,

00:31:19: Connor: how is it structured as part of a response

00:31:22: Connor: to an AI engine Could be open AI, could be

00:31:24: Connor: here we're partnering with, could be any of

00:31:27: Connor: them.

00:31:28: Connor: So now, when you say select AI, what are

00:31:30: Connor: sales for the last year, it sends all for

00:31:33: Connor: your situation.

00:31:35: Connor: This is the SQL we would run.

00:31:36: Connor: You know, select some sales from using.

00:31:40: Connor: One of our sample scheme is shsales.

00:31:43: Connor: But it knows what you've got in your own

00:31:45: Connor: database.

00:31:46: Connor: So that says your sales was $700,000.

00:31:48: Connor: Because now it knows it can run an SQL that

00:31:51: Connor: you never wrote.

00:31:52: Connor: You asked an English language question.

00:31:54: Connor: It said because I now know the tables and

00:31:56: Connor: columns, here's the SQL rerun.

00:31:58: Connor: You can show you the SQL if you want, or it

00:32:00: Connor: can just give you the answer.

00:32:01: Kai: So is this technically working in the

00:32:04: Kai: database or on the client, or is it

00:32:07: Kai: translation and stuff working on a cloud?

00:32:10: Connor: It's a combination of things.

00:32:12: Connor: So obviously the data source is your own

00:32:15: Connor: database, so we have it in our autonomous

00:32:17: Connor: database.

00:32:18: Connor: I would do a demo here.

00:32:20: Connor: I had planned to do a demo this year at the

00:32:21: Connor: conference here, but the Wi-Fi is a little

00:32:23: Connor: bit challenging to get to a server in

00:32:25: Connor: Phoenix.

00:32:26: Connor: And then what happens is, obviously we want

00:32:28: Connor: to leverage the chat GPT style of facility

00:32:32: Connor: where you can take an English language

00:32:34: Connor: request and turn it into something

00:32:35: Connor: technical.

00:32:35: Connor: You don't want to do that in the database,

00:32:37: Connor: because now you're loading a full AI engine

00:32:39: Connor: in the database and next thing you know,

00:32:40: Connor: you've got a 500 terabyte database, which

00:32:43: Connor: one terabyte of which is your data, the

00:32:44: Connor: rest is AI facilities.

00:32:47: Connor: But we just do a rest call out to a

00:32:49: Connor: provider AI facility which says here's the

00:32:51: Connor: English text, but here's the context as

00:32:53: Connor: well.

00:32:53: Connor: Here's our tables, here's our columns,

00:32:55: Connor: here's how they're related together.

00:32:56: Connor: That all goes as one request, so basically

00:32:58: Connor: your text and the context.

00:33:01: Connor: And then it comes back and says oh well,

00:33:03: Connor: given that I now know what your data looks

00:33:05: Connor: like, here's how the SQL you would write to

00:33:07: Connor: do it.

00:33:08: Connor: And so the nice thing there is, we get to a

00:33:10: Connor: point where ultimately, you know, a lot of

00:33:13: Connor: facilities in IT are really designed to

00:33:16: Connor: help business users.

00:33:17: Connor: The CEO, you know.

00:33:18: Connor: He's the person that gets on the phone or

00:33:20: Connor: gets an email and says you know what was

00:33:22: Connor: last week's sales?

00:33:24: Connor: And he gets sent off and everyone in the IT

00:33:26: Connor: department gets told drop everything.

00:33:27: Connor: Something's come from the CEO.

00:33:29: Connor: This is priority one.

00:33:30: Connor: Someone go write a query, someone do

00:33:32: Connor: whatever, and what?

00:33:34: Connor: For the last 20, 30, 50 years in IT, we

00:33:37: Connor: tried to solve this by saying, well, we

00:33:38: Connor: need to give the CEO a dashboard, something

00:33:42: Connor: that they can understand, because you know

00:33:44: Connor: we give them a screen.

00:33:45: Connor: You click here, you click here and here's a

00:33:47: Connor: pre-canned sort of answers based on your

00:33:49: Connor: questions you've asked in the past.

00:33:51: Connor: But now this thing is now.

00:33:53: Connor: You know, just ask it.

00:33:55: Connor: And eventually they'll probably do it into

00:33:56: Connor: a microphone, but at the moment they'd

00:33:58: Connor: simply type in some text saying what were

00:33:59: Connor: last week's sales versus the same period

00:34:01: Connor: for last year and they'll get an answer,

00:34:03: Connor: not of some fictional data source.

00:34:05: Connor: They can answer from your actual database

00:34:08: Connor: using an AI engine to build the SQL.

00:34:10: Connor: That's super cool.

00:34:12: Caro: How do you so, when I think it sounds great,

00:34:16: Caro: first call.

00:34:18: Caro: But then I remember, for example, some of

00:34:21: Caro: my data models of my customers.

00:34:25: Connor: I can see where this is going.

00:34:29: Caro: How should it work there?

00:34:30: Caro: So how do you feed this thing with

00:34:32: Caro: information?

00:34:33: Caro: So what table does what mean and which

00:34:36: Caro: column does?

00:34:38: Connor: And I totally agree what I've spoken about.

00:34:40: Connor: There is a utopian vision in the perfect

00:34:43: Connor: world, because we all know that in the real

00:34:46: Connor: world, in our sales table, we have a column

00:34:49: Connor: called total sales and when you say, does

00:34:51: Connor: that say total sales?

00:34:52: Connor: I go no, no, no, don't use that column.

00:34:54: Connor: It's a column here called unaccounted for

00:34:56: Connor: XY3.

00:34:57: Connor: That's where the total sales is.

00:34:58: Connor: That's where you need to get it from.

00:35:00: Connor: We need to take that, add it to the total

00:35:02: Connor: sales, divide it by the margin.

00:35:03: Connor: Where's the margin?

00:35:04: Connor: Well, we don't store that in the database.

00:35:06: Connor: We go ask Betty in finance.

00:35:08: Connor: She tells us that it's 11% this week.

00:35:10: Connor: So you know, reality will obviously come

00:35:13: Connor: into play as well.

00:35:16: Kai: I mean, this is definitely what I was

00:35:18: Kai: trying to say beforehand.

00:35:20: Kai: So the developers are not anymore just

00:35:22: Kai: writing everything like SQL statements,

00:35:24: Kai: programs and whatever.

00:35:25: Kai: They are ensuring that the model can train

00:35:27: Kai: based on the stuff we are doing right now.

00:35:29: Kai: So what we saw with the testing frameworks

00:35:32: Kai: in the database was that people started to

00:35:34: Kai: use annotations for tables and columns Like

00:35:37: Kai: what is this column for?

00:35:38: Kai: So I could imagine that, if this is

00:35:40: Kai: something you have to do to train the model,

00:35:42: Kai: to have each and every column edited or

00:35:45: Kai: extended by a comment, like an English

00:35:48: Kai: sentence, which is kind of an understanding

00:35:50: Kai: before the AI, and some keywords that are

00:35:53: Kai: additionally to that, not directly, the

00:35:55: Kai: constraints or something, so GPT knows

00:35:58: Kai: which table is connected to another.

00:36:01: Kai: But this is something the AI can understand,

00:36:03: Kai: but this is something the AI cannot do by

00:36:05: Kai: themselves.

00:36:07: Kai: So we are getting busy then going to our

00:36:09: Kai: data model and try to document it right, so

00:36:12: Kai: the GPT can read the documentation, or GPT

00:36:15: Kai: in general can read the documentation and

00:36:17: Kai: understand the data model by your

00:36:19: Kai: documentation, and I think this is where it

00:36:22: Kai: leads to.

00:36:23: Connor: I agree and funnily enough, there is a tool

00:36:25: Connor: he puts his marketing hat on again.

00:36:27: Connor: There's a tool in that Oracle that has been

00:36:29: Connor: doing that for about 20 years the concept

00:36:31: Connor: of, rather than generating code, you

00:36:33: Connor: describe the requirements of what you want

00:36:35: Connor: to achieve and that course, that tool is

00:36:37: Connor: called Apex.

00:36:39: Connor: Apex is pretty much what you've just said

00:36:40: Connor: there.

00:36:40: Connor: You describe in metadata what you want to

00:36:43: Connor: achieve and Apex, you could say, has led

00:36:46: Connor: the way, is a pioneer in that sense,

00:36:48: Connor: because the database is now catching up, as

00:36:50: Connor: you say, with things like annotations and

00:36:51: Connor: even just simple comments.

00:36:53: Connor: But that is an Apex.

00:36:55: Connor: Developer doesn't write code.

00:36:56: Connor: They do a little bit of code here and there

00:36:58: Connor: and, god forbid, they're into CSS and

00:37:00: Connor: JavaScript.

00:37:01: Connor: But, putting that aside, you got it on

00:37:02: Connor: record now.

00:37:03: Connor: But, putting that aside, the vast majority

00:37:06: Connor: of an Apex developer's time is describing

00:37:08: Connor: the functionality they want.

00:37:10: Connor: This is where I want the regions, this is

00:37:11: Connor: where I want the items, this is how I want

00:37:13: Connor: the page flow to work.

00:37:14: Connor: They're not coding it, they're describing

00:37:16: Connor: metadata, and so that's one of the cool

00:37:18: Connor: things.

00:37:18: Connor: I think Apex will actually really benefit

00:37:21: Connor: from AI facilities, because people will be

00:37:23: Connor: able to talk to AI and say I want to build

00:37:26: Connor: an Apex app and it's not going to write

00:37:27: Connor: code.

00:37:28: Connor: It's going to build Apex metadata, because

00:37:30: Connor: the last thing you want is an AI to be a

00:37:32: Connor: code generator.

00:37:33: Connor: Everyone talks about oh AI is going to

00:37:35: Connor: write my code.

00:37:36: Connor: I think that's a nightmare scenario,

00:37:38: Connor: because the AI is going to write beautiful

00:37:40: Connor: code, several million lines of it, and then

00:37:43: Connor: you're going to say, ok, now who's going to

00:37:44: Connor: maintain that?

00:37:45: Connor: Another instance where a copy and paste the

00:37:47: Connor: code and say optimize it, find the error

00:37:49: Connor: right Exactly and so, yeah, I view Apex as

00:37:54: Connor: being a maybe lucky that it was a solution

00:37:58: Connor: that suits AI, in the sense that you never

00:38:00: Connor: really generate the code, it's just

00:38:02: Connor: interpreted metadata, because that's what I

00:38:04: Connor: also love.

00:38:05: Connor: I also love metadata to work with.

00:38:07: Kai: I mean we heard from the Apex corner at

00:38:10: Kai: Oracle that they are heavily interested in

00:38:13: Kai: AI and starting to bring them to the app.

00:38:15: Kai: So if you have, any secret information for

00:38:17: Kai: us from the team?

00:38:18: Kai: No, just kidding.

00:38:20: Kai: I think they need to do so much stuff to

00:38:22: Kai: bring it out as a declarative way inside

00:38:25: Kai: Apex too, right?

00:38:26: Kai: So we as developers or IT I think people we

00:38:31: Kai: are getting frustrated if we try a new

00:38:32: Kai: feature and it's not working at all.

00:38:34: Kai: So Apex goes AI and you open Apex and try

00:38:38: Kai: to speak like, create a table and a form on

00:38:40: Kai: it, and what it does is like can you repeat

00:38:42: Kai: that?

00:38:43: Kai: I didn't get that.

00:38:44: Kai: I'm not speaking Spanish by now.

00:38:45: Kai: Can you repeat that in English?

00:38:47: Kai: Can you speak louder?

00:38:48: Kai: No, I'm not able to do that.

00:38:49: Kai: Here's your button, right, and this is

00:38:52: Kai: something I would say OK, you shipped

00:38:54: Kai: something which is not even close to usable,

00:38:57: Kai: right?

00:38:57: Kai: Yeah, we would see where this leads to.

00:38:59: Kai: But the question how this works like is the

00:39:02: Kai: model in the database and the calculation

00:39:06: Kai: maybe spread from client to database is the

00:39:08: Kai: second.

00:39:09: Kai: We got the top co-pilot and all the

00:39:11: Kai: assistance stuff.

00:39:13: Kai: Everyone said hey, it's a great thing to

00:39:15: Kai: try on your private projects and things,

00:39:17: Kai: but if it leads to a customer's workspace,

00:39:21: Kai: I cannot send the code and the metadata and

00:39:23: Kai: the context to another service in the

00:39:25: Kai: internet.

00:39:26: Kai: This can't leave my environment.

00:39:28: Kai: So they said co-pilot is nice, it's cool,

00:39:30: Kai: it's a future, but I cannot use it for my

00:39:31: Kai: daily work.

00:39:32: Kai: Instead of create a table, amp and so on,

00:39:35: Kai: and then I use this, prep this code snippet

00:39:39: Kai: and replace the column names and so on with

00:39:41: Kai: my context from the customers In this case.

00:39:43: Kai: I would say if we have AI and the customer

00:39:46: Kai: knows that anything leaves their

00:39:48: Kai: environment to a third party service, they

00:39:51: Kai: would stop it, even if it's inside Oracle

00:39:53: Kai: Cloud.

00:39:53: Kai: I think there would be people saying this

00:39:56: Kai: is not the right way.

00:39:57: Kai: We don't want something to leave the

00:39:58: Kai: database.

00:40:00: Connor: And I think AI is just a magnification or

00:40:02: Connor: those kind of services.

00:40:03: Connor: It's just a magnification of the existing

00:40:05: Connor: security challenges we've already had or

00:40:07: Connor: always had.

00:40:08: Connor: Plus, the example is the amount of

00:40:10: Connor: applications out there that use open source

00:40:12: Connor: JavaScript libraries, because there's a

00:40:15: Connor: bazillion of them and they're awesome.

00:40:17: Connor: Every problem you have, someone has written

00:40:19: Connor: a JavaScript library to solve that problem.

00:40:22: Connor: You could almost say there are no new

00:40:24: Connor: problems or all the typical IT related

00:40:27: Connor: problems.

00:40:28: Connor: I want to validate an email address.

00:40:29: Connor: I want to call the rest of it.

00:40:31: Connor: There's a JavaScript library out there and

00:40:33: Connor: it's open source.

00:40:34: Connor: The problem is, a lot of these JavaScript

00:40:36: Connor: libraries were written to solve a

00:40:37: Connor: particular problem by a particular person

00:40:39: Connor: who, in all good intentions, said I'll make

00:40:42: Connor: this open source so the world can use it.

00:40:45: Connor: They had good intentions and, of course,

00:40:46: Connor: then someone says, well, I don't want to

00:40:48: Connor: look after this anymore and someone says

00:40:50: Connor: I'll take over it for you, and we have no

00:40:52: Connor: idea who that person is and you just

00:40:55: Connor: blindly download these open source

00:40:57: Connor: facilities which, as I said, for the vast

00:40:59: Connor: majority, are going to be wonderful.

00:41:02: Connor: But that concept of we already have this

00:41:04: Connor: implicit trust in something that came from

00:41:07: Connor: somewhere else that we have no idea where

00:41:08: Connor: from?

00:41:09: Connor: That's a worry, because it opens up so many

00:41:12: Connor: opportunities for bad actors.

00:41:14: Kai: This is a content of, I think, one of our

00:41:17: Kai: first podcast episodes from the.

00:41:18: Kai: Devs on Tab.

00:41:19: Kai: It was called Library Nightmares, Episode 1

00:41:22: Kai: and 2.

00:41:23: Kai: We talked about exactly this.

00:41:24: Kai: What happened in a real, real example where

00:41:27: Kai: some Nodejs package in MPM repository was

00:41:30: Kai: flawed and was injected like code injected,

00:41:32: Kai: and what happens right after that?

00:41:34: Kai: And I think it was like for lock for J.

00:41:37: Caro: Yeah, lock for J and another one.

00:41:40: Kai: Yeah, I mean we do that, but the difference

00:41:42: Kai: is, if we have an AI engine, it needs so

00:41:45: Kai: much power to calculate on.

00:41:46: Kai: So we don't have a problem in the Oracle

00:41:48: Kai: Cloud because you can just get limitless

00:41:50: Kai: hardware.

00:41:51: Kai: You have to pay for it, obviously, but you

00:41:53: Kai: cannot use the on-premises server right now

00:41:55: Kai: at the customer's place to say now you're

00:41:57: Kai: doing AI, so here one terabyte of model at

00:42:00: Kai: least, and then each and every question the

00:42:02: Kai: CEO is asking, every systems has to be shut

00:42:06: Kai: down to have the power to answer the

00:42:09: Kai: question.

00:42:09: Kai: So this is something that has to be solved,

00:42:12: Kai: definitely, and you cannot give everyone a

00:42:14: Kai: very, very high-profile laptop to do it on

00:42:16: Kai: the machine.

00:42:17: Kai: I'm very, very looking forward to see how

00:42:19: Kai: the on-premises thing is working.

00:42:22: Kai: Then, besides getting a new another Oracle

00:42:24: Kai: Cloud service bound to our system, we see

00:42:27: Kai: OK, I've actually prepared another

00:42:30: Kai: technical question for you here we go.

00:42:32: Caro: Where I want to know your opinion, but it's

00:42:34: Caro: not as fancy as AI.

00:42:36: Connor: That's OK.

00:42:37: Connor: We've been diving in all sorts of

00:42:39: Connor: directions this morning.

00:42:41: Kai: I think Connor will answer this question

00:42:43: Kai: using at least one time.

00:42:44: Kai: Ai right Could be.

00:42:46: Kai: Maybe We'll see.

00:42:48: Caro: Ok.

00:42:48: Caro: So it's a bit of a story of this conference

00:42:51: Caro: actually.

00:42:51: Caro: So I met one of my customers and they

00:42:55: Caro: talked about the 8.23 C features and how

00:42:59: Caro: they like this whole JSON stuff and how it

00:43:02: Caro: would change their development, et cetera,

00:43:05: Caro: fixing some problems, and they said they're

00:43:08: Caro: already in the cloud.

00:43:09: Caro: And they said something like OK, the next

00:43:12: Caro: year in autumn when we turn it we see 23 C

00:43:17: Caro: might be released.

00:43:18: Caro: No, no, they just want to update there,

00:43:21: Caro: they directly want to update their

00:43:22: Caro: environment.

00:43:23: Caro: And I was surprised, because most time a

00:43:26: Caro: bit more hesitating.

00:43:28: Connor: Cautious.

00:43:29: Caro: Yeah, Thanks Cautious about this whole new

00:43:31: Caro: stuff and we already know your YouTube

00:43:34: Caro: channel.

00:43:34: Caro: You're doing a lot of videos also from a

00:43:37: Caro: new features et cetera, and I just wanted

00:43:39: Caro: to ask you of your opinion would you also

00:43:42: Caro: update instantly when the 8.23 C is

00:43:45: Caro: published, or would you do some other

00:43:47: Caro: strategy?

00:43:48: Connor: I think probably since about 19 and 18 C

00:43:51: Connor: came along, we recognize the fact that no

00:43:54: Connor: matter what we do, there's always going to

00:43:56: Connor: be two types of customers.

00:43:58: Connor: There's the customers going.

00:43:59: Connor: For God's sake, oracle, I need this feature.

00:44:02: Connor: You told me it was coming.

00:44:03: Connor: I need it.

00:44:04: Connor: It's going to give me a business

00:44:05: Connor: competitive advantage.

00:44:06: Connor: Where is it?

00:44:07: Connor: There's another half of customers going.

00:44:09: Connor: Just leave me the hell alone.

00:44:10: Connor: I'm happy with my version of the database.

00:44:13: Connor: I don't want to upgrade ever.

00:44:14: Connor: The only time I actually want to do any

00:44:15: Connor: kind of patching which is an impact to me

00:44:17: Connor: is if there's security issues.

00:44:19: Connor: And there's always going to be both types.

00:44:22: Connor: I feel and don't get me wrong, I think

00:44:24: Connor: Oracle in the past have perhaps not done a

00:44:27: Connor: great job with one or the other.

00:44:30: Connor: You go back to the Oracle 10, 11 days where

00:44:32: Connor: there was like five, six years between

00:44:34: Connor: releases.

00:44:34: Connor: People who don't like upvoting they're

00:44:36: Connor: going.

00:44:36: Connor: This is the greatest thing ever.

00:44:38: Connor: The people that are waiting for a

00:44:39: Connor: particular facility are going.

00:44:41: Connor: What the hell?

00:44:42: Connor: I've got to wait five years just to have

00:44:43: Connor: petitioning Come on.

00:44:45: Connor: And then we sort of went to maybe almost

00:44:48: Connor: maybe too far the other way with 18, 19.

00:44:50: Connor: There was going to be a 20 and a 21, where

00:44:53: Connor: by the time people sort of were familiar

00:44:55: Connor: with 18, we're going you should be using 19.

00:44:58: Connor: They go OK, I'll upgrade to 19.

00:45:00: Connor: We go yeah, you should look at 21.

00:45:02: Connor: Yeah, and so people got frustrated at that.

00:45:04: Connor: We've gone too far the other way, trying to

00:45:06: Connor: satisfy the customers and saying, get us

00:45:07: Connor: features fast.

00:45:09: Connor: And then the people that don't like

00:45:10: Connor: upgrading going.

00:45:11: Connor: I'm spending two thirds of my IT budget

00:45:13: Connor: doing regression, testing and upgrades.

00:45:15: Connor: Give me a break.

00:45:16: Connor: So that's why you would have seen this

00:45:18: Connor: thing over the last couple years, this

00:45:19: Connor: thing between innovation versus long term

00:45:21: Connor: support releases.

00:45:22: Connor: So with that background in mind, one of the

00:45:24: Connor: things I do, like we've done in 23C, is

00:45:26: Connor: because we got a product out to people

00:45:28: Connor: early.

00:45:28: Connor: We got 23.2 out with that developer 3

00:45:30: Connor: release.

00:45:31: Connor: If you download it today, you get 23.3.

00:45:33: Connor: So we've actually updated that release

00:45:35: Connor: Internally.

00:45:35: Connor: We're now working with 23.4.

00:45:37: Connor: I don't know if that will become the

00:45:39: Connor: general available release, that's 20.4 is

00:45:41: Connor: the one you get.

00:45:42: Connor: I think if you use the Oracle base release

00:45:44: Connor: on cloud at the moment, ga will maybe be

00:45:47: Connor: 23.5, could be 23.6, et cetera.

00:45:50: Connor: I view that it's different to 19.

00:45:52: Connor: Because 19 came out as 19.3 and it came out

00:45:55: Connor: on cloud and very quickly GA etc.

00:45:57: Connor: So there weren't several iterations of

00:46:01: Connor: patches, bug fixes, security fixes, etc.

00:46:04: Connor: By the time you get access to 23C for real,

00:46:07: Connor: it's probably going to have been through

00:46:09: Connor: three or four RUs.

00:46:11: Connor: That doesn't mean I would say everything

00:46:12: Connor: will cost much to just be leaping on board.

00:46:15: Connor: But what does give you is increased

00:46:16: Connor: confidence that, if you are thinking well,

00:46:20: Connor: normally I wouldn't upgrade so early

00:46:23: Connor: because I know with 19 it was fairly raw.

00:46:26: Connor: I think Oracle has learned some of those

00:46:30: Connor: errors and we've changed direction a little

00:46:32: Connor: bit.

00:46:32: Connor: So I think 23C by the time it goes GA, will

00:46:35: Connor: be pretty damn stable.

00:46:37: Connor: And don't get me wrong, I've enjoyed

00:46:39: Connor: monitoring our database forum for the free

00:46:42: Connor: release because when we came out with 23.2,

00:46:45: Connor: there's all these people that much like

00:46:46: Connor: myself we love going in and trying to find

00:46:48: Connor: edge cases where we can break things and

00:46:50: Connor: someone's going I tried to do a Boolean on

00:46:52: Connor: a null constraint on this and put an index

00:46:54: Connor: on it and then I got an Oro 600.

00:46:56: Connor: But we applaud that.

00:46:58: Connor: Anyone that wants to go hunting to try and

00:46:59: Connor: find things.

00:47:01: Connor: Every bug makes our product better, because

00:47:03: Connor: by the time we go GA it's one less bug that

00:47:05: Connor: is found out in the wild.

00:47:08: Connor: So my advice for your customer would be for

00:47:10: Connor: me, if I was a business, when 23C goes GA,

00:47:14: Connor: I would generally wait maybe one or two RUs

00:47:17: Connor: and then start looking at it.

00:47:19: Connor: Because the challenge with GA is, up until

00:47:21: Connor: then, every single genuine production user

00:47:25: Connor: will be running it on Linux and probably on

00:47:28: Connor: Oracle Cloud, where we have total control

00:47:29: Connor: of the OS, the stack, whatever.

00:47:31: Connor: When it goes GA, you're now, yes, you're

00:47:34: Connor: running on, say, a supported Oracle

00:47:35: Connor: platform, but maybe your OS patches aren't

00:47:38: Connor: as up-to-date as ours, or maybe you're more

00:47:40: Connor: up-to-date than ours.

00:47:41: Connor: There's these little variations.

00:47:43: Connor: So you run that risk and that's why we do

00:47:45: Connor: GA less.

00:47:46: Connor: Last we do developer free first, let's get

00:47:49: Connor: people out there familiar with it.

00:47:50: Connor: Then we go cloud first, because we know the

00:47:52: Connor: entire stack, and then we go the most

00:47:54: Connor: volatile, which is GA.

00:47:56: Connor: People are maybe running the same Linux as

00:47:58: Connor: us, maybe they're not.

00:47:59: Connor: They may be running the same hardware as us,

00:48:01: Connor: maybe they're not.

00:48:02: Connor: They've got some good deal with some cheap

00:48:04: Connor: Chinese arm manufacturer that does sort of

00:48:08: Connor: budget almost correct chips.

00:48:10: Connor: There's so many scenarios when it goes GA

00:48:12: Connor: that the risk of encountering something

00:48:14: Connor: that no one else has ever seen it goes up a

00:48:16: Connor: bit.

00:48:17: Connor: So I would.

00:48:18: Connor: What's the saying?

00:48:19: Connor: They say it's great to be a pioneer, but

00:48:21: Connor: you've got to remember that in the early

00:48:23: Connor: days most of the pioneers died, and so I

00:48:27: Connor: like to follow the pioneers.

00:48:31: Kai: I was searching for the right way to finish

00:48:34: Kai: because we're already late so we don't have

00:48:37: Kai: our categories with us because they are the

00:48:40: Kai: same question.

00:48:40: Kai: So I mean, in case if there's any

00:48:42: Kai: technology that appeared during the year we

00:48:46: Kai: didn't speak, you can tell us.

00:48:47: Kai: If you want to undo some technology Trend

00:48:52: Kai: from the last year, go ahead, but we would

00:48:54: Kai: say that the categories are not necessary

00:48:56: Kai: at this part.

00:48:57: Kai: You already gave our or her customers and

00:49:00: Kai: the people in advice and if I would say how

00:49:02: Kai: to close the episode of this, this Devs on

00:49:06: Kai: tape thing properly, do you have another

00:49:08: Kai: advice for our listeners from Devs on tape

00:49:12: Kai: or do you have any very high secret

00:49:14: Kai: information from Oracle?

00:49:15: Connor: to share with us.

00:49:17: Kai: I have to do it again.

00:49:19: Connor: I do have many high level Oracle secrets?

00:49:21: Connor: No, I have none.

00:49:22: Connor: I would love to create the impression that

00:49:25: Connor: the place I sit inside the Oracle

00:49:28: Connor: organization is where they go.

00:49:30: Connor: We should let Conor know about this stuff.

00:49:32: Connor: But the reality is, if you drew the Oracle

00:49:34: Connor: Award chart and you start at Larry and

00:49:36: Connor: Safra at the top, you go down one level and

00:49:39: Connor: then about 13 more levels and then five

00:49:41: Connor: more and at the bottom that you'll find a

00:49:43: Connor: bucket and a cabbage and me next to those

00:49:45: Connor: two.

00:49:46: Connor: That's where I sit.

00:49:48: Connor: So rest assured, any secret things that

00:49:51: Connor: have been worked on Oracle it's not that I

00:49:53: Connor: can't tell you, it's because they don't

00:49:55: Connor: tell me.

00:49:57: Kai: This is the guy who was doing TikToks,

00:49:59: Kai: YouTube and social media stuff and he's

00:50:01: Kai: working around each and every conference

00:50:03: Kai: around the world.

00:50:04: Kai: Please do not give him anything that could

00:50:06: Kai: be dangerous.

00:50:08: Connor: Exactly right.

00:50:08: Connor: But in terms of advice, my parting advice

00:50:11: Connor: for anyone listening would be enjoy this

00:50:14: Connor: exciting time in IT.

00:50:17: Connor: I've been watching Twitter just the last

00:50:18: Connor: few days.

00:50:19: Connor: You've got Sam Altman.

00:50:20: Connor: He's a CEO of an OK, now he's not a CEO now

00:50:23: Connor: he's a CEO.

00:50:24: Connor: Now he works for Microsoft.

00:50:25: Connor: Now he works for OpenAI again.

00:50:26: Connor: Now the board's gone, now the board's back.

00:50:28: Connor: It's an exciting time to be in IT.

00:50:31: Connor: I would also say because of that excitement.

00:50:34: Connor: Whenever anything exciting is going on,

00:50:35: Connor: it's easy to get swept up in it.

00:50:38: Connor: So it was the term trust, but verify.

00:50:40: Kai: This is a fine last word of the podcast.

00:50:42: Kai: I will not destroy it with a big outro, but

00:50:44: Kai: thank you, karo, for attending us again.

00:50:46: Kai: See you next year, I guess.

00:50:48: Kai: No worries, I'd love to be back.

00:50:50: Kai: Thank you, karo.

00:50:51: Kai: And also, of course, thank you to our

00:50:53: Kai: listeners who tuned in for this episode of

00:50:56: Kai: Devs on Tape.

00:50:57: Kai: See you in two weeks and we're going to

00:50:59: Kai: enjoy the rest of the conference today, I

00:51:00: Kai: guess.

00:51:01: Kai: Thank you very much.

00:51:05: Caro: Bye.