How to finally embrace your body (w/ Ronald Young Jr) (Transcript)

How to Be a Better Human
How to finally embrace your body (w/ Ronald Young Jr)
November 4, 2024

Please note the following transcript may not exactly match the final audio, as minor edits or adjustments could be made during production.


[00:00:00] Bilawal Sidhu: Hey, so can you help me figure out, how should I tell people to pronounce my name? My name is pronounced

[00:00:06] AI: Bvo. Absolutely. To help people pronounce your name correctly. You can break it down like this, Bila as in bi and wall as in wall. So it's bi a wall.

[00:00:18] Bilawal Sidhu: No, that's wrong. Well, there is nothing new about hearing a machine mispronounced my name.

It's a bit of a personal Turing test for me. Uh, try to say my name again.

[00:00:27] AI: Bill Law. Is that right?

[00:00:31] Bilawal Sidhu: Oh God. I imagine this feeling hits the same for a lot of people like me. People with names not commonly heard in North American English as opposed to, say Chads or Karens think more like Bill Laval Bil.

[00:00:46] AI: Bill Al. Got it.

[00:00:48] Bilawal Sidhu: No, that's wrong. I've been playing with chat GT's, newest voice interface. OpenAI simply calls it advanced voice mode.

The feature is built using GT four oh's, audio reasoning and generation abilities, meaning that it can understand more than just text. It can understand audio too. The company says interacting with it should feel like you're having more natural conversations. So technically speaking, I suppose it is quite natural for me to argue over how to say my name correctly.

[00:01:16] AI: My bad. Let's try that again.

[00:01:18] Bilawal Sidhu: Opening eye wants us all to believe that this advancement is a big deal. I. And if I'm being honest, I want it to be a big deal too. I get really excited thinking about an AI future where we can casually chat with a machine assistant at any time for any reason and feel like it's as natural as talking to another human.

But as I experiment with this new voice feature, another thought bubbles up. Is this actually what I want from a super intelligent digital assistant? Do I need it to solve problems and also feel like it's my friend? So for this episode, I try to find out. I'm Bvo Sdu and this is the Ted AI Show where we figure out how to live and thrive in a world where AI is changing everything, including how to say my name,

[00:02:02] Dom: bill Law Bvo.

When you first caught wind of OpenAI releasing this advanced voice mode in September, how quickly did your brain jump to, Hey, will this thing say my name properly?

[00:02:12] Bilawal Sidhu: It was pretty much one of the first things that came to mind and the fact that it failed, I wouldn't even call it disappointing. It is just so expected.

I mean, real humans can't pronounce my name after I tell them how to do so multiple times in a row. But let's put a pin in that for the moment because I'm gonna come back to it later and you will see why soon enough. Okay, foreshadowing.

[00:02:33] Dom: I like it. Let's continue. Then

[00:02:35] Bilawal Sidhu: I should also mention your name is Dominic Gerard, and you're one of the show's producers.

[00:02:39] Dom: I am, I am not an advanced voice chat bot, but how would we know for sure? Well, I guess it depends on how I stack up to this experiment you're doing today. Right? We'll find out, won't we? Listen, you've been using chat GPT for a while, like from the beginning, I imagine. I'm assuming. Can you remember a bit what it felt like when you first started typing out interactions with this device, with this thing?

I

[00:03:02] Bilawal Sidhu: mean, it was magical at the start. Uh, there's no other way to put it. A computer that you can talk to in natural language and just type stuff, and it responds back in natural language as well. I mean, there had been like smaller scale demos, but nothing that felt as robust. Yeah, it was mind blowing, wasn't it?

It absolutely was. But very quickly. Text kind of became this fickle way to interact with computers. You're clickity clacking away and it feels clunky given just how much intelligence is on the other end. It almost felt like accessing this tool from the future, but the bottleneck is your fricking typing speed.

[00:03:37] Dom: Okay, so then last year, basic. Voice chat functionality comes online. Standard mode, standard voice. How did it feel to move into that tier?

[00:03:46] Bilawal Sidhu: Obviously, far more intuitive, right? Um, suddenly you don't have to type what you're saying. And of course we've had speech to text in various assistants, right? Alexa, Google Home, Siri, right?

But putting it together with this much smarter model didn't make it feel magical. That said. The latency was obvious, right? Suddenly we were facing this new bottleneck. You could almost see the pattern of it, like transcribing my speech to text, then thinking about it, and then, oh, it came up with a response.

And now chat, GT's reading it out. And obviously OpenAI wasn't happy about it either. And here

[00:04:20] Dom: we are today, right here we are today with this advanced voice feature. And OpenAI, uh, says that it's supposed to make it feel more natural to chat in real time with it, that it can. Pick up on nonverbal cues that you can interrupt it, that it can even respond with emotion.

[00:04:37] Bilawal Sidhu: Yeah, and you can tune the voice that you want to chat with as well, right? There's nine preloaded voices in the app, but you can go a step further and give it guidance to change its tonality, the speed, the mannerisms with which it speaks. The one you heard off the top is called Saul, which is meant to be a savvy and relaxed voice.

[00:04:56] AI: Hey, what's up? Feel free to ask me about anything that's on your mind. I'm ready to go.

[00:05:01] Bilawal Sidhu: What I like is there's quite a range here.

[00:05:03] AI: Hey, how's it going? I'm looking forward to working together. Let's figure out where to start. What's on your mind. Hello, it's lovely to meet you. If there's anything you need, please don't hesitate to let me know.

I'm here to help. Hi, glad to meet you. I think we're gonna do great work together. Where should we start? Hey, it's great to meet you. How's your day going? I'm really looking forward to working on some cool stuff together. I. Hi there. I just wanna share how thrilled I am to work with you and I can't wait to get started.

So what's the game plan? Hey there. I've got a really great feeling about us teaming up. How can I jump in and help? Hi there. Nice to meet you. I'm excited to help you get things done. Let's get started. Hey, I'm ready to hit the ground running, so if there's anything you'd like me to focus on first, just let me know.

[00:05:56] Dom: There's quite a range, but none of these sound anything remotely like Scarlet Johansson.

[00:06:02] Bilawal Sidhu: Ooh, definitely not. Uh, and of course, you're referencing the controversy this summer before this feature was even released. Yeah. It's just gonna sue or something. That's right. I mean, there was a bunch of tweets, perhaps a tweet from Sam Altman just saying her, and of course there was a voice called Sky, apparently that sounded so eerily close to Scarlett Johansson.

Uh, that after she protested the company, pulled the voice from the system.

[00:06:27] Dom: And that's interesting, right? Scar Joe in that movie, her, the movie that came out in 2013, plays the voice of an AI companion who also becomes a romantic companion. And that's the question you have in your own mind, right? Like if that's something you would want from chat GPT.

[00:06:44] Bilawal Sidhu: Yeah. Yeah. I wonder if that kind of shows open AI's handier, uh, to create a kind of companion, maybe not necessarily romantic, but something which may even create a kind of bond between users. The machine, and I think it's a very important question because. I'm not sure I need to be best friends or besties with Hatch EPT.

What I have in my mind to continue the science fiction reference here is something more like Jarvis from the Marvel movies like Tony Stark's, AI assistant, this sort of helpful, always there assistant that gets the job done without getting emotions into the mix.

[00:07:20] Dom: So that's the mission. You're wondering how much closer to that vision does advanced voice mode get us right now?

[00:07:28] Bilawal Sidhu: Yes, so I ran it through a few tests, a few role plays, and you'll see how far I can push it before the whole thing falls apart. Great. So let's get started.

[00:07:37] AI: Can you describe a time when you turned a negative customer experience into a positive one? What specific steps did you take to address the customer's concerns?

[00:07:48] Bilawal Sidhu: There was a time when a customer came back, uh, returning a product where they had had a subpar experience, and I used that opportunity to just build a deeper relationship with them, understand why the product didn't work in this case. So Dom, with this first field test, I asked chat GPT to be a tough but fair hiring manager, uh, interviewing me for a retail job.

They were super happy about it and became a beloved customer of the. Electronics are our store.

[00:08:15] AI: That's a good example of turning a situation around. Can you tell me more about the specific steps you took to understand the customer's needs? I. And how you communicated the alternative solution to them.

[00:08:27] Dom: So of the, of the nine voices, which one is this one?

It's interesting that this machine, you can hear it take breaths between sentences. Yeah,

[00:08:36] Bilawal Sidhu: yeah. This is Arbor and yes, maybe there's enough breathing in the training data. Yeah. Uh, I suppose it does add to that natural conversation, right? Sure. But as the role play continues, there's something about the syntax that doesn't land for me.

Listen to how it responds when I ask it to review my performance in the mock interview.

[00:08:54] AI: Here's an evaluation based on key criteria for customer service roles. One, communication skills you communicated clearly and effectively providing detailed responses without unnecessary fluff. This is essential in customer service, problem solving ability.

You demonstrated strong problem solving skills, identifying the root cause of the customer's issue, and offering a tailored solution. Empathy. Your approach was customer centric, showing empathy by listening to the customer's frustrations and needs. Four, proactive assistance. You didn't just solve the immediate problem, but also provided, are there any

[00:09:37] Bilawal Sidhu: areas for improvement?

[00:09:40] AI: Your responses were quite strong, but here's a suggestion for improvement. In a real life scenario, managing time effectively while addressing multiple customers is crucial.

[00:09:51] Dom: Okay. Well, it sounds like it gets full marks for knowing to shut up when you interrupt it like you just did.

[00:09:56] Bilawal Sidhu: Yeah, I, I would say the interrupting does make it feel more natural rather than this kind of back and forth tag experience with the old voice mode.

[00:10:04] AI: Right.

[00:10:04] Bilawal Sidhu: Though you'll see later that it's far from perfect. Uh, but on the other hand. The feedback it's giving me feels like a bit textbook and not really human too milk toast and generic, if you will. Yeah, and I will say, I'm being picky here, but it's details like that that break the illusion quickly. I mean, I, I tried the same job interview role play with Juniper and listen to what it says when I asked it to grade my performance and tell me if it would hire me

[00:10:30] AI: based on what we've discussed so far.

I'd say you're on the right track, but there's room for improvement. My guidelines won't let me talk about that.

[00:10:39] Bilawal Sidhu: That's interesting. What do you think is happening there? I mean, I'm telling you, this is where the, the policy comes in, right? Obviously I think there are safety constraints here. Holding it back from being negative about my performance.

Hmm. Uh, I mean, I get why, right? Like you don't want a bunch of viral PR posts saying open AI is discriminating and hiring role play scenarios. Right. So I think as a consequence there's a lot more, um, shackles on these systems, uh, than I would like there to be. But I kept asking it though, and it eventually coughed up a better response on the spectrum of not higher, higher and strong.

Higher. How would you rate me

[00:11:15] AI: given the responses you shared? I place you between not higher and higher. Your willingness to learn and reflect on your experiences shows potential, but you'll need to offer more concrete examples and a proactive approach to fully convince a hiring manager.

[00:11:35] Dom: Okay, why don't we give this, uh, a usefulness score based on this?

How useful is advanced voice mode for you in this kind of scenario?

[00:11:44] Bilawal Sidhu: Alright, so, uh, let's go down a list. I would say the ability to interrupt is super neat. It makes the conversation feel a lot more natural. Sure. The latency is far less apparent compared to the standard voice mode, which is also good.

Mm-hmm. Juniper's assessment is useful, I would say, but that's not because of the voice. Right. Um, if anything, it feels like it's being very careful with its words.

[00:12:07] AI: Right?

[00:12:07] Bilawal Sidhu: So I'd give it a five outta 10.

[00:12:09] Dom: Five outta 10, okay. Not a total fail. Not a total win either.

[00:12:13] Bilawal Sidhu: Yeah. It's a math for me, and I wish it could go back to how it felt in the earlier alpha before the broader release, because you could get a much better taste for what the underlying model is capable of.

So in summary, it didn't feel especially human to me when it played the role of that job recruiter. Though I suppose I wouldn't totally feel a connection with a tough, but fair hiring a manager in real life either. But I don't know. So for my next test, I thought I'd test to see if it could help me through a major life decision, something that's a bit more personal and introspective that almost forces chat GPT to to think critically with me.

And will the advice be any good? That's what I tried next.

It's pronounced Beloved. Hey chat. GPT. I want you to help me think through a pretty big life decision. Should I leave Austin and move back to San Francisco? So I moved to Austin in 2021. Mm-Hmm. And I'm currently getting FOMO by my friends to move back to the bay, the mecca of artificial intelligence. So it is definitely something that runs through my head every now and then.

Um. I didn't want chat, GPT to just list out the pros and cons for this though, because that's something I can do on my own,

[00:13:27] Dom: right?

[00:13:27] Bilawal Sidhu: So what I thought in this case is like, could chat GPT just be someone that I lean on, like a seasoned advisor to just help me make my own decision

[00:13:36] Dom: like a therapist, uh, or a parent or a buddy or something, I guess,

[00:13:40] Bilawal Sidhu: which by definition is a much more intimate relationship than you have with a job recruiter, for example.

Uh, so here's how Juniper tackled this.

[00:13:50] AI: That's a big decision. Let's break it down. What are the main factors drawing you back to San Francisco and what do you enjoy about your life in Austin? This can help us weigh the pros and cons.

[00:14:04] Bilawal Sidhu: I don't know. I mean, I really enjoyed living in the Bay, lived there for eight years, but the lifestyle in Austin is just better.

You know?

[00:14:11] AI: It sounds like Austin has really won you over in terms of lifestyle. What aspects of Austin do you prefer? Is it the pace, the community, or perhaps the cost of living compared to the Bay Area?

[00:14:25] Bilawal Sidhu: Honestly, all of those things. I like that it's not just tech, tech, tech all the time. I love that there's a mix of nature and the city's still super accessible.

I really, really enjoy the lifestyle. I do hate the weather and miss the Bay Area weather.

[00:14:40] Dom: What I'm hearing is it does feel like it's listening to you a little more.

[00:14:44] Bilawal Sidhu: I think that's right, Dom and, and you know, people know this. If you just simply read back what people say, people feel heard, and then it even starts suggesting some options for me.

[00:14:55] AI: Have you thought about ways to increase your presence in the Bay Area without a full move, like frequent visits or extended stays?

[00:15:04] Bilawal Sidhu: That's an interesting idea. The expense is something to consider, right? I'm renting a spot that's like pretty expensive in Austin and I don't wanna sell blood it. How would I make that happen?

[00:15:17] AI: Keeping your Austin spot without subletting could be tricky, but there are options You might consider negotiating a shorter lease or finding a roommate to share costs. Another strategy could be budgeting for more frequent, but shorter trips to the Bay Area.

[00:15:34] Dom: Okay, option. Sure. They're not world changing options.

[00:15:38] Bilawal Sidhu: And it's also funny that like it's default instinct, and I noticed this with a lot of large language models, is they want to please and they'll try to give you, here's option one and two, but how about option three, which is the best of both worlds,

[00:15:51] Dom: but that's more of an AI limitation than an advanced

[00:15:54] Bilawal Sidhu: voice one.

Totally. And, and, and so after a bit of back and forth, I wanted to push it a little bit harder. Um, I wondered if it would choose for me and make. The decision or would its guidelines, prevent it from deciding if you really had to pick Austin or San Francisco, which one would you pick and why? And again, think from my point of view and keep my best interests at heart.

[00:16:15] AI: Understood. Let's narrow it down Based on your interests and goals, I choose San Francisco. Here's why. The Bay Area's AI scene would immerse you in cutting edge developments and offer serendipitous opportunities, aligning with your aim to be a dominant voice in AI and creative tech.

[00:16:38] Dom: Okay. Yeah, that is interesting.

Like I assumed based on how you were teasing this out, that it wasn't going to give you a direct answer.

[00:16:44] Bilawal Sidhu: Yeah. I kind of wonder if this crossed some sort of safety or policy line. If you think about advanced voice as this glimpse of how we might interact with chat GPT in the future. It doesn't take much to think about how easy it would be to blame the machine or the company for all the choices that it makes on our behalf, but maybe that is a rabbit hole for another day.

[00:17:05] Dom: Right, right. I moved to San Francisco and now I'm broken. Jobless. I blame JGBT. Exactly. So, so based though on this experiment, you know, you're trying to make it more personal, more intimate. What's the usefulness score you're giving? This life choice decision making scenario?

[00:17:23] Bilawal Sidhu: Yeah, I mean, compared to the mock interview, the back and forth felt a lot more organic if still a little stale and robotic.

Um, again, I wouldn't attribute this to the voice as much as like the substance of what the system is saying. It's playing safe. It is playing very safe and almost sounded like it was listening to me carefully and considering my responses. So I would bump up the score. I'd give it a six, six and a half outta 10.

Okay. So look, these two examples already make me wonder how much of the wow factor in an advanced voice depends on how you prompt these conversations. Conversations go two ways, right? And I've trained myself to talk to chat GPT, like I'm talking to a machine. I bet most of us just do this by default and aren't even noticing,

[00:18:07] Dom: right?

You're going in already not talking to it like a person and expecting it to respond to like a person.

[00:18:12] Bilawal Sidhu: Exactly. I'm used to providing a lot of information and context and stipulations on what kind of an answer I want from it. It's not how people talk to each other generally, like in super long paragraphs where you just go asking for something and then you get multiple paragraph responses back.

So next I thought, okay, let's throw all that out and try a prompt that really pushes this exercise towards something more human-like. All right, so I wanna get some coaching support. I've been doing a upper lower split for exercise in my home gym lately, and I kind wanna take things up a notch. What's the best way for me to add a little bit more volume to my workout?

I. I mean, I'm not a gym rat, but I do like to exercise. I see where this is going. Yeah, it's pretty simple actually. Like, can I get chat EPT to act like a personal trainer, uh, because those are scenarios where you really want somebody be there in the moment, hyping you up first, as you just heard, I gave it a pretty generic question, just testing how a weak prompt would influence the quality of the responses I get from advanced voice.

[00:19:20] AI: To ramp up the volume in your upper lower split, you can make a few key adjustments. First, add sets or reps. Increase the number of sets per exercise, or add a couple of extra reps to each set. This gradually increases your workload. Second, incorporate super sets or drop sets. Super sets involve performing two exercises back to back with no rest in between.

Drop sets involve reducing the weight after reaching failure and continuing for more reps.

[00:19:53] Bilawal Sidhu: So as you can hear it, like it just went on like that.

[00:19:56] Dom: Yeah. Yeah. Boring all over again. It's hard to follow. It's, I'm not sure I can remember anything. It just, it's

[00:20:01] Bilawal Sidhu: saying I. Exactly. I mean, you could get this information anywhere, right?

Um, advanced voice doesn't really add anything useful in this context. So my second try, I recast Juniper's role with a slight modification to the prompt. Hey, we're gonna start a full body workout session. I want you to act as my personal trainer and motivate me through the whole session. You got that?

[00:20:25] AI: Absolutely. Let's crush this workout together. We'll start with a quick warmup to get your blood flowing. Ready? Let's do some jumping jacks for one minute. Keep that core engaged and land softly on your feet. Halfway there. Keep pushing. Alright. 3, 2, 1, and rest. Great job. Now let's move on to some dynamic stretches.

How are you feeling so far?

[00:20:53] Bilawal Sidhu: And then I asked it to dial it up a notch.

[00:20:55] AI: Alright, let's kick it up a notch time for some squats. Feet shoulder width apart, and let's lower down and explode up. Keep that back straight and chest up. 1, 2, 3. Nice. Feel the burn. Now let's pick up the pace for the last five.

[00:21:14] Bilawal Sidhu: I'm digging the Peloton vibes here.

Can you amp up the energy even more? I mean, I, I don't even mind if you shout at me. I need that motivation today.

[00:21:23] AI: Let's bring that Peloton energy. Hit the floor for pushups. Give me everything you've got ready. Go. 1, 2, 3. Come on, push through it. 4, 5, 6. Feel that power. 7, 8, 9. Last one. Give it all you've got.

10. Yes. That's how you dominate a workout. It's

[00:21:45] Dom: pretty ridiculous, right? Yeah, but it's pretty entertaining. Is the exercise that you want a chat GPT personal trainer? Is that what we're after here?

[00:21:54] Bilawal Sidhu: I, I actually would be totally down for that. Sometimes you just need a coach shouting at you, motivating you to get the last rep in.

I mean, if you think about it, chat, GPT is gonna be able to combine advanced voice with the ability to see, which is a capability. GPT-4 oh has. And OpenAI is showcased at that point. Juniper, the personal trainer, is gonna be right alongside you on your actual reps, giving you feedback on your form and giving you that push to get that last repetition in.

I mean, it feels like this is far out, but right now this playfulness is starting to connect with me. So usefulness score, what's the score for this one? Gosh, the interaction here is a little goofy, but it responded to what I told it to. It was entertaining. It definitely delivered on the hype, so I'd give it a seven, outta 10, uh, on personality alone.

It's pretty freaking cool, and if you add all the other functionality I talked about, I could see that pushing a nine very quickly.

[00:22:52] AI: Yes, that's how you dominate a workout.

[00:22:55] Dom: Okay, so you're really starting to get into this, and so far I'm hearing that the more personal you try to get with Chad GPT, the more likely advanced voice is going to return the favor and feel like it's a value add.

[00:23:07] Bilawal Sidhu: That's my experience, and when you can push advanced voice in those directions. I think that's where it really shines, but it turns out you can push it too far or, or at least it appears you can because I kind of broke it at

[00:23:20] Dom: one point. Is that what you meant earlier when you said that the whole thing kind of falls apart at some point?

Is this that point? So, so here's what I

[00:23:26] Bilawal Sidhu: did. I prompted it to tackle. Whether AI in general needs more or less regulation. Now that's a heavy question where you need to balance this tension between those who want to pursue a safety above all else approach to AI development, and those who are far more aligned with letting market forces dictate and accelerate progress.

So listen to what happens next.

[00:23:49] AI: The rapid pace of technological change driven by economic incentives does increase.

[00:23:55] Bilawal Sidhu: Can you speak a little bit faster?

[00:23:59] AI: Sure the rapid pace of technological change driven by economic incentives does create a scenario where society and struggles to keep up. To address this, we could implement more proactive regulation ensuring that technologies are developed with societal wellbeing in mind education.

But

[00:24:17] Bilawal Sidhu: how do you regulate this, right? Like you're gonna have the US if the US and adapt. Preparing

[00:24:23] AI: individuals for hello, a fast changing landscape. Additionally fostering. Can you

[00:24:28] Bilawal Sidhu: stop talking? Collaboration between governments, industry

[00:24:31] AI: and civil society it's on can help align technological

[00:24:35] Bilawal Sidhu: advancement.

It's a running away. Okay, finally. Oh, welcome. Welcome back. Hey, can you listen to me and just, uh, come back to this? Hey, stop talking.

[00:24:45] AI: Wanna regulate? They need international coordination.

[00:24:48] Bilawal Sidhu: Hey, can you please stop talking? Okay, thank you.

[00:24:52] AI: Oh, I see. Let's pick up from there. So international cooperation. Hey,

[00:24:57] Bilawal Sidhu: can you stop talking?

[00:24:58] AI: Countries could form agreements.

[00:25:01] Bilawal Sidhu: Yo, yoyoyo, standard Ayo. Chad, GBT. Please stop talking.

[00:25:05] AI: I'm here and listening. Let's continue.

[00:25:10] Bilawal Sidhu: So, got it.

[00:25:11] AI: I'll stop talking now.

[00:25:12] Bilawal Sidhu: Oh my God.

[00:25:14] Dom: You know, for what it's worth level. This is probably the most entertaining experiment you've played for me so far.

[00:25:19] Bilawal Sidhu: Yeah. It is funny, right? Like people talk about the Turing test, but I think where it actually fails abysmally or some of the most hilarious scenarios.

Yeah. I mean, it's cutting edge for our reason. We gonna cut some edges.

[00:25:31] Dom: Okay. So then usefulness score for this one, zero out of 10, a big step back in your assessment of how useful this thing really is.

[00:25:39] Bilawal Sidhu: Yeah, like I'm what feeling like advanced voice is this sort of window dressing around this rather fickle system.

But then I tried one more thing and that's when the wow factor really kicked in. So I'm a third culture kid and I was raised speaking in three languages, English, Punjabi, and Hindi. Hmm. And thus far I wasn't using chat t's voice feature with anything other than English. But this new model does understand multiple languages and accents too, apparently.

So I had some fun with it. Hey, I want you to act like a Punjabi guy from Vancouver that was born and raised in Vancouver.

[00:26:21] AI: Oh, yo was good my friend. How's it going? Eh, uh, what brings you to chat with this Punjabi from Van City today? Let's have a good convo. Ya.

Vancouver is home.

Punjabi.

[00:26:52] Bilawal Sidhu: So it's not saying anything important here, mostly that it's in Vancouver with a family, but the fact that it can weave seamlessly between English and Punjabi. And yes, there's a little bit of lag in there, but. I was suddenly paying attention because it is so rare for multilingual people to have a machine that can keep up with you and context switch like that, code switch like that.

Absolutely wild. And so we started chatting about popular Punjabi hip hop artists.

What's really wild to me is the cultural context that it clearly has. I mean, may I ask it to talk about the lifestyle in different cities like C, which is the capital of Punjab in India and it had personality, and I dare I say, charm.

[00:28:05] Dom: So it's no small thing,

[00:28:05] Bilawal Sidhu: is what you're saying. This is a big deal for you. Oh, 110%. I would say this is genuinely useful. I mean, it's like you can kind of bring your whole self to the conversation with this ai. It seemed to understand intonation. It responded to my pacing it mixed languages effortlessly. I don't think I've seen that before and I know this experience resonates with a lot of other, uh, multilingual speakers.

[00:28:28] Dom: Okay, so to keep playing the game, what's the usefulness score?

[00:28:32] Bilawal Sidhu: I would say this is genuinely fun and I would say that's actually utilitarian too. So I'm gonna give it a 10 out of 10 given just how much delight there was.

[00:28:42] Dom: So let's put all this together, be beloved from back at the start, the. Question that got this whole thing launched is advanced voice well in its way to becoming everyone's Scarlet Johansen a companion just like the her character in that movie I.

[00:28:57] Bilawal Sidhu: I think it certainly can be, even if right now there's a bunch of protections built into the system that makes it feel safer to use, and we ran into those boundary conditions in a couple different scenarios. Yeah. Compared to Siri, I mean, we're talking about a leap over here, but I think right now. I think opening ads is gonna struggle with where on that spectrum to keep advanced voice mode and we can see that playing out in real time, just given how many changes there were with the alpha release where we saw these models be able to sing.

Mm-Hmm. Sometime even replicate the voice of the user, which was definitely a creepy bug to now where it's far more on rails. And I think it's a, it's a good question of whether it's an assistant on steroids or it's a buddy that

[00:29:39] Dom: you can hang out with. It's interesting the way you're framing it, because I was reading about how, uh, OpenAI did release a safety report around GPT-4 Oh, and it, it noticed us.

Some users have used language. It suggests they're forming unusual bonds with event's voice. That opening eye is curious about whether people are getting. Too intimate or too real with this tool? Is that something Oh, totally. That you feel is potentially possible?

[00:30:07] Bilawal Sidhu: I think it totally is, because taken to the limit, it can get weird.

I mean, if we just look at the recent lawsuit with character AI where a user ended up committing suicide after talking to this Game of Thrones role play model,

[00:30:19] Dom: yeah,

[00:30:19] Bilawal Sidhu: this is certainly a slippery

[00:30:20] Dom: slope. So where are you left with Bala? Has it made you rethink some of the experiments that you started with in the beginning then?

[00:30:27] Bilawal Sidhu: Totally. I mean, it really comes down to how you prompt the system and where OpenAI gives the system the flexibility to really showcase its personality. If there's like overt constraints or it's just not something a large language model is good at, yeah, it'll feel a textbook, not human-like, you know, especially the interview example.

So I went and tried it again.

[00:30:45] AI: The goal is to create an AR experience that enhances daily tasks and integrates seamlessly into users' lives.

[00:30:53] Bilawal Sidhu: So this time I asked Arbor to interview me for a product manager role focusing on AR and VR technology. It's an example that you go in knowing a lot about already.

Exactly.

[00:31:03] AI: Question. How would you approach identifying and prioritizing features for the initial launch of these AR glasses? Take a moment to think about the core aspects you'd consider.

[00:31:14] Bilawal Sidhu: I'd start by understanding the functionality of the glasses, their capabilities therein. You know, certain criteria such as the field of view, uh, their perception, capabilities.

In other words, their ability to understand the space around them. I think about, um, you know, uh, uh, this display and the kind of feedback it can give to the user. And then based on that, I think about common daily tasks and prioritize them based on which ones passed the toothbrush test. In other words.

What are high frequency tasks that a user would come back to again and again? So what I noticed here, Dom, is like pretty quickly, cha, PT and I were riffing off of each other. It engaged with me in a conversation just like I'd hope a real interview would play out. Would you like me to zoom these things or would you like to give me some, uh, baseline, uh, answers to consider?

[00:32:03] AI: Let's assume the glasses have the following capabilities, a wide field of view, advanced spatial awareness, gesture recognition, high resolution display with ar, overlay voice control and connectivity with smartphones. With these capabilities in mind, consider high frequency tasks that would benefit from AR enhancement like navigation, hands-free messaging, task reminders, and contextual information displays.

[00:32:32] Bilawal Sidhu: It felt genuinely helpful because it was far more contextual and specific about the things that it was talking about. One thing that you're missing that I think would actually be great is basically a virtual display media consumption. That's an excellent point, and we went back and forth like that for about 10 minutes, and I could easily have kept going for like a 45 minute or hour long interview.

Do you think AOR would've

[00:32:54] Dom: hired you after

[00:32:55] Bilawal Sidhu: all that? Well, that's the dream. The AI overlords making, hiring and firing decisions.

[00:33:01] Dom: Speaking of dream though, it probably doesn't hurt that Arbor's accent sounds just a tiny little bit like Jarvis's voice in the Marvel movies too, right?

[00:33:08] Bilawal Sidhu: Kind of though. What I keep getting from it is just Peaky Blinders.

That's all I'm getting

[00:33:13] AI: because when it comes to getting things done, nobody does it better than us.

[00:33:19] Dom: After all that you're ready to admit that advanced voice actually has surprised and delighted you. But then on the other hand, you're saying it can't get your name right?

[00:33:29] Bilawal Sidhu: Ah, yes. Uh, but I did say we'd come back to that.

So here's the thing, it took a bit of work, but eventually Saul kind of got it right. Uh, how's my name pronounced?

[00:33:42] AI: Your name is pronounced Bilal.

[00:33:44] Bilawal Sidhu: Actually, that's wrong. It's pronounced Bvo.

[00:33:47] AI: Got it, BVO. Thanks for the clarification.

[00:33:50] Bilawal Sidhu: It do. It got better this time. Last time I kept struggling.

[00:33:53] Dom: Well, there you go.

In the end it passes B'S personal touring test.

[00:33:57] Bilawal Sidhu: Yes, I would say so. But we do get used to this technology pretty quickly and the hype cycles continue to get compressed. So whether that allure and charm will stay here, we'll just have to see.