Commercial Real Estate Podcast

Bronwyn Smith on AI-First Real Estate Tech: Compliance, Cost, and Human Oversight

41 min Listen to the original Episode page
Artificial IntelligenceReal Estate TechnologyComplianceVenture CapitalCanadian MarketSoftware DevelopmentRisk Management

Summary

Bronwyn Smith, founder of MyAbode, discusses building an AI-native platform for residential real estate brokerages to address compliance and operational fragility. She contrasts MyAbode's rapid, AI-first development cycle with legacy competitors, noting that while AI compresses coding timelines, regulatory nuance remains the primary bottleneck. Smith emphasizes that humans must retain final liability for legal decisions, positioning AI as an augmentation tool rather than a replacement. The episode also covers cultural barriers to AI adoption in Canada, investor mindsets, and practical use cases for automation in real estate and personal productivity.

Key takeaways

  1. 1
    Adam Powadiuk claims the performance gap between AI adopters and non-adopters will be enormous initially but standardize in five years, suggesting early adoption offers a temporary competitive advantage.

    Adam Powadiuk · 0:51

  2. 2
    Smith argues that while AI can provide legal/compliance assistance, humans must remain accountable and 'put their name on the line,' suggesting AI is a tool for augmentation rather than replacement of professional liability.

    Unidentified speaker (speaker_3) · 10:28

Episode sections

  1. 0:00 – 3:39

    Podcast Intro and Adam's Personal AI Use Cases

    Hosts introduce the episode and guest Bronwyn Smith. Adam Powadiuk discusses using AI for market surveillance and personal tasks like mediating screen time disputes, noting the gap between early adopters and laggards will widen.

  2. 3:39 – 7:32

    Bronwyn Smith's Background and MyAbode's Origin

    Smith details her career from contract surety to tech roles at Shopify, Airbnb, and Ada (merged with DialPad). She explains founding MyAbode in 2024 to solve compliance and operational fragility in residential real estate brokerages using AI.

  3. 7:32 – 10:11

    AI-First Development vs. Legacy Infrastructure

    Smith contrasts MyAbode's AI-first rebuild with legacy on-prem competitors, highlighting how AI reduced development time from months to weeks. She identifies compliance and deal nuance, rather than technology, as the primary bottleneck.

  4. 10:11 – 11:25

    Compliance, Liability, and Human Oversight

    Smith argues that while AI can assist, humans must retain final judgment and liability for legal/compliance decisions. She emphasizes that agents will not disappear but will become 'super-powered' by AI tools that ensure compliance.

  5. 11:25 – 13:51

    Market Saturation and the '80% vs 100%' Product Problem

    Smith notes that while many companies claim to use AI, few have deeply integrated it due to legacy infrastructure. She warns that the low barrier to entry creates 'AI slop' (80% products), making the 100% reliable product the key differentiator.

  6. 13:51 – 15:50

    Cost Constraints and Model Confidence Issues

    Smith identifies high token costs for vast documents (e.g., pre-construction deals) and AI overconfidence as current limitations. She describes using competing models to detect when the AI 'doesn't know' to prevent dangerous hallucinations.

  7. 15:50 – 18:35

    Data Security and Employee AI Usage

    Smith compares AI risk to self-driving cars, arguing that providing approved, secure internal tools is safer than employees using free public accounts. She highlights the risk of data leakage and the need for corporate governance on AI usage.

  8. 18:35 – 27:19

    Skill Erosion vs. Creative Enhancement

    The hosts and Smith discuss the risk of skill atrophy (e.g., writing emails) versus the creative boost AI provides. Smith advocates for teaching proper AI usage rather than banning it, comparing it to the adoption of calculators and GPS.

  9. 27:19 – 28:15

    Regulatory and Cultural Barriers to AI Adoption

    Speaker_3 states that regulation and rules are trying to catch up with AI, adding time to processes, while Canadian culture is more fearful of AI adoption compared to other countries. Adam Powadiuk notes that Canada ranked last in a chart of 40 countries regarding population excitement about AI.

  10. 28:15 – 29:54

    Canadian vs. US Investor Mindsets

    Speaker_3 contrasts Canadian investors, who ask about downside protection and avoiding losses, with US investors who focus on rapid growth to become billion-dollar companies. She notes her company is fully Canadian-backed and contrasts this with her previous experience working with US financing.

  11. 29:54 – 32:31

    AI-First Infrastructure and Future Capabilities

    Speaker_3 predicts that most tech will be rewritten as AI-first, including Outlook. She forecasts that within a year, standard lease reviews will take hours instead of days, and AI will be able to vet ideas and self-build features for brokerage partners.

  12. 32:31 – 35:11

    Personal AI Use Cases and Automation

    Aaron Cameron uses 'Muse' for email summarization, nightly news scrubs, and scheduling. Adam Powadiuk uses AI agents for property, company, and bond deal monitoring, and for real-time auction price monitoring to identify wine deals.

  13. 35:11 – 41:14

    Societal Impact and Effective Altruism

    The hosts discuss the 'brain drain' from AI, with Aaron Cameron questioning if skills like essay writing are still valuable. They touch on 'effective altruism,' a concept originating at Oxford, and note that some major AI owners are effective altruists concerned about AI's potential to decouple from human control.

Claims worth verifying

Adam Powadiuk believes the gap between AI adopters and non-adopters will be enormous in the early days but standardize in five years.

Adam Powadiuk ·0:51

How to verify: No primary source; this is a subjective forecast/opinion expressed by the speaker. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

MyAbode was founded in 2024.

Unidentified speaker (speaker_3) ·7:08

How to verify: Verify via MyAbode's corporate website, LinkedIn company page, or Canadian corporate registry (e.g., Corporations Canada) for incorporation date. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Ada merged with DialPad.

Unidentified speaker (speaker_3) ·5:15

How to verify: Verify via press releases from DialPad or Ada, or SEC filings if either is publicly traded, or major tech news archives. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Primary source

MyAbode will close 33,000 deals in the current year.

Unidentified speaker (UNKNOWN) ·9:17

How to verify: Verify via MyAbode's annual reports, press releases, or investor presentations detailing transaction volume. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

MyAbode uses models from Anthropic and OpenAI.

Unidentified speaker (speaker_3) ·24:21

How to verify: Verify via MyAbode's technical documentation, API usage disclosures, or patent filings listing model providers. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Primary source

Adam Powadiuk uses CoStars for market information.

Adam Powadiuk ·3:11

How to verify: Verify via CoStars' marketing materials or Adam Powadiuk's public profiles/social media endorsements. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Adam Powadiuk uses Plaud for AI note-taking.

Adam Powadiuk ·1:53

How to verify: Verify via Plaud's website, app store listings, or Adam Powadiuk's public profiles/social media endorsements. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Wilson Lee is a managing partner at Round 13.

Unidentified speaker (speaker_3) ·5:35

How to verify: Verify via Round 13's website, LinkedIn profiles, or press releases announcing his role. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

MyAbode's token costs have decreased by 50% from version 1 to current.

Unidentified speaker (speaker_3) ·26:39

How to verify: Verify via MyAbode's internal financial records or engineering logs, which are private and not publicly verifiable without access. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Bronwyn Smith worked at Shopify and Airbnb.

Unidentified speaker (speaker_3) ·4:55

How to verify: Verify via Bronwyn Smith's LinkedIn profile or professional biography. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Regulation and rules are trying to catch up with AI, adding time to processes.

Unidentified speaker (speaker_3) ·27:19

How to verify: Primary source: Podcast transcript. Secondary: Review regulatory filings or press releases from Canadian financial regulators (e.g., OSC, CIRO) regarding AI guidelines in real estate or finance during 2026-11-20. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Primary source

Canadian investors ask about downside protection, while US investors focus on speed to a billion-dollar company.

Unidentified speaker (speaker_3) ·28:34

How to verify: Primary source: Podcast transcript. Secondary: Compare pitch deck questions or investor Q&A transcripts from Canadian vs. US VC firms in 2026. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Most tech, including Outlook, will be rewritten as AI-first over time.

Unidentified speaker (speaker_3) ·29:54

How to verify: Primary source: Podcast transcript. Secondary: Monitor Microsoft's product roadmap and release notes for Outlook for AI-native features in 2026-2026. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Within a year, AI will review standard release leases in a couple of hours.

Unidentified speaker (speaker_3) ·30:23

How to verify: Primary source: Podcast transcript. Secondary: Verify myAbode's product capabilities and case studies for lease review speed by November 2026. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

A US brokerage automatically created training videos when their platform changed.

Unidentified speaker (speaker_3) ·31:22

How to verify: Primary source: Podcast transcript. Secondary: Identify the specific US brokerage mentioned and search for their press releases or case studies on automated change management training. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Aaron Cameron uses Muse for email summarization and news scrubs.

Adam Powadiuk ·33:04

How to verify: Primary source: Podcast transcript. Secondary: Check Aaron Cameron's public social media profiles or professional bio for mentions of using Muse. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Adam Powadiuk uses AI for real-time auction price monitoring for wine.

Adam Powadiuk ·34:17

How to verify: Primary source: Podcast transcript. Secondary: Check Adam Powadiuk's public social media profiles or professional bio for mentions of AI tools used in wine auctions. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Effective altruism originated at Oxford.

Adam Powadiuk ·37:07

How to verify: Primary source: Podcast transcript. Secondary: Verify the history of the Effective Altruism movement, which is widely documented as originating from Oxford University philosophers. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Major owners of AI are effective altruists concerned about AI decoupling from human control.

Aaron Cameron ·38:25

How to verify: Primary source: Podcast transcript. Secondary: Review public statements, funding disclosures, or organizational affiliations of major AI company founders/owners regarding effective altruism. For recent events, verify 2026-08-20 through 2026-11-20. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify MyAbode's 2024 founding date and current funding status via Crunchbase or PitchBook.
  • Check SOC2 compliance certifications for MyAbode and competitors like DotLoop or AppFolio.
  • Review the 33,000 annual deal volume claim against MyAbode's public case studies or press releases.
  • Investigate the 'chart of 40 countries' referenced by Speaker_3 regarding AI excitement levels, likely from the Oxford Insights AI Readiness Index or similar.
  • Assess the regulatory landscape for AI in Canadian real estate brokerage firms through provincial real estate council guidelines.
  • Compare the cost structure of token usage for large documents against traditional SaaS pricing models for real estate tech.
  • Look for independent reviews of 'Muse' or similar AI summarization tools for commercial real estate news aggregation.
  • Check for any disclosed conflicts of interest between the hosts and MyAbode or other AI tool providers mentioned.
Full transcriptexpand

0:00Unidentified speaker (speaker_0): Welcome to the CRE podcast, 100% Canadian, 100% Commercial Real Estate. Now here are your hosts, Aaron Cameron and Adam Powatik.

0:11Adam Powadiuk: Welcome to the Commercial Real Estate Podcast powered by First National. I am Adam Powtak sitting in office with Aaron Cameron. We have a topic today that I'm uh pretty excited about at a personal level, a professional level. We're gonna get in deep on implementing a AI with somebody who's at the forefront of the industry. It's Bronwyn Smith, CEO, my abode. She's gonna share with us the highs and lows of trying to revolutionize the way that uh we do business. Uh welcome Bronwyn.

0:40Unidentified speaker (speaker_0): Hi, thanks.

0:41Aaron Cameron: Before we go there, before we get to Bronwyn, what makes you such an AI enthusiast, Adam? Like what is it about it that just gets you so excited?

0:51Adam Powadiuk: Well, it's so if if anything, but it could be almost crass about why I'm enjoying it currently. It's human nature is what it is. Uh or industry or big laggards. I think there's a a period here where the gap between those spending time on it versus those who don't will be enormous versus five years from now when it will just be standard Your practice. And so I think the advantages in the early days will be more pronounced because the gap will be that much larger.

1:15Unidentified speaker (UNKNOWN): And so

1:16Adam Powadiuk: you're a superhuman player in industry. And yes, everybody's using it extensively, but imagine theoretically a universe where you were the only person that had it, how unbelievably successful you would be.

1:26Unidentified speaker (UNKNOWN): There's

1:26Adam Powadiuk: in a smaller scale that does exist because there's many people that are not using it or using it properly. Before we hit record, Braun and I were joking about you know, people using Chat GPT to check movie times tonight to go to a movie. which is not the best use of technology. And I love the potential for where it can go. You see something as unfinished as it is now and still so useful. Imagine the polished version that's just around the corner.

1:51Aaron Cameron: Sitting in front of Adam is an AI note taker.

1:53Adam Powadiuk: Yes, Plaud. I highly recommend it. P L A U D. It's very cool. It's the memory I always wanted, but I don't have.

2:00Aaron Cameron: What's the coolest use of AI you've done?

2:04Adam Powadiuk: I can tell you anecdotally a short one. My daughter and I were arguing about her screen time. So I said, Well, you need to convince me and AI is gonna be the judge. And so we both we both pitched our cases to the AI. We had a minute to rebut each other's arguments at the end, and then the AI made its verdict on who was more compelling in their argument. That's like the most kind of isolated.

2:25Unidentified speaker (UNKNOWN): Who won?

2:26Aaron Cameron: Yeah.

2:26Unidentified speaker (UNKNOWN): So

2:27Adam Powadiuk: I did, but actually interestingly is said the only reason I won was during my rebuttal. I had a stronger rebuttal. This she had very strong opening arguments. Mine were a little weaker, but I had a very good rebuttal on the other side. So I did win that one. And I'm not a total, I know, animal. I did give her some screen time for to participating in this exercise with me. I'm not heartless, of course. But then I guess on a more professional level. it's just having complete market surveillance. You're never not up to speed on absolutely anything that you can think of, anything that needs constant detail and attention, specified entirely to you. You know, like a program like CoStars, great. You know, you can get market information at your fingertips. But if I want something that is just personalized entirely to me, that's where I'm my chat GBT is. So right now that's how I'm liking it. I know there's other areas where it's gonna be strong like underwriting, things like that. I'm not so focused on that because that's not my job. But for me it's just having complete market knowledge. If I had to spend four hours every morning, you know, doing this work, I would only be half as up to speed as I am now.

3:31Unidentified speaker (UNKNOWN): So

3:31Aaron Cameron: Bronwyn, let's get to you. We're gonna go background. We're gonna go background first, but I'm actually curious, let's just stay on this topic. What's the most interesting use of AI that you use

3:39Unidentified speaker (speaker_3): For me, I'm obviously very pro-AI. And it's never been more fun to be a builder. Like for us, if you're building technology and software, it's you can just imagine whatever you want. And within a day, you have a working prototype. And um, like design is a whole new thing. It used to be it would take months and you have to like spend all this time in advance on this process. perfect design. And now you're like designing as you go. And so the ability to just take something into imagination and get a sort of working prototype version in days, weeks, hours is just super fun.

4:16Aaron Cameron: I built

4:17Unidentified speaker (UNKNOWN): an

4:17Adam Powadiuk: app

4:17Aaron Cameron: to

4:18Adam Powadiuk: manage my kids' chores at home, specialized to them with all the inputs I wanted. It was

4:23Unidentified speaker (UNKNOWN): a good idea.

4:23Unidentified speaker (speaker_3): Exactly

4:24Unidentified speaker (UNKNOWN): how

4:24Unidentified speaker (speaker_3): you want.

4:27Aaron Cameron: So let's go backwards. Let's start with your career. You assume you graduated elementary school wanting a career in AI. Yeah.

4:35Unidentified speaker (speaker_3): Exactly that. No. Actually, my first job was I was in contract surety, so financially backing construction companies out in BC. I went back, did my MBA. I was a strategy consultant at Bain for a number of years. And then I spent the last 10 years or so in tech. You would know Shopify. I did Airbnb for a bit. So kind of that tech for entrepreneurs. And then I worked at a Canadian company called Ada, conversational chatbot company. And so I was in AI before AI became sort of the thing. And so I got early exposure to it. Ada just actually recently announced they merged with DialPad. so success story for Canada there. And I was in tech, and then I got tapped on a shoulder from a mentor of mine, Wilson Lee. He we had worked previously together at an early startup. He is now a managing partner at Round 13, which is a venture capital fund in Toronto that does a

5:35lot in Canadian investing. And they had been investing. in real estate brokerages, residential real estate brokerages, and thought there was an opportunity to run these brokerages better with technology and just asked my opinion on it. So I started to hang around and was just kind of blown away with the problem. Like you have for most Canadians, the biggest investment you'll ever make in Toronto, it's an average of a million bucks. I know it's gone down, but it's still a million bucks people are spending. And when you look how it's often managed, it's just so fragile on the back. And compare that to, you know, at Ada, we had the chat bot to help like Peloton sell a t-shirt in it. That's so buttoned up and SOC2 compliance and all of this stuff. And and just real estate just was not at the same place. And so that was super interesting. intriguing. And then also with where AI was going and where it it's still to go, it was a problem we could now solve. So previously because it's so regulated and every area has different rules and every case is somewhat a unique case, that's a really hard problem to solve with technology, but with AI becomes a lot more solvable. So I got intrigued and then we ended up starting MyAbo, which we build tech primarily for the back office for

6:54residential real estate brokerages. We own, operate, and invest in brokerages too. So we get to try firsthand what we build and see what works and what doesn't. And uh we've been going ever since.

7:06Adam Powadiuk: What year was My Abode uh founded?

7:08Unidentified speaker (speaker_3): Uh 2024.

7:09Adam Powadiuk: Okay. So it was fully in the in the AI era. In

7:12Unidentified speaker (UNKNOWN): the

7:12Unidentified speaker (speaker_3): AI, era. yeah.

7:13Adam Powadiuk: That was the original vision, mission statement. is to take this as uh as far as you can.

7:18Unidentified speaker (UNKNOWN): And

7:18Aaron Cameron: since you started in 2024, I mean the technology has made huge advancements. So maybe just talk through that first vision

7:27Unidentified speaker (UNKNOWN): of

7:27Aaron Cameron: what you thought here's how we can solve this problem to where it is today in that delta.

7:32Unidentified speaker (speaker_3): Yeah. So when I first joined, they had made some initial investment before I got there. They had used a sort of third. party to build, and it was not an AI sort of centric building, which made sense. Like nothing had been to date. And the way I would say tech was built previously or how they tackled that problem before was okay, I'm gonna follow a deal processor and I'm gonna just think about ways to make that person's life a little bit easier. And I think we sort of rebuilt everything and brought in a whole new team and a different sort of style of building. And you're able to just rethink in an entirely different way the problem. So now it's okay, how do I close a deal with the least work for an agent, the most compliant and most cost effective possible? And that's just a different way you solve. And so that the just the entire problems we solved and how we solved it changed entirely once we were able to solve. switch to AI. But I think what was almost more exciting was just

8:32there is a lot of upfront work in terms of making sure your data is safe and accessible and all of that. But then once you're sort of set up, we call it shipping product in technology, how fast you can produce product and get it out. It's just what would have taken six months is a month, what would have taken three months is a week. Like it's just unfortunate. unbelievably how fast you can get things out now.

8:53Unidentified speaker (UNKNOWN): So

8:53Adam Powadiuk: the tech's not the choke point in this, not the bottom line.

8:56Unidentified speaker (speaker_3): No tech's gonna become free at some point.

8:58Adam Powadiuk: It's it's when it's gotta be uh compliance and I'm sure uh lawyers weighing in on liability and more traditional ways of doing business, probably slowing down the process more than uh chip's ability to uh enhance a the AI experience.

9:13Unidentified speaker (speaker_3): And I think that was part of the. beauty of us having

9:17Unidentified speaker (UNKNOWN): like

9:17Unidentified speaker (speaker_3): we'll close 33,000 deals this year. So we're like a lot of deals are going through our systems. And learning for me is there's no standard deal. Like every deal has its own nuance. The amount of handwritten deals is like still pretty significant, right? And so being able to actually test and learn and have people in the process and compliance experts and really train. your system and your data as you go was a pretty critical part of this. And it's also there's a lot of nuance. I think that was another learning. Rico has rules. You think it's black and white, but actually every deal processor looks at that role and applies it in a subtly different way. And so really understanding what are we solving for here? What actually do we need to do here was, you know, there was some learning in that. So we created a very robust engine and tweaking that took a bit of time. But so yeah, the compliance took a little extra. But yeah, being able to build is not the bottleneck.

10:11Adam Powadiuk: Do you see the liability side and the legal side we'll call it becoming institutionalized, where it's it's we readily accepted where you could have say legal advice offered on a closing provided by AI? I mean, I know today you probably cannot. Do you see that framework? getting easier?

10:28Unidentified speaker (speaker_3): So I think that it's critical that humans are still taking the judgment. And if I hire a lawyer, you better have reviewed this and you better be putting your name on the line and you're still accountable. That said, I probably wouldn't hire a lawyer not willing to use an AI tool to make them better. Like I think you have the world's knowledge at your fingertip to make you powered in a superhuman way. Like why would a smart person not want to be doing that? And so that's how I look at it in this space. I don't think agents go away. I don't think that agents can't, I mean realtors can't be really looking through things at the very end and doing the right things for their customers, but why wouldn't you be super powered and have the right tools checking and making sure you're in compliance along the way.

11:15Unidentified speaker (UNKNOWN): When

11:15Adam Powadiuk: you look at competitors, call it like legacy way of selling these assets, how far is it penetrated into their process versus where you are?

11:25Unidentified speaker (speaker_3): So not far that what we're seeing. I think every company at this point has puts AI in and their strategy statements somewhere. Like I'm I'm certain if we looked at all the public. board materials that go out, AI must be pretty prominently featured in almost every single thing. But in terms of actually deeply using it, um, not as much because what we had the ability to do, which was an advantage, is build from the ground up entirely AI first. Uh it's harder if you have old infrastructure. Some of our competitors are still like on-prem software. Like it's just a different thing to build. So we could move a little bit faster because of that. That said, they all have their own pitch of it. I think what's been harder, less like competing with legacy software. I think most people see like where the world's going. They understand AI has power. What's been interesting is like you said, you made an app in 30 minutes. The amount of people who can get a sort of half-baked idea out and pitch it is enormous. So every day there's some new person in their garage who has something they're pitching. It's very, very, very easy to get an 80% product out. It's very, very hard to get a hundred percent product out. And that

12:45helping distinguish yourself that you are that 100% a hundred percent product, not the 80% eighty percent product, I think. has created a lot of noise and it's kinda like the AI slop. I'm sure there's a bunch of AI podcasts out there, but they're not necessarily good along the way. But using AI to help you make your podcast better probably works incredibly well.

13:03Aaron Cameron: We have had this conversation before about using blockchain on registries and how much easier that would make 10 years ago we had the conversation. A lot of these a lot of these transactions much more secure, safe, and less onerous. Does that come across your desk?

13:17Unidentified speaker (speaker_3): I think I right before here I was looking at X and formerly Twitter and there's been hundred companies impacted with some sort of fraud, something or the other, this week. So it's definitely top of mind. I think it's so new at this point that there's a gazillion ways the world could go. And I think we're all trying to muddle model our way to the best, safest way. Right now, it's using the top models has advantages, being able to use the speed that comes with it has definitely advantages. And so I think that's where most tech companies are building right now. But

13:51Adam Powadiuk: what's the biggest hindrance you've had or the success you thought you would see that wasn't realized? Where does the tech fail in its current iteration?

14:00Unidentified speaker (UNKNOWN): So

14:01Unidentified speaker (speaker_3): fail is a strong word.

14:03Adam Powadiuk: Hasn't succeeded yet. Is that uh better phrasing?

14:06Unidentified speaker (speaker_3): So two answers to that. There's definitely places today where the cost to use AI is still so high that I think it will get there in time, but it just doesn't make sense. And so certain pre-construction deals or even commercial deals are so vast that the cost of going through that is just like enormous. And so

14:25Adam Powadiuk: having a human go through it is. cheaper than having tokens to go through it.

14:29Unidentified speaker (UNKNOWN): Interesting.

14:29Adam Powadiuk: Okay.

14:29Unidentified speaker (speaker_3): And so there are still moments where it is cost prohibitive or or the cost has been surprisingly higher. Not a lot. And that will get better with time. I think one of the things that was really hard at the beginning and we had to get really tight on is knowing when the AI doesn't know. Because that's the scariest thing, is if it's overconfident in itself and it's saying, yes, I have 99% confidence level and then it's wrong, like that's really scary. So we built a lot of competing models testing each other and trying to call out when like one was saying it's confident, another saying it wasn't, so we could try to measure and monitor that. But that piece of like, I don't need it to do everything. That's okay. But I need to know when it doesn't know how to do it so I can bring in the appropriate judgment and human. along the way. And where it does does not instill confidence is if it's saying it has a high degree of precision and then it's wrong. So that took a little bit of time.

15:26Adam Powadiuk: Like the outright hallucinations that I

15:29Unidentified speaker (UNKNOWN): think

15:29Adam Powadiuk: were.

15:30Unidentified speaker (UNKNOWN): But if it says, you

15:30Unidentified speaker (speaker_3): know what, I'm likely to hallucinate with this answer, then you're like, okay, that's fine. I know why.

15:36Unidentified speaker (UNKNOWN): But if

15:36Unidentified speaker (speaker_3): it's like I'm ninety 99% nine percent sure this answer is right, that's like a little scarier.

15:41Unidentified speaker (UNKNOWN): I'm

15:41Adam Powadiuk: gonna try that my next human to human interaction. Go. I might be hallucinating, but I think that uh I'm accurate.

15:48Aaron Cameron: Or it could be the gummies I just took.

15:50Unidentified speaker (speaker_3): Why are you more AI nervous? I get this.

15:55Aaron Cameron: Particularly from the lending perspective. When you're using AI and building out these tools,

16:02Unidentified speaker (UNKNOWN): who

16:02Aaron Cameron: owns

16:02Unidentified speaker (UNKNOWN): the other thing?

16:03Aaron Cameron: who owns the the servers, who owns the the the large language models, where is that coming from? What access points do they have? Right? Like there's just

16:13Adam Powadiuk: compliance though, essentially.

16:14Aaron Cameron: Yeah. But you know, like we're so I can tell you, like, you know, we're not allowed to use certain AI tools right now simply because we're just not confident that it's got the right protection mechanisms to it, right? And that there isn't ways to get in there. I can't remember the the name of the company, but there was a company out there that had used one of the AI tools to build some sort of tool for them. And it was used it was outward facing and one of the regulators or one of the sort of the compliance people came in and said, Okay, let's watch this. And they went and asked that same AI agent, how can I hack this thing that you just helped me build? And it was like here, here's how you hack

16:48Unidentified speaker (UNKNOWN): this thing. Yeah. yeah. Right.

16:49Aaron Cameron: Because it it it knew how where all the loopholes were of the thing that it had just built, right? So that, you know, there's a lot of risk

16:55Unidentified speaker (UNKNOWN): there.

16:55Unidentified speaker (speaker_3): I and the way I think about it too, and how I've talked about this is it's the same problem the self-driving cars have, where it's like, could they have an accident? Yes. But the question is, is it still safer than people driving on themselves? And that's what you need to measure it at. And I was talking about this last week, but the reality is the vast majority of realtors. or folks are using AI in some way, shape, or form. Is you providing a tool that is leveraging the proper models with the proper authority and not training on your data and all of this type of stuff significantly safer than them going on some random open for each Hat GBT account and throwing in stuff, like without a doubt along the way. And so of course we have to keep building and making it safe. And it I think it is, but I also think it's important that companies know like their employees are using this stuff. They're trying things. And you should give them safe tools that you approve to use in AI because I think a lot of them have they are doing things on the side.

17:52Unidentified speaker (UNKNOWN): 100% they

17:52Aaron Cameron: are.

17:53Unidentified speaker (UNKNOWN): Yeah,

17:53Adam Powadiuk: I read a stat recently from some poll that 60% of employees had done something with AI that kind of knew to be offside from the regulatory bodies of their industries. That's a very high number.

18:04Unidentified speaker (speaker_3): Well, even like they're checking how to write an email because they they want help with it, right? But like there's probably things on that email you don't want just sitting in a free OpenAI account, right? And so you need to provide the t these tools to your employees.

18:16Unidentified speaker (UNKNOWN): I

18:16Aaron Cameron: didn't think it impacts like the brain drain, a bad term, but you know what I mean. Like I'm already finding myself, it is much easier to go, hey, I want to write an email to so and so. And so I wanted to say these five things and off it pops. versus there's an art and it takes some intellect and some thoughtfulness to draft a really good email. That skill set will effectively be gone.

18:35Unidentified speaker (speaker_3): Yeah, so it's so funny because I'm obviously this is my world and I'm so excited where it's going. This is probably the one area that gives me so much cringe is the AI written emails, and it's so obvious.

18:49Unidentified speaker (UNKNOWN): So obviously, obvious. done poorly.

18:51Unidentified speaker (speaker_3): And for me, like I really do try to write everything myself. And could it edit? Could it give me suggestions? Do I argue with it to make my point stronger or where it's weak? Totally. Really helpful for that. But I really do try to get people to write their own thoughts because like it cannot replace your thinking. It's here to like enhance or to argue, to debate. I think that's a great use case. But yeah, like the LinkedIn. like just slop coming out is

19:17Aaron Cameron: well when I think about like kids in university writing essays or like I see it, like memos or reporting and all sorts of stuff that I see in like my day-to-day now where it they've put 10% of the thought that they should have or would have put in in the absence of AI, which fine, great. You produce something that that is probably good or even maybe even better, because AI is very smart, but we're all gonna get dumber. as a result.

19:40Unidentified speaker (speaker_3): Or we need to teach people how to use it, right? And it I think this is a tool. It's like then the calculator came out, right? Like don't fight it. But like student needs to learn, like you still have to think, you still have to have a structure, you still have to know what you're trying to say. And then this is how you enhance it. If you just dump in a bunch of garbage, it's gonna look like garbage coming out. And so, but that's an another example where I think it's gone a little better. But at first, I think that the school's approach was often like, I don't want them touching it at all. They're just doing their homework on it on the side versus like, okay, it's gonna be happening. How do I teach them to do it safely? How do I teach them to do it properly? Like it more embracing it. Yeah.

20:16Adam Powadiuk: Make them great at it. Make

20:18Unidentified speaker (UNKNOWN): them great

20:18Unidentified speaker (speaker_3): at it. Yeah.

20:19Unidentified speaker (UNKNOWN): Yeah,

20:19Adam Powadiuk: I was stunned at that. I mean the analogy I always think of is I had a job in uh I guess it would have been end of high school,

20:26Unidentified speaker (UNKNOWN): driving

20:26Adam Powadiuk: a flower truck delivery. Very nice job. Everybody's very happy to see you. I used Pearlie's then. The area that I drove in Toronto where I learned everything with like a paper map. I know inside and out to this day. But then I think of other areas when I started driving around for real estate in areas this city that I spent no time in previously and I just used GPS, I didn't learn anything. Like I don't know any intersections or any of that even like

20:50Unidentified speaker (UNKNOWN): an app. You don't

20:50Aaron Cameron: have follow the purple line and

20:51Unidentified speaker (UNKNOWN): follow the

20:52Aaron Cameron: arrow as it

20:53Unidentified speaker (UNKNOWN): leads you through.

20:54Adam Powadiuk: So I do worry about that. But maybe it does it opens up new areas of your brain when you learn how to effectively pilot these things.

21:00Unidentified speaker (speaker_3): So I would say I'm like more creative than I've ever been, right? Because before you could think of something, but you couldn't take it that far, right? You can take things so much further. And so yeah, I think it can open up different areas, or like I feel like I get involved with marketing in a way I didn't used to, or design a way I did you know I wouldn't have before because it's so easy to start just get it started. And you're like, well what if it looked like this or what if I tried this or how do I test that? And I wouldn't even touch those areas before. I

21:26Aaron Cameron: think we've come full circle because Adam started the conversation which about there's there's sort of a the super users and then everyone gets left behind. And I feel like there is gonna be this sort of I don't know this there are people that like yourself that are really going gonna to lean in and take it to grow and expand. I think there are others that are gonna use it because it's the path of least resistance. I think that's where my hesitation comes from, right? They don't care. Well I needed to write an email it was a lot faster, a lot easier and send an email sent, right? Without really caring about the outcome. I don't know. I think there are some people that find it that way, that look at it that way.

21:58Unidentified speaker (UNKNOWN): But

21:58Adam Powadiuk: both of those people are rewind rewinding the clock forty 40 years and you the people that are trying to master something versus people that are just happy taking the easiest route will probably end up in the same place that they'll end

22:07Unidentified speaker (UNKNOWN): in the current

22:07Adam Powadiuk: environment.

22:08Unidentified speaker (UNKNOWN): Yeah, yeah, yeah.

22:08Aaron Cameron: That's fair.

22:10Unidentified speaker (UNKNOWN): This

22:10Adam Powadiuk: is just another iteration of kind of the attitude and approach you have to work getting you the outcome.

22:15Unidentified speaker (UNKNOWN): Yeah.

22:15Unidentified speaker (speaker_3): Not to

22:16Aaron Cameron: let

22:16Adam Powadiuk: it

22:16Aaron Cameron: and so we all collectively become efficient, more efficient as a society.

22:20Unidentified speaker (speaker_3): But every time we've got more efficient on something, you just spend time on other stuff. things.

22:23Aaron Cameron: Yeah, yeah. Think about when Excel came out

22:25Unidentified speaker (UNKNOWN): and

22:25Aaron Cameron: how much more efficient businesses became as a result, right?

22:29Unidentified speaker (speaker_3): And particularly in our industries, right? Like often I've talked with folks and I know the narrative is like, oh, there's gonna be no real estate agents or it's gonna shrink dramatically. I just don't see that's the case, right? Where at the heart of it, yes, they want you organized and they want you to be in compliance, but this is an emotional decision. You need judgment, you need someone to help you tell nuance, you need to know. neighborhood correctly. Like all of these things are not something that you're gonna ask your muse to tell you and you're gonna feel comfortable with. And so I think the people who are good at those soft skills and then embrace that they can be more organized and more powered and all of that with the tools, I think will do exceptionally well. And yes, I think it will power a bunch of people. And some people I think should embrace it along the way. But I I don't see these roles going away. I think they change a little bit, but the judgment becomes just that much more important. And well

23:20Aaron Cameron: in buying and selling real estate, it's a human thing, right? Like

23:24Unidentified speaker (speaker_3): a once in a while transaction, it's not a commodity. You're not doing it every

23:28Unidentified speaker (UNKNOWN): time. No, and it's a

23:29Aaron Cameron: touch and feel. You've got to be there, you've got to feel the space, you've got to walk the neighborhood like that. You can't replace that.

23:34Unidentified speaker (speaker_3): But wouldn't you love something to be like, Yeah, by the way, this has been like thoroughly checked and like you're in compliance when you sign your paperwork?

23:42Unidentified speaker (UNKNOWN): Like,

23:42Unidentified speaker (speaker_3): yes.

23:43Adam Powadiuk: You mentioned uh Muse a second ago. Currently, it seems like every two weeks that one of the platforms has the latest and greatest and they're the best. And then two weeks later they're eclipsed by somebody else. Who primarily powers your platform? Or is it multi-tools, or how do you approach that from a platform uh basis?

24:01Unidentified speaker (speaker_3): So we and I would say a lot of tech companies, we do build it like agnostic to the fact that these tools change all the time and we're constantly testing and figuring out the best solution. And there's even like certain things one model might be better at than others, or even like the cost might make sense to use a cheaper model than the others. And so you tend to be a little bit more agnostic. We're definitely using a lot of anthropic and OpenAI now. And whenever a new model comes up. out, we we tend to go with them. What I think it'll go is that where we won't compete is like being AI agent for real estate agents. I think everyone will have their own AI agent. And I think the big companies have the best shot at sort of playing that space. But what we will have is the compliance rules, the rule engine, the ability to know like all the data that we've done to know that if you ask your AI agent to go to our Abbey to make sure the deal is in compliance and closed properly, we can do that properly. And so that we're over investing in the areas we think we can win at. I do think that big players have a better shot at some of those like personal AI agents.

25:05Unidentified speaker (UNKNOWN): How

25:05Adam Powadiuk: do controls your token expense every month?

25:08Unidentified speaker (speaker_3): You know what? It's not so bad.

25:10Adam Powadiuk: Okay, good.

25:11Unidentified speaker (speaker_3): Yeah, yeah. Yeah,

25:12Aaron Cameron: that was one of the arguments or I've heard this that AI is great and can be efficient, but it's more expensive than the way you're running your business today.

25:19Unidentified speaker (UNKNOWN): Even

25:19Adam Powadiuk: if true, I mean that would get solved in six months.

25:21Unidentified speaker (speaker_3): I think the beauty, so with our business, we're profitable, we have to live within our means. This is actually one of the reasons I was really excited because I had come from like the venture world and it got a bit crazy there in like the boom times and the raises people were getting and the spend that was happening out. Like Silicon Valley was based on a lot of truths. The TV show, I don't know if you've seen it. And so it led to just bad behaviors. For us, we live within our means. It causes just like critical thinking and good discussions. I have like an amazing engineer and product leader, and they're constantly like, this seems like a bad ROI. Like, why would I spend money on this? And so, just like smart thinking and like, or let's do an isolated test. Oh, the spend is out of control and it's not looking promising. Let's stop doing that. Like they've been really, really good about it. So we've managed it like quite okay.

26:08Unidentified speaker (UNKNOWN): And

26:08Adam Powadiuk: how many what per I mean just roughly, what percentage of projects you're looking at would fall into that category where expense of tokens limits it rather than the tech's ability to execute?

26:19Unidentified speaker (speaker_3): Not many. And even the ones where the expense limits it, we put it in the like let's solve that later. category because it will get cheaper. There'll be cheaper models or again, you don't have to use the most expensive model for all parts of it, or you find better ways to structure the data. And so like from our first version to now, we've put so little time and optimizing costs. It's already like fifty 50% percent cheaper, right? And so if we actually put effort into it, it'll take down from there. So I don't think token costs in the medium to long term will be what stops. I think there is definitely an adoption curve. And we're talking about mainstream businesses. We're talking about people like I have conversations with people who maybe they're right now, but like they didn't even like the cloud. So I hadn't even got them to cloud-based software. And I'm trying to jump them to AI, right? So you're dealing with humans, you're dealing with the emotions, you're dealing with a lot of change. It's changing faster than

27:19a lot of people can sort of soak in themselves. And so that's taking a bit of time. And I also think regulation and rules are trying to catch up, but they're not quite sure how to do it yet. So they're throwing different things against the wall right now. So that adds a little bit of time. I think it's less, it's not the tech, it's not the token costs. cost. I think it's more the the other factors that. sort of slow it down.

27:40Unidentified speaker (UNKNOWN): How

27:40Adam Powadiuk: much does Canadian culture factor into that, given that Canadians are more fearful than other countries? How often do you just run into outside of the technology, another human is just I'm fearful of this, I don't want to engage with this, and that kind of hinders your business.

27:55Unidentified speaker (speaker_3): It's huge. It's so interesting. We were saying before uh you came in, I was looking at a chart this week and it was showing the percent. of the population who was essentially excited about AI and thought it was gonna better their lives. And it looked at like 40 countries and Canada was last on that list. And so we're definitely a little bit more timid in that regard. I would say the hurdle of excitement or are people just being a little bit more scared or pessimistic is definitely there. Also on the investing side, it's very interesting. So previously most of the companies I'd worked at had raised from US funds. In Canada, when I talk with investors, and I I have an amazing investing partner with round 13, but generally when I'm talking to different investor groups, I've never been asked more like, tell me your downside protection. How do we make sure we don't lose money here? Don't have that conversation with American investors, like at all. It's just like what's the fastest way to be a billion dollar company? There's pros and cons to both. Like you can also make dumb decisions trying to grow super fast and it doesn't make sense. But it is shockingly different culturally, the two areas. And it's been really interesting navigating the financial environment with this is my full like Canadian backed, fully Canadian company versus in before I always worked with as sort of US financing.

29:14Unidentified speaker (UNKNOWN): So

29:14Adam Powadiuk: it's it's funny when we talked about your business uh you know starting in 2024. And in terms of AI, that's a really long time ago. In terms of how many times you had to, you know, brush your teeth and go to bed, it's not that long ago. It's uh because everything's changing so rapidly. And Aaron and I do like to end off on you know, kind of forward looking, but with AI in particular, usually we say to people, hey, in five years, what but where are you gonna be

29:35Aaron Cameron: in five weeks? Yeah, where's the tech gonna be? Yeah,

29:37Adam Powadiuk: let's shorten short that timeframe. time frame, you know, right down. We have you on next year. Where do you think the company is just from a growth trajectory from uh tech just having gotten better and better, tokens having gotten cheaper, Canadians having no longer afraid of the boogeyman under the bed? What happens a year from now?

29:54Unidentified speaker (speaker_3): I think overall most tech gets rewritten sort of AI first over time. And I really hope like Outlook and all of that falls onto that category.

30:03Unidentified speaker (UNKNOWN): Finally

30:03Unidentified speaker (speaker_3): along the way. But I think most of that gets written over time. We see ourselves a world where like just a significant portion can just be AI reviewed. You really are using humans for the tricky cases, the unique cases, the extremely expensive cases or whatnot. But if you see a lease, I think in a year we'll be at a place where like a standard release lease, you could just have AI do the checks and balances and get that done within a couple hours versus taking days and weeks in time for something standard like that. So the speed of accuracy, the speed it's going, I think we can get to that place pretty quickly. Where I see some of it going, which is super fun, is as you build sort of your whole infrastructure AI first, one of the things we've been experimenting and doing is that anyone, whether it's an agent, whatever one a of brokerage, our brokerage partners, can like submit ideas in and the AI itself is able to sort of look at the idea, vet it, and then sometimes just self-build it. We, of course, will still check before it goes into shipment and go, but that idea that you create this like ecosystem or world that everyone's part of creating it together, I think is a really cool thing that is not far away. I think within a year we'll be able to be in a place like that. And that's going to be gonna for a lot of different technology. So I think that's really fun.

31:18Aaron Cameron: The user build is a very interesting concept.

31:22Unidentified speaker (speaker_3): Yeah, there's a brokerage actually. This has already happened last year that I was following in the States. They're a little further ahead of us. They had built this system where every time they made a change in their platform, which was like a self-built self-filled platform, the platform recognized it and then created training. videos and materials automatically that went out to all their users to like update them what's going on. Because the change management piece is such a hard part of that. So you see this ability where it starts to understand oh, like they've raised this bug. Oh, they want this feature. Oh, and it's able to just self determine that this is the right thing to do with the right rules and checks and balances in place.

31:59Aaron Cameron: Fascinating. Run of time. Thank you so much for coming on. I really appreciate you taking the time to come join us. Thank you to First National for Power of the Podcast. And stay tuned for the after show.

32:11Adam Powadiuk: Welcome to the commercial real estate podcast after show where Aaron and I discuss the discussion. You kind of caught me uh at the beginning when you asked me, you know, my favorite uses of AI. Do you run agents on your phone? Do you have automated process? at a personal level? I'm gonna ask you two lines of questions here. One is just personal, just running the Cameron household, your life. What are you using unfettered by regulations you'd have in your professional life about what you can implement? And the technology is this the sky's the limit. What are you doing, you know, to make your life better with AI? How deep have you gone?

32:46Unidentified speaker (UNKNOWN): Yeah,

32:46Aaron Cameron: well I've got Muse on my personal email. So it summarizes. kind of everything and and keeps me on top of everything I need to respond to and manage. It spits out a nightly commercial real estate news scrub from all sorts of different places that I've pointed it.

33:04Adam Powadiuk: Yes.

33:05Aaron Cameron: And then it will do like scheduling. Here's what you got scheduled this weekend and here's how long it takes you to get places and you're gonna need a grandma to come and help tell you. take one

33:15Unidentified speaker (UNKNOWN): of your kids someplace or whatever.

33:17Aaron Cameron: Yeah.

33:18Unidentified speaker (UNKNOWN): I yeah.

33:19Adam Powadiuk: So F SM is similar. I've got you know a bunch of different agents that scrub for various information, whether that's just property monitoring, different company monitoring, bond deal monitoring. So yeah, it is just hyper-personalized, you know, delivered exactly when you need it. Even the small differences, you know, there's groups that will email out aggregated information to you at 8 a.m. every day. But maybe your schedule as your kids means we better have a seven. You can even personalize down to that level. Just

33:43Unidentified speaker (UNKNOWN): the little things

33:43Adam Powadiuk: that

33:44Aaron Cameron: I get it. It's actually, it's my idiot'll even say it's 9:30, 9:45. Your kids are in bed. Here's the things you need to read. Yeah.

33:51Adam Powadiuk: That's

33:51Aaron Cameron: fair. And

33:52Unidentified speaker (UNKNOWN): you

33:52Adam Powadiuk: go, no, it's not just still screaming and running around

33:54Unidentified speaker (UNKNOWN): the room.

33:54Aaron Cameron: Yeah. Yeah. Can you not see this? Yeah.

33:57Adam Powadiuk: As you know, I do enjoy uh collecting wine. So I've got one set up that I use in real time when I'm at auctions that is monitoring all the auction prices, looking at average prices for those bottles to identify the deals. It obviously knows when all these auctions are going to close. So, you know, nor my normal process pre-AI was to have 10 different windows open on my computer and I'm frantically going back and forth and trying to research in real time as the countdown on the auction is 45 seconds to go and it's 60. there's six different bottles from six different estates, and I'm trying to price them all and figure out where I'm at. This is way better and way cleaner. It's uh way less stressful. You know, this this things like that. And

34:36Aaron Cameron: now you know you get the good deal.

34:37Adam Powadiuk: Yeah, yeah, exactly. You can kind of spend with confidence and also gets me out of my comfort zone areas because then

34:42Unidentified speaker (UNKNOWN): I

34:43Adam Powadiuk: found just by default, because I know certain areas better than others, I I'd would only buy the areas I know because I was more comfortable bidding on it,

34:49Unidentified speaker (UNKNOWN): which

34:49Adam Powadiuk: then just leads to you only. knowing that area even further is broadened my horizons.

34:54Aaron Cameron: There's a thousand use cases.

34:56Unidentified speaker (UNKNOWN): Yeah,

34:56Aaron Cameron: infinite,

34:57Unidentified speaker (UNKNOWN): actually. The

34:58Adam Powadiuk: uh the brain drain. Do you really feel like that's that's real? You know, like I mean, that you don't have any horse riding skills, but your ancestors did. Do you find your life is diminished because your the Cameron families lost their ability to ride horses?

35:11Unidentified speaker (UNKNOWN): I

35:11Aaron Cameron: don't know. I just don't know. Maybe I'm so Canadian. I'm just concerned about these things, right?

35:17Adam Powadiuk: Way out there on the the right of the fear curve is uh where you're looking at.

35:21Unidentified speaker (UNKNOWN): It's not it's not

35:21Aaron Cameron: fear. It's more just like how does this impact sort of society at large, right? Like, I don't know. Writing an essay, is that a valuable skill set? Maybe not. Who maybe who gives a shit? Who cares? Right? That my kids, your kids won't really know how to write an essay.

35:36Unidentified speaker (UNKNOWN): Like

35:37Aaron Cameron: truly, like sit there and go to the library. and do the research

35:40Unidentified speaker (UNKNOWN): and

35:40Aaron Cameron: the citations and you create the thesis, and then you block out how you're gonna do defend your argument. And it would take weeks to really write a thoughtful essay, particularly if it had to be like twenty five, thirty pages. Like you had to have like some deep research done, right? But they'll

35:55Unidentified speaker (UNKNOWN): be

35:55Aaron Cameron: really good at writing

35:56Adam Powadiuk: instructions for agents, a skill that we did not learn.

35:59Unidentified speaker (UNKNOWN): Yeah,

35:59Aaron Cameron: it would

35:59Unidentified speaker (UNKNOWN): just read it at the that same time. way.

36:00Aaron Cameron: Yeah, no,

36:01Unidentified speaker (UNKNOWN): exactly.

36:01Aaron Cameron: So maybe it's irrelevant, right?

36:03Adam Powadiuk: Yeah. I mean, I did use the analogy of pearlies. I definitely know those areas way better than the ones that I learned just GPSing around, you know, for my job. But either way, I navigated both areas of the city effectively. You know, by virtue of the fact that I have GPS, maybe means I don't need to know street by street certain areas.

36:20Unidentified speaker (UNKNOWN): I

36:20Aaron Cameron: keep going down, like when before Google, when you didn't know something, you had a debate with your friends for hours or like someone was going to the encyclopedia and in their dad's office or something, right?

36:31Unidentified speaker (UNKNOWN): Yeah,

36:31Aaron Cameron: I think we're all better off because we can just Google the answer to who scored the most points in the nineteen ninety-three NBA finals. Right? Like, I don't know, right? Like

36:40Unidentified speaker (UNKNOWN): And right now

36:41Adam Powadiuk: I think you talk about impacts on society. Right now I think it's great because it's all it's all assistance, it's all tailwind. Clearly the data is showing to date, no significant job losses.

36:50Unidentified speaker (UNKNOWN): The

36:50Adam Powadiuk: downsides, my mind, is a short list, but it would be extinction of the human race, of course, the Canadian fear of uh extinction, and then job loss. Neither of which currently have happened, but we have realized a lot of benefits. So so far the pros have been realized, the cons haven't, but the cons are very significant.

37:06Aaron Cameron: Have you come across effective altruism?

37:07Adam Powadiuk: Define that for me. I mean define that for the listener. I know what you mean.

37:13Aaron Cameron: I actually don't even know. This concept of effective altruism where it started off in Oxford. Here's the concept. It's about truly the betterment of mankind and humans. Here's the logic. You're walking down the street and you see a little boy in a river and he's drowning. And what do you do? Well, you you jump and you save him. them. And however, maybe you ruin your shoes, you gotta buy new shoes, you gotta go get dry cleaning because. you get your clothes all dirty, but who cares? Like you're not gonna think twice about the fifty bucks or the hundred bucks it's gonna cost you. Yet why wouldn't you spend that money to donate to to children that are that need need that sort of investment, whether they live in sort of you know third world countries or not? And so it's this concept that proximity and what drives that desire to donate and to help others. I'm not doing a very good job. You should just Google effective altruism.

38:04Unidentified speaker (UNKNOWN): But it leads to it.

38:05Aaron Cameron: Yeah, it leads down to a lot of the major owners of AI are these sort of effective altruists. And there's a whole really interesting line of what they think is going to transpire. And, you know, in the future, do AI agents deserve rights and freedoms? And where does it go? with their ability to take over and and destroy mankind and displace workers of the agents? Yeah. And it's it's got

38:33Unidentified speaker (UNKNOWN): it's

38:33Aaron Cameron: got it's it's really kind of bizarre. It's a very, very, very bizarre concept, but it's wrapped up in AI to the point where the Pentagon and Donald Trump's party have a very anti effective altruist, because they really want to curtail where AI goes, or a as Donald Trump calls it. super intelligence. These effective altruists are very, very worried about AI's ability to kind of decouple from human control and just start doing their own thing. That's very, very interesting.

39:04Unidentified speaker (UNKNOWN): Very

39:04Adam Powadiuk: dark if you take it far enough.

39:05Aaron Cameron: Yeah, right. Yeah, yeah, yeah. So there are also some crazy conspiracy, or not conspiracy, some effective altruists have these sort of bizarre beliefs about how effectively, altruistically ants are more important humans than humans because there are a bazillion ants on earth and only eight billion humans. And so therefore, by just scale, they have more value, right? Like it's just some very bizarre, some bizarre theories. That's not I don't they don't believe that. It's not like this is not like a belief system, it's Of just the way that they debate and argue these things. Google effective altruism. I did not do it.

39:38Unidentified speaker (UNKNOWN): It

39:38Adam Powadiuk: sounds like a deep dive is uh is needed. So I'll ask you maybe a more straightforward question then in your professional life. So I have to think, given your role, that you would be part of the overall AI strategy.

39:51Aaron Cameron: Just like everybody else, like there's so much value. I mean, like Bronco was describing, like there's just there's so many iterations of efficiency and and the ability to scale that you just never get anywhere else. It's just doing it safely.

40:05Unidentified speaker (UNKNOWN): Yes.

40:05Adam Powadiuk: That's the I

40:06Unidentified speaker (UNKNOWN): think

40:06Aaron Cameron: the big one it really is more like cyber related and like cyber threat related, right? Is if you build out all of this platform, does it make it more susceptible to cyber attacking?

40:17Adam Powadiuk: It must, I mean the more the transfer away from your human brain into into computer, the more accessible it is. But it's all worth it.

40:25Aaron Cameron: It's all worth it.

40:27Unidentified speaker (UNKNOWN): Anyways,

40:27Adam Powadiuk: I think we've gone

40:29Aaron Cameron: way weird enough. That was the least CRE podcast in

40:33Unidentified speaker (UNKNOWN): podcast we've had in a while. In

40:35Adam Powadiuk: a while, definitely.

40:36Aaron Cameron: Did not say the words cap

40:37Unidentified speaker (UNKNOWN): rate

40:38Aaron Cameron: once.

40:39Adam Powadiuk: Didn't mention

40:40Aaron Cameron: discount to nav

40:41Unidentified speaker (UNKNOWN): or

40:42Adam Powadiuk: class other than residential. Yeah. It's uh yeah. It's good for us though. It's probably good for both of us to get out of the

40:48Unidentified speaker (UNKNOWN): a

40:48Aaron Cameron: little bit different. I like

40:48Unidentified speaker (UNKNOWN): that. Well

40:49Adam Powadiuk: thanks everybody for uh listening to the end.

40:56Unidentified speaker (speaker_0): Thank you for listening to the CRE podcast. The information from this broadcast is not to be relied upon as financial investing, professional accounting, or legal advice. First National Financial LP holds Financial Services Commission of Ontario license number 10514 and 11252.

This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.
All insights