The Iferia Techcast
00:09:32 · Ezekiel Iferia, Anne Cantera
[00:09:32] S0: Oh, that's close. [00:09:34] S1: It's close. It's close. I say that only because I'm part of a group called the Agentyx Foundation, and they are the best builders in this space that I've seen. And the things they build are both brilliant and terrifying, and they are, like, two years ahead of the majority. So I'm I'm I'm guesstimating about a year, eighteen months, two years based on their work because I think they're, you know, eighteen months, two years ahead of most. So, and I belong to a lot of groups. I probably belong to a 100 different AI groups. That particular one has some of the most futuristic building, period. So, yeah, I I... But I also... So it also will depend on what you do. Right? But if if your AI can replace an intern, again, you still need interns because you're still going to need humans working and building the workforce of the future. And since the very beginning, my mantra has been assistance, not replacements. But that leads us to another really short sighted strategy of tech leadership where they want what's called the super IC. Have you heard that terminology? [00:10:58] S0: No. Not sure about that.
00:10:58 · Ezekiel Iferia, Anne Cantera
[00:10:58] S0: No. Not sure about that. [00:11:00] S1: Okay. So super IC means individual computer... Contributor, super individual contributor, meaning you can have a team of 20 and get it down to two super ICs with a c... With AI. Right? So if two people with the augmentation of AI can do the work of 20, they salivate at the idea of laying off those 18 people, paying those two a little bit more, and pocketing all that money. That's... I... I've literally seen them salivate over that idea at a CTO conference that I went to. And, again, shortsighted. You have two people doing the work of 20. What happens when one of those people is sick or hit by a bus or quits or, again, super shortsighted. You're not... It it... It's like it's like a garden. Right? Plant a beautiful garden. You harvest everything from it, and you don't plant another one. You're going to go hungry. But they... But because you have the harvest in front of you right now, you're not worried about going hungry next year because you're blinded by what's in front of you in the moment. And that's typical. That's typical C suite thinking. [00:12:24] S0: Alright. So... And, you know, let's go into the topic of ethical AI. Know, that that last word, it's become a buzzword these days. What exactly does it mean, you know, when you strip it off all the technical jargons and, you know, remove the tech... Terminology. So what exactly does it mean in actual design of a product or system?
00:12:50 · Anne Cantera
[00:12:50] S1: So it's interesting. I think every single podcast I've been on has asked me this question. I've refined my answer, so you'll prob... Hopefully, you'll get a good answer. It means different things to different people, for one. Right? There is the the the one camp who is like, none of it's ethical. You shouldn't use it. We should shut it down. That's not gonna happen. Right? This group of people have so much power and money that we, the people, are not easily going to shut down that technology and make it go away. The the people whose books and art and music and everything who's been stolen and and used, I think, should be compensated. I have a friend who works in that space. He helps people sue and get settlements and get paid for their IP that's been stolen. There's the environmental issue. I think every single tech company should be made to do water conservation, environmental conservation. I think it should just be par for the course. That's not happening, but I hope it will. For me, what it meant was, how can I use this technology to help people? And my answer to that was by consulting for underserved businesses, whether the business is small, minority owned, female owned, veteran owned. So for me, that's what I do. I left enterprise, and now I have... My consultancy works as much as possible with exactly those types of businesses. Because in my mind, a small business can use AI as an equalizer and an elevator. Right? They can do things with it that they probably couldn't afford to do by hiring. If they can't afford to hire and they need help, what are they supposed to do? Also, I work I work with some founders who work 80 a week because they can't just yet bring people on. There aren't a lot of people who are skilled the way they're skilled. And if I I give them hours of their life back, and that... Especially to, like, a single mom who's trying to run her own business, those hours of her life back are really a gift for her, and I feel good about that at the end of the day. So that's what it meant for me. Like, how can I work in this space and feel good about it and and feel ethical? So when you ask me what ethical AI means, it... For me, it means not building tech that's gonna put people out of work and helping people who are otherwise overlooked.
00:15:40 · Ezekiel Iferia, Anne Cantera
[00:15:40] S0: Alright. So... And, you know, can you can you maybe give us an example of, you know, an AI system that looks ethical on the surface but really isn't when you, you dive deeper into it? What's the tale? [00:15:55] S1: I have to think. That's a great question. Best question of the week. Oh. Alright. So it looks ethical on its face, but on this... Underneath, it's not. Man, you got me off guard. I need to really think about that. I think anything that is designed to hustle data without people knowing, and there's a lot of that. Right? People are having to kind of trade off data for convenience. Right? Like, when I'm using any AI, I know not to put personal information. I know not to to to do certain things. I don't think people know that. I don't think people generally know that. Like, they'll take a picture of their resume and upload it. Right? Once it's uploaded to ChatGPT or Claude or anything like that, it's theirs. You don't... That's not yours anymore. Right? So I think I think that silent stealing of data is probably the most obvious to me. I I... If I thought about it longer, I could probably come up with something better. But off the top of my head really quick, I think that that that quiet stealing and harvesting of data is is probably... It's the the one that comes to mind first. How you get them is theirs, and anything they give you is theirs. Keep that in mind.
00:15:55 · Anne Cantera, Ezekiel Iferia
[00:15:55] S1: I have to think. That's a great question. Best question of the week. Oh. Alright. So it looks ethical on its face, but on this... Underneath, it's not. Man, you got me off guard. I need to really think about that. I think anything that is designed to hustle data without people knowing, and there's a lot of that. Right? People are having to kind of trade off data for convenience. Right? Like, when I'm using any AI, I know not to put personal information. I know not to to to do certain things. I don't think people know that. I don't think people generally know that. Like, they'll take a picture of their resume and upload it. Right? Once it's uploaded to ChatGPT or Claude or anything like that, it's theirs. You don't... That's not yours anymore. Right? So I think I think that silent stealing of data is probably the most obvious to me. I I... If I thought about it longer, I could probably come up with something better. But off the top of my head really quick, I think that that that quiet stealing and harvesting of data is is probably... It's the the one that comes to mind first. How you get them is theirs, and anything they give you is theirs. Keep that in mind. [00:17:34] S0: Oh, awesome. Awesome. That was a really interesting response and an insightful one, I'd say. So, you know, you talk also about agentic AI. So, you know, when you, you know, break it down in plain language, what exactly does it mean? You know? And why should a regular person, you know, someone who doesn't really work and take care about it?
00:18:00 · Anne Cantera
[00:18:00] S1: Well, I think, originally, when we would build agents, it started as a chatbot or a voice agent, right? You would call and you would get a voice agent, you know, an IVR kind of scenario. And then what it evolved into was specialized agents, right? It was like the marketing team has an agent, the sales team has an agent, this team has an agent, and they all kind of work back and forth with each other. And then it turned into, alright, there's gonna be like a queen bee or an orchestrator agent, and she's gonna have a whole swarm of agents, and they're all gonna do different jobs. And the thing with that is the narrower the job of the agent, the less likely it is to hallucinate and go off the rails. So now we have we have hives full of agents doing little tasks and little jobs, and it's evolved into loop engineering and and self self improving. And why should people care about that? Well, twofold. They can work for you or they can work against you. So it's good for everybody to learn my favorite word, foundational AI and how it works. So you can protect yourself, be part of the conversation, and know what to pay attention to. So one of the things that keeps coming up is like, in my mind, this is gonna become a deeply dividing issue economically because he will... Who with the... He with the most money will have the best AI and be able to have the power and control, whereas you and I might have little agents we can spin up or build ourselves. But can we compete with the Googles of the world? I don't think so. But, also, it's gonna be huge AI against huge AI. Right? Like, China's AI versus The US AI, and that's the part that really I... I'm deeply concerned about security because whoever has the the best agents will win. The best AI will win. So the the the the takeaway in general, I think, is is you can use AI to teach you about AI. The best use to me about ChatGPT or Claude or any... You can learn anything. You can teach yourself anything for very little cost, or at least... Maybe not very little cost, but very little money. And that's... Everybody needs foundational knowledge about this stuff. Just like using a computer. Right? Like, when when home computers first came out, you had to build foundational knowledge about how it works. This is the same thing. You need to you need to learn foundational knowledge about AI, period.