Transcript excerpt of a conversational interview with Hamel (Hamel Husain), an independent AI consultant who builds AI products and teaches evals. The speaker answers quick-fire questions about AI’s impact on jobs, trust in AI vs politicians, human oversight for AI-driven government, AI companions, and whether AI could be sentient or have a soul.
On this side. That's all you need.
Wow.
Yeah, right, dude, this camera, I'll send it to you. This is the way my social. My new social team was like, you'll get this. And it is the way, bro. So, Hamil, tell me about yourself. Yeah. How do you spend your days?
So I'm currently an independent consultant. I help people build AI products, and I spend my days doing that. I've worked with over 30 companies so far helping them build AI products. Also teach people about applying data science techniques to building LLMs, so, particularly around evals lately. So, yeah, that's a little bit about me.
Hamill's understanding it. He runs the hottest AI evals course in the game right now, and he makes you a wizard after you take it. That's my thinking is, I feel like a wizard after that.
Well, it just is Robert, that has become a wizard.
No, no, I think everybody else has been. I'm just giving it a name. Okay. So, Hamill, quick question that I'm asking everybody to get a pulse from people. Do you think AI is going to kill us all? Take our jobs? Thoughts?
Oh, no, I definitely don't think it's going to kill us all. I definitely don't think it's going to take all the jobs. Things just shift the jobs a bit. You know, I personally feel like I have more things to do because of AI. Like, I just have way more ideas, not enough time to do those. And I suspect all the people I talk to, they feel that way, especially software engineers. I think it'll shift the kind of work you do for sure, but I don't think it's going to, you know, I think there'll be some, like, shifting of roles that you'll have to adapt to, potentially. I don't think it's gonna end in a bad place, is my feeling.
You think it's good for the future?
I think so, yeah.
Cool. Okay. Awesome. And you said that it's shifting the types of stuff you do, and you got a lot of ideas now. Would you say that it gives you higher leverage?
Yeah, yeah, for sure. Yeah. Yeah. There's a lot of things that just wouldn't be feasible to do that are suddenly feasible. Whether that's like making tools of various kinds and automating things that, you know, that you don't. You wouldn't do normally. So, yeah, it's definitely just. I feel like it. Just get more leverage. Cool.
Very cool. Okay. Another question is, I've been asking people, do you trust AI more than you trust politicians?
That's a really good question. It's hard to say which one I trust more. I view both very skeptically.
And it.
Depends on the politician also, I guess.
Yeah. And the model.
I guess just have to verify either one. Honestly, like quite a bit. I'm not sure like there is a winner for me. Yeah, yeah, yeah.
It's case by case.
Yeah.
Okay. Okay. What about AI President versus current president?
I like that there's a human there. I'm not ready for this. Completely autonomous government, AI overlords. I'm not sure, like I have faith and no human in the loop.
Got it. Okay. Okay. So having the human there, having the human connection gives you some level of.
Faith in the way things are ran.
Yes, got it. Do you think that answer might change over time the way that things are drastically evolving on the AI side?
No, but the nuances. I think human in the loop is important. I would hope that whoever is in charge is using AI to make better decisions, just like anyone, any other kind of work and like, you know, do things more efficiently. But I definitely want a human to be checking things and making sure that the AI is aligned with, you know, whatever there is. I just don't think. I personally feel uncomfortable with just saying the idea that pure autonomous is going to do what we want without any kind of oversight.
No human in loop is a scary thought. Yeah.
And it's like this is a spectrum. Like how much oversight? I don't know. I just don't think like full autonomous. Just hand it over and just close your eyes.
Yeah. You know, not good.
Yeah.
We're finding that's not good for LLM apps. Probably not good for running a whole country.
Yeah.
Great. Okay. So that's interesting. Eval's problem to think about another thought. AI companions thought on AI companions. That's the one number one use case in industry right now. Companions life coaching, therapy. What are your thoughts on that? If. If somebody had an AI boyfriend, AI.
Girlfriend, you know, I have no idea. I haven't looked into that. I don't understand that. I've never tried to take a look. Yeah. So I really don't have an opinion on that at all. Like, I feel like. Yeah, I don't know enough to comment totally.
What if? Cuz so it's not something I think about. I want kids someday. We have kids. What if your kids ended up dating an AI companion? What would you do?
I feel really uncomfortable. Yeah, I feel really uncomfortable. I don't feel like. Does this sound healthy to me? Yeah. So yeah, I wouldn't. I would try to figure out what's going on, like, have some conversations. Do you need. That is, can you have a human connection? Yeah.
Yeah. Okay. Okay. Another one is, do you think that AI could ever have a soul, actually be sentient or conscious?
I'm not sure I have a precise, grounded definition of what soul and consciousness is.
I don't think anybody else does actually say.
Yeah, but, yeah, I think that I need to. Yeah. Yeah.
Thank you, Hamil. Appreciate it. Okay, thanks.
