Subscribe
Share
Share
Embed
The WP Minute brings you news about WordPress in under 5 minutes -- every week! Follow The WP Minute for the WordPress headlines before you get lost in the headlines. Hosted by Matt Medeiros, host of The Matt Report podcast.
Eric Karkovack (00:00)
Hi everyone and welcome to the WP Minute. I'm Eric Karkovack. Today's episode features a segment from Matt's discussion with Beaver Builder's Robby McCullough. Robby stopped by to tell us about Beaver Builder's AI integration and the strategy behind it. The guys also dig into how this technology fits into WordPress and the greater open source landscape. Now you can catch the entire episode over at our WP Minute Plus channel.
Visit the WP Minute.com for all the details.
Matt Medeiros (00:32)
You know, speaking of open claw and watching some of these technologies, I'm of the mindset that this stuff is going to get more expensive and we're already starting to see it. Right. You're starting to see it with the different, ⁓ you know,
management of your tokens, the rate limiting is the word I was looking for, like the rate limiting and spend on like Opus 4.7 versus 4.6. A lot of people are now like, I'm just gonna go 4.6. Probably just like you, I've been dabbling in much more open source models because I'm looking at this going, hey, we're gonna need open source in like five years time for AI for anyone to run this and be able to afford it because
The models are going to get pretty amazing in say five years, but they're going to be so damn expensive. ⁓ For the reason of they're going to get faster, they are going to get way better. And those companies are going to say, if you want this power, you're going to have to pay for it ⁓ to the point where you'll be considering, do I hire somebody for, let's say 100 grand to work on in the business, or do I spend 100 grand on this agent?
And the difference is, is I don't have to pay health insurance and they don't take a vacation. Right. Those will be serious things that people. But, you know, listen, as another aside, the insurance companies are going to screw this all up and you're to have to be paying insurance on your agents to to the point to the point where you're just be like, what the hell are we doing? We're back to hiring humans again. ⁓ Yeah. Liability. Right. Yes. Certainly. Liability for sure. ⁓ You know, so I think, you know, I'm looking for the.
Robby (01:58)
Yeah.
Yeah, liability.
Matt Medeiros (02:13)
the successes now early on and what's happening in open source land for AI. ⁓ Because I think we're going to need it. think open source is a necessity, not even a business plan anymore. ⁓ Like we're going to need open source in order to just survive. And I think people are going to change their tune on like what open source means in about five years. Your thoughts on the trend of open source in AI? ⁓
Robby (02:39)
You know, admittedly, I haven't tinkered with open source models all that much or local models. I have been suffering from this, you know, shared experience where the pace that everything is moving right now is so fast. And, you know, I'll go to bed one night thinking like, okay, tomorrow I'm going to do this. Like, I'm going to play with this. I haven't gotten to do this yet. I have this thing on my to-do list. And then I woke up and, ⁓
OpenAI announces their pet feature, and you can hatch a pet in codex, and you can give it a name. I was like, ⁓ there goes my day. ⁓ I haven't jumped on the Mac mini train yet. I have my old gaming rig down here with a graphics card. I know I could be playing around with some of the local models and throw it on the GPU. ⁓ That just has been one of those areas that I haven't had the time. I think that the...
Matt Medeiros (03:11)
Yeah.
Yeah.
Robby (03:37)
the token demand that we're seeing right now. I've heard it put in a way that any company that is generating tokens, the demand is outpacing the supply. And the bottlenecks right now are compute. And then eventually it might be energy. The demand is just skyrocketing. again, it's wild. Another term I heard.
Matt Medeiros (03:58)
Yeah. Yeah.
Robby (04:00)
just today is like a once in a generation technology that's coming out right now and it feels like we're living through something. Something like that, know, short of like maybe the internet or like personal computing. I haven't experienced in my lifetime. So like to answer your question, yeah, I don't know. I think that like one of the areas I'm looking into that I haven't really had to before is routing. Like I've been running, so I've been running like the $20 codex a month.
subscription and then the $100 Claude and I switch off. I have my OpenClaude running through Codex. I mostly use Claude code for my coding projects. But then I've upgraded one and downgraded the other for a month. For the most part, yesterday was the night that Claude reset its weekly. So I was like, okay, let's load it up. Let's try and get as close to that 0 % usage as I can and maximize my tokens. What I'm curious to play with
you know, once that kind of subsidization, subsidization, sorry, I have baby brain. Once tokens being subsidized by VC, like once the cost starts to catch up with what we're actually paying, I'm curious to play with tools like OpenRouter or, you know, explore some of these ways to use different models for different tasks and not have your, you know, Opus 47 working on like, you know, copying files or,
just doing the kind of grunt work tasks, but I haven't gotten there yet.