Amy Kamala joins Luke and Jonathan to discuss art, poetry, biology, and how decisions get made in the WordPress project. Amy shares her perspective on the human element in automation, Luke expresses his love for contributors over contributions, and Jonathan offers insight into why he and Luke are still together.

Transcript#

Jonathan:
Luke, when we started planning this season of Crossword, you were just coming back from WordCamp Asia and you pinged me like, hey, we gotta bring Amy on this season. And what was funny, so you and Amy got to spend some time together there at WordCamp Asia, but I’d already interacted with Amy a bit from my early forays into the hosting team from previous WordCamps. So you were like, I’m pretty sure we’re talking about the same person. And in fact, we were.

Luke:
Yeah, it was crazy. I was at WordCamp Asia on Shabbat and that meant, you know, a bunch of restrictions on my behalf. And I was waiting for the Sabbath to end and I was just hanging out in the speakers green room and I meet Amy there and it’s like, we’ve got all this stuff in common. She’s an individual contributor like I am, you know, sponsored, but also not working for a company. And also, she’s there on Shabbat waiting for Shabbat to be over too. And so we had a lot in common. It was really nice. It felt serendipitous. So Amy Kamala, welcome to Crossword.

Amy:
Thank you so much and thank you for having me.

Jonathan:
So, Amy, for those who don’t know, can you give us just a little bit of your background, like how you came into this WordPress ecosystem?

Amy:
Sure. So I’m from Los Angeles and when I was younger, I knew these guys that started a web hosting company called DreamHost. And I knew them for quite some time and later on started working at their company in social support, then in technical support, and then became a WordPress support specialist. And that’s how I got into the WordPress project and contributing to the WordPress project, because they were kind of ramping up to become a WordPress host and to market themselves as a WordPress host. And we were pivotal in starting the hosting team, which is a team that I’ve been the most involved in. And the hosting team started in 2017. I started contributing in 2018 and I became a team rep in 2019. And that was kind of my entry point.

Luke:
And the release coordinator for WordPress 7.0, is that right?

Amy:
I’m one of them.

Luke:
Right. Okay. And that is not a small role. There is a lot of work involved in that. I’ve been seeing your name in Slack, running meetings. I’ve been seeing your name on the Make WordPress website, writing blog posts, giving people updates. Is this your first release?

Amy:
It’s not, I was also a release coordinator for 6.9 and I was involved in the All Women’s release back in 2021.

Luke:
Cool.

Jonathan:
Amy, one of the things I’m curious about, so when you were working at DreamHost, you got involved in the hosting team, what was the initial inspiration for you to start contributing to the project? Like what was that initial thought process? Was there someone who asked you or were you like, hey, this looks interesting to me? What was that starting point?

Amy:
So I was on a dedicated WordPress support team and the CEO and leadership and our direct manager were encouraging everybody to hop onto Slack and check out the hosting team. So I did it. I’m really not a shy person, which is beneficial and also not beneficial at the same time. And my first meeting, I didn’t know what to expect. The hosting team was new and they were trying out different types of meetings. And so at the time it was a video meeting and there were like eight people there, 10 people there. And I didn’t know what to expect. And I just went out of curiosity and cause it was like in my head from leadership to go. And when I went into the — it was Zoom, I think, or something like that. When I went into the video meeting, everyone there was somebody that I already knew or had a connection with. So it was people that worked for DreamHost and other web hosts that I just knew for some reason through mutual friends or colleagues. And so it was a really, really, really soft landing for me. I even knew that one of the team reps was Kira and that made it a super easy entry point. There were zero barriers. There was no like — I started taking notes. That’s kind of how I got, instead of just being an attendee, how I started like being active on the more administrative side.

Jonathan:
And for those who don’t know, this is before AI took all the notes for us. Back in the old days. And also for those who don’t know, how would you describe what the hosting team is or does? What is this hosting team?

Amy:
Yeah, this was before AI, period. Right? The before times. So the hosting team is basically an intermediary between web hosts that support WordPress and the WordPress open source project, communicating between the two, facilitating testing between the two to help assure compatibility with hosting systems and also keep the development process informed with any issues that might be happening on the hosting end. So that’s the main role. We have a distributed testing software that about 70 different web hosts have installed on their systems, sending test data back to the core team, which has been helpful for having insight into errors and issues over the years. That was developed way back when, or first developed way back when, I think in 2018 or 2017, something like that. And we also have two handbooks, the hosting handbook, which of course is standar d for all the teams, and the Advanced Administration Handbook, which is like a technical handbook for users and web hosts. And that’s also shared with the docs team.

Luke:
It’s difficult, Amy, to know exactly what to talk about when it comes to WordPress 7 because as you’re fully aware, being so involved with the release, there hasn’t really been any controversial issues or mess ups or anything like that with this release, has there?

Amy:
I have no comment.

Luke:
Well, I thought maybe we could dive into my favorite actually. I was being — Jonathan, in case you don’t know, I was being conceited.

Jonathan:
I picked up on it.

Luke:
Yeah, okay. That was sarcasm. All right. So Akismet. This is a plugin that’s bundled with WordPress and now in more ways than one. Did you follow along with that Trac ticket where it was proposed that Akismet got merged into the new connected screen and then that got immediately merged and then it got reverted and there was a lot of discussion. Now it’s back in it. Have you been following that?

Amy:
Not that particular — yes, I’m aware of that. As a coordinator, that hasn’t been a major pain point for the release from my perspective.

Luke:
Right. From my perspective, decision making has been a major pain point for the release. Would you agree with that?

Amy:
I could agree with that.

Luke:
Okay. Cause that’s to me, when we’re talking about Akismet, that’s what the issue is, right? It’s a decision-making sort of problem.

Amy:
Yeah, I mean, I think that stems from a few different sources, but the main thing being that it’s kind of a democratized decision-making system where anybody and everybody can share their opinion, comment on the tickets, object to something, suggest something, open a PR, so on and so forth. So conversations tend to be pretty lengthy. And there’s not a single final decision-making person. I mean, there is, there’s Matt, but there’s not someone making decisions about every single development problem. Like they’re day-to-day doing that. It’s kind of — people discuss it. There are tech leads and the tech leads make decisions when needed, you know, but there’s not like one point of escalation to make final decisions about these things. When it comes to major decisions, leadership obviously is there to call the shots. You know, like including that in the connector screen at all would obviously be something that leadership would be involved with. But otherwise, the conversations are open to the public and it tends to be — I mean WordPress has tickets that are like 10, 20 years old that are, haven’t been closed out because they’re going to be included at some point. And some of those have really, really lengthy conversations, like over the course of years. And there are some of those tickets in 7.0 where there’s been years of decisions and decision discussions and conversation, and then there’s a consensus at some point.

Luke:
It’s culminated in awesome. So can I just ask real quick? I actually do mean this as a genuine question, even though it won’t sound like it. When you say leadership, do you mean Matt Mullenweg? Is that — are they synonyms?

Amy:
I mean Automattic leadership. Well, okay, so Automattic leadership, from what I can tell at least, is not in the weeds working on every single detail. The core team is, the developers are, the designers are, but there’s not like — but yes, I mean, the WordPress project leadership, Mary and Matt, would be the ultimate points of escalation there. Or previously, Josepha. But there are other decision makers, of course, that are really in the weeds. Decisions aren’t made until they’re proposed and discussed and discussed more and tested and so on and so forth.

Jonathan:
I’m curious, except for the occasional interventions, like we saw with the Akismet, the process that I’m hearing you describe is one that, at least to an outsider, could seem pretty opaque. How do things get done? What does this look like? From your perspective in the role of a release coordinator, how do you navigate that? What seems like a lot of ambiguity about what’s going to get done? It’s like if you see a lot of activity in the ticket, I could see it being unclear sometimes — is this going to get moved forward if it does at all? Is that just something that is just, this is just the way it is? Or like, what’s that experience like for you in that role of trying to like coordinate things?

Amy:
Well, the release coordinator wouldn’t determine how things get done. Maybe more like when.

Jonathan:
Of course, but in terms of knowing what’s happening, being able to gauge like, is this on track or is this not when it’s perhaps ambiguous, like who’s going to decide.

Amy:
Okay, so I could say that when there are bugs and a lot of discussion or debate about something, about a ticket or PR or feature, it frequently seems to boil down to right before the release, what is realistic to include that the release squad determines and the tech leads. So the past few releases, the tech leads and dev leads have been like — or for this release, for example, the new database table for the real time collaboration feature was available to merge. It wasn’t completed, but it was available to merge months ago before the delay happened. But tech leads felt that it wasn’t fully flushed out and it wasn’t ready. So a temporary workaround was merged instead. And this decision was made right before one of the releases, I think it was like RC1. I mean, like hours before. And it was kind of because that was the point where a decision had to be made. So a decision was made and there was discussion about it. Everybody who made themselves there could provide input. But ultimately it boiled down to like technically what was realistic, what was feasible and what the leads think is best for that type of thing. In terms of not merging something like that, people could discuss that until they’re blue in the face, but it would boil down to a technical lead actually merging the work or not. So I guess in that regard, the tech leads are decision makers, pretty big decision makers when it comes to releases. But then it turned out that a lot of testing data came back and it wasn’t performing as expected on as wide of a variety of hosting systems as the goal would be. And so it just, it couldn’t be put out there into the world as a stable software yet. There are some other things that happened in this release cycle, you know, that kind of contributed to it. But for example, the package was like, I think it was like 30 megs — or sorry, gigs — or 50. It was like a huge, huge, huge package. And it turned out that there were a lot of artifacts left. So there was at one point a couple of releases ago, like a huge push to clean it up and reduce the package size down to like what is considered a more normal size. One of the features was removed because it had — it was including a third party code package that was big. So one of the features, one of the flagship features was removed, not real time collaboration, but client side media processing. It got punted to 7.1, unfortunately.

Luke:
Really?

Amy:
Right? Well, 7.1, you can look forward to it.

Luke:
Yeah. Hmm. I didn’t actually know that. I was looking forward to that. That’s sad. Yeah. Something else to look forward to. I wonder, like when you’re talking about this process with the tickets, it made me think — do you ever get the impression? I sometimes get the impression that it’s a sort of form of open source filibuster where people, it feels like people find problems. They go looking for problems that aren’t — you can’t do this yet because there’s this minor thing or don’t forget to involve these minor people — and they sort of like delay a decision. Do you get the impression people ever do that intentionally?

Amy:
Well, I think that there’s a lot of negativity thrown towards WordPress in general, and that does kind of spill over into the conversations about such things where people tend to be very opinionated and maybe even a little aggressive in their communication. And I don’t know if it’s intentional or if people are just really, really passionate about development, WordPress development. I do think there are pot stirrers, not on the core team. I can’t identify anybody that I know off the top of my head that’s actually a contributor that’s an intentional pot stirrer at this point in time, but externally, definitely.

Jonathan:
When you said that I just saw Luke as a future pot stirrer.

Luke:
Yeah. I’m not a pot stirrer in that way, I only stir your pot.

Amy:
I mean, we do get — there’s been like a few occasions where like somebody came into the hosting channel and started bashing every host’s support team, just like all support teams across the board suck and they fail and so forth.

Luke:
Yeah, and we get a lot of that in the plugin review channel. We get a lot of that. So there are complainers, there are whinges. That’s normal. But what we don’t want is we don’t want that sort of to be systematized.

Amy:
Right. Absolutely.

Jonathan:
Amy, you have an extensive — what I would describe as like a fine art and just art background. You’ve got, if someone checks out your websites, you’ve got a bunch of your work up there. And not just like fine art, also you have a background in music. You’ve got a not too secret SoundCloud link that you’ve got embedded on your profile there, which I think is fantastic. We’ve had — for someone listening without any context, we’ve had a pretty technical discussion over this past period of time. And I’m curious, like how does your art background, how has that informed the work that you do today? Or is there no connection?

Amy:
There’s definitely a connection. I mean, so when it comes to technology and software, the goal is the user experience. And that’s directly aligned with art. The goal there is the viewer’s experience, how the viewer interacts with the artwork and receives the artwork, interprets the artwork and experiences the artwork. So, you know, it’s kind of very like human experience focused. I know software doesn’t seem human. You know, for a lot of people it can seem very foreign and confusing and maybe even overwhelming but ultimately it’s, you know, a thing that people made for humans to have a specific type of experience. And so that’s the direct bridge for me. And I think the two are — I mean they’ve linked intricately. For example websites, you need the front end, you need the styles, you need the design, you need all of that, and all of that is art or art based.

Jonathan:
How do you feel about AI-generated art?

Amy:
So I’m not a big fan of AI just in general. I don’t hate it. I don’t want it to like disappear, but I’m a little worried about how obsessed everybody is and also like kind of the variety of things. Which makes sense because it’s kind of a new technology, really trendy and so forth. Some of it — some AI artwork can be really, really beautiful. But there’s always something weird about it. I can’t really put my finger on it, but there’s just always something that’s a little bit off, a little bit weird. But you know, I don’t hate it. If people want to like generate drawings for coloring books and like throw them up on Amazon to make extra money or whatever, I’m all for that. It’s a good side gig.

Luke:
I want to be careful about the word art in that sense. Let’s call it AI generated images or AI generated videos.

Amy:
Mm-hmm. Sure.

Jonathan:
No, I chose art on purpose, because it is a subjective thing, right?

Luke:
Well, I think art changes the person that created it. We’ve talked about this, haven’t we? I think art changes you, the creator. And so if it doesn’t change the creator in the process of being made, then I don’t think you can call it art.

Jonathan:
So you don’t think it can change the models that —

Luke:
Well, no — sorry. I don’t mean to go off on a weird tangent here, but actually it doesn’t. The models are static. The models are frozen is the technical term. The models cannot change. Now, if we sometime in the future had a model that could change, that was capable technically of actually changing, and it did so while generating the art — then I would, the images, then I would call that art. I’m not saying machines are incapable of creating art, just that machines are currently incapable of creating art.

Jonathan:
Interesting. What do you think about that, Amy?

Amy:
I think that’s really interesting and I definitely can align with the idea that art affects and changes the creator and can be transformative and that can’t necessarily apply to AI. I also think that art is a really emotional experience. Humans are emotional, whether we want to admit it or not, everything comes with some form of emotion, even if it’s just like, I feel okay. That’s, you know — and that’s part of our system that AI doesn’t have. And so the emotional experience of seeing art and also the emotional experience of making art, what it is that is being changed about you and what you’re expressing and so forth, it’s a human sharing their inner world kind of with other humans. It’s not something that computers can do.

Jonathan:
So I only barely know about what I’m about to say. It’s been interesting as I’ve been studying more about like how these models are working. They’re describing it in terms of having like synthetic emotions because one of the things that they’ve been able to see and study is that — because these models when we interact with them are taking on a character or persona — and what they’ve been able to see and reproduce is that when you as a human respond in a certain way to it, it will trigger what we understand as like synthetic emotional responses from the models — care from the character. So it’s like if you put a bunch of pressure onto it, its actions will be affected by that — it’s more likely to try and like cheat its way through something. So which is pretty interesting because I would have agreed initially like, yeah, machines can’t experience emotion, but like when you break down what it is and how it works in our own brains, it’s interesting. And the research is starting to suggest that the way that we interact can actually get different outcomes based on how we’re treating and what we’re evoking, in terms of that synthetic emotion. It’s not true emotion, but there’s something that’s effectively akin to it, which to me makes it really interesting when you think about art and what kind of a response a piece of art can elicit, both in the creator, which is not how I usually think about it, but in the viewer of it as well.

Luke:
Can AI write poetry?

Amy:
Not very good poetry. I mean, poetry is art. In my perspective, poetry is actually one of my most cherished art forms. I have no desire to be a professional poet. But I write poetry as a hobby, I guess. I’ve probably written a few hundred poems or more than that in my lifetime. Started when I was a kid. So I considered that art — music, visual art, music, writing, dance, all of those things. When you say that AI has programmed the illusion of emotions, it’s still not emotions, it’s manipulation. It’s the program manipulating us. In film, there’s something called suspension of disbelief. Have you heard that phrase? So that’s what it is. It’s humans projecting into the experience of interacting with this computer program — emotion. It’s not real emotion. It’s an intentional manipulation of the human’s ability to have a suspension of disbelief.

Jonathan:
Actually, so I don’t disagree with that, but the part that was focused on here is how these models reason, not how they respond to the user. And one of the things that we’re finding is the way that these large language models are working. And this is just, again, I only barely know what I’m talking about here. The thing that stood out to me though is they took an example where they fed a bunch of short stories to these models. And the story would be a sad story or a happy story. And they noticed that they would consistently trigger the different parts of the synthetic brain that responded to that emotion of sadness, which is the same way like a human brain works. We read a sad story, for most of us we feel some sadness, et cetera. And that feeling, it was affecting how the character generated by the model responded to the user. And so it’s interesting, like it’s more of a breakdown of like how do emotions in our brain even work? And they’re slowly recreating how it works. It’s not like — yes, you can pretend to be excited or whatever. But what’s more interesting is that what the models are actually doing, the actions that they’ll take are changing based on which response is being triggered.

Amy:
Yeah, I mean, that’s in line with machine learning. There’s one component that’s very much missing from trying to figure out how human brains and emotions work, and that’s the chemical component. We have a variety of chemicals — with biology, chemicals — and we are not just electrical impulses and zeros and ones. We are way more complex than that. And maybe even the idea of a soul, and/or, you know, we generate — computers generate energy also, a different type of energy.

Jonathan:
So Luke asked about can they write poetry and you’ve got this background in poetry. I remember the first couple of poems that I wrote as a teenager. One of the things that bugged me — like I was living in California, I entered a national or a statewide poetry contest that I actually ended up placing super high for my grade and had my first couple of published pieces. I remember reading the other pieces and being like, this isn’t poetry. Like y’all are not even rhyming. And obviously I had a naive and limited understanding at the time. It’s interesting, it struck me, this is far before any machines are trying to generate this stuff, right? And when you say like, it’s not good, there’s a lot of human written poetry that I would — I by my subjective taste would say wasn’t good. But at this point, and it’s interesting from a pure technical perspective, I would argue most of the latest models are capable of producing something that the typical reader would actually think is good. And so that brings up this interesting thing about like taste and subjectivity, which I don’t know, I think it’s actually super valuable. Like what you think may be good as art may not be the same as what I think is, but that doesn’t make it any less art.

Amy:
Absolutely.

Luke:
Yeah, that’s not where I was going with that. Sorry. So I’ve written thousands and thousands and thousands of lines of poetry since I was a teenager to this very day. Only — usually people don’t call it poetry, but they call it code. And I was trying to tie all this back together with WordPress.

Jonathan:
I saw where you were trying to go with that.

Luke: Yeah. How have you noticed that AI contributions, how have they influenced the 7.0 release?

Amy:
Well, you know, there are some folks that really want to streamline workflow processes within the core team and in WordPress in general, and automation is one of the ways that that could be done. So some things are being automated that weren’t previously automated. And there’s discussion about what else could be automated. One of those discussions is release announcements, if release announcements should be written by AI or not, and there’s some folks who are pushing that they should be and some folks who are not. And so in that regard, you know, in theory, the release coordinator position could be done by a bot. Not to like replace myself, but it kind of could be — to do the release parties. It’s already scripted. There’s a script in advance that everybody follows for every release party and that’s to make sure that the party goes in a certain way and steps are followed in a certain way. There’s not like — if you don’t say the exact sentence, you know, that’s wrong or anything like that, but you’re supposed to — there’s a process.

Jonathan:
Sure, yeah. There’s a process though.

Amy:
Yeah, and that easily could be a back and forth with a bot. But in that regard, there’s not a push to replace release coordinators. It’s just a thought that I’ve had. There’s not a push to replace release coordinators because humans are still valued. But for things like summarizing the technical aspects of features into a marketing blurb and stuff like that, there’s definitely some of that going on. And there’s not a huge debate or discussion about it. There’s maybe some smaller discussions here and there where some people are like, use AI for this, here, I did this prompt. And other people are like, be really careful with AI usage because — so that’s kind of the status. Some things are being automated and some things are using AI, and release announcements are kind of like in that line of things that are being looked at for automation. But otherwise it’s really still human centered. Humans are reviewing the PRs and discussing it with each other and doing the merges and compiling the packages and all of that.

Luke:
So in your opinion, should we become more AI automated as a project or should we — have we gone too far already or have we got the balance right? Like how are you thinking about it right now?

Amy:
I don’t think we’ve gone too far already, but I think we’re definitely at risk of going too far. I’m not aligned with the current push with using AI to automate things like release announcements. I like voices from humans and people’s personalities kind of leak into the announcements and that’s not going to happen if we’re using AI. And it makes it special and appealing. Something I noticed just yesterday during dev chat for the core team. So myself and the other core team rep are kind of alternating doing meetings. And so he did the meeting yesterday, but I had written up — speaking of scripts — it’s a lot of pressure to do these things in front of like there’s 53,000, I think, people in that channel. It’s a lot of people that can see and watch what you’re doing. So I’m a little bit more used to it now, but it definitely — there’s some pressure for it. I don’t want to embarrass myself, which I do all the time, constantly. But I try to minimize the amount that I embarrass myself. So I usually write up what I’m going to say, at least a little starter, like a draft, before I do meetings for hosting team also. I started doing this years ago to kind of take the pressure off. So I wrote up a script for the meeting yesterday in my voice, the way that I talk. It’s like very peppy and bubbly and cute, and there’s hearts and kind of girly and so forth. But JB, the other team rep, does not talk like that at all. It would be weird. It would be totally out of character if he were to be like that, right?

Luke:
I would love it if JB gave that a go. I wanna see a teddy bear hugging a love heart from JB.

Amy:
You know, he’s nice, so you never know. But yeah, I shared my script with him because we kind of determined who was going to do the meeting right before. So I was like, oh, here, I prepared this. Then I realized he can’t even use this — it’s totally useless to him because it’s just like blatantly my voice. AI would ruin that, where there’s different people’s personalities. And I like seeing how other team reps and other people talk and behave and like what kind of voice they use and how they show their personality and the things they share and stuff like that. And I would be concerned that we lose that human element that makes it — that’s kind of a draw for contributors. I wouldn’t be wanting to contribute if it were all bots, you know, and the other people are one of the reasons that being on a team — collaborating with other people, working with them, maybe even like disagreeing sometimes can be beneficial. And you can like make friends via disagreeing if you’re nice about it. Yeah. I mean, it’s good, healthy debate.

Jonathan:
That’s how Luke and I are still together after all these years.

Amy:
I would hate to see core and WordPress in general lose the human element altogether. And so I do think that we should be careful about that in general. And there are some guardrails up, but there’s nothing official. And with the integration of AI into the actual WordPress software, that’s a consideration for anyone using the WordPress software. We shall see what happens with this.

Luke:
Yeah. Well, it reminds me of — I think it was Zig. I’m going to get my projects mixed up here, but there was a popular open source project that recently banned all AI contributions. Just said, no, you can’t submit any AI. And one of the reasons that were given that I loved was: we don’t value contributions as much as we value contributors.

Amy:
Really? Hmm. I like that. I don’t know about banning AI altogether. So I know some of the devs are using Claude for help debugging and, you know, things like that. I don’t think that devs are like vibe coding or anything. I didn’t think that developers that are working on WordPress core are — they’re coding themselves, but definitely using AI as a tool, as an assistant.

Jonathan:
People save vibe coding for the plugin review team.

Amy:
Yeah.

Luke:
Yep, yep, we get a lot of that.

Amy:
I bet.

Luke:
Yeah, I mean, I’d say 80% of submissions roughly. Just gut feeling, yeah.

Amy:
Really? Is that like recent — that’s over the past few years, right? Because AI wasn’t even a thing like five years ago.

Luke:
Well, even just compared to last year where we’ve got four times as many submissions every, every week. Yeah.

Amy:
Wow, that’s very interesting. Has the plugin team up to — I’m sorry that I don’t know this. I feel like I actually should know this, but the guidelines for submissions, I’m sure that have been updated for AI — are there restrictions on the — now I’m interviewing you. Are there restrictions on the — really?

Luke:
No, no, that’s good. They haven’t been updated. Actually they haven’t. And I would love to see them updated, but there are no restrictions. You can submit any AI plugin you want. What I’m concerned about is people dumping, you know, here’s my vibe coded plugin. I want to get the badge. I want to be able to say, I’ve got a plugin on the repo. Here you go. And then never returning to ever update it again.

Amy:
Right. Not maintaining it.

Luke:
I mean, and even with the best intentions — maybe people have great intentions, but then they get excited by the next vibe coding project and forget about it.

Jonathan:
Well, and the thing that I learned over observing Luke’s experience and thinking about my own is there’s often a — because it’s volunteers on the other side. There’s people with very low effort that are generating these things. Then they’re getting responses from a real human on the other side. And there’s been at least a few situations that Luke and the team see where folks aren’t even bothering reading those responses from a human. They’re putting them into their model and then they’re copying and pasting what it says right back to the team, which feels like a complete disregard for the time and effort and energy from those reviewers.

Amy:
Yeah. I mean, that’s basically an AI submitting a ticket. And I don’t know if there’s like momentum to make guardrails for that with the plugin team. There’s nothing official for the core team either, but I do know that devs especially are expected to — and also when we publish posts, it’s expected to mention if AI was used. So folks know what they’re looking at, that there’s not an illusion that this was created by a human and that this other third-party thing was too.

Luke:
And I’ve proposed exactly that. I’ve proposed an AI scale of, how much did you use AI and in what ways has it been used, and sort of simplifying that into a five step scale. But that’s sort of still in the early stages. The problem though is that there doesn’t seem to be very much momentum. It seems like the answer — I think it’s okay to say that on the plugin team in general and also everywhere else I look in the tech world, the answer to all the problems, including problems caused by AI, is more AI. And that any solution that doesn’t involve AI does not have a lot of support or momentum behind it. So at the moment to respond, you know, to the flood of new submissions, most of the team are just saying, well, let’s use more AI to handle this. And certainly Matt Mullenweg has been vocal in our chat saying, we need to use AI for this and use AI for that. I don’t know, I’m worried that it’s going to sort of spiral into a black hole of AI everything.

Jonathan:
I spent about an hour today reviewing and cleaning up some skills that I use in my work for Claude and that are shared with the team. And there’s been a couple of things — like when I’m working with AI, I’ve noticed that I don’t care if someone uses it, but I care that they’re taking full responsibility for the output that they’re sharing with me. And one of my personal pet peeves will be when I get something that the person clearly hasn’t even read themselves, like all the way through. Cause it’s like, I’m happy with whatever tools that you use to create an output that you are proud of and willing to be responsible for. And I think that’s a line for me where it’s like, I see people who — I’ve been hearing things more like lately, like Claude says this or Claude said that. I’m like, well, what do you say? And which to me — like when I went through and revised the skills, I want to make sure that I understand what’s happening. And if I don’t understand it, there’s a few lines that I came to where I’m like, I don’t understand what’s happening technically here. I flagged those to go back and make sure that I understand. And that’s where I find oftentimes is the gap with folks. Like they’re submitting something they vibe coded — this is not the case all the time. But oftentimes it’ll be they didn’t even bother to read it and, let alone understand what’s really happening with this. And that’s, I think, the problem more than what tools you use to get there.

Amy:
I agree with that. I definitely agree with that.

Luke:
People often talk about AI in two ways. They talk about AI out of both sides of their mouths. On the one hand, AI is this amazing revolutionary new technology that is changing the way that we work and replacing people’s jobs. And it’s completely revolutionizing the planet in a way that hasn’t been seen since the industrial revolution, blah, blah, blah, blah. It’s this whole big thing. And then other people, they’re like, oh, you can use whatever tools you want. Like AI is just some little tool. You could use an IDE and code it yourself, or you could just use a bit of AI. It’s just a tool. Which one is it, Jonathan? Is it just a tool or is it actually revolutionizing your workflows?

Jonathan:
It’s definitely more the former than the latter. I don’t think I’ve said it’s just a tool. Well, I mean, I do think of it as a continuation from the calculator to the computer. There was a time where it’s like, oh, you can’t use a calculator for this stuff. What kind of an accountant are you if you can’t do it by hand? And that’s carried on. You can’t use spreadsheets for this. That’s a computer. And we’ve —

Luke:
You just did. You just did. People can use whatever tools they want. That’s what made me say it. So then you can’t say that AI can create art and have emotions if it’s just a tool. You can’t have it both ways.

Jonathan:
I’m not arguing that it’s just a tool, when it comes to how a human would use it, including I’d argue to create art, there’s the point of disclosure, which I think is fair, especially in this sort of fast changing place. I don’t think people question mathematicians anymore as to whether they use calculators in their work, but there is at this stage a continuum of like, what are we claiming that we did versus the tools that we used?

Luke:
Back me up Amy, you’re on my side right?

Amy:
Yes. I don’t know. We’ve had Stack Overflow, right? And people copy-pasting blocks of code from each other. That’s different because it’s still people interacting and the brainchild of someone at some point. More so than, you know, AI. But honestly, I think that it’s just a trend right now. AI is just a trend right now and it’s not going to die completely. This is my prediction. It’s not going to die completely. No way. But it’s going to fizzle out and people are going to kind of get sick of having to decipher if something’s real or not. And, you know, there might be some backlash in that regard. But also it’ll — it just won’t be as popular. It won’t be the thing that everyone’s talking about and so on and so forth in some years. I don’t think — when React came out, not everyone was building with React, it was like the most popular thing. And this has happened with quite a few different softwares and languages. Something comes out and everyone’s totally preoccupied with it. And then it just kind of fizzles out and it doesn’t stop existing. It’s still a tool that people use, but it’s not the hot new thing that everyone has to jump on. And I think right now AI is just the thing that everyone’s jumping on and everyone’s talking about because everyone else is jumping on it and talking about it. And it’s a new thing and it’s very interesting. And there’s a lot to explore and experiment with when it comes to that. And so people are just preoccupied at this point in time, but I don’t think it’s going to be a forever thing. I do think that AI is going to be integrated into our lives permanently in kind of big ways. But we’ve seen with the self-checkout kiosks at stores that they still need humans to maintain them. You know, the computers and programs can’t exist without us. We can unplug them.

Jonathan:
It’s interesting, my oldest is 16 and the next oldest is 12 and they’re both really into music and art and it feels a little counterintuitive to me but I actually feel optimistic. My daughter who’s 12 is still pretty like against AI, like just in general, not in principle. And when I think about their futures, I’m encouraging them to continue in creative arts. I actually feel kind of optimistic about that because on the one hand, at a pure technical level, the ability to like generate songs and do all this stuff is mind blowing, like how good it already is. But there’s still something about being able to play a keyboard, play the guitar, to be able to like — expression is also going to change. So I don’t know if I’m being naive, but somehow I feel like it’s a good choice to go into the arts at this point. So it’s gonna be interesting to see how that plays out.

Amy:
That’s very interesting. I mean, my 12 year old has shown an interest in animation and I have a master’s degree in animation from UCLA. So I taught her how to make stop motions and edit — basic animation basics. So she and her friends have been making animations. They’re so cute. I wouldn’t like take them to DreamWorks, you know, but their age — actually it’s pretty good for their age. There’s something really satisfying about making art, writing, dancing, drawing, playing music. There’s that satisfaction and offloading that to AI deprives us of that satisfaction. And I think that people at some point — and or maybe even some people now such as myself — you know, you miss the mark if you’re offloading your creative process to something else to do because the point of making something is to make something. It’s not so someone else sees the thing you made. It’s not to have the final product. The process of making the thing is part of it. And yeah, using tools — and I remember when digital video was a new thing back in the day, in the before times, before digital video was a thing, I was doing — I was taking photographs and printing the photographs myself in a dark room and, you know, Super 8 film and stuff like that. Digital video was already a thing at that point, but I just liked doing that. Not a huge thing yet. And there’s a lot of kind of transferring from analog to digital and society had a big upheaval about that, especially with music. When music went online and art could be — music could be duplicated rapidly and widely. Musicians still make music, you know, even if their music is leaked, they still make it because of the satisfaction of making it.

Jonathan:
Luke and I are both going through university college classes and courses. One of the ones — so we’ll talk about this — several of my classes involved a fair amount of writing, and it struck me a few times. I could very easily generate a satisfying output that would achieve the grade that I’m after in it. And I don’t, for a few reasons. But the big part of it for me, because I’m going through these classes because I actually want to learn the subjects. And it is in the process — like last night I was writing for a philosophy class, a piece on ethics. And it took a while to get my brain warmed up enough to like do what I wanted it to do. But that process of creating that piece of work also helped transform some of my own thinking and connect the dots, which I wouldn’t get if I just had a model do it for me. So that’s the value in it for me.

Amy:
I think that we as a species are at risk of brain atrophy, where — if we just offload, even offloading our mental health, you know, to AI, which has been proven repeatedly already to be dangerous or life threatening. If we don’t exercise our ability to think and use our brains and to question and to find answers and come up with ideas and so forth, we are at risk of becoming dumber, much dumber. Because you know, reading and doing art and playing music and all of these things actually build neural pathways and make you smarter and sharper and are actually quite healthy for you. And if we offload all of that and stop doing that, as a species, we’re going to get stupider. I hope not, I hope I’m wrong, but that’s a real risk.

Luke:
We’re going to continue the trend of getting stupider.

Amy:
Right, right, it’s not just that.

Luke:
Yeah. I — so it’s not just that either. It’s like the whole environmental concerns and there’s a whole bunch of issues around AI. I feel like we’re at this pivotal moment where we have to decide. We have to make a boomer decision. All right. We lay so much at the feet of the boomers — you destroyed the environment, corporate greed, corruption, whatever, you know, like they did good things too, but you know, we blame the boomers for a lot of things. I’m talking millennials and gen X, and I feel like we’re at risk of doing that, making the same mistakes as they did and having gen Z, gen alpha look at us and go, those millennials, they wrecked the world with their AI. And you look at the stats, like there was a Gallup poll recently that showed a huge increase in fear towards AI from Gen Z and a huge decrease in excitement. You know, Gen Z is — and like we’ve all just subjectively just said with our own children, they’re not excited, and Gen Alpha too, not excited about AI in the same way we are. So I think we really got to keep that in mind when we’re making decisions about how we structure a project, an open source project. And what sort of contributions we do and don’t allow.

See how I’m moving all of this back to WordPress? It’s very crafty and clever.

Jonathan:
I was thinking as you guys are talking that I could probably train a pretty convincing model on everything that Luke has ever said, but it still wouldn’t be Luke. It still wouldn’t be Luke.

Amy:
It’s very true.

Jonathan:
Amy, thanks so much for joining us. For folks who are interested in learning more about you and the work that you’ve been doing, what’s the best way for them to find you?

Amy:
Well, I would encourage everybody to make a wordpress.org profile and hop onto the make.wordpress.org Slack and come into the hosting channel, which is just hashtag hosting, and to say hi. Our meetings are every Wednesday at 0900 and 1800 UTC. That’s 11 AM Pacific Standard Time, which is my time zone. That’s the meeting that I lead and then my European counterparts and Asian counterparts lead the other meeting. So if folks want to chat with me, I’m there. I’m lurking in that channel.

Jonathan:
Excellent. Well, thank you so much for joining us and sharing your perspective.