Whiskey Web and Whatnot

The authoritative voice of AI, programming, and the modern web. Also whiskey.

217: Open Source, Agents, and the Next AI Wave w/ Angie Jones

Recorded live at All Things Open, Robbie Wagner and guest co-host Jason Lengstorf talk with Angie Jones about AI, coding models, and the evolving landscape of open source. They unpack the importance of community-driven tooling, skill gaps in using AI, the futu...

Creators and Guests

RobbieTheWagner
Jason Lengstorf
Angie Jones

Show Notes

Recorded live at All Things Open, Robbie Wagner and guest co-host Jason Lengstorf talk with Angie Jones about AI, coding models, and the evolving landscape of open source. They unpack the importance of community-driven tooling, skill gaps in using AI, the future of agentic systems, and why open communities matter in shaping technology.

In this episode:

  • (00:00) - Intro
  • (01:18) - Wine rating & review
  • (03:30) - Hot Take: Claude vs ChatGPT
  • (04:20) - Hot Take: Is HTML a programming language?
  • (04:56) - Hot Take: Is CSS a programming language?
  • (05:31) - Hot Take: Is AI to blame when people do dumb shit with it?
  • (09:50) - How open source shaped Angie’s career
  • (14:54) - The problem with AI training on open source data
  • (23:02) - What is MCP and why does it matter?
  • (25:10) - Using MCP when you don’t follow every part of the spec
  • (26:25) - Are we building new walled gardens inside MCP?
  • (31:16) - Building DevRel in a rapidly-changing AI landscape
  • (36:06) - Conference season stress & preparing new talks
  • (39:41) - Where to find Angie online

Links

Connect with Jason

Connect with Angie

Connect with the hosts

Subscribe and stay in touch

Whiskey Web and Whatnot Merch
Enjoying the podcast and want us to make more? Help support us by picking up some of our fresh merch at https://whiskey.fund.

Episode Transcript

Intro: Welcome to Syntax. Welcome to a brand new episode of the Front End Happy Hour podcast. Welcome to this week’s JS Party. Live from Ship Shape Studios, this is Whiskey Web and Whatnot. With your hosts, Robbie the Wagner, and me, Charles William Carpenter III. That’s right Charles. We drink whiskey and talk about web development.

Intro: I mean, it’s all in the name. It’s not that deep. This is Whiskey Web and Whatnot. Do not adjust your set.

Robbie Wagner: Hey,

Robbie Wagner: what’s up everybody? Welcome to Prosecco Programming and Patter.

Robbie Wagner: You might have heard of Whiskey Web and whatnot, but we’ve changed today. , We’ve got some, some lovely Prosecco. We have guest host Jason Langsdorf and our guest Angie Jones. How’s it going? Oh, okay. So you’ve been on the show before, but for anyone who has not heard that one and has not heard of you, could you please give a few sentences about who you are and what you do?

Robbie Wagner: Yeah,

Angie Jones: Yeah, so I’m Angie Jones. I am VP of engineering over AI Tools and Enablement at Block.

Angie Jones: worked in various roles throughout tech over the decades, but yeah, I love, teaching and speaking and talking about tech.

Robbie Wagner: Nice, nice. All right, so the thing we’re all here for, we’ve got this Prosecco. It is a hundred percent Prosecco.

Robbie Wagner: And it is from Costco, so you know, it’s good. Yeah, let’s get into it.

Angie Jones: Bubbly,

Robbie Wagner: I guess I should be pouring this here. I don’t know if you could even hear it if I did. Oh, you kind of

Robbie Wagner: can

Angie Jones: Oh,

Jason: Yeah, there it is.

Jason: Cheers.

Jason: You

Jason: got

Jason: a toast for us?

Robbie Wagner: I do.

Robbie Wagner: Okay.

Robbie Wagner: To MCP, the protocol that finally made our ais talk to each other instead of just hallucinating alone. Cheers to context, connection, and controlled chaos.

Angie Jones: love that.

Angie Jones: Oh, Alright,

Robbie Wagner: so this one tastes remarkably more like wine than the last one.

Jason: Yeah,

Jason: last time we had bubbly juice.

Jason: This time it tastes like wine and it tastes like Costco wine.

Robbie Wagner: Is that high praise or

Angie Jones: ah, yeah.

Angie Jones: May

Jason: are we starting

Robbie Wagner: a headache? Yeah. Okay. yeah. Not a lot of flavor notes or anything to write home about. , It’s a Prosecco.

Robbie Wagner: Alright. Our rating system is zero to eight tentacles, zero being the worst, eight being the best.

Robbie Wagner: Four and a half-ish being middle of the road. What do you think, Jason?

Jason: know.

Jason: I, I think I gotta be, , maybe more gentle because Prosecco is not typically my favorite.

Jason: But this one

Jason: in particular, it’s a little metallic.

Robbie Wagner: So the thing about Italian wine though, is that it all has to be like approved by some guy who says, this tastes like Prosecco, check mark. yes, it’s DOC labeled.

Robbie Wagner: I ripped it all off. But,

Robbie Wagner: yeah.

Jason: Well then

Jason: my apologies to

Jason: Italy cause I’m rating this at two.

Angie Jones: oh my God.

Robbie Wagner: oh, what do you think,

Robbie Wagner: Angie?

Angie Jones: You know, I like,

Angie Jones: , My Prosecco and wines on the sweeter side, and this does give me those sweeter notes, so I

Angie Jones: really like that.

Angie Jones: I usually will drink something like this as a mimosa with some orange juice. So I’m thinking about it in that regard. And it would, I think it would, I think it would be really good. So I’m gonna go with like,

Angie Jones: maybe like a four

Angie Jones: and a

Angie Jones: half.

Robbie Wagner: Okay. Alright.

Robbie Wagner: Yeah. I’m gonna say, I’m gonna say five.

Robbie Wagner: It’s, it’s

Robbie Wagner: a Prosecco.

Angie Jones: It’s Prosecco.

Robbie Wagner: Prosecco.

Robbie Wagner: Simple Pro.

Robbie Wagner: We

Robbie Wagner: could mix the other one in, which is kind of like

Robbie Wagner: a

Jason: That’s it there. Yeah, then it’s, then it’s kind of a

Angie Jones: Then I get a mimosa.

Robbie Wagner: Alright.

Robbie Wagner: got

Robbie Wagner: a few lukewarm takes here. Claude or

Robbie Wagner: GPT

Angie Jones: as

Angie Jones: far

Angie Jones: as

Angie Jones: models.

Robbie Wagner: Yeah, no context at all as to when you would be picking. Which, just which one? Oh, wow.

Angie Jones: I can’t

Angie Jones: go

Angie Jones: the, it

Angie Jones: depends

Angie Jones: route.

Robbie Wagner: No, You can, I’m,

Robbie Wagner: I’m just kidding. I’m just

Angie Jones: okay. Okay. So it

Angie Jones: definitely depends on what I’m trying to do. So if I’m trying to code, do anything like really technical, I’m gonna grab Claude all day. , If I’m having a discussion, I’m trying to brainstorm something like that, I’m gonna go with GPT.

Angie Jones: You know, they say the GPT five is pretty good at coding. , I don’t know. I haven’t tried it. I don’t like it so much for even discussions. So I definitely don’t wanna try it

Angie Jones: for coding.

Angie Jones: , But yeah,

Angie Jones: , That’s

Angie Jones: my

Angie Jones: take. Okay.

Robbie Wagner: Okay. Yeah, I’ve not had much success with GPT and coding. It gives me a

Robbie Wagner: lot of trash. Yeah.

Robbie Wagner: Yeah.

Angie Jones: yeah.

Robbie Wagner: alright, So,

Robbie Wagner: I won’t

Angie Jones: waste my time.

Robbie Wagner: Yeah.

Angie Jones: Yeah.

Robbie Wagner: Is H-T-M-L-A programming language?

Angie Jones: No.

Angie Jones: It’s a

Angie Jones: language,

Angie Jones: it’s a

Angie Jones: markdown Language.

Angie Jones: There’s specific criteria that a programming language has. Things like conditional statements or iterations and things like that,

Angie Jones: that

Angie Jones: does

Angie Jones: not exist.

Angie Jones: And HTML. There’s some structure. Yes, there’s some rules, yes,

Angie Jones: but not quite

Angie Jones: programming.

Angie Jones: That

Angie Jones: does

Angie Jones: not

Angie Jones: mean

Angie Jones: I’m a

Angie Jones: gatekeeper.

Angie Jones: Let the record show.

Robbie Wagner: Okay,

Robbie Wagner: so follow up is CSSA programming language, You know

Angie Jones: that one I would’ve said no a while back, but they have gotten. Really deep. And I’m not a CSS wiz

Angie Jones: at

Angie Jones: all, so I

Angie Jones: don’t

Angie Jones: even

Angie Jones: cooking

Angie Jones: these

Angie Jones: days. They

Angie Jones: are cooking. like

Angie Jones: there’s a lot that you can do, or other people

Angie Jones: can

Angie Jones: do.

Angie Jones: I

Angie Jones: cannot do.

Angie Jones: But there’s

Angie Jones: a lot

Angie Jones: you

Angie Jones: can do

Angie Jones: with CSS, you know, nowadays as far as

Angie Jones: like logic

Angie Jones: and

Angie Jones: hierarchy

Angie Jones: and stuff

Angie Jones: like

Angie Jones: that. So,

Angie Jones: I don’t know.

Robbie Wagner: now?

Jason: they got

Jason: functions. they

Jason: got

Jason: if else

Jason: now.

Jason: Okay.

Jason: Like it’s, we’re, I think we’re

Jason: touring

Angie Jones: at this.

Robbie Wagner: I think so. Yes. All

Angie Jones: Alright.

Robbie Wagner: You heard it

Robbie Wagner: here first.

Robbie Wagner: Is AI to blame when people do dumb shit with it?

Robbie Wagner: Like,

Robbie Wagner: I don’t know. I guess, , like should there be guardrails or, I don’t know. I don’t know what, what the context is,

Robbie Wagner: but you know.

Angie Jones: you know?

Angie Jones: I

Angie Jones: found myself this weekend,

Angie Jones: uh,

Angie Jones: telling

Angie Jones: some

Angie Jones: reply

Angie Jones: guy

Angie Jones: skills

Angie Jones: issue.

Angie Jones: You

Angie Jones: know,

Angie Jones: like

Angie Jones: I, I

Angie Jones: do

Angie Jones: think that

Angie Jones: this is a

Angie Jones: non-deterministic tool.

Angie Jones: AI is

Angie Jones: right?

Angie Jones: Yeah. This is non-deterministic.

Angie Jones: So when you say it

Angie Jones: did

Angie Jones: something

Angie Jones: wrong, like

Angie Jones: What does that mean? Like there

Angie Jones: there is no

Angie Jones: right.

Angie Jones: You

Angie Jones: know

Angie Jones: what I mean? ,

Angie Jones: And

Angie Jones: so the

Angie Jones: way that

Angie Jones: you

Angie Jones: get

Angie Jones: it to perform

Angie Jones: what you want it to do is

Angie Jones: very

Angie Jones: much

Angie Jones: so

Angie Jones: a

Angie Jones: human

Angie Jones: skill.

Angie Jones: Mm-hmm.

Angie Jones: Right?

Angie Jones: Yep.

Angie Jones: ,

Angie Jones: And so

Angie Jones: I would

Angie Jones: say

Angie Jones: it’s

Angie Jones: not the

Angie Jones: AI’s fault,

Angie Jones: , But I don’t wanna blame humans either. Like this is new to all of us. We’re all trying to figure out how do

Angie Jones: we work with non-deterministic.

Angie Jones: Very

Angie Jones: powerful.

Angie Jones: Systems now,

Angie Jones: right? Yeah. And so

Angie Jones: it is a skill issue. I, I wasn’t trying to be

Angie Jones: mean, but

Angie Jones: it

Angie Jones: really

Angie Jones: is a

Angie Jones: skill issue.

Jason: Well, I, I think too, it is really important to keep in mind that like AI can’t be blamed because AI is a tool.

Jason: Like if you, if you hit yourself

Jason: in the

Jason: head

Jason: with a

Jason: hammer,

Jason: it’s

Jason: not the

Jason: hammer’s

Jason: fault,

Robbie Wagner: but should there have been a warning on the hammer that said, if you hit

Jason: Well, and

Jason: and this is sort of the question is like, you know, is it, necessary to put a, the contents hot label on a cup of coffee or things like that? And, you know, it, it sounds silly like that McDonald’s, , lawsuit is always pointed to as like the frivolous lawsuit.

Jason: But if you actually

Jason: look into it,

Jason: that coffee was so dang hot that it gave that woman, like, it fused her skin together. It was so hot, right? So it wasn’t

Jason: really

Jason: a frivolous lawsuit.

Jason: But it, it does kind of point to like, okay, where does responsibility end? Like if, if I have a

Jason: knife in

Jason: my house, do I need to have a posted placard on my front door that says caution sharp things contained within this house.

Jason: and I think , the same is kind of true with digital tools like LLMs are

Jason: not

Jason: intelligent,

Jason: We don’t have. They, well, they’re anthropomorphized, which is a really dangerous thing because

Jason: we think of them

Jason: as people,

Jason: despite them just

Jason: being complicated

Jason: calculators. And so if you go to your complicated calculator that’s been explicitly prompted like in the GPT-4 era

Jason: to be a sycophant, and you say, I don’t like what you said to me, tell me something else.

Jason: And it says, you’re absolutely right. Here’s the thing you wanted to hear.

Jason: And

Jason: then people are like, oh, it’s

Jason: the AI’s fault. It’s

Jason: like, no, it’s not

Jason: like if, like this is the, this is the the billionaire problem. If you only surround yourself with people who tell you what you want to hear, and they give you bad advice because you push everybody who gives you good advice away,

Jason: that is not their fault.

Jason: That’s your fault. Right. And if you won’t listen to a tool and it tells you the thing you already wanted it

Jason: to say,

Jason: that’s still

Jason: on you.

Robbie Wagner: yeah,

Angie Jones: That’s spicy.

Angie Jones: Jason.

Jason: I don’t even

Jason: think that’s spicy. That’s like, like if I, if I hold my calculator

Jason: upside

Jason: down and

Jason: write boobs,

Jason: my

Jason: calculator is

Jason: not being inappropriate. Right.

Robbie Wagner: Yeah,

Robbie Wagner: I don’t know. It’s a hard problem because I feel like there’s not enough education around, like for the casual, non-technical person, they just pick up some AI and they’re like, oh, this is like super smart. It knows what it’s talking about. You tell it to make you a plan and that plan sounds

Robbie Wagner: smart.

Jason: this is the thing,

Jason: that I think needs a warning is I think whenever you start a ChatGPT 4 session, ? It should come with an acknowledgement that says like before you start this, you know

Jason: this

Jason: isn’t

Jason: alive, This

Jason: is a calculator. You

Jason: are using

Jason: a

Jason: tool.

Jason: This is not a person, it’s not a friend, it’s not

Jason: a conscious entity.

Jason: It’s a tool.

Jason: And that to me seems like a reasonable thing because it is masquerading as a, conscious person,

Robbie Wagner: Yeah.

Angie Jones: But it’ll tell you,

Angie Jones: it’ll say I’m

Angie Jones: not

Angie Jones: intelligent,

Angie Jones: I’m

Angie Jones: not human.

Jason: so well sometimes

Jason: until you ask it, like, do you have feelings? And it’s like, I have feelings. And it, like, it’ll do, it’ll do stuff that’s really interesting

Jason: where

Jason: it

Jason: it, you know,

Jason: it puts on,

Jason: its, its

Jason: anthropomorphized

Jason: face.

Angie Jones: My

Angie Jones: ChatGPT Don’t

Angie Jones: play that

Angie Jones: Like it knows. Stay

Angie Jones: your place.

Angie Jones: You are not

Angie Jones: me.

Angie Jones: You are not one

Angie Jones: of

Angie Jones: us.

Robbie Wagner: Yeah.

Robbie Wagner: All right, so we’re at all Things Open. So I want to ask open source things. , How has Open source influenced your

Robbie Wagner: life?

Angie Jones: Oh

Angie Jones: my

Angie Jones: God.

Angie Jones: I

Angie Jones: think

Angie Jones: it

Angie Jones: gave

Angie Jones: me

Angie Jones: a great

Angie Jones: , Pathway

Angie Jones: to,

Angie Jones: I

Angie Jones: guess like

Angie Jones: work

Angie Jones: out loud,

Angie Jones: When you’re working at a company on enterprise software, like you can’t

Angie Jones: really

Angie Jones: get into the nitty gritty of like.

Angie Jones: What

Angie Jones: you’re building

Angie Jones: or

Angie Jones: the techniques and things that you all using just a little

Angie Jones: bit, right? You can share

Angie Jones: a

Angie Jones: little bit, but not the whole story and open source.

Angie Jones: ,

Angie Jones: Once I started working on

Angie Jones: open source

Angie Jones: projects that

Angie Jones: gave

Angie Jones: me something that I can publicly talk about, , and, you know, talk with

Angie Jones: others about,

Angie Jones: you

Angie Jones: know,

Angie Jones: now

Angie Jones: we have like

Angie Jones: this common thing that we use or we invest

Angie Jones: in

Angie Jones: or we

Angie Jones: love or we hate,

Angie Jones: or whatever,

Angie Jones: you know.

Angie Jones: But that gave

Angie Jones: us something,

Angie Jones: , To talk about in the public domain, right? And

Angie Jones: so,

Angie Jones: that

Angie Jones: was

Angie Jones: really big for

Angie Jones: me because

Angie Jones: it gave me a platform and a stage and, you know, , the

Angie Jones: opportunity

Angie Jones: to

Angie Jones: make myself known in this space.

Angie Jones: But

Angie Jones: more

Angie Jones: than

Angie Jones: that, like I actually

Angie Jones: studied,

Angie Jones: ,

Angie Jones: Open

Angie Jones: source

Angie Jones: while doing my master’s.

Angie Jones: We had a open source, class.

Angie Jones: And

Angie Jones: so

Angie Jones: I

Angie Jones: learned

Angie Jones: a lot

Angie Jones: about

Angie Jones: just

Angie Jones: like

Angie Jones: governance

Angie Jones: of projects and, , maintainers and, how to be

Angie Jones: nice

Angie Jones: to

Angie Jones: contributors and

Angie Jones: welcome it and, how not to do this with your

Angie Jones: project.

Angie Jones: For those listening, I’m doing my arms like I’m hoarding

Angie Jones: something.

Angie Jones: basically, it was

Angie Jones: different.

Angie Jones: Like, you know, I grew up in the era of like Microsoft software where. Everything was so proprietary no way could you like look at it, , let alone contribute to it. And, you know, everything

Angie Jones: was

Angie Jones: a pay to play

Angie Jones: model.

Angie Jones: And so to see like open source come and like really shake that up and just change

Angie Jones: the minds in how

Angie Jones: we

Angie Jones: software

Angie Jones: that’s been

Angie Jones: I

Angie Jones: don’t know,

Angie Jones: like

Angie Jones: just groundbreaking

Angie Jones: for me

Angie Jones: personally.

Jason: I mean, and

Jason: not, and

Jason: not

Jason: just

Jason: shake

Jason: it

Jason: up, but

Jason: like

Jason: almost

Jason: completely extinguished. The

Jason: other model. Yeah.

Jason: Like even the, even the hardcore closed source people are still, they’re using and often shipping

Jason: open source, like Microsoft was the

Jason: number one opponent. And here they are.

Jason: They’re like

Jason: the champions

Jason: of open source. these days. Yes.

Robbie Wagner: Yeah.

Angie Jones: oh,

Angie Jones: so

Angie Jones: Yeah,

Angie Jones: that’s exactly

Angie Jones: right.

Angie Jones: and that,

Angie Jones: that’s been really interesting to

Angie Jones: watch.

Angie Jones: So

Angie Jones: much

Angie Jones: so

Angie Jones: that

Angie Jones: they

Angie Jones: acquire GitHub,

Robbie Wagner: and

Robbie Wagner: NPM

Angie Jones: and

Angie Jones: NPM and NPM. So, yeah. That’s been really interesting to watch. I, I love the open source community, I work on an open source,

Angie Jones: ,

Angie Jones: AI agent now named Goose, and I

Angie Jones: did

Angie Jones: a

Angie Jones: talk

Angie Jones: a

Angie Jones: couple of weeks ago.

Angie Jones: Just

Angie Jones: talking

Angie Jones: about the open source nature

Angie Jones: of

Angie Jones: it

Angie Jones: and

Angie Jones: how important open source is right now when

Angie Jones: we

Angie Jones: look

Angie Jones: at like

Angie Jones: the agentic

Angie Jones: AI movement,

Angie Jones: so

Angie Jones: we’re starting

Angie Jones: to see

Angie Jones: these kind of

Angie Jones: walled

Angie Jones: gardens again, where you have proprietary companies

Angie Jones: that are

Angie Jones: controlling both the models

Angie Jones: as

Angie Jones: well as the

Angie Jones: agent.

Angie Jones: call

Angie Jones: me

Angie Jones: a

Angie Jones: conspiracy

Angie Jones: theorist,

Angie Jones: but

Angie Jones: if

Angie Jones: I’m

Angie Jones: gonna benefit from the agent taking the long way to solve your problem because

Angie Jones: you’re

Angie Jones: spending

Angie Jones: more money

Angie Jones: on

Angie Jones: tokens.

Angie Jones: I don’t

Angie Jones: know. I might bake

Angie Jones: that

Angie Jones: into

Angie Jones: Yeah.

Angie Jones: You

Angie Jones: know

Angie Jones: what

Angie Jones: I

Angie Jones: mean?

Angie Jones: I don’t

Angie Jones: know.

Angie Jones: I’m

Angie Jones: not saying

Angie Jones: anybody’s

Angie Jones: doing that,

Angie Jones: but just things that cross my mind, you know what I mean? And so, , I think that even

Angie Jones: right

Angie Jones: now

Angie Jones: as

Angie Jones: we are like.

Angie Jones: Looking at ai, we’re trying to figure it out, is really important for the open source players to have a seat

Angie Jones: at

Angie Jones: this table

Angie Jones: and to help, basically guide where we’re

Angie Jones: going,

Angie Jones: which is

Angie Jones: what I think

Angie Jones: that, , goose does, right? So Goose is not , the most used agent

Angie Jones: or

Angie Jones: anything

Angie Jones: like that, but I’ll tell you what it

Angie Jones: is.

Angie Jones: It

Angie Jones: was

Angie Jones: the

Angie Jones: reference implementation for

Angie Jones: MCP from

Angie Jones: the client

Angie Jones: side,

Angie Jones: and

Angie Jones: because it’s open

Angie Jones: source.

Angie Jones: The community decides

Angie Jones: what Ag

Angie Jones: agentic AI will be for them because

Angie Jones: we

Angie Jones: have

Angie Jones: the

Angie Jones: most

Angie Jones: permissible

Angie Jones: license. People can fork it, people can contribute to it, and , they contribute the features they want to see.

Angie Jones: And

Angie Jones: a

Angie Jones: lot of

Angie Jones: that is

Angie Jones: transparency.

Angie Jones: and

Angie Jones: this forces

Angie Jones: the other big players to then adopt that. So Goose usually

Angie Jones: is

Angie Jones: the

Angie Jones: first one that’s

Angie Jones: gonna

Angie Jones: have things

Angie Jones: like.

Angie Jones: Huh? Why are we stuck with one model and one provider? What

Angie Jones: if

Angie Jones: we

Angie Jones: opened

Angie Jones: this

Angie Jones: up

Angie Jones: and you bring

Angie Jones: any model, you

Angie Jones: want to this, right?

Angie Jones: People

Angie Jones: weren’t

Angie Jones: doing that, and now

Angie Jones: you’re starting

Angie Jones: to see this is now being forced because consumers are saying, Hey, I don’t want just one model.

Angie Jones: I might want GPT for my planning and reasoning,

Angie Jones: and I

Angie Jones: want to Use

Angie Jones: Claude when

Angie Jones: it’s time to execute.

Angie Jones: Or

Angie Jones: even

Angie Jones: a cheaper local model for execution

Angie Jones: or

Angie Jones: something like that, right? And so, , the community,

Angie Jones: the

Angie Jones: open

Angie Jones: source

Angie Jones: community is the one that is pushing,

Angie Jones: what

Angie Jones: agen AI will be. And I, I’m

Angie Jones: proud that

Angie Jones: Goose is a

Angie Jones: one of the

Angie Jones: vehicles

Angie Jones: that

Angie Jones: allows them

Angie Jones: to do that.

Robbie Wagner: So somewhat related to open source and ai. Lemme think how I can say this. Like, you know the thing recently, everyone’s been all.

Robbie Wagner: Like, oh, if you can’t take art for free and use it in ai, then it’s like AI would die like you. We need all this free stuff. How do you think that relates to like code and like all the code it’s getting from GitHub and , is there a problem there and how do we like

Robbie Wagner: have

Robbie Wagner: some sort of funding back from that or something to keep sustaining open source?

Robbie Wagner: ‘cause if there’s nothing to pull from, then where does it get its data from?

Angie Jones: That’s

Angie Jones: a

Angie Jones: hard

Angie Jones: problem.

Angie Jones: I am not a fan

Angie Jones: of.

Angie Jones: How these models

Angie Jones: were born

Angie Jones: and trained on,

Angie Jones: ,

Angie Jones: Data,

Angie Jones: essentially without people’s

Angie Jones: permission.

Angie Jones: Right. I,

Angie Jones: don’t

Angie Jones: like that

Angie Jones: the

Angie Jones: world

Angie Jones: seems to have decided though, that

Angie Jones: this is

Angie Jones: so

Angie Jones: valuable

Angie Jones: that we are willing to

Angie Jones: overlook

Angie Jones: that

Angie Jones: you

Angie Jones: stole

Angie Jones: all

Angie Jones: of the

Angie Jones: data

Angie Jones: Yeah.

Angie Jones: From

Angie Jones: the internet

Angie Jones: to

Angie Jones: build these

Angie Jones: tools.

Angie Jones: Right?

Angie Jones: Yeah.

Angie Jones: So

Angie Jones: I

Angie Jones: don’t know.

Angie Jones: where we go from here on. I

Jason: there’s some precedent for ways to make this stuff work. Because basically what we’re looking for is, if a private company is gonna go into a public space and do something that benefits that company, it shouldn’t be solely for the company’s benefit. And so like an example of, of one way that we’ve seen this done is in Vancouver, for example, if you’re gonna build a building, you have to include a certain portion of that building that is affordable housing, because they’re worried that if people come in and just build luxury condos,

Jason: it’s

Jason: gonna ruin the city

Jason: because there won’t, you know, it won’t be possible for people to live there.

Jason: You know,

Jason: look at San Francisco for example. so , the people building. Can either put in that percentage of, of affordable housing or they can fund public works like a public park, a public library, something like that. So in some way, the company has to allocate some of its profit toward benefiting the community

Jason: that it’s going to benefit from. Right. And so one of the things I would love to see, and this is where I think the, like why they’re fighting so hard against regulation. I think that, , they should be taxed at a higher rate and that that should go back into funding, art, education, funding, actually paying for these programs that they benefit from.

Jason: if they need that to exist, if art is so important that if we made them pay for it, the company folds,

Jason: kind of

Jason: sounds like,

Jason: they need to build that into their profit

Jason: model. Right? because

Jason: otherwise,

Jason: I mean, the other, the other outcome is like, if we don’t do this, then they’re gonna continue to use this stuff for free.

Jason: Which is gonna disincentivize anybody from doing it because they know their work’s gonna get stolen, which means the input

Jason: suffer, which

Jason: means that eventually

Jason: this

Jason: dries up

Jason: and everybody’s far worse off for it.

Jason: So

Jason: this is

Jason: one of those things where, well, and

Jason: I

Jason: mean this is sort of the thing I’m worried about is that we, we’re very bad as an industry, as a country, as a, as a species at thinking about second and third order effects of things.

Jason: And the first order effect of AI is that everybody gets access to all this great information really fast. The second order effect is that it disincentivizes literally everybody else from creating because they know that their work is gonna get chewed up by the AI and they won’t get compensated for it.

Jason: So the third order effect is we stop getting good information and the AI’s are frozen at whatever level they were at,

Jason: if the companies are smart, then putting 10% of their profit into sustaining the creativity that makes their product good is a really wise investment. But I don’t

Jason: think

Jason: they’re smart.

Jason: I think they’re greedy

Jason: and I think

Jason: they’re,

Jason: not going

Jason: to do it. And

Jason: I think

Jason: it’s gonna cost

Jason: us a lot in the long run Unless we get regulation.

Robbie Wagner: when they first started, Sam Altman was all,

Robbie Wagner: oh

Robbie Wagner: yeah. Like this is

Robbie Wagner: a non-profit and like

Robbie Wagner: all

Robbie Wagner: this

Robbie Wagner: bullshit

Jason: I mean between, between like Sam Alton and Peter Thiel and Mark Andrees, and I think

Jason: we’ve realized

Jason: that the people in

Jason: charge are

Jason: not

Jason: good people.

Robbie Wagner: Yeah.

Robbie Wagner: Yes.

Robbie Wagner: Yeah,

Robbie Wagner: ‘cause I think if you did that and you took like all the money made and you put it in funds for like, if you put us all out of work, we get some of that money.

Jason: I, that’s,

Jason: the, I

Jason: think

Jason: this is,

Jason: sorry, this

Jason: is your

Jason: episode. I’m gonna rant.

Angie Jones: no, go.

Jason: the the thing that makes me the most mad is when you see an interview with these folks and they say something like, well, yeah, so AI is gonna make it so that people don’t have to work. we are gonna make it so that you live these lives of leisure and it’s gonna drive the prices down.

Jason: And I’m like, motherfucker, have you ever, ever

Jason: seen

Jason: a

Jason: company

Jason: voluntarily, Lower prices? Have you ever seen a

Jason: society make it possible for somebody to

Jason: exist

Jason: without having to make money?

Angie Jones: And the same people who are saying this

Angie Jones: are the

Angie Jones: same

Angie Jones: people

Angie Jones: that

Angie Jones: you

Angie Jones: know

Angie Jones: vote

Angie Jones: against.

Jason: right?

Angie Jones: The types of programs that help

Angie Jones: people

Angie Jones: in

Angie Jones: need. So

Angie Jones: now

Angie Jones: all

Angie Jones: of

Angie Jones: a

Angie Jones: sudden we’re all,

Angie Jones: in need and we think

Angie Jones: that these

Angie Jones: billionaires are going to say,

Angie Jones: oh,

Angie Jones: you

Angie Jones: know what?

Angie Jones: Here’s a

Angie Jones: couple of

Angie Jones: million

Angie Jones: dollars,

Angie Jones: Jason, go live

Angie Jones: your

Angie Jones: best life.

Jason: Yeah.

Jason: Like we’ve, we’ve literally never seen Peter

Jason: Thiel part

Jason: with anything to help somebody

Jason: else out. Like he is

Jason: out to amass the

Jason: most he can.

Jason: And if he wins the anti-regulation stuff and he, he becomes one of the, you know, the prime investors who has all the AI bucks coming in and he becomes the, you know, one of the world’s first trillionaires.

Jason: He’s not

Jason: gonna turn around

Jason: and pay for me

Jason: to go to art

Jason: school.

Jason: Yeah.

Jason: Right.

Jason: He’s gonna find a

Jason: way to monetize

Jason: my therapist.

Jason: Like,

Jason: you know what

Jason: I mean?

Jason: Like, he’s gonna

Jason: make it

Jason: so I

Jason: gotta

Jason: pay

Jason: to

Jason: go

Jason: outside.

Jason: Like,

Robbie Wagner: Yeah.

Jason: it’s, it’s, it’s just very frustrating that we’re, we’re saying like, oh no, don’t, don’t regulate innovation. Don’t regulate innovation. And then the promise is if we have unbounded regulation or unbounded innovation,

Jason: it’s

Jason: gonna

Jason: benefit

Jason: society.

Jason: We have. Zero proof, zero precedent, no plan, no promise on how they’re actually going to make anybody’s lives better. So the, end result most likely is that wealth disparity.

Jason: increases. I

Jason: don’t even

Jason: know

Jason: what

Jason: happens

Jason: next,

Jason: because

Jason: if we, if we don’t, well, I mean , the

Jason: last time we had

Jason: wealth

Jason: disparity

Jason: this bad,

Jason: right?

Jason: We

Jason: had, it was like the Great Depression, the end of the gilded age. The last

Jason: time

Jason: we saw that

Jason: in Europe,

Jason: they started cutting people’s heads off like it’s

Jason: historically,

Jason: this

Jason: does

Jason: not

Jason: end

Jason: well

Jason: for the very, very rich people.

Jason: Yeah. Right. And So interest, again, lack

Jason: of

Jason: understanding,

Jason: of second

Jason: and third

Jason: order

Jason: effects. Yeah.

CTA: This just in! Whiskey.fund is now open for all your merch needs. That’s right, Robbie. We’re hearing reports of hats, sweaters, and T-shirts, as well as a link to join our Discord server. What’s a Discord server? Just read the prompter, man. Hit subscribe. Leave us a review on your favorite podcast app and tell your friends about our broadcast. It really does help us reach more people and keeps the show growing. All right, back to your regularly scheduled programming.

Angie Jones: I

Angie Jones: almost feel like people

Angie Jones: think,

Angie Jones: companies

Angie Jones: think

Angie Jones: that

Angie Jones: they can’t even

Angie Jones: afford

Angie Jones: to

Angie Jones: think

Angie Jones: about that.

Angie Jones: Like

Angie Jones: the

Angie Jones: race

Angie Jones: is so

Angie Jones: tight

Angie Jones: on

Angie Jones: a global scale

Angie Jones: of.

Angie Jones: we

Angie Jones: all know we are in a

Angie Jones: bubble. It’ll burst.

Angie Jones: Some

Angie Jones: will

Angie Jones: fall, some will stand.

Angie Jones: So everybody’s racing to be

Angie Jones: the one that stands

Angie Jones: in morals,

Angie Jones: ethics,

Angie Jones: everything

Angie Jones: is

Angie Jones: like,

Angie Jones: I have

Angie Jones: no time

Angie Jones: to think about that.

Angie Jones: I

Angie Jones: just gotta

Angie Jones: win the

Angie Jones: race.

Angie Jones: Yeah.

Angie Jones: You

Angie Jones: know,

Angie Jones: that’s

Angie Jones: what

Angie Jones: I

Angie Jones: feel

Angie Jones: like,

Angie Jones: that’s

Angie Jones: where

Angie Jones: we are.

Jason: ,

Jason: And

Jason: honestly,

Jason: this is to,

Jason: to

Jason: bring this

Jason: away

Jason: from the precipice of

Jason: a,

Jason: of

Jason: like

Jason: the most depressing episode of Whiskey Web and whatnot ever made. I

Jason: I I do wanna say this is one of the things that I really admire about you, specifically, Angie, is that

Jason: when

Jason: I saw

Jason: this coming up, my instinct was to back away.

Jason: I was like, I don’t want to be involved in this, because this is gonna be exactly what

Jason: you just.

Jason: Said,

Jason: and

Jason: you

Jason: dove

Jason: right in.

Jason: I looked at this

Jason: and I was like,

Jason: this is gonna be exhausting. These are not the people that I want to be surrounded with.

Jason: This is not the culture that I want to be a part of. And you looked at it and you said, i’m not gonna

Jason: not

Jason: be in that

Jason: room.

Jason: Like, I’m not gonna

Jason: let them make

Jason: decisions

Jason: without us

Angie Jones: they will.

Jason: I

Jason: admire

Jason: you.

Jason: because I don’t,

Jason: I don’t have the, whatever it is that, makes you strong enough to do that.

Jason: ‘cause I, I’m

Jason: like,

Jason: whatever, just

Jason: put me

Jason: in

Jason: the,

Jason: put

Jason: me in

Jason: the

Jason: data

Jason: mine, I

Jason: guess.

Jason: I’ll just work away when, when all the jobs are gone.

Robbie Wagner: Yeah. Let’s, let’s pivot back a little. Um,

Robbie Wagner: let’s

Robbie Wagner: talk about MCP in general. Still ai, but a little lighter. What is

Robbie Wagner: MCP.

Angie Jones: Okay,

Angie Jones: so MCP stands for Model Context Protocol. It’s a open standard that was developed by anthropic. It’s pretty

Angie Jones: new,

Angie Jones: less

Angie Jones: than a

Angie Jones: year old.

Angie Jones: It

Angie Jones: was

Angie Jones: released in November, 2024.

Angie Jones: So what

Angie Jones: MCP does

Angie Jones: . All the stuff we complained about in the early days of, , using AI where they’re hallucinating a lot and just kind of making things up and stuff like that.

Angie Jones: The reason they do that is ‘cause they don’t have context. So I mean, it’s literally they’re not failing. That’s how they were designed, right, is to generate a response, predict what should be the next word, like that. Is what they do, right? And

Angie Jones: so

Angie Jones: when

Angie Jones: you

Angie Jones: don’t

Angie Jones: have context

Angie Jones: in front

Angie Jones: of you.

Angie Jones: Yes,

Angie Jones: you maybe pick the wrong thing.

Angie Jones: , MCP

Angie Jones: provides

Angie Jones: a

Angie Jones: way

Angie Jones: to

Angie Jones: get

Angie Jones: your

Angie Jones: models

Angie Jones: context.

Angie Jones: And

Angie Jones: with that,

Angie Jones: the

Angie Jones: models now

Angie Jones: also

Angie Jones: have function calling abilities.

Angie Jones: Couple

Angie Jones: those two.

Angie Jones: And

Angie Jones: now

Angie Jones: you get to

Angie Jones: agents,

Angie Jones: right?

Angie Jones: Ag agentic ai,

Angie Jones: ,

Angie Jones: Where

Angie Jones: you can actually

Angie Jones: build

Angie Jones: software that can

Angie Jones: talk

Angie Jones: to

Angie Jones: these models.

Angie Jones: And then take actions on your behalf.

Angie Jones: And

Angie Jones: so

Angie Jones: in order for

Angie Jones: an agent to take action on

Angie Jones: your behalf, one, they need to

Angie Jones: be connected

Angie Jones: to

Angie Jones: your systems

Angie Jones: and

Angie Jones: they

Angie Jones: need

Angie Jones: data and

Angie Jones: context.

Angie Jones: So

Angie Jones: MCP is

Angie Jones: a

Angie Jones: standard

Angie Jones: that

Angie Jones: provides

Angie Jones: a

Angie Jones: way

Angie Jones: for

Angie Jones: the

Angie Jones: agents to get that context right. And.

Angie Jones: It’s, a nice, clean way so that you don’t have to do integrations with every single

Angie Jones: API

Angie Jones: that you might want your AI to talk to.

Angie Jones: Instead, if that

Angie Jones: API, is exposed via an MCP server,

Angie Jones: for example,

Angie Jones: you

Angie Jones: can just

Angie Jones: plug that into

Angie Jones: any agent. And MCP is,

Angie Jones: , LLM agnostic. It’s agent agnostic, so it truly is, this open

Angie Jones: standard and

Angie Jones: it’s

Angie Jones: been great

Angie Jones: for.

Angie Jones: us

Angie Jones: getting value

Angie Jones: out of ai.

Robbie Wagner: I’m

Robbie Wagner: trying to equate this to, so like in Ember, what I know, you used to have like serializers for your data, right? Or in anything where you need to munch some data, you’ll have a serializer.

Robbie Wagner: And if anyone didn’t quite get the, , standard right, the spec, right, they didn’t quite follow the spec. They have a lot of problems. So I’m wondering if that is a problem with MCP, if like some people are doing a little bit of their own flavor, not following the spec, like, and it doesn’t work everywhere or like what kind of challenges are you seeing around that?

Robbie Wagner: Or is it

Robbie Wagner: a challenge? I

Robbie Wagner: don’t know.

Angie Jones: It’s

Angie Jones: pretty simple. Um, , the spec is wide and so you don’t have to implement like every part of the spec, but

Angie Jones: there’s

Angie Jones: like certain parts you definitely do if you’re gonna. Like work

Angie Jones: at all. I guess most common part

Angie Jones: of that

Angie Jones: is

Angie Jones: the

Angie Jones: tools part of

Angie Jones: the spec. So

Angie Jones: as

Angie Jones: long as you follow

Angie Jones: that, you should be good.

Angie Jones: So

Angie Jones: your server might not have,

Angie Jones: all

Angie Jones: of

Angie Jones: the other

Angie Jones: pieces, but

Angie Jones: the tools part is critical in

Angie Jones: order

Angie Jones: for

Angie Jones: it

Angie Jones: to

Angie Jones: work. So as long as

Angie Jones: you

Angie Jones: follow

Angie Jones: that

Angie Jones: part,

Angie Jones: and

Angie Jones: it’s

Angie Jones: not like really hard,

Angie Jones: it’s

Angie Jones: like.

Angie Jones: Okay,

Angie Jones: describe

Angie Jones: the

Angie Jones: name like this, the description, and then what

Angie Jones: do you wanna do in the body?

Angie Jones: So

Angie Jones: very much like a regular function. So there’s

Angie Jones: syntax

Angie Jones: there of course,

Angie Jones: but

Angie Jones: , Other than

Angie Jones: that,

Angie Jones: yeah,

Angie Jones: it should be good to go.

Jason: I think I saw you, you post about this recently where you were saying that people are, are going out there saying, oh, this is our MCP agent for x.

Jason: Our agent four y are, are you seeing that, is that marketing or is that, are, are people trying to build

Jason: like little

Jason: walled gardens

Jason: inside of

Jason: MCP?

Angie Jones: Okay,

Angie Jones: so

Angie Jones: it, it’s the servers that they’re seeing that for. So they build a MCP server

Angie Jones: and a MCP server.

Angie Jones: Let’s say you make a MCP server

Angie Jones: for,

Angie Jones: let’s

Angie Jones: just

Angie Jones: say YouTube. All right.

Angie Jones: So I got MTP server for YouTube,

Angie Jones: which means if I

Angie Jones: plug this

Angie Jones: into,

Angie Jones: ,

Angie Jones: Any agent.

Angie Jones: Then it should be able to interact with YouTube

Angie Jones: on

Angie Jones: my

Angie Jones: behalf, meaning logged

Angie Jones: in

Angie Jones: as me, Angie, and

Angie Jones: do

Angie Jones: whatever I needed to do, right? So I can say something

Angie Jones: like,

Angie Jones: oh, find y’all the videos on

Angie Jones: MCP and make

Angie Jones: me

Angie Jones: like,

Angie Jones: , A playlist or something like that, right?

Angie Jones: Anyway,

Angie Jones: so

Angie Jones: if I

Angie Jones: were to make that server

Angie Jones: and

Angie Jones: I

Angie Jones: said, Hey.

Angie Jones: I

Angie Jones: made a

Angie Jones: new YouTube MCP server

Angie Jones: for

Angie Jones: goose,

Angie Jones: or

Angie Jones: for claw

Angie Jones: code,

Angie Jones: or

Angie Jones: something like

Angie Jones: that. Sounds silly.

Angie Jones: ,

Angie Jones: And the reason

Angie Jones: it

Angie Jones: sounds

Angie Jones: silly

Angie Jones: is because

Angie Jones: the

Angie Jones: MCP servers are supposed to

Angie Jones: be

Angie Jones: agent

Angie Jones: agnostic,

Angie Jones: meaning

Angie Jones: like,

Angie Jones: as

Angie Jones: long as

Angie Jones: it, the agent adheres to the spec, your

Angie Jones: MCP server

Angie Jones: works

Angie Jones: with all

Angie Jones: of

Angie Jones: them.

Angie Jones: So

Angie Jones: can’t see DODs, uh.

Jason: the, that’s

Jason: sort of

Jason: the

Jason: underlying question is

Jason: like,

Jason: are you

Jason: seeing,

Angie Jones: people say

Angie Jones: that I don’t.

Angie Jones: know

Angie Jones: if it’s marketing

Angie Jones: or

Angie Jones: just a misunderstanding.

Angie Jones: So for example,

Angie Jones: If

Angie Jones: you use claw code,

Angie Jones: right?

Angie Jones: You

Angie Jones: might

Angie Jones: make

Angie Jones: an MCP server, you tried it

Angie Jones: with

Angie Jones: claw code and you say, yeah, I’m

Angie Jones: just

Angie Jones: gonna

Angie Jones: advertise

Angie Jones: it

Angie Jones: as this

Angie Jones: ‘cause I

Angie Jones: haven’t even tried

Angie Jones: it with the

Angie Jones: rest of the agents.

Angie Jones: I don’t know if it

Angie Jones: works or

Angie Jones: not,

Angie Jones: or,

Angie Jones: you

Angie Jones: know, call code is popular.

Angie Jones: you tag that

Angie Jones: onto yours for SEO value.

Jason: Yeah, know Why

Angie Jones: they do

Angie Jones: it, but it upsets me because, uh, Kenzie dos, he, he quote tweeted that in, in that

Angie Jones: my

Angie Jones: tweet

Angie Jones: and

Angie Jones: said,

Angie Jones: that’s like

Angie Jones: saying I

Angie Jones: made a website for

Angie Jones: Google

Angie Jones: Chrome.

Angie Jones: And

Angie Jones: that’s

Angie Jones: exactly

Angie Jones: what

Robbie Wagner: it, like website

Angie Jones: works

Angie Jones: on

Angie Jones: any

Angie Jones: browser.

Angie Jones: Who well, and, and this

Jason: is, this

Jason: is

Jason: the

Jason: part that

Jason: always had

Jason: me worried is because

Jason: in the

Jason: old

Jason: days,

Jason: right, there

Jason: was like internet

Jason: Explorer and Netscape, and you

Jason: would

Jason: run

Jason: into the

Jason: websites

Jason: that were like,

Jason: oh, you’re on Netscape.

Jason: You need to open

Jason: Internet

Jason: Explorer

Jason: to

Jason: see

Jason: this

Jason: one. Yeah.

Jason: Right.

Jason: And so,

Jason: That was my That’s kind of true to see the same

Angie Jones: thing

Angie Jones: happen.

Angie Jones: Okay.

Angie Jones: Well that’s

Angie Jones: kind

Angie Jones: of true.

Angie Jones: ,

Angie Jones: Like I was talking about with the spec. So if I implement tools, yes. I work in like pretty much any.

Angie Jones: these agents that adhere to the spec are called MCP clients, right? So

Angie Jones: I

Angie Jones: work with any client if I support tools.

Angie Jones: Now,

Angie Jones: there’s other features within the spec that you can add to your MCP server, but

Angie Jones: that

Angie Jones: doesn’t mean

Angie Jones: all of the agents support those yet, right?

Angie Jones: For example,

Angie Jones: MCP

Angie Jones: UI is something

Angie Jones: that’s, relatively new, is

Angie Jones: getting A lot

Angie Jones: of community,

Angie Jones: Interest in it,

Angie Jones: but

Angie Jones: it’s where

Angie Jones: you

Angie Jones: know how

Angie Jones: you’re talking to your agent and it’s just text-based. Right? And maybe they even ask you a question saying like, do you want me to do option one, option two, or option three?

Angie Jones: And you have to type

Angie Jones: that in, right?

Angie Jones: That would be

Angie Jones: better

Angie Jones: as

Angie Jones: like

Angie Jones: a radio, a

Angie Jones: set of

Angie Jones: radio button chat by something like that. Right? So something you can interact with. So that’s one of, the features that’s being proposed to the spec. And of course it’s so new that a lot of agents have not adopted that yet.

Angie Jones: So if you MCP server has that, then yeah, you might.

Angie Jones: Say this does not work with this particular agent. In fact, goose is like the only one that has implemented that. Again, because it’s open

Angie Jones: source. A community member added that,

Angie Jones: not

Angie Jones: The core team. Right.

Angie Jones: And so,

Jason: So

Jason: it’s, it’s

Jason: early

Jason: browser days. We’re,

Jason: we’re inventing things. The features are,

Jason: We

Jason: haven’t, we haven’t unified on like a a W three C for, for MCP oversight.

Angie Jones: Okay.

Jason: Well,

Jason: that’s

Jason: what I’m, oh, that’s,

Jason: kind of what

Jason: I’m asking. ‘cause like, you know, right now the browsers have the, the

Jason: sort of like

Jason: core group that’s like, these are the

Jason: features.

Jason: We’ve, we’ve as a group decided these are gonna ship and then all the browsers will, will

Jason: put an

Jason: intent

Jason: to ship. Yeah. Yeah.

Jason: And that’s the signal that like, yeah, it’s coming so you can safely use this.

Jason: Whereas other ones are like, well, only Google is gonna ship this, so, you know, safari’s not gonna

Angie Jones: this. Yeah. don’t

Angie Jones: use this in

Angie Jones: your

Jason: website

Jason: unless you’re

Jason: comfortable

Jason: with

Jason: non-standard stuff.

Angie Jones: I

Angie Jones: think we’re,

Angie Jones: we’re kind of there. I would say that it is adopted

Angie Jones: and

Angie Jones: everybody

Angie Jones: has done

Angie Jones: the whole

Angie Jones: intent to ship

Angie Jones: thing.

Jason: Okay. haven’t gotten to it

Angie Jones: yet. Right.

Angie Jones: Okay. well, no,

Angie Jones: that’s, that’s

Angie Jones: actually great to hear. Yeah. Yeah, yeah. Oh yeah.

Angie Jones: I think it’s pretty mature for, to be

Angie Jones: such

Angie Jones: a.

Angie Jones: A

Angie Jones: young model.

Angie Jones: There’s, , so

Angie Jones: much

Angie Jones: energy around

Angie Jones: it. There’s

Angie Jones: like all of these working groups and things that are like really focused on,

Angie Jones: ,

Angie Jones: On making

Angie Jones: sure

Angie Jones: that this

Angie Jones: is,

Angie Jones: a big

Angie Jones: boy standard.

Angie Jones: You know?

Jason: Is

Jason: it my question?

Robbie Wagner: Oh, it can be. Yours

Jason: Because I feel like It’s

Jason: a good

Jason: follow up

Jason: here. one of the things I was really curious about is, because you’ve wade into this space that is brand new and you’re working in, enablement, you’re running the Derel team for Goose, , you’re doing a lot of the outreach stuff. How are you building a, , educational and advocacy program around

Jason: concepts that

Jason: like didn’t exist

Jason: last month?

Jason: because, you know, like Dere, it feels like took many years to mature to the point. there

Jason: was

Jason: anything

Jason: resembling like

Jason: a

Jason: standard or a plan

Jason: that you could follow. So how have you approached this and, and how are you looking at this space and with how quickly it’s moving to build a program that is

Jason: resilient in the face of how

Jason: fast

Jason: everything

Angie Jones: You know, I get a lot of kudos

Angie Jones: about

Angie Jones: my team and how

Angie Jones: they are. Essentially teaching the masses about this brand new technology, and it goes far beyond our product, right? So we’re like teaching concepts that are applicable to our competitor, like any, product. And that’s totally fine, right? I think that’s how it should be done.

Angie Jones: but I’ve always been one to kind of work on the cutting edge and then bring everybody along with me. I feel like that.

Angie Jones: My purpose in tech. I’m pretty good at like figuring out new complex stuff and then breaking it down so that we all understand together. I shouldn’t be the only one who gets this, so we’re, we’re building.

Angie Jones: Out loud

Angie Jones: and learning

Angie Jones: together. So it’s very obvious. This is new. We’re not experts. We’re very clear about that. , And we say, but hey, I tried it out. This is what I’ve learned. This is what I learned, doesn’t work so great. And just sharing those lessons. And because, , we work in DeVere, we have. The privilege of working with this stuff day in and day out.

Angie Jones: Whereas other people who are working on like production software, engineering teams, like may not be able to dive that deep ‘cause they have other deliverables they need to meet. Right. the other part of my role is, , internally with our engineering org. And so I work with them. To, they use all sorts of tools, goose claw cursor, they use ‘em all.

Angie Jones: So I, I’m very aware of everything that’s out there and what works and what doesn’t work in real enterprise environments. Right. And so, ,

Angie Jones: my

Angie Jones: job is to

Angie Jones: help them go beyond just using like code generation, but to embed,

Angie Jones: ,

Angie Jones: AI

Angie Jones: and agents into

Angie Jones: like.

Angie Jones: Their repos into ci, into like all the things, right? And so, I see

Angie Jones: a

Angie Jones: lot,

Angie Jones: I learn a lot, , on real,

Angie Jones: production, software development.

Angie Jones: And so I feel it’s my duty

Angie Jones: to

Angie Jones: like

Angie Jones: share

Angie Jones: those

Angie Jones: lessons,

Angie Jones: ,

Angie Jones: More

Angie Jones: broadly

Angie Jones: externally.

Jason: And so

Jason: so

Jason: your

Jason: strategy

Jason: is

Jason: kind of

Jason: just

Jason: go all in on

Jason: learning in

Jason: public. Oh yeah.

Angie Jones: Go

Angie Jones: all

Angie Jones: in

Angie Jones: on learning probably. I like,

Angie Jones: I don’t know any of this and you

Angie Jones: don’t either.

Robbie Wagner: You gotta

Robbie Wagner: learn.

Robbie Wagner: I’m one for

Angie Jones: the team.

Angie Jones: here’s what I learned,

Angie Jones: kind of thing. Yeah,

Robbie Wagner: Yeah, I think that’s a great way to do it because something you’ve just learned, you’re gonna be able to explain

Robbie Wagner: better ‘cause

Robbie Wagner: that person’s

Robbie Wagner: in the same exact situation you were in a minute ago.

Robbie Wagner: Like. They wanna learn it. You just learned it.

Robbie Wagner: Perfect.

Jason: And it’s, it’s

Jason: kind of the only way to get out of, , what I think the major hangup for me with a lot of the AI stuff for so long was that every

Jason: discussion, every

Jason: demo was theoretical. People

Jason: were just telling me

Jason: what they

Jason: thought it would do or, or what they

Jason: think it could

Jason: solve. And I was like.

Jason: Build it.

Jason: Show me, like, show me anything.

Jason: Like, I don’t, I don’t, believe you because you

Jason: keep telling

Jason: me

Jason: and

Jason: nobody’s

Jason: showing me and

Jason: it

Jason: doesn’t

Jason: seem real.

Angie Jones: I did, uh, the Dere Khan. ,

Angie Jones: So

Angie Jones: it’s funny.

Angie Jones: All my friends, Jason

Angie Jones: included.

Angie Jones: Laugh about

Angie Jones: Dere Kahan, and

Angie Jones: they’re like,

Angie Jones: what is

Angie Jones: that?

Angie Jones: Like

Angie Jones: y’all just a

Angie Jones: bunch of dere

Angie Jones: like

Angie Jones: selling

Angie Jones: your products to

Angie Jones: each other. But it’s,

Angie Jones: it’s,

Angie Jones: not

Angie Jones: that

Angie Jones: at all

Angie Jones: Dere, Kahan is very much about the craft of developer relations and like how to do these sorts of things, right?

Angie Jones: And so I did the keynote there and I talked about how Dere is now on its biggest stage yet. And I talk about like, this is a new movement where. All engineers, no matter, like what they’re doing, have like a mandate from like their CEOs that they have to learn these tools and these concepts and these software engineering patterns.

Angie Jones: and

Angie Jones: so

Angie Jones: this

Angie Jones: just goes way beyond

Angie Jones: your

Angie Jones: product or

Angie Jones: your tool.

Angie Jones: Like

Angie Jones: your,

Angie Jones: everything

Angie Jones: that you’re doing is uplifting the entire software engineering community. You know what I mean? And so,

Angie Jones: ,

Angie Jones: Y’all should

Angie Jones: check that

Angie Jones: talk out.

Angie Jones: If you have a, have some time. And

Jason: And just

Jason: to

Jason: be

Jason: clear,

Jason: I don’t

Jason: think

Jason: DeVere Con is

Jason: Devereux

Jason: selling

Jason: other

Jason: DeVere sinks. Oh, Okay.

Jason: Okay. That’s what I think

Jason: every other

Angie Jones: Oh,

Angie Jones: Oh,

Angie Jones: yes.

Robbie Wagner: You’re not

Robbie Wagner: wrong.

Robbie Wagner: Speak.

Angie Jones: Shame. Shame.

Robbie Wagner: Oh. Speaking of

Robbie Wagner: other

Robbie Wagner: conferences, you’ve been, uh, at a few and we’ll be at a few more,

Jason: one

Jason: or two.

Robbie Wagner: yeah. I wanted to get your take on like, how do we solve this problem that every single conference is like, you know, when’s a good time,

Robbie Wagner: October,

Jason: For,

Robbie Wagner: like,

Robbie Wagner: please, My, my my sleep

Jason: schedule

Jason: is

Jason: begging for

Jason: an

Jason: answer to this question.

Angie Jones: yeah.

Angie Jones: Yeah.

Angie Jones: And it’s draining. I’m on tour

Angie Jones: right now. ,

Angie Jones: I

Angie Jones: literally will

Angie Jones: not see

Angie Jones: my house

Angie Jones: in

Angie Jones: the month of October.

Angie Jones: Um, so it’s

Angie Jones: just like back to back trips. The part that gets me is, , people don’t realize like how much it goes into

Angie Jones: Speaking at a conference, like the major

Angie Jones: part, of course

Angie Jones: is like preparing my talk and

Angie Jones: all

Angie Jones: of that.

Angie Jones: But I mean, each conference is like, oh, can we jump on a call? Oh, can I see your slides? Oh, can we meet, can we do? And it’s like, I have too

Angie Jones: many

Angie Jones: conferences to

Angie Jones: do this. Like, I’m gonna

Angie Jones: show

Angie Jones: up,

Angie Jones: gonna do a

Angie Jones: great

Angie Jones: job,

Angie Jones: And

Angie Jones: then,

Angie Jones: and

Angie Jones: then

Angie Jones: let’s

Angie Jones: just call

Angie Jones: it

Angie Jones: a day. You

Angie Jones: know,

Jason: I hit my breaking point

Jason: with this because I,

Jason: I, you

Jason: and

Jason: I are on,

Jason: on tour together

Jason: here for a little while. I think we

Jason: got like

Jason: five

Jason: events

Jason: over the next few weeks that we’re gonna both be at.

Jason: I was getting those same emails and I,

Jason: kind

Jason: of

Jason: hit my breaking point

Jason: and I just

Jason: sent an

Jason: email that was like,

Jason: listen,

Jason: I’m so

Jason: flattered that you wanted me to be part of this. I’m not doing any of that. Uh. I understand if you’re uncomfortable with that, I will happily yield my

Jason: spot. but

Jason: like, I can’t. I

Jason: can’t,

Jason: I can’t

Jason: do it.

Jason: I’m

Jason: gonna

Jason: show

Jason: up. I’m gonna do

Jason: a

Jason: great

Jason: job.

Jason: I’m

Jason: gonna

Jason: bring

Jason: my

Jason: computer

Jason: and

Jason: I’m gonna

Jason: plug

Jason: it in.

Jason: And

Jason: that’s all it is gonna

Jason: be

Jason: like, I’m not

Jason: sending you slides

Jason: ahead

Jason: of time.

Jason: We’re

Jason: not

Jason: doing

Jason: a content

Jason: review.

Jason: We’re

Jason: not

Jason: doing,

Jason: I

Jason: don’t

Jason: need

Jason: a speaker

Jason: coach. Like,

Jason: I

Jason: like,

Jason: I, I

Jason: appreciate

Jason: all of these resources.

Jason: I’m

Jason: not doing it

Angie Jones: Yeah.

Angie Jones: yeah, that’s

Angie Jones: true. I mean, and right

Angie Jones: now, I mean,

Angie Jones: developer

Angie Jones: advocates

Angie Jones: Are

Angie Jones: mostly

Angie Jones: the ones that people are paying to come to this,

Angie Jones: right? Yeah.

Angie Jones: Like a lot of these

Angie Jones: conferences. They’re community driven or whatever. They

Angie Jones: don’t

Angie Jones: have the budget

Angie Jones: to like

Angie Jones: fly everybody

Angie Jones: in

Angie Jones: and pay. You know,

Angie Jones: so companies

Angie Jones: are sponsoring

Angie Jones: that,

Angie Jones: which

Angie Jones: means

Angie Jones: that

Angie Jones: those people

Angie Jones: are overbooked.

Angie Jones: right?

Angie Jones: And

Angie Jones: so

Angie Jones: they

Angie Jones: don’t, like

Angie Jones: someone

Angie Jones: asked

Angie Jones: for

Angie Jones: my

Angie Jones: slides. It’s like

Angie Jones: I

Angie Jones: got five

Angie Jones: conferences

Angie Jones: before

Angie Jones: I get to your conference.

Angie Jones: Like,

Angie Jones: I

Angie Jones: do not have,

Angie Jones: You’re,

Angie Jones: I

Jason: can’t

Jason: think your

Jason: slides will

Jason: exist

Jason: roughly 36 hours before your yeah. phenomenal,

Angie Jones: promise you that, you know?

Robbie Wagner: Yeah. Yeah. that’s tough. I do think it is helpful on the flip side, if you’re like speaking at one conference and you haven’t spoken before, you might want those resources. So I

Jason: Well, why they do it, that’s

Jason: exactly

Jason: what it is.

Jason: And I

Jason: I, think

Jason: the part

Jason: that’s,

Jason: that’s tricky

Jason: is like

Jason: if I’m running

Jason: the

Jason: event.

Jason: I

Jason: wanna

Jason: make

Jason: sure

Jason: that the speakers

Jason: are,

Jason: are ready. I

Jason: wanna make sure that that,

Jason: you know, they’re

Jason: not gonna

Jason: show

Jason: up

Jason: and do something that violates the code of conduct. If they’re

Jason: just gonna stand up

Jason: there

Jason: and like do a 30 minute ad for their company.

Jason: I want to have a talk with them beforehand. So I, I know from the, conference organizer’s perspective

Jason: why

Jason: every

Jason: single one

Jason: of those steps

Jason: exist. it is no judgment on the conference organizers themselves. and with Angie it’s like, it’s Angie’s job.

Jason: I

Jason: don’t get

Jason: paid

Jason: to speak.

Jason: Right? Like I get paid when I work on code TV stuff. So this for me is,

Jason: is because

Jason: I really like to speak

Jason: and

Jason: there’s

Jason: networking opportunities, but like,

Jason: it’s kind

Jason: of

Jason: a cost center

Jason: for

Robbie Wagner: yeah.

Jason: Like it’s

Jason: almost a marketing

Jason: expense

Jason: for me. So

Jason: if you add

Jason: six hours of

Jason: meetings,

Jason: I

Jason: just

Jason: can’t, I

Jason: can’t afford

Jason: to do it.

Jason: Right.

Robbie Wagner: they need speaker pre-check. Like if you’re on the list, you don’t have to do the meetings.

Robbie Wagner: Oh

Robbie Wagner: my God,

Robbie Wagner: that’s a great

Robbie Wagner: idea.

Angie Jones: I love that

Jason: I love

Angie Jones: idea.

Jason: a,

Jason: We

Jason: should make

Angie Jones: Yeah, yeah,

Angie Jones: yeah.

Angie Jones: yeah.

Robbie Wagner: Okay,

Robbie Wagner: I think that’s about all we had. Uh, what would you like to plug or where can people find you, et cetera?

Angie Jones: yeah,

Angie Jones: so

Angie Jones: check me out. , My website is Angie jones.tech.

Angie Jones: All my social channels are

Angie Jones: there.

Angie Jones: I hang

Angie Jones: out

Angie Jones: mostly

Angie Jones: on

Angie Jones: LinkedIn

Angie Jones: these days,

Angie Jones: ,

Angie Jones: As

Angie Jones: well

Angie Jones: as Twitter.

Angie Jones: I’m gonna call it Twitter for life.

Angie Jones: Hmm. cesspool,

Angie Jones: but

Angie Jones: hey, you

Angie Jones: know,

Angie Jones: um,

Angie Jones: I can, it, it’s my cesspool. I can I can find some good content there, but yeah,

Angie Jones: I really highly encourage if you are a software engineer right now, you are, find yourself in the boat that I’m talking about where you’re being kind of pushed to learn how to use these tools effectively.

Angie Jones: Definitely come outside of your company. There’s so many other people who

Angie Jones: are doing this.

Angie Jones: Who

Angie Jones: are learning out

Angie Jones: loud and

Angie Jones: sharing

Angie Jones: the

Angie Jones: things

Angie Jones: that

Angie Jones: they’re

Angie Jones: learning. So you wanna

Angie Jones: kind

Angie Jones: of

Angie Jones: tap into the community right

Angie Jones: now. It’s a good time to do that.

Robbie Wagner: All

Robbie Wagner: right. Uh,

Robbie Wagner: thanks for coming on Angie. Thanks everyone for listening. If you’d like to please subscribe, leave us some ratings and reviews. We appreciate it, and we will catch you next time.

Outro: You’ve been watching Whiskey Web and Whatnot. Recorded in front of a live studio audience. What the fuck are you talking about, Chuck? Enjoyed the show? Subscribe. You know, people don’t pay attention to these, right? Head to whiskey.fund for merchant to join our Discord server. I’m serious, it’s like 2% of people who actually click these links. And don’t forget to leave us a five star review and tell your friends about the show. All right, dude, I’m outta here. Still got it.