What happens when the model CAN'T fix it? Interview w/ software engineer Landon Gray [Podcast #213]

[00:00] Welcome back to the freecodecamp [00:01] podcast. I'm Quincy Larson, teacher and [00:03] founder of freecodecamp.org and today [00:05] I'm interviewing software engineer [00:07] Landing Gray who's famous for proving [00:09] that rag pipelines can be written in [00:11] Ruby and for popularizing Ruby as a [00:13] language for building machine learning [00:15] projects. Before we hear from him, some [00:17] community news. Freecodecamp has [00:18] published a new YouTube course that will [00:20] teach you beginner front-end development [00:22] skills like HTML, CSS and JavaScript. [00:24] You can code along at home and build a [00:26] variety of projects, your own [00:27] interactive quiz game, a currency [00:29] converter and even a Trello-style Kanban [00:32] board. Along the way you'll learn how to [00:33] use APIs and local storage to extend the [00:35] functionality of these bite-size apps. [00:37] It's a 12-hour 19-project-long [00:40] course free on the freecodecamp YouTube [00:42] channel. Check it out after you finish [00:44] this tutorial. We also published a [00:46] course that will teach you how to test [00:48] your software [00:49] and ensure it doesn't break when you add [00:51] new features. Prolific freecodecamp [00:53] instructor Beau Carnes teaches this [00:54] course. He'll introduce you to the [00:56] testing pyramid and show you how to [00:57] balance fast unit tests against complex [01:00] end-to-end user journeys. You'll also [01:01] learn how to automate some of the [01:02] testing using an open-source library [01:05] called Playwright and LLM testing tool. [01:08] It is a 1-hour course on the [01:09] freecodecamp YouTube channel. We also [01:11] published a tutorial that will help you [01:15] with the fact that more and more apps [01:17] are relying on probabilistic [01:19] LLM output alongside deterministic API [01:22] calls. This makes life harder for devs [01:24] who now need to ensure that [01:25] hallucinations don't escape to end [01:27] users. [01:28] Freecodecamp has published this advanced [01:30] use observability tutorial that will [01:33] teach you emerging best practices and [01:35] architectural patterns for dealing with [01:36] this. It's about a 40-minute read. Check [01:38] it out. Link it in the description and [01:40] learn how to containerize your ML ops [01:43] pipelines. That's machine machine [01:45] learning operations. This tutorial is [01:47] the result of hard-won deployment [01:49] wisdom. The author spent 3 weeks [01:51] debugging a Python library error [01:53] due to dependency conflicts. His [01:55] eventual answer? Containerize the entire [01:57] project with Docker. This tutorial will [02:00] show you how to structure your [02:01] containers with multi-stage builds. [02:03] You'll also learn how to experiment and [02:06] track with ML flow, and you'll learn [02:09] versioning with DVC, GPU passthrough, [02:11] and a bunch of other advanced [02:12] techniques. 40-minute read. [02:14] Link in the description. Song for this [02:17] week is 2006 Everybody by UK producers [02:20] Basement Jaxx. If you're familiar with [02:22] their work, you know you're in for a [02:24] psychedelic, look at silly romp uh [02:26] between the swoons, the bongos, the [02:28] swooning chorus, and the song just feels [02:32] like it's held together by duct tape, [02:34] but it works. I love it. Link in the [02:36] description. [02:38] If you want to support free Code Camp, [02:39] we're a charity. We're tax-exempt. You [02:42] can donate to us, deduct it from your US [02:44] taxes, and you can support our mission. [02:46] donate.freecodecamp.org. [02:49] 10,338 [02:51] kind folks donate each month, and it's a [02:53] huge help. It helps us keep producing [02:54] these open-source learning resources. [02:57] You can also pick up a free Code Camp [02:58] shirt, [02:59] shipped anywhere in the US for $20. If [03:01] you're outside the US, just take our [03:03] assets, screen print them on a shirt, or [03:05] use a like heat transfer something like [03:07] that, make your own free Code Camp [03:08] shirt. Free for personal non-commercial [03:12] usage. [03:13] Check out shop.freecodecamp.org if [03:14] you're in the States you want to order [03:15] one. Now for my interview with Landon [03:17] Gray, he's a software engineer who [03:18] worked at agencies for years. Then he [03:20] taught himself AI-assisted software [03:22] development, and now he's helping other [03:23] devs do the same. He works as an AI [03:25] engineer at an enterprise software [03:27] company and runs a popular newsletter. [03:29] We're going to talk about how large [03:31] language models are just the raw fuel [03:32] and harnesses and harness logic are the [03:35] real engine to get things done. Why [03:37] building your professional network is so [03:39] helpful for finding clients and landing [03:40] job interviews. And why Landon helped [03:42] port Python machine learning libraries [03:44] to Ruby. And what he why he thinks that [03:47] now that AI is just an API call away, [03:49] the Ruby ecosystem is better positioned [03:51] than ever. Landon Gray, welcome to the [03:53] freeCodeCamp podcast. [03:55] Thanks for having me. Happy to be here. [03:58] Well, I'm happy to have you here. Uh [04:00] been really enjoying reading your work. [04:02] Uh you're one of the many developers out [04:04] there who doesn't just keep their heads [04:06] down and do their work, but actually [04:07] shares in public what they're building [04:09] and and contributes to open source and [04:11] gets involved in conferences. So, you [04:14] have moved into AI engineering [04:17] uh as a field. And this is a new field [04:20] where essentially you combine software [04:22] engineering skills with [04:24] leveraging to the fullest all the uh [04:27] LLM-focused tools and and other tools [04:30] that are broadly described as air quotes [04:32] AI, which is a term so vague that it's [04:33] basically meaningless. But uh I wanted [04:36] to ask you [04:37] uh since you're doing this, [04:39] everybody can learn something right off [04:40] the bat. What is a harness [04:43] in the context of large language models? [04:46] Yeah, definitely. So, [04:48] um [04:49] I I posted uh a post on social uh [04:53] reposting about harnesses and [04:56] uh like a harness, when when I think [04:58] about LLMs, it's it's sort of the raw [05:00] fuel. [05:02] Um you can't [05:04] I mean, you can you can chat with it. [05:05] You can you can have a conversation. It [05:07] can do a lot more now. [05:09] Um but [05:11] every company that's using a a large [05:13] language model is using the model as [05:15] fuel and they're building tooling and [05:18] infrastructure around it to get it to [05:21] perform and behave in ways that um [05:25] achieve the results that they're trying [05:26] to achieve. So, so recently um and I [05:29] love this company, [05:31] but I I stopped subscribing to them [05:32] because the pricing tiers have gotten [05:34] kind of wild, but I've been using [05:35] Perplexity a lot for my deep research. [05:38] Um and I had been using Claude also and [05:41] their their research capability. But my [05:44] my my theory is that Perplexity has a [05:46] really great harness. So, they they're [05:48] they may be using Claude, which I'm [05:50] using the raw materials to do research [05:53] and and their tooling. Um But but they [05:55] have a really good infrastructure built [05:57] up around large language models to do [06:00] deep research and um that that [06:02] infrastructure and their product and [06:04] their tooling is the harness in which [06:06] you know, I'm getting better results [06:08] from them than I am the Claude model [06:10] directly. So. Yeah, so So, that might be [06:13] a [06:14] a way to define harness when people use [06:16] this buzzword is essentially it's all [06:18] the tooling around the LLM that somehow [06:21] improves the output and like reduces [06:23] hallucinations or structures the output [06:26] in a way or organizes it or asks [06:30] you know, additional queries and stuff [06:31] to further clarify or um you know, maybe [06:34] in the case of getting code back uh has [06:37] a bunch of like fuzzing that it does and [06:38] things like that. [06:39] >> Yeah. Yeah, exactly. And And so, it's [06:42] interesting because I I feel like this [06:44] race is getting played where you know, [06:47] on one side you have machine learning [06:49] engineers who are working on large [06:51] language models and they're trying to [06:52] make the model as good as possible. And [06:54] then on the other side, you have people [06:56] who are using these models and they're [06:59] mitigating the flaws at how not great [07:01] they are or terrible they are by like [07:04] writing software to kind of constrain um [07:08] sort of the range of capabilities or [07:09] kind of improve the outcomes. Um and I I [07:12] see a lot of that as [07:14] I guess one side of the AI engineering [07:17] coin in a sense. Yeah. Yeah, so a lot of [07:20] people might just be making, you know, [07:22] naked API calls to an LLM and then just [07:24] taking the output at face value and [07:27] thinking this sucks. Like why would [07:28] anybody ever use this? But what they're [07:30] not doing is they're not getting like [07:32] the nice [07:33] I guess force-directed flow that come [07:35] from something coming through a harness [07:37] on its way to your app. Yeah. And and I [07:40] I see a harness as more Have you heard [07:42] of like a GPT wrapper? Yes. It's just [07:46] they have ChatGPT or they have like uh [07:49] GPT-40 or some model, and then they have [07:52] one prompt, and they have a layer on top [07:55] that just serves the response of that [07:57] prompt and uh [07:59] the the model response call, and and I [08:02] see the harness is, "Hey, we're building [08:04] a layer of complexity on the in between [08:06] the between the model and the user of [08:08] software to really get [08:11] stronger results, though." [08:13] Yeah. Well, it's good to know that uh [08:15] like obviously it's extremely expensive [08:17] to improve the underlying foundation [08:19] model. [08:20] Uh it it often takes like hundreds of [08:22] millions of dollars worth of training [08:24] process to get a new version of GPT or a [08:27] new version of Claude. Uh but [08:30] uh a lot of these improvements you can [08:32] do to the harness can be pretty simple [08:35] software changes. Like you might be [08:36] working with a traditional code base, [08:38] and uh you can you can really iterate [08:40] quickly uh without this long life cycle. [08:43] Exactly. And and you can tweak the [08:45] models now [08:47] for a little bit of cheaper, so you [08:48] don't have to like [08:50] like retrain a model, you know, from [08:53] scratch if you want to improve ChatGPT. [08:55] Now we have like like so so between, you [08:58] know, writing software and retraining a [09:01] model from scratch, you have all this [09:02] space in between where so you can you [09:04] can fine-tune the model, [09:06] um and and you can get a lot of really [09:08] really good results that outperform the [09:09] base model, um and it's it doesn't have [09:13] to be as expensive, but that that's just [09:16] like another layer of the I guess the [09:18] stack. [09:19] Yeah. Well, since you're working as an [09:21] AI engineer and you worked as a [09:22] traditional software engineer as well, [09:24] and you've studied uh a lot of machine [09:26] learning to understand what's actually [09:28] going on in the hood uh under the hood. [09:31] For developers who are getting into AI [09:33] engineering, how would you differentiate [09:35] AI engineering from other data-centric [09:37] fields like data science and data [09:40] engineering. [09:41] Yeah, so [09:43] um [09:44] let's start I think when I think about [09:46] data science, I think a lot of [09:49] uh [09:50] there's a lot of focus on some of the [09:52] statistics and you might be looking at [09:56] um [09:57] linear regression, Bayesian sort of [09:59] algorithms and these are these are a lot [10:01] of things that you'll kind of cover if [10:03] you if you study math pretty adept. Um [10:07] and so I think [10:10] I think when I think data science, I [10:12] think of people who have like a really [10:13] strong [10:15] uh like statistical background and [10:18] they're they're they're working to sort [10:20] of solve a solution through like that [10:24] sort of lens. [10:25] Um when I think of data engineering, I [10:27] think of that is sort of the [10:30] uh [10:31] how do I want to say it? Like maybe [10:34] the [10:35] plumber or the preparer of like data. [10:37] Like data data is like a raw material. [10:40] It's not in the form you need it. It's [10:41] not structured as you need it. And so [10:44] uh [10:45] like the data engineer is really working [10:47] to try to build out systems to get the [10:49] data in the form that it's needed and [10:52] and to transfer it and plumb it to [10:54] wherever it's needed so that it's [10:56] reliable for [10:58] uh someone like a machine learning [11:00] engineer or data science so they can [11:01] start looking at it. [11:03] Um like that there's a lot of like I [11:05] won't I won't say it's like [11:07] front where it's probably the most val- [11:08] some of the most valuable work is like [11:10] prepping the data. People spend 80% of [11:13] their time on the data. There's like the [11:14] 80/20 rule. So 80% of your time on the [11:16] data, 20% actually like building the [11:19] models like you know, if you're doing [11:21] like machine learning or or using [11:23] picking the appropriate model. Um so and [11:26] then the data engineering piece data [11:29] engineering is a really uh [11:32] in- interesting uh a field that I feel [11:35] like is very [11:37] not well defined. I think the first [11:40] person who really did a great job at [11:41] like [11:43] trying to like define sort of that world [11:45] is Swyx, and he has an article that [11:48] talks about data engineering. And so [11:49] like there there's all these different [11:51] layers. So so so my entry into to AI [11:54] engineering I might have said data data [11:56] engineering into AI engineering is I [11:59] come from a software background. Like [12:01] I've done a lot of software consulting. [12:03] I worked as a software developer early [12:05] in my career. And so like when I was [12:07] coming towards, you know, AI I was like, [12:10] "Okay, [12:11] I don't yet have the skill set [12:14] and I have never trained a machine [12:16] learning model um yet at that point. And [12:20] so I was like, [12:21] you know, how do I [12:24] Yeah, so so my strength was more [12:25] software side. So how do I kind of [12:26] approach process AI field [12:29] from from that perspective? And so I [12:31] found things that were more adjacent to [12:33] kind of my skill set. Um and then I also [12:36] had to try to reinforce my skill set by [12:39] taking some coursework. I did some [12:41] coursework with UT Austin to kind of [12:43] bolster my skill set, learn how to train [12:46] and and train models and things like [12:48] that. Um but but I I think I was [12:50] deviating from my earlier point. Like [12:52] there's different layers of the stack. [12:53] There were like machine learning [12:54] engineers who can sometimes be [12:57] differentiated from AI engineer. [12:59] Sometimes when a company's saying AI [13:01] engineer they mean sheet machine [13:03] learning engineer. And sometimes they [13:04] mean a software engineer who's really [13:06] good at using the model. And so it's [13:07] it's very confusing. You have to look at [13:09] the job description to see like, "Okay, [13:11] what brand of AI {slash} ML engineer is [13:15] this is this job?" So um [13:18] yeah, I don't I don't think there I [13:20] think there is a good definition. I [13:21] don't think companies necessarily like [13:23] honor it when they're trying to find AI [13:25] engineers. [13:26] So there's a definition which, you know, [13:29] Swyx, I guess, [13:30] uh who's been on the freeCodeCamp [13:31] podcast three times now, most recently I [13:34] actually just interviewed him yesterday. [13:36] Uh and that uh probably came out a few [13:38] weeks before this came out. Uh but uh [13:40] check out my interview with him. Uh but [13:42] he he coined the term AI engineering and [13:45] set up the AI engineer conference [13:48] series, which now has events in Europe [13:51] and Asia and everywhere, right? Mhm. And [13:54] uh basically what you're saying, that [13:56] term hasn't been fully like standardized [13:59] in the eyes of employers. They may have [14:01] kind of like they may slap that label on [14:03] jobs that are not necessarily AI [14:05] engineering roles, but are actually more [14:07] like ML engineering and vice versa, [14:09] depending on the level of sophistication [14:11] of the employer. Um [14:12] >> Yeah. [14:14] Yeah. So, uh UT Austin, real quick, [14:17] University of Texas Austin, one of the [14:19] best public universities in the US and [14:22] by extension the world. Uh I'm from [14:24] Texas and uh did you grow up in Texas, [14:26] too? [14:27] Uh I I grew up in [14:29] uh Rochester, New York, uh upstate New [14:32] York. And yeah, then I ended up moving [14:34] down to Texas. I went to Abilene [14:35] Christian University and got my [14:37] bachelor's in computer science there and [14:40] kind of that's where I got my start in [14:42] my career. Very cool. Well, we're going [14:44] to get into your career in a second, but [14:45] I just have to ask, like why spend all [14:47] this time and energy learning how the [14:49] models actually work under the hood? Why [14:51] learn machine learning like you did when [14:54] for most developers, like using an LLM [14:57] is just making a series of API calls? [14:59] Yeah, time, energy, and money. I I [15:01] dropped a lot on on some of the [15:03] programs. So, for me, it's [15:07] a lot of it comes down to like first [15:09] principles and [15:11] you know, if you want to use these [15:13] tools, you can just use them by, you [15:16] know, [15:17] playing with the models, chatting with [15:19] them all day, and you will get a lot of [15:21] value from that. [15:23] But then you will hit [15:25] potentially a wall and you'll be like, [15:27] "Well, I wanted it to do this." And [15:30] like [15:32] when that moment happens for you, [15:35] you have to dig deeper. Like you have [15:37] you have to understand [15:39] why is the model producing this output [15:42] instead of some other output? And [15:45] it like it's difficult if you don't have [15:48] a foundation to either even know how to [15:50] like debug or think about the problem in [15:52] such a way that you can solve it. Um [15:56] things like reliability and things like [15:58] latency. Like why is the model taking so [16:01] long to get a response? Like [16:03] having a fundamental understanding of of [16:05] how these models are working, [16:07] um which I must say I I'm still [16:09] learning. I'm still growing every day to [16:11] try to understand different layers of [16:13] the stack such as like inference and and [16:16] where these bottlenecks can be. But [16:17] having that understanding [16:19] when you're in the workplace can help [16:21] you find solutions to to really [16:23] difficult problems. [16:26] Yeah, and I will once again quote the uh [16:28] the attract mode from Mortal Kombat 3 [16:30] arcade game. There is no knowledge that [16:32] is not power. [16:34] You guys know, the more you know. Uh [16:36] yeah, like knowing knowing is half the [16:38] battle, right? [16:39] So like [16:40] So it sounds like you pursued these It [16:43] was a huge investment of time, energy, [16:44] money, going back to school, and you [16:46] were able to do this because you have [16:49] faith that in the future this knowledge, [16:52] understanding how these systems actually [16:53] work and not just having it be some [16:55] black box that it's some other guy's job [16:57] to know how it works, right? Like it's [17:00] above my pay grade. I can't stand when [17:01] people say that. Like [17:03] You know what I'm saying? But but it [17:04] it's just the ultimate punt. [17:06] Uh but but you were not satisfied merely [17:07] to just use these tools. You wanted to [17:09] understand how they actually work. Cuz [17:11] this is one of the more important [17:12] innovations of our lifetime most likely. [17:15] Exactly. And I I didn't want to get [17:17] stuck in this moment where you know, [17:20] there there becomes that moment where [17:22] you get stuck and and like for a lot of [17:24] people when they get stuck [17:26] they then ask the model like how to fix [17:29] it and like there are some problems that [17:31] the model will not know how to fix. Like [17:34] and and we I like I [17:36] I've been on teams that have [17:37] experimented with this. So we've had [17:39] really badly latency problems in an [17:41] organization related to So we were [17:44] building a chat application. We had [17:45] really bad latency problems and you [17:49] know, if you're just [17:50] if you're just relying on the model to [17:52] like write your code and to do [17:54] everything, eventually you're going to [17:55] hit a bottleneck and you're going to ask [17:56] it, "Hey, look at all my code, Claude, [17:59] like and and try to fix the latency [18:00] issues." And it won't. Like at the time [18:03] it won't and it didn't and and now what? [18:06] Are you going to tell your CEO like we [18:07] can't fix the latency problems, you just [18:09] have to wait a long time for the [18:11] response to come back to the user or you [18:13] going to think creatively and go back [18:15] to, you know, like people have written [18:17] about latency. Are you going to look at [18:18] those white papers? Are you going to [18:20] explore the internet to find information [18:22] to help you solve the problem cuz the [18:24] model can't solve the problem? Like I [18:26] know we always want the model to solve [18:27] the problem and most of the time it [18:29] does, but that [18:30] when you've reached the edge and it's [18:33] you and your competitor, one of you is [18:35] going to be able to solve it or both of [18:36] you. But yeah, and you want to be able [18:38] to solve it. And so [18:40] um [18:41] you know, I I I really like um [18:44] Kel Kelsey Hightower talks about like [18:46] growing skills and adding to your skill [18:47] belt [18:49] or your tool belt. Like I'm trying to do [18:50] that all the time. Like, you know, like [18:52] I have some skills in machine [18:55] like like using like ML models. I have a [18:58] ton of skills in using LLMs cuz I just [19:01] work in that so much more. [19:03] Um but then I have like a coworker who's [19:05] like really skilled in like, you know, [19:07] using the ML model and you know, like we [19:09] have different competency but like as [19:11] you solve problems, you you grow [19:14] different skill sets, and you can tackle [19:16] different types of problems. [19:18] Um [19:18] Yeah. [19:19] >> Yeah. Engineers are paid to think. [19:21] They're paid to solve problems. They're [19:23] not paid to throw up their hands and [19:25] say, "This is just how it is. There's [19:26] fundamentally there's this latency. [19:28] Nobody knows what's going on." Yeah. [19:29] You're paid to figure out what's going [19:31] on and figure out how to fix it. Right? [19:33] >> Exactly. So, keep your intellectual [19:35] curiosity about you. Your career depends [19:37] upon it. Let's talk about career advice [19:40] because you have had a pretty [19:43] variegated career. You've done a lot. Um [19:45] and you are relatively young. [19:48] Uh and in just your, you know, decade or [19:50] so working as a dev, you've accomplished [19:53] a lot more than a lot of devs have. Uh [19:56] and I want to figure out how that is so [19:57] we can all learn from that. So, first, [20:01] AI assistance, AI-assisted software [20:04] development, uh freeCodeCamp published a [20:05] popular book on that written by Mruguesh [20:07] Mohapatra who runs most of our devops [20:09] and stuff. He's a heavy user of a lot of [20:11] these tools like cloud code. He's got [20:13] fleets of agents getting things done for [20:15] him. [20:16] Uh and pretty much everybody on the team [20:17] does use these tools. [20:19] Um even though, of course, you have to [20:21] be careful with them and they're not [20:23] perfect. [20:24] Um but how does AI assistance change [20:28] software development? Like how how do [20:29] you think it's impacting your career, [20:32] and how are you seeing it [20:33] impact the careers of people around you? [20:36] I think it accelerates everything. Um [20:39] just hyper speed. Uh [20:42] which which people think, "That's good." [20:44] But if you're doing the wrong thing, it [20:46] just accelerates and proliferates that [20:49] like everywhere as well. So, if you if [20:51] if you can if you're writing code or if [20:54] you're using the LLM to write code, [20:56] you know, [20:58] say there's a bad pattern in your system [20:59] and you're using AI-assisted [21:01] programming, that bad pattern's going to [21:03] persist, and you're going to you're [21:04] going to eventually the system that [21:06] you're building's is to crumble because [21:07] you can't [21:09] you know, it's not designed designed [21:11] well. So, so AI assistant programming [21:14] has definitely helped kind of [21:17] you know, practitioners build more, but [21:19] you still have to have the fundamentals. [21:21] You still have to know how to design [21:22] software and think about software, but [21:25] it's more of from like a conceptual [21:27] standpoint where if you can [21:29] if you're in your mind palace, you can [21:31] architect a system and communicate [21:33] exactly what needs to be done [21:35] effectively via prompting, like you're [21:38] you can build that system, but like most [21:40] people [21:41] or many people don't have that [21:43] experience and haven't seen multiple [21:46] different systems. Like, you know, I've [21:48] been to big organizations. I've I've [21:50] been that have billions of dollars in [21:52] revenue. I've been to small [21:53] organizations and worked in small [21:55] organizations, like one or one or two [21:56] people, a couple people, and [21:58] you know, like [22:00] that is what I'm leaning on when I'm [22:02] using these [22:03] when I'm doing AI system programming. [22:05] I'm not just starting with a blank slate [22:07] and being like, "Do this thing." [22:09] Um [22:09] so [22:10] I don't know if I answered your question [22:12] if I deviated, but [22:13] >> Yeah. [22:14] So, so essentially it can accelerate the [22:15] rate at which you make mistakes and it [22:17] can accelerate the decay of a code base [22:20] if you don't know what you're doing. So, [22:21] you do need to know what you're doing at [22:23] the end of the day. [22:24] Uh [22:25] one thing I often say is like if you are [22:27] just using LLMs to write all your code, [22:29] if you're just hitting tab tab tab and [22:30] not really thinking about it, if you're [22:31] just vibe coding everything, [22:33] uh [22:34] why are you not just a completely [22:36] commoditized widget? Where Why can we [22:38] not just drop [22:39] you know, uh low-paid worker from XYZ [22:43] country in there to essentially do what [22:45] you're doing? Why can we not automate [22:46] what you're doing completely? [22:47] >> Mhm. Right? Uh and and those skills, [22:49] those fundamentals, I think are what [22:50] differentiate a dev who's able to wield [22:53] these tools with great efficacy from [22:55] somebody who's just stamping out a bunch [22:56] of uh technical debt all day. [22:59] All right? [22:59] >> Yeah. [23:01] Definitely. Yeah, and [23:03] and and a lot of people are talking [23:04] about, and I agree, it's like you're [23:07] you're coming in to your work with a [23:09] taste, a perspective, and [23:12] you know, we all like hopefully, you [23:14] know, you have strong opinions. What is [23:16] it What's that quote? Strong opinions [23:17] loosely held or something like that? [23:19] Like hopefully you've developed some [23:22] sense of, you know, this is the way I [23:24] think things should be done, and you [23:26] don't have to hold to it, but you should [23:28] have some perspective. Like you should [23:30] say, [23:31] you know, I want to do this might be [23:33] good at this and not at not at this, or, [23:35] you know, like this is how I build build [23:38] systems. Like that's something that you [23:39] you're not going to get from the model. [23:42] Um I I I kind of think that like the [23:43] model's trained on all the internet. Do [23:45] you want the average or the you know, [23:48] the mean of everything everyone said? [23:50] Like no, I want, you know, [23:53] I I want a curated, you know, I like I I [23:56] like following like YouTube streamers or [23:58] or uh Twitch stream. You like they have [24:00] a a very curated taste, opinions. [24:03] Sometimes they're very polarizing, but [24:05] you follow those people because [24:07] it's a perspective, you know? But you [24:09] don't want the average perspective over [24:10] Right. You want to cultivate your own, [24:13] and that's what makes you valuable. [24:15] Yeah. [24:16] >> One one thing that makes you valuable. [24:18] For sure. So, let's talk a little bit [24:20] about the uh job search and uh the [24:23] freelance gig search. Uh which you Mhm. [24:26] both sides of that. Uh you're currently [24:27] working full-time, but uh you previously [24:29] freelanced, and you've worked at [24:30] agencies, which means you've worked for [24:32] dozens of organizations. Uh one of the [24:34] things you've said that I I love, "Stop [24:36] going to people to find jobs. Force [24:38] people to come to you." [24:41] Yeah. [24:42] I I I think that's very true. I mean, [24:44] early in my early in my career, you [24:46] know, I've had professors, uh I took a a [24:49] web dev two course, and and everyone's [24:52] like, "Hey, build a portfolio. You know, [24:54] build a portfolio. Start blogging. Start [24:56] making YouTube videos. Start making [24:58] little toy dummy apps. Posting them on [25:01] GitHub and see if anybody stars it or [25:02] hearts it or likes it. You know, and and [25:06] that's a really great way to get [25:07] noticed. I think the worst way to get [25:09] noticed is to [25:11] put a piece of I I think I have my [25:13] resume somewhere. Yeah, put a piece of [25:15] paper in a pile of a thousand other [25:17] people and hope that like you can beat [25:20] out everyone for like one role. Like [25:23] Like that that is a way and I have [25:24] applied to jobs traditionally, but some [25:26] of the best jobs I've had, including my [25:28] current one, I've gotten reached out to [25:30] by recruiters who were like, "Oh, we see [25:32] all the work that you're doing. You [25:33] know, do you want to interview here?" [25:35] And like [25:37] Like for me, it's it's nice because it's [25:38] like [25:39] out of 10 people who reach out or 20 [25:42] people who reach out, like one of them [25:43] I'll be like, "Yeah." And the other ones [25:45] I'll be like, "No, thanks." Like [25:47] Um and it's like to have the option of [25:49] what like what kind of work you choose [25:52] it is great, but I think on being on the [25:55] flip side, it's really hard cuz you're [25:56] you're competing. And I say, "Stop [25:58] competing." Like find find what you're [26:00] good at. Find what makes you exceptional [26:02] and play to your strengths. Don't Don't [26:04] try to play the game everybody else [26:06] plays. Don't try to play like [26:08] you know, this finite game. Try to play [26:10] this infinite game where you you define [26:12] what it is that you're best at and and [26:15] and and showcase those skills, you know, [26:18] on social media or [26:20] in in-person events and things like [26:22] that. [26:23] Um [26:24] Yeah. Yeah. Very practical advice. Uh [26:27] you tell people to build in public. What [26:29] does that entail? [26:31] Yeah, I I I say post what you're doing. [26:33] Like if if you're working on a project, [26:35] post about it. If you're studying, stop. [26:38] Take a picture. [26:39] Post about it. Like I think people get [26:41] self-conscious that like [26:43] uh they get self-conscious about the [26:45] work that they do, which I also [26:48] did and also sometimes do. [26:50] Or you're sitting there and it and [26:52] you're like, "Well, if I post about this [26:53] concept, people are going to think I'm a [26:55] fraud because you know, maybe I should [26:58] have this concept down." I mean, there's [26:59] so many things to learn in life. [27:01] >> Yeah. To go around expecting that [27:03] everyone's been exposed to everything is [27:05] is like absurd. Like I went to one [27:08] conference talk [27:09] and or I went to one conference [27:12] uh and [27:13] I was talking with someone who had a [27:16] stronger math background. I think he had [27:19] a master's in mathematics, maybe had [27:21] tried to get a PhD, and was also running [27:24] a company for him at the time, and he's [27:27] like, "Are you familiar with this [27:29] concept?" And I was like, "No." And he's [27:31] like, "Really?" And he thought that was [27:32] interesting. He's like, "That's okay." [27:34] And then he he proceeded to like explain [27:36] it to me. We engaged about it. I could [27:37] ask him questions, and it's like it [27:39] turns out [27:41] you know, a lot of times in [27:43] you know, in high school, everybody like [27:44] didn't want to answer the question cuz [27:46] they didn't want to look stupid. But [27:47] like for me, a lot of times I've looked [27:49] very stupid, but I've gotten like a lot [27:51] of people to take notice, and I've [27:53] improved the things I got wrong, and [27:55] it's it's only like helped me. Like I've [27:58] said, [27:59] I've mispronounced words before, and I'm [28:00] just like, "Nobody cares. Nobody cares." [28:02] Like play your game. Don't like And I [28:05] don't mean game like [28:06] you know, like everybody's got a game, [28:08] but I'm just like [28:09] do your own thing. Like learn, find [28:11] people who are fun to like play with and [28:13] work with, and and I think things will [28:16] be kind of better. Like you'll you'll [28:18] find your tribe. You'll find people that [28:20] you mesh well with. You'll build things. [28:22] You'll have fun. Like [28:23] you know, it's an open world. Yeah, in [28:26] high school everybody's afraid to look [28:27] stupid. [28:29] Do you like what I mean? Dare to be [28:30] stupid. Yeah, definitely. Do it. [28:32] >> That's going to That's going to like [28:34] give you lots of opportunities. [28:35] Uh you said don't be afraid of [28:37] mispronouncing words. Like I gave a talk [28:40] in front of like 300 people at the [28:41] inaugural Code Land conference. So Ron I [28:45] organized it and had me come out to be a [28:47] speaker. Uh uh Scott Hanselman is in the [28:50] audience, a legend and one of my big [28:53] Yeah, one of my big role models and uh [28:55] Me, too. [28:56] >> I would I I don't remember what I said, [28:58] but I was like uh there's this word [28:59] adage. It's like some short pithy kind [29:02] of like uh [29:03] I guess [29:05] trick like rule of thumb, I guess. And I [29:07] I for some reason I thought it was like [29:09] a like a a French word or something, so [29:10] I said "adage". [29:13] And afterward he was very polite. He [29:15] always said "It's adage." And I was like [29:17] "Oh man, I feel so dumb." And he said uh [29:19] I think he said something like [29:21] you know, don't ever [29:23] uh feel dumb when you mispronounce a [29:24] word because it means you learned it [29:27] from a book. [29:29] And that you're [29:29] >> Yeah. you know, well read essentially. [29:31] Like you're you're probably flexing all [29:32] kinds of vocab that other like people [29:34] that don't read wouldn't even have on [29:35] the radar, right? So um yeah, like that [29:38] was a really cool experience. I'll never [29:40] mispronounce it adage again. Yeah. It's [29:42] a very salient memory uh to solidify how [29:45] to pronounce that word. Um That's the [29:47] important thing is like once you have [29:48] that moment, you learn it and then you [29:50] just like move on. But but some people [29:52] like they stay stuck and they don't [29:54] expose [29:55] um they don't expose like [29:58] what they know. It like the benefit is [30:01] you get feedback. Like if you never get [30:03] feedback, okay. [30:05] Low-key I'm really into the book 48 Laws [30:07] of Power, not from like a Machiavellian [30:08] sense, but I think it like gives people [30:10] with empathy more leverage in a [30:13] in a world often controlled by people [30:16] who don't always have a lot of empathy. [30:17] You like you understand the thinking of [30:19] others, but one of the concepts is um [30:23] Oh man, I I lost it. It was [30:24] >> This is that famous book that like cult [30:26] leaders use. I I think there was like a [30:29] Yeah, I'm familiar with the book. [30:30] >> banned from prisons and all that stuff. [30:32] But but there's a there's a there's a [30:33] there's a there's a One of the 48 Laws [30:35] of Power is isolation is dangerous. So [30:37] like what what that's saying is like [30:39] that engineer who is a secret rockstar [30:42] engineer who's in the corner building a [30:44] massive thing is doing the most [30:46] dangerous thing and is really hurting [30:48] his career because what he actually [30:50] needs to be doing or she actually needed [30:52] to be doing is exposing their work to [30:54] the rest of the org so people can see [30:55] it, give feedback so that they can do [30:57] better. Like isolating yourself, not [31:00] sharing your opinions and thoughts [31:03] like gives you a a misconstrued version [31:05] of the world. Like you're living in your [31:07] own world kind of like Plato's cave and [31:09] you think the shadows on the wall are [31:11] the world. But like on the outside it's [31:13] very very different. And and this is [31:15] something a lot of people struggle with [31:17] in their careers like [31:18] you know, they think they got it going [31:19] on, they think they're good at [31:20] programming and then they go to a Google [31:22] or they go to some other company and [31:24] they're just on kind of a different tier [31:26] and it's like [31:27] if you expose yourself, you can get [31:29] feedback from some of the best people [31:30] and that will make you better. Like it's [31:32] all to like improve and and to get [31:35] better at your craft, you know, but you [31:36] can't do it just within a vacuum. I [31:39] mean, some people can't but like I'm not [31:40] one of those people. Yeah, there's [31:42] always someone cooler than you. So [31:44] >> Yeah. [31:45] understand where you are the top 10% and [31:49] where you enjoy working on something and [31:52] are willing to outwork everyone else. [31:54] That that is kind of like your like a [31:56] quote from you that kind of summarizes [31:58] your approach of like how to specialize. [32:00] And I I just want to call that out. Uh [32:02] but [32:03] you know, where you are in the top 10% [32:06] like there are lots of places you could [32:08] become top 10% and and you tell people [32:11] to go deep into just a few areas [32:14] but it takes time to identify what is [32:17] worth your time, money if you're going [32:18] back to school, uh your energy. Uh [32:21] people have finite careers, they have to [32:23] decide. They have to at some point make [32:24] a decision. They jump, right? They can't [32:27] just stand on the bridge the whole time. [32:29] They have to at some point jump in the [32:29] river. They have to figure out where, [32:30] right? Yeah. [32:31] >> Um [32:32] how do you identify areas worth going [32:35] deep with? [32:36] Yeah, and [32:38] that that's a good question. I think for [32:41] me, I think [32:43] finding what you're good at, you [32:46] uh [32:47] family member I always quoted to me, [32:49] um [32:49] is actually my mother. She says, "Know [32:51] your strengths, [32:52] know your weaknesses." Like and and and [32:55] for me, I would I'm very very clear [32:58] about the things that I'm really good at [33:01] doing. I'm very very clear about the [33:03] things that I am not good good at doing [33:05] or the things I hate doing. So, I try to [33:09] take all the things that I'm really [33:11] great at doing and love doing, which are [33:13] also the things I love doing, and maybe [33:15] I don't know if it's a chicken or egg, [33:16] like I love doing good good good good [33:17] good good good good good good good good [33:18] good. But but I try to find those things [33:20] and I try to to to see if I can package [33:24] those up, [33:25] you know, really really well. Um and so [33:28] so for me, if I'm looking at me, I don't [33:30] just mean technical skills. For me, my [33:32] childhood, a lot of it was spent in [33:33] service. So, in service to the world, [33:37] you know, volunteering, you know, my [33:39] high school had this motto, men for [33:41] others. I went to an all boys school. [33:42] You know, serve serve serve. You know, [33:44] and and so like when you're serving [33:46] others and you're helping others and [33:48] you're mentoring others and giving back, [33:50] you end up interacting with a lot of [33:51] people. So, like I got to, you know, [33:55] work with all sorts of different people [33:56] and so like, okay, that's a skill that I [33:58] have. And then but but then you want to [34:00] combine that with, you know, another [34:02] skill. So, like software development. I [34:04] actually was not a really great [34:06] programmer. I really struggled. I don't [34:09] even want to talk about how I did on the [34:10] AP in high school. But like I worked at [34:12] that skill and then I got better. But I [34:14] realized early, I said, "I'm not going [34:16] to be uh there's some podcast that talk [34:19] about like the coding machine at meta. [34:20] I'm not going to be the coding machine, [34:22] but I can be just good enough to really [34:24] make a good impact and be a solid [34:27] engineer. But now, if I combine that [34:29] with my skills of working with others [34:31] and collaboration, then I'm going to be [34:34] really good. And oh, by the way, I took [34:36] some like personality assessments and I [34:38] took this one called like [34:39] Strengthfinders. It's helped me find [34:40] your strengths. And my strengths have [34:42] changed, but some of my highest [34:43] strengths have been empathy. So, I [34:45] combine that and then one of my other [34:47] highest strengths, my highest strength, [34:48] this is a new one cuz I retook the [34:50] assessment, was like being able to look [34:51] at other people and identify what their [34:53] strengths are. [34:54] Super great when you're on a team [34:56] because if you're struggling with [34:57] something, you look across your org and [34:59] you say, "Hey, you know this person in [35:01] this corner on this other project [35:03] unrelated? [35:04] They're really good at this one thing [35:06] that we're terrible at. [35:07] And can we bring them in and can we [35:09] create a team that could be really [35:11] successful at solving X or Y problems? [35:13] And so so you start looking at your [35:15] strengths and you're saying like, "Okay, [35:17] how do I [35:18] fit into, [35:20] you know, an org or different [35:22] situations?" Well, and then it starts [35:25] defining what sorts of organizations are [35:28] best for you. Um and then on the [35:30] marketing piece, it's even better [35:33] if you start posting [35:34] on social media or writing in areas of [35:37] your strengths, [35:38] um [35:39] people see those traits in you and they [35:42] say, "Wow, you match a pattern of all [35:44] the people in our org." So, then they [35:46] end up pulling you in and that's why the [35:49] some of the companies I've liked the [35:50] most, I've gotten pulled in because I [35:52] was talking about what I did and they [35:54] saw like, you know, this is a high [35:56] empathy person who like [35:59] posts in a way on social media that like [36:02] is in alignment with our values. I [36:04] wonder if they're any good. So, then [36:05] they [36:06] say like, "Hey, you want to do an [36:08] interview?" And I do an interview and [36:09] like I nailed really well with the [36:11] interview. Like uh for Deal Maker, you [36:14] know, the the VP of engineering is my [36:17] direct manager. Uh [36:19] he he is really great. He really cares [36:22] about the team. He really cares about [36:23] the people. Like he's always supportive. [36:25] My favorite thing about him is he says [36:27] in many meetings he says, "Layman, [36:30] I'll I'll I'll I'll rant and stand up [36:32] and he'll say after my rant he'll be [36:34] like, "Okay." Or our our our one-on-ones [36:37] he'll be like, "Is there anything I can [36:38] support you with?" And I'm like, "Huh, [36:40] that's a good question. I'm not sure. I [36:42] think I'm good. I think I I think I got [36:43] this under control. Like, thank you. [36:45] Like, thank you. Like, how often do [36:47] people ask like like, "Is there anything [36:48] that you need? Is there anything I can [36:50] support you with?" Like, that's that's [36:51] super great. But, like [36:54] like to find this culture, you know, I [36:57] had to like I've had to show my value so [37:00] that the people with the values can be [37:02] like, "Hey, kind of join join what we're [37:04] doing." [37:05] >> So, essentially like uh like mindsets [37:09] attract like like mindsets. So, it's [37:11] essentially So, it's it's kind of like [37:14] you are very service-oriented. That's [37:16] like a huge part of your upbringing. Uh [37:19] there's an entire category of [37:20] management. I don't know if it's still [37:22] in vogue [37:23] uh in management circles, but it's [37:24] called like service leadership. Yeah. [37:26] Servant leadership. And And the idea is [37:29] that essentially the manager is at the [37:30] bottom of the organization trying to [37:33] support everybody else. And everybody [37:34] else goes to the manager and tells them [37:36] like, "Hey, I need this." Or like, "Oh, [37:37] this went wrong and and it's because of [37:39] this and stuff like that." And the And [37:40] the manager's trying to essentially [37:42] react and facilitate rather than [37:44] dictate. Um [37:45] >> Yeah. Yeah. So, it sounds like you found [37:48] that kind of leader. And because you're [37:50] so service-oriented, [37:52] uh you you were attracted to that [37:53] organization. They were attracted to [37:55] you. [37:55] Yeah, exactly. And same is true and I [37:57] worked for another company called Test [37:59] Double. [38:00] The CEO got up one day in a meeting and [38:03] we were fully remote and [38:05] he posted the org chart, you know, which [38:07] is like a hierarchical chart, but it I [38:09] think it was like flipped. And it was [38:11] like CEO at the bottom just like you're [38:13] saying. And it like blew my mind. I was [38:16] like, "Wow. Like, you are supporting all [38:19] of us." And then all the little people [38:21] are supporting those people and then I'm [38:23] on the ground floor, but I'm at the top [38:25] now. It's such a great analogy. Yeah, [38:27] and if you think about it in terms of [38:28] who actually needs support and who needs [38:30] help, it's often like the newcomer [38:32] people that are at the bottom of the [38:33] proverbial pyramid. Well, those are the [38:35] people that need to be lavish with the [38:36] most support and energy and time. [38:40] The people who've been there for a long [38:41] time, especially like the people in the [38:42] upper management that are running the [38:44] organization probably been doing that [38:46] for a long time. They need the least [38:48] context, the least support. And they [38:50] they probably also have the means [38:51] through which they can acquire whatever [38:53] they need to get their job done anyway [38:54] and they have like the know-how and the [38:56] tacit knowledge and everything like [38:57] that. So, it does make sense [39:00] to to think of it that way if you want [39:02] to run a small company. Now, you've [39:04] worked at a lot of different size [39:06] companies and I want to drill into that [39:09] because you have the benefit of having [39:11] worked with dozens of different teams. [39:13] And you've even said that like in the [39:15] past that one of the reasons you enjoyed [39:17] consulting so much is because you're not [39:20] actually living in the house that you're [39:22] stepping into. You're you're kind of [39:23] like a guest stepping into a house [39:24] trying to figure out what's wrong with [39:26] that family there, what's dysfunctional, [39:27] what what's wrong with the house itself, [39:29] and and fix it and then you get to leave [39:32] and you don't have to sleep in the bed [39:33] there. You don't have to stay in that [39:34] organization. You you can just be like [39:36] the the friendly kind of [39:38] dinner guest essentially. And then you [39:41] get to leave and go and get into another [39:42] house and do the same thing in a [39:44] different organization. Kind of like an [39:46] organization whisperer so to speak like [39:48] like helping to solve and that that's [39:49] the nature of consulting. That's to an [39:50] extent of freelancing. Can you talk a [39:53] little bit about what drew you to that [39:56] lifestyle instead of just working [39:58] you know, at the same company for many [39:59] many years as many devs prefer to do. [40:02] Yeah, I think I think boredom honestly. [40:04] Like I [40:05] I always say I get bored fast and then [40:08] what I mean by that is like [40:11] I will [40:12] come into a new environment, I will [40:14] learn a ton, [40:16] I will get 80 90% confident in what I [40:19] do, then I'll look at like other things [40:21] I can do within that organization. I'll [40:22] take up those. [40:24] And I feel pretty comfortable. Like I [40:25] feel like I can make an impact and [40:27] that's often sometimes when I know when [40:29] to leave cuz I like I need to keep [40:31] learning. Like I am like fundamentally I [40:34] I just I love learning. [40:36] Test taking, another thing. But I love [40:38] learning new things and so [40:40] like I was finding that pattern [40:43] um was was emerging and and then when I [40:47] discovered consulting it was like, "Oh, [40:49] well, it's the same thing. You get to [40:51] leave companies and come to different [40:52] companies, but you you have like a main [40:55] house that you can go to. You have main [40:56] people that you can [40:58] communicate with and work with and and [41:00] build, you know, [41:02] uh like like have have a sense of [41:04] belonging with, but you also get to [41:06] explore all these other places, [41:07] different industries, um and so that was [41:10] like really really appealing to me at [41:12] that stage of my career because, you [41:14] know, it was a little bit earlier and I [41:15] just I wanted to see everything there [41:17] was. Like I was like, "I want to work in [41:19] healthcare. I want to work in fintech. I [41:20] want to work in automotive." Um I spent [41:23] a huge portion of my career in [41:24] automotive, which is ironic cuz it was [41:26] like I got in consulting. I had a [41:28] two-year automotive gig and another [41:30] two-year one. So it was like it was kind [41:32] of like [41:33] I did get so I did get to see other [41:35] things, but it was kind of like, "Oh, [41:37] this is this is ironic. Like this is [41:39] like the universe being like, 'Haha, you [41:41] thought you were going to'" Um but but [41:43] honestly I I really loved it and it it [41:45] it was a lot of fun. [41:47] Yeah, and one question I have for people [41:49] that are getting into it. I mean, like [41:51] just to pair back what you said, uh know [41:53] when to leave so you can keep so you can [41:56] go elsewhere and keep learning. Once you [41:58] get them to 80 90%, they can probably [42:00] take it from there, right? You've [42:00] already set them up for success and you [42:03] can politely excuse yourself from the [42:05] table because the the family chemistry [42:07] is healing or whatever like analogy you [42:09] want to if you want to use the dinner [42:10] party like the [42:12] the cousins that hate each other are now [42:13] talking again. And so your job is done. [42:16] You can leave and trust that it'll get [42:17] done by itself and you can go find some [42:19] other dysfunctional family to support [42:21] essentially. Yeah. [42:22] So [42:23] let's let's talk about the [42:25] landing clients, right? Like when you [42:26] were at the consultancy they probably [42:27] helped place you Test Double. Of course [42:30] we had one of the founders on, Justin [42:32] Searls. [42:33] Very cool conversation that should [42:35] already be live by the time this goes [42:36] live. He was he was your direct chart or [42:39] like he was your boss. Yeah, he he had a [42:41] lot of different roles. He wasn't my [42:43] direct. I think I I think most of the [42:45] people at the time direct reported [42:46] directly to the CEO. Technically the way [42:49] the orgs are we had a support person [42:51] but Searls had like different like [42:54] technical leadership roles around like [42:56] evangelism. I think at one point he held [42:58] like CTO and things like that to kind of [43:01] and he was he was he was a part of like [43:03] I think sales pipeline helping get more [43:05] work in and clients and things like [43:06] that. [43:07] >> So so you got to work with him and and [43:08] they would kind of tell you where to go? [43:11] Yeah, so a lot of times it was like a [43:13] conversation where it was like, "Hey, we [43:15] have this opportunity and these are your [43:17] skill sets so you're very forthright [43:18] about like these are all the languages I [43:20] know, this is you know, I'm a really [43:21] strong communicator. [43:24] I work really well on teams, you know, [43:26] and like here are certain these are the [43:28] challenges that we're having across our [43:30] client accounts or here's a new a new [43:32] client that's coming up and we think [43:33] that you fit that profile like you that [43:36] you can or you can help them in some [43:38] way." [43:39] So you would you you have you have [43:41] freedom at least at the at the [43:43] consulting companies I worked for like [43:45] Paravirtual and Test Double. You have [43:46] freedom to say like, "You know, I I [43:48] really don't think that's a great place [43:49] for me." Or you can say, "You know, [43:51] yeah, I want to I want to I want to work [43:53] there." Sometimes it's feast and famine, [43:54] so right? So like when the economy's [43:56] great and money's just flowing it's [43:57] like, "Yeah, you can be like, yeah, I [43:59] don't know if this is the one for me." [44:00] But when it's scarce sometimes you got [44:02] to be like, "Okay, I I got to just do [44:03] this one cuz you know, there's there's [44:06] only so many people can do it and we we [44:07] need money as a as a as an org to keep [44:10] keep going, but yeah, it's it's a really [44:12] cool way to [44:14] get new opportunities, for sure. [44:16] Well, let's talk about like the uh when [44:18] you're outside of the organization, [44:19] you're you're hunting alone, so to [44:20] speak, and uh you are trying to land [44:23] freelance clients yourself. Now, you've [44:26] done this, and one of the piece of [44:28] advice you always have is talk about [44:29] real problems that you've solved. Now, [44:31] you have the benefit of having done a [44:33] lot of work prior to going out and [44:35] sitting down with that potential client. [44:37] Um [44:37] how would you conduct a [44:39] a client meeting with a prospect? Uh [44:42] maybe take us through like kind of a [44:43] scenario from end to end, how you would [44:46] you know, prospect them, and how you [44:47] would ultimately close and get that work [44:50] lined up and get the contract. Yeah, so [44:53] I'm like pretty bad at like prospecting [44:55] from the from the standpoint of like [44:58] like cold calling or finding out like [45:01] where like reaching out to people and [45:03] being like, "Hey, can I help you in any [45:05] way?" Like, most of my work has been [45:08] inbound, so what will happen is like [45:10] people see the wide net that I'm [45:12] creating with my blog and Substack and [45:15] conference talk, and then they say, "Oh, [45:17] wow, this person looks like they can [45:18] help us." And then I get a message in my [45:20] email saying like, "Hey, can you help [45:22] us?" or an introduction or something. [45:24] Then we just have a conversation, and [45:26] it's mostly just listening. It's like, [45:27] "Hey, what are you trying to do? Like, [45:30] what are the problems that you're [45:31] facing?" And then they'll tell you a [45:33] little bit about it, and you'll ask [45:34] questions, you'll you'll you'll say [45:37] like, "Well, [45:38] you know, you know, you kind of got it." [45:40] And you a lot of it's like trying to vet [45:42] like motivation. Like, sometimes they [45:44] just want to talk to someone, and that [45:46] might be the end of the engage like the [45:47] conversation. It's like, you talk for an [45:49] hour or 30 minutes, and then [45:51] you you offer some advice, maybe, and [45:53] then they they don't come back. But [45:55] sometimes it's like it's it's a deeper [45:57] problem, and they they're like, "We're [45:58] looking to, you know, pay someone to [46:01] help us solve this problem and when [46:03] that's the opportunity then you you kind [46:05] of like you you get into this phase of [46:07] like you're pitching different [46:08] solutions. And so that was that was kind [46:10] of challenging for me early on because [46:12] it was like I don't know how to put [46:13] together these proposals and things like [46:15] that. So I had to do a lot of reading. [46:17] Um one person I like is Alan Alan Weiss [46:20] million-dollar consultancy. I believe it [46:22] was one of the books he wrote and he [46:24] kind of talks about how to [46:26] you know, structure [46:28] um [46:30] kind of [46:31] how to take in information from the [46:32] client and kind of structure how you [46:34] want to start the engagement. So I put [46:36] together sort of like a project [46:37] requirements like this is what we're [46:39] trying to solve. You know, I'll put the [46:41] costs in there and then they would sign [46:43] it and say hey like kind of [46:45] hop aboard and we would we would begin [46:47] the engagement. So I've done like [46:49] project management projects, I've done [46:51] exploratory AI projects where we [46:54] you know, they they they they they were [46:56] experimenting with AI. They didn't know [46:58] exactly what they wanted to do. They [46:59] wanted to see what AI could do for them [47:01] and so like I [47:02] was on a short project where I kind of [47:04] helped them do that. Um [47:06] and and things like that. So [47:10] Awesome. And when you're helping these [47:11] people like there's you give them like [47:13] kind of like a free consult essentially. [47:15] You meet with them and and it's kind of [47:16] like a sales job but it's also like [47:17] trying like a fact-finding mission. [47:18] You're trying to figure out okay, is [47:20] there actually work here that makes [47:21] sense for me to parachute in and help [47:23] them with it. Um [47:26] at what point do you identify like okay, [47:27] like me if somebody were doing that, I [47:30] would try to sponge as much information [47:31] out of them as I could before I had to [47:33] actually write them a check or send them [47:34] a wire or sign any contracts or [47:36] anything. But you, the onus is on you [47:38] the consultant to kind of approach all [47:40] right, well, now we're kind of heading [47:42] into the territory where I've got to [47:43] really start working in earnest. How [47:45] would you like get somebody to actually [47:47] give you money? [47:49] Yeah, so [47:53] it's funny it's funny now I help people [47:55] I don't I often don't [47:57] take pay like recently I had a project [47:59] where someone came and they said we have [48:01] money want to give you this money and so [48:03] that's the easiest way to get money is [48:04] they say they have money and they want [48:05] to pay you for your experience that's [48:07] like the easiest way. But yeah, to your [48:09] point like sometimes people reach out [48:11] frequently and they don't always pay and [48:14] so [48:15] you have to and I've been bad at this [48:17] honestly I just talk to people. But [48:20] at some point you have to constrain you [48:22] have to look at your time and say like [48:23] this is really costly [48:25] and then like not every buyer you'll [48:28] find out is like a motivated buyer some [48:29] people just want free advice some people [48:32] actually have this project but that I [48:33] think I think that's the biggest thing [48:35] you need to figure out is like do they [48:36] have the capability to pay me if not [48:39] then I'm going to have you on these long [48:40] calls and nothing's going to happen like [48:42] I I was approached by [48:45] uh an individual and there was some work [48:49] to be done [48:51] basically from like a private equity [48:52] firm and we had like these long drawn [48:55] out like month long [48:57] conversations then we started having [48:59] meeting sessions I never took a dime and [49:01] that was like totally a mistake like it [49:02] was such a [49:04] I think it was kind of a waste of time [49:05] and I should I should have said like hey [49:08] I'll give you two free call or like [49:10] I mean you shouldn't give you really [49:12] shouldn't give a lot of free stuff away [49:13] but you say hey I'll do one call two [49:15] calls but like after that you got to pay [49:16] me X amount of dollars for the rest of [49:18] the time um because otherwise it I think [49:21] it'd be kind of exploitative you know uh [49:24] for you [49:26] What are some other best practices that [49:27] you've picked up like I'm definitely [49:29] going to link to that book Alan Weiss [49:30] multi uh million dollar consulting I'm [49:32] going to read it myself that's a very [49:33] strong endorsement [49:34] >> it's great he's he's [49:35] >> just want to have more perspective when [49:37] I [49:38] coach people within the free code camp [49:40] community [49:41] one one thing I guess this is like [49:42] tangential this is not a plain question [49:44] but with all the LLM power tooling and [49:47] developer productivity potentially [49:49] skyrocketing and some people would argue [49:50] well it's just a bunch of repair work [49:52] that's going to need need to be done [49:54] because it's going to be low-quality [49:55] slop code. And and I've had like [49:58] Internet of Bugs and some other creators [50:01] who are I guess more skeptical [50:03] of LLM code generation on the the [50:05] channel, too. So, I don't want to say [50:07] that I'm excited with one or the other, [50:08] but like assuming that [50:10] developers do [50:12] dramatically improve their individual [50:13] productivity, and a given company needs [50:16] fewer developers. [50:17] And perhaps some companies that are like [50:20] small to medium businesses who would [50:21] have previously hired a developer, now [50:23] they're just contracting [50:25] on a kind of case-by-case basis. Maybe [50:27] they have like like legal work. Like [50:30] freeCodeCamp does not have a lawyer on [50:32] staff. We have different specialists [50:35] whom we reach out to like, "Hey, we have [50:37] an intellectual property-related [50:38] question. We need like a you know, like [50:40] we need to get this or a contract [50:42] lawyer. We need to get this contract [50:44] proofed for this grant that we're doing [50:46] or you know, international law." Like we [50:47] we have all these different specialists [50:49] in our stable, so to speak, whom you [50:52] know, I have their wire information and [50:53] I just pay them basically by the [50:54] billable hour with law. It's possible [50:57] that for a lot of small to medium [50:58] businesses, maybe a similar arrangement [51:01] comes up with developers where they're [51:02] like, "Okay, we need a solution that [51:04] does XYZ." [51:05] And [51:06] you know, you're specialized in this [51:08] particular type of work. [51:10] How much would it be for you to [51:12] implement a solution that does XYZ? Like [51:14] my theory is that a lot more [51:17] development work in general is not going [51:18] to be getting like a normal W-2 job, [51:21] like what we call in the US, but it will [51:23] be like kind of like [51:25] case-by-case project-oriented [51:28] you know, contracts. [51:29] Do you think that that is likely to come [51:31] to pass? [51:33] I think so, and I I definitely hope so. [51:35] Um [51:36] like I would love to see sort of the [51:39] bigger tech companies shrink down and [51:41] figure out they don't need as many [51:42] people, and maybe we're seeing it now [51:44] where they're just pushing out [51:46] developers into the workforce, but [51:48] there's this opportunity [51:49] where the people who are out there might [51:53] have bright ideas and they might create [51:55] more smaller companies tackling niche [51:57] challenges [51:59] and the people in the market can now [52:01] work for those smaller companies and I [52:03] think it de-risks us as a as a as a [52:06] nation frankly like like [52:09] when the when when the big companies [52:11] fail and they let go of 10,000 people [52:13] that that that could have an impact on [52:15] unemployment like on people being able [52:17] to take like keeping this economic [52:19] machine running and if you have a lot [52:21] more companies that's like oh you know [52:24] 10,000 people go out but there's all [52:25] these tiny companies of like [52:27] two to 20 to 50 people that need people [52:31] you know we don't have to rely on these [52:33] these big boys to kind of you know [52:36] shoulder the brunt of you know the [52:38] system we've created so I I really I [52:41] really like that idea and I want to live [52:42] in that world personally. Oh okay cool. [52:45] Well, I'm I'm glad to hear that you're [52:46] warmed to the notion like [52:48] if if that comes to pass [52:51] uh a lot more people are going to be [52:53] having to go out and prospect. And [52:55] they're going to need to understand [52:57] sales and they're going to need to [52:58] understand uh essentially customer [53:00] service and project management and a lot [53:03] of things that in theory if you're in [53:05] like the Dilbert-esque software [53:06] engineering role somebody just throws a [53:08] big you know phone book worth of specs [53:11] over the cubicle wall and you just start [53:12] working on it right like Yeah. Do you [53:14] think the days of that are largely gone [53:17] or how how much do you think like the [53:19] big companies are still going to retain [53:20] big staff of engineers based on your [53:22] experience? I mean you are working for a [53:23] company as a developer how many [53:24] developers do they have? [53:27] Uh how many developers do we have? I'm [53:29] not sure the number actually. It's [53:33] around probably 100 plus or minus sounds [53:35] right. Yeah. And those are dedicated [53:37] developers and I and I think like the [53:39] series D you know giant startups and [53:42] stuff like that and and of course public [53:43] companies are still going to have big [53:45] staffs of developers. But I'm saying [53:46] like for the small to medium businesses, [53:49] uh [53:50] and and maybe to an extent startups. [53:52] Yeah, I mean, if if you're if you're [53:55] fast-growing company, [53:57] I mean, yeah, I think agents, AI, all [53:59] this stuff could somewhat offset some of [54:01] the work, but sometimes you just need [54:03] people to do work and you need you need [54:06] people collaborating and [54:09] and and solving problems. So, I think [54:10] the smaller size [54:12] agencies, like they can't they can't [54:14] have like a whole staff of like people. [54:17] So, I think they'll outsource to like an [54:19] individual or a smaller team to to to do [54:23] some of their internal functions. Um [54:25] but then those companies might consult [54:28] with them to solve other problems. And [54:30] so, you get this like [54:32] really healthy environment where it's [54:34] like, [54:34] you know, eventually like small teams [54:36] get saturated. Like as a one-person [54:39] consultancy, you know, I might have to [54:41] say, "I can't take your work, but here's [54:43] someone else I've worked with in the [54:44] past." And now that person has food to [54:47] feed their family. And then they'll get [54:49] saturated cuz if you do a good job, it [54:51] turns out you get more work. Like people [54:53] spread your name around, you know? And [54:56] so, [54:57] eventually you'll get sat- saturated. [54:58] And not everybody who gets saturated who [55:00] says, "I want to scale to, you know, a [55:03] massive company and doing this thing." [55:05] It's like not everybody wants that. Some [55:06] people just want to like wake up, take [55:08] care of their family, make money, grow [55:10] their sales, like and keep it as like a [55:12] one-person or a small business, and [55:14] they're happy to give away work to [55:16] trusted part- like like honestly, I like [55:20] like I I I I have a friend who does like [55:22] consulting, and someone else approached [55:24] me and was like, "Hey, you can you do [55:25] this work?" And I'm like, "I don't have [55:26] ti- like I right now don't have time to [55:28] do [55:29] a lot of like projects." I'm like, "Hey, [55:31] can you do this project? And I'll bring [55:33] you in and I'll bring you in, you know, [55:35] and and say like, 'Hey, introduce you, [55:37] You'll do the work and then I don't have [55:39] to really be a part of it. Maybe I get a [55:41] finder's fee, which is nice. Like, you [55:43] know, I could all use a vacation, but [55:45] like I don't I don't have to do the [55:47] work. Sometimes people don't want to do [55:48] the work." Like, I I personally a lot of [55:50] the engineering work [55:52] like [55:53] like if I had a if if I was like [55:55] full-time consulting again, I would not [55:57] I probably wouldn't do what I'm doing [55:59] now for my current company like for [56:02] other companies. I'd probably want to do [56:04] something different and rely on [56:05] strengths like, you know, I'd probably [56:07] want to do strategy work or help people [56:09] like figure out in terms of speeds that [56:11] they can't. Like, that that plays to [56:12] some of the strengths that I are [56:14] underutilized. [56:16] Um so so I think it's like kind of an [56:17] open world. There's there's there's tons [56:20] of opportunities. [56:22] Um but but I think people are in a [56:23] really difficult place right now because [56:26] it's like we went to school for this [56:28] thing and that's like [56:30] maybe our identity or that's what we've [56:32] been trained to do, but the world isn't [56:34] like the world that right now is not [56:36] honoring that. Like, AI, LLMs, like how [56:39] the world is shifting is not honoring [56:40] that. So, we have to like change of [56:42] change how we think about what we do and [56:44] how we add value. Um and that's like [56:47] important. Like, I did um like a couple [56:49] months ago, I was doing some strategy [56:51] work [56:52] uh for a company, a nonprofit, and like [56:56] it was no it was no software [56:58] development. I was just talking with the [56:59] founder, basically saying, [57:02] you know, he'd say, "Hey, here's some [57:03] problems that I'm having." And I'd be [57:05] like, "Well, here's how you solve these [57:06] problems." And he's paying me. And I had [57:08] all this anxiety cuz it was very [57:10] different than some of the work I've [57:11] done. I had all these anxieties like, [57:13] "How am I even getting this value?" So, [57:15] I asked him. I was like, "Are you [57:16] getting value from this?" He's like, [57:17] "Oh, this has been super valuable." [57:19] Um and I'm like, "What?" Like [57:21] so [57:22] you know, if that's true, then I think [57:24] there are a lot of skills that other [57:26] people have that are trying to find work [57:29] that are super valuable, but they don't [57:30] know that it's valuable yet. And [57:33] this kind of dovetails into what I was [57:35] saying before is like isolation's [57:36] dangerous. You got to get out there and [57:38] show people that you can do different [57:40] things even though you don't think it's [57:41] valuable, other people may pay a lot for [57:44] it. So. [57:45] >> Yeah. And and I mean, what I just heard [57:47] was basically a 3-minute uh [57:51] sales pitch for building a network. [57:53] Building a reputation essentially. And [57:56] uh to to plug my own book that I [57:57] published it and it's free. Uh just [57:59] Google learn to code book. [58:01] Um but basically [58:04] people need to build their skills, but [58:05] they also need to build their network [58:07] and their reputation. And I think [58:09] for the longest time people just [58:10] focused on skills cuz I mean that's what [58:13] you know, all if you in your in school, [58:14] nobody cares about your network or your [58:16] like it's all about like, okay, well [58:17] what can you actually do? What are your [58:18] skills? You go to the hackathon, what do [58:19] you build? You know, and and the skills [58:22] are so important. That's like the [58:23] prerequisite for even getting in the [58:24] door, right? Nobody wants to hire a dev [58:26] who can't dev. Uh but uh once you have [58:28] your skills to a certain level, then you [58:30] really do need to make sure you pivot to [58:32] building and addressing those network [58:34] and building a reputation which you [58:36] know, as you said, if you have a good [58:37] reputation, people will come to you. [58:40] Uh and like how many jobs did you apply [58:41] for in your most recent job search? [58:44] Um in my most recent jobs for search, uh [58:48] I [58:48] I don't I've applied to a lot of jobs. [58:51] Like [58:52] like I think magnitude like sometimes [58:54] hundreds of companies I'll I'll like [58:56] apply to and I'm doing doing like a [58:58] resume in a stack of resumes as you said [59:00] or how are you doing [59:01] >> Yeah. Like I I've done it I've done both [59:02] ways. I don't I don't think like [59:04] you you can apply to companies [59:07] and you could you could win in the [59:08] battle of resumes in a stack of resumes. [59:11] I've gone to I've I've applied to [59:13] companies, gone to like four rounds and [59:16] it's just like [59:17] they decide to go in a different [59:19] direction or something something happens [59:21] and it's it's very challenging. I mean, [59:23] I think the resume is supposed to get [59:24] you in the door, so I think I've already [59:26] won. Like I consider a win if I get a [59:29] job interview. Like, the resume did what [59:31] what it did. But, when you look at [59:33] LinkedIn and it's like 100 people [59:36] apply, like, there's probably more than [59:37] that on like other sectors. Like, it's a [59:39] lot of work and it's like [59:42] like [59:43] like I I've had interviews with [59:44] companies that didn't don't align with [59:46] my values and I don't want to work here. [59:48] Like [59:49] like like I know this is the company [59:51] that I a big-name company that I've [59:53] seen, you know, in the insurance space [59:55] or whatever where I'm like, "Oh, that'd [59:56] be so cool to have that on my resume." [59:59] Then I go there and I'm like I don't [60:00] want to be here. Like, the people like [60:04] we're not of the same makeup and like we [60:07] kind of self-select and they're like, "I [60:09] don't want you either." And it and it's [60:11] not always like a hard skill thing. It's [60:12] just like you just don't fit that [60:14] culture, whatever culture means. So, but [60:17] that's a lot of work. Like that's a lot [60:19] of effort to like try to figure out what [60:22] works. [60:23] >> I don't know, hours and hours of your [60:24] life like [60:25] submitting your application and then [60:27] going to the venue where you're going to [60:29] be interviewed and interviewing and then [60:30] only to discover this is one of many [60:32] many jobs that is not a good fit for me. [60:34] So, like, my heart goes out to anybody [60:36] who's doing the job search. But, is [60:38] there an easier way? Like, have you [60:39] gotten any jobs purely through your [60:41] reputation where somebody's reached out [60:43] and like, "Hey, we'll make a role for [60:44] you." Has that started happening to you [60:45] yet? [60:46] Yeah, definitely. I mean, I think I [60:48] don't know if anybody's made a role for [60:50] me per se, [60:52] um but like they had a role and I was [60:54] like a good fit and then I got in and it [60:56] turned out to be a wonderful thing and [60:59] like the honestly, a lot of the places [61:01] that I've applied to [61:03] and got in, like [61:06] I ended up like not liking it and and [61:09] being like [61:10] this is not a good place for me. And [61:13] that is like the worst thing that could [61:14] probably happen. Like, I think it's [61:16] almost worse than a layoff. Like [61:19] and then that's [61:20] and the reason I say that is because you [61:22] are now you kind of have this like light [61:24] golden shackle where you're working at a [61:26] place you really don't like and you [61:27] can't leave because you don't have [61:29] something lined up. Where where if [61:30] you're let go then you're just on the [61:32] race. You can't leave because you don't [61:33] have a place lined up and it's like and [61:36] you realize you don't want to be there [61:37] for much longer. That's like a terrible [61:39] feeling and and so because of certain [61:42] experiences I've had I've I've had to [61:44] really like I'll I'll research companies [61:46] really deeply. Like I'll use AI, I'll [61:48] use Glassdoor, I'll read up all about [61:50] them and and like their CEO and like [61:53] their their engineering directors and [61:54] like what are the real experiences [61:56] people having here? And it may not be [61:58] that it's a bad company but it may be [62:00] like bad for me. Like the way I operate, [62:03] the way I use my strengths [62:05] it will be a weakness in that company [62:06] and my strengths will not be valued in [62:08] that company. So you got to find [62:09] companies where it's like your strengths [62:10] will be like [62:12] you know, top tier. Like this is what [62:14] they want and it will be easier to make [62:17] you know, it will be easier to grow your [62:18] career there than some place that [62:20] doesn't value what you bring to the [62:22] table. So what I'm hearing is uh [62:24] do your research before you start [62:26] applying to companies because you're [62:28] just wasting time if you apply to a [62:29] company where like like I'll often go [62:31] into like uh GPT-5 too and I'll just be [62:34] like, "Hey, here's the name of this [62:36] person. Like [62:37] find their LinkedIn, find their blog, [62:39] find their like come through their [62:40] social media and create like a dossier [62:42] on this person." [62:43] Yeah. And often if they're if there's [62:46] enough stuff on the internet they'll [62:48] they'll assemble this and then of course [62:50] sometimes they're hallucinations or [62:51] they're getting like somebody else [62:52] that's named John Smith or something [62:53] like that. So you have to be really [62:55] specific with this but click through all [62:57] the different links and just use that. [62:58] Like it essentially gathers all the [62:59] links so you can do that. You can get [63:01] like a a litmus. Okay, is this somebody [63:03] who I would actually be interested in [63:04] for example interviewing on the [63:05] freeCodeCamp podcast or is this somebody [63:07] who like is this a company that [63:09] freeCodeCamp would accept a grant from? [63:11] Right? Like we've never accepted a grant [63:12] from a tobacco company for example. [63:14] That's a pretty obvious thing that you [63:15] don't want to do. [63:17] But, but sometimes there's like more uh [63:19] like vague like like okay, how do they [63:21] treat like their customers? Like do they [63:23] have a good reputation on Reddit? Stuff [63:24] like that, [63:25] so like like these tools are very [63:27] powerful for doing research and there [63:30] are a lot of people out there that [63:33] are pretty toxic, but you can survive [63:34] for a long time [63:36] as a toxic individual or as a toxic [63:38] company, like an entire company of like [63:40] toxic douchebags, right? [63:42] >> Yeah. [63:43] And [63:44] you know, the market will allow you to [63:45] survive like that and [63:47] you don't want to find yourself working [63:48] for one of those companies, right? Yeah, [63:51] definitely. And if you're among and I [63:53] think everyone should be the best [63:55] talent. Like I think you should find a [63:56] set of skills that you're really good [63:58] at, wrap those up and you become the top [64:00] 10 20% in that space, which sounds like [64:03] when I say that to people, they're like, [64:04] "Oh, you can't just say that, man." I'm [64:06] like, "No, I'm telling you be the best [64:08] at whatever it is you're doing and and [64:10] look inward and figure out what you love [64:12] doing and what gets you [64:15] excited to wake up in the morning and [64:17] and just go off and be the best at that. [64:19] But I I think I think that's the way to [64:21] do it and it's like I think those people [64:24] I think companies [64:27] companies are trying to find great [64:29] talent to help them solve problems and [64:31] it's like [64:32] they they will start losing some [64:35] extraordinary people [64:37] if if if they don't, you know, build a [64:40] culture that attracts extraordinary [64:42] people. Like you go look on Glassdoor [64:44] and it's like some of those Fang [64:45] companies that they're they're rating [64:47] like 4.4. You know, they make sure that [64:49] that's really high because [64:51] you know, they want really good people [64:53] to work there. Like I will you know, if [64:55] I got an offer from Fang and an offer [64:57] from like a [64:58] medium-sized company doing the same [65:00] thing and the Fang company offered me [65:02] and like if they offered me $200,000 [65:05] more [65:06] like I would not take it if I thought [65:09] the company not offering me $200,000 [65:11] would be much more enjoyable and I'd [65:13] probably build a career I love more at [65:15] there, you know? And And to To other [65:18] people it's like that's absurd. Like [65:20] Yeah. That's like $200,000. You can You [65:22] can You can retire 20 years early. [65:24] >> and retire early, right? You know, [65:26] that's what a lot of people would think, [65:27] but until you've actually been there and [65:29] worked in like a hostile environment, I [65:31] think people under appreciate the toll [65:34] it takes. [65:34] >> Really. And And people don't count up [65:36] the costs. Like I think that's a Stephen [65:37] Covey thing. Count up the costs. I think [65:39] it's Stephen Covey. [65:40] The cost is not I don't like it there I [65:43] find another place. The cost is that [65:45] it's your health, it's your [65:47] relationships, it's your marriage, it's [65:48] how you interact with the world. You're [65:51] angry, you're you're upset because the [65:53] work's not good. You're taking it out on [65:55] your dog, you're taking it out on your [65:56] wife, your husband, you know, you get [65:59] colon cancer. Like Like I was watching [66:01] an interview from And I think Okay. I [66:04] think Netflix is probably a great place, [66:06] but there was this one guy who's who's a [66:08] director at Netflix. He was in an [66:09] interview um in a podcast and he was [66:12] saying he was working all the time and [66:14] working really hard and [66:16] and then he he had some health scare [66:18] where it was like you have cancer now. [66:20] And it's like [66:22] Could how much you were working [66:23] contribute to that result? And also, you [66:27] know, some of this is like self-work, [66:29] like [66:30] I think we have to set boundaries for [66:31] ourselves and it's not always the [66:33] company's fault about how hard we work [66:35] or what we do or how we feel. Like we [66:37] have to We have to have some control [66:38] over that. But it shouldn't be to the [66:40] point where though you're pushed. We [66:42] were subtly pushed in that direction by [66:44] some employers. Like I I read this book [66:47] Apple [66:48] It's called Apple in China. I read it [66:49] recently. It's a really good book if [66:50] you're if you're interested in China and [66:52] especially if you're interested in Apple [66:53] and like manufacturing and the migration [66:56] of manufacturing to China. But like [66:57] there's this very [66:59] you know [67:00] poignant scene in the book where they're [67:02] talking about like a lot of the people [67:04] that were flying over to China to like [67:05] run the fact like the software and the [67:07] hardware engineers were dying. And like [67:09] 45 50 from smoking a lot, drinking a [67:12] lot, from The stress. [67:13] >> properly. And uh they were having like [67:16] one of the, you know, engineering [67:17] managers was at the funeral of another [67:19] engineering manager on like a Sunday [67:21] afternoon and then he had to like leave [67:23] early to go to an all hands meeting [67:25] among the engineering team. Like he [67:26] didn't even have time to attend his dead [67:28] friend's Yeah. [67:30] you know, funeral on a Sunday because [67:32] the the work culture was such that they [67:33] couldn't slow down at all. They had to [67:35] keep moving to meet these, you know, [67:36] publicly stated deadlines and stuff like [67:38] that. And so [67:39] uh you know, like again, I don't know [67:40] anything I'm not a doctor. I'm not I [67:42] want to be clear that we're not a [67:44] medical None of this is medical advice [67:46] or anything like that. Uh I don't know [67:48] about the instance you're talking about [67:49] uh with the person getting cancer. [67:51] Uh that could have been genetic. It [67:52] could have just been incidental. It [67:53] could be the lifestyle related stuff. Uh [67:55] preventable cancer. It could be they [67:57] were too busy to go get a colonoscopy. [68:01] As soon as your insurance will pay for [68:02] you to get a colonoscopy, get a [68:04] colonoscopy. [68:04] >> Get it. I've gotten one. Right? Get your [68:07] prostate checked. Yeah. I know it's [68:09] humiliating. Get your prostate checked. [68:11] Like if you have family history, argue [68:13] with the doctor that you need these [68:15] procedures early. Get them approved [68:16] because you do not want to die at 40 [68:20] from a preventable disease that was in [68:22] your family history because Yeah. [68:24] >> too busy to get a So, that is my my [68:26] public service announcement for the [68:28] podcast. I always have to fit one in. [68:30] That does turn it That does That does [68:32] kind of tune into work though because [68:34] it's like if you're working at the place [68:35] where you place where you can't step [68:37] away from your keyboard for an hour [68:39] and the company like they're just going [68:40] to be like, "Oh my god." Like and you're [68:42] worried about getting fired. Like you [68:44] probably won't take more You won't [68:45] probably won't go to the doctor's [68:46] appointment. You probably won't go to [68:47] your kids little league game a little [68:49] bit earlier. Like that's not it doesn't [68:52] promote life, I think, you know. Um [68:56] >> So, And and at the same time I want to [68:58] be empathized with anybody who's [69:00] watching this is like, "Damn, I like [69:01] I've my rent is due, my kids are hungry. [69:03] Like, I'll take literally any job." [69:05] Right? Like, I I totally empathize with [69:07] that. And uh I don't want people to like [69:10] turn away like if they really need [69:11] income to turn it away on principle [69:13] grounds. No, I'm going to I'm going to [69:14] wait around and find a job that [69:15] resonates with me. I know that they were [69:16] speaking from a position of relative [69:17] privilege having these options available [69:19] to us. Uh so but uh one of the things [69:22] that you said there that I think is [69:24] really interesting is [69:26] just being selective. Uh [69:29] and you've had dozens of consulting and [69:31] freelance engagements. [69:33] Uh you've said that you really believe [69:34] in people. Your modus operandi is to [69:37] drop in, learn, get consensus, get the [69:40] team working together. [69:41] Uh what does this process usually look [69:43] like in practice? Like, let's say you [69:45] hypothetically get a client and you drop [69:48] in. [69:49] How do you manage that relationship? How [69:51] do you how do you scope it? Uh I just [69:53] want practical advice for people that [69:55] are uh planning to do this. [69:58] Yeah, so [70:00] Can you Can you state the question one [70:01] more time, Stephen? Yeah, what is what [70:02] is the deal flow? What is the workflow [70:04] associated with uh helping a client uh [70:08] and getting paid for it? [70:10] Yeah, so I guess is this is this after [70:12] the contract's been signed and you you [70:14] start going over it? [70:15] >> maybe if you can walk us through the [70:16] entire flow, like how you would approach [70:18] it. Okay. I think from the beginning [70:20] >> that you have from having worked in [70:22] consulting and with uh doing freelance. [70:25] Yeah, I I think from the beginning it's [70:29] they reach out to me, you reach out to [70:31] them, and then you have a conversation, [70:32] and then if the conversation goes well, [70:35] you know, then [70:37] it it may turn into engagement. If it [70:39] doesn't go well, then nothing happens. [70:41] But if it goes well, it might be able to [70:42] turn into an engagement, but I think you [70:45] just have to ask a lot of questions. And [70:46] one question you have to ask yourself is [70:48] like, "Can I Can I help these people?" [70:51] You know? And sometimes the answer is [70:53] no, but sometimes the answer that no is [70:55] not obvious. Like, sometimes it's I I [70:57] technically don't have the a technical [70:59] capability to help them, and that I [71:02] would say is very rare. Like [71:05] it's very It's very rare. Like and I'm [71:07] not saying it's rare for me. I think a [71:09] lot of people would be surprised at how [71:12] little you have to not little, but like [71:15] people think you have to know so much to [71:17] be able to help a company. Uh you know, [71:19] you just have to know a little bit more [71:21] about the problem or and a lot of it's [71:23] pattern recognition. Like I've been in [71:24] two companies, I know what problems they [71:26] have there, I know how we solve them. [71:28] I'll just apply the same thing. Like [71:29] you're not recreating the wheel a lot of [71:30] times. [71:31] So, can I help them? So, the technical [71:34] thing, yeah, I can help them [71:36] technically. Can I help them [71:38] org-wise? Like will they allow me to [71:41] come in and teach them and to help them [71:44] and support them to solve the problem [71:46] that I know how to solve? That's a [71:47] different question. Sometimes you're [71:49] getting into places [71:50] and you can solve the problem in a week [71:53] or two. Sometimes the culture is such [71:55] that [71:56] you need to [71:58] influence people and build trust. [72:03] Almost every org you work at, you're [72:04] going to have to build trust because [72:07] you you you won't be able to do the [72:08] technical piece without without the [72:10] people there. [72:11] So, it's not So, cuz cuz you won't have [72:13] the permissions to the systems that you [72:15] need to access. You don't know how the [72:17] system works cuz you're you're you're [72:19] day one. So, if if you can't build [72:22] trust, [72:23] they're not going to tell you or make [72:25] your life easier to figure out how to [72:27] help them, [72:28] which which seems counterintuitive cuz [72:30] it's like they hired you to be there, [72:32] and now they're not going to help you [72:33] solve the problem? They're going to [72:34] waste 10 to 50,000 dollars [72:38] to sabotage your And they will They may. [72:42] Like if you're a jerk and you you show [72:43] up and they're like, "I know everything. [72:45] I build every system I work for Google." [72:46] And that [72:48] And like they won't want to they won't [72:49] want want with you. So, you got to be a [72:51] person you want to work with. But, once [72:52] you get to that stage, it it's it's [72:55] trying to figure out like, okay, [72:57] like, what are the challenges they're [72:59] they're having? A lot of times, [73:01] and this is [73:03] this is like status quo for consulting. [73:05] They'll come you in they'll bring you in [73:06] to bring solve one problem, [73:08] and you either solve that problem or you [73:10] find out that's not the problem that's [73:11] needs solving. And so, then you have to [73:13] roll it back and say like, "Hey, and I [73:15] know you said this is the problem, but [73:16] this is a symptom, and the actual [73:19] problem is much deeper. The actual [73:20] problem is there's this process that's [73:23] broken that's causing this pain, [73:26] um or people are unhappy, and so, you [73:29] know, maybe because they're overworked, [73:31] because of some [73:33] pager duty alarm or something. So, it's [73:35] it's trying to make people's lives [73:37] better. Like, that's that's what you're [73:39] trying to do. You're trying to make [73:40] you're trying to reduce organizational [73:42] pain. You know, someone's getting calls, [73:44] someone's getting hammered all the time. [73:46] So, that's kind of how you move in the [73:47] organization. I think I deviated again. [73:49] Yeah, let me recap what you said because [73:50] I think it's important. So, you need to [73:52] influence the people in the [73:53] organization. Closing the sale is just [73:54] step one. Then you need to get in and [73:56] really then you get access and hopefully [73:58] you get access because you're nice to [74:00] all the gatekeepers within the [74:01] organization, and they respect you, and [74:03] they give you easy access to things you [74:05] need, and they don't they're not [74:07] paranoid like, is this person going to [74:08] replace me at some point? Like, are they [74:09] going to try to outsource my job to am I [74:11] training my replacement? Like, I'm sure [74:12] that sort of paranoia runs especially [74:14] among developers who tend to be more [74:16] paranoid than the median human being. [74:18] Uh but, basically, you figure out what [74:21] challenges they are. Often, [74:23] uh they bring you in to solve one [74:23] problem. It's just a symptom of a much [74:25] larger problem. And then, like, are you [74:28] looking at potentially having to rescope [74:29] the contract? Like, "Hey, like, I can [74:31] help you with this, too, but this is [74:32] beyond the scope of my $10,000 contract. [74:34] This is going to take a few months." [74:36] Yeah. So, you get on the ground floor, [74:37] you add value immediately, you help ease [74:39] pain of the people, and once you do that [74:44] and solve the problem at hand, then [74:46] there's this [74:47] mo- most of the time they'll have other [74:49] problems, so you just stay there. Like [74:50] you you solve that problem, you solve [74:51] this problem, and you keep working in [74:54] that organization as problems pop up. Um [74:57] cuz and [74:58] cuz there are no more problems and [75:00] you're done. Like you're like, "Okay, [75:01] I've helped you with everything I think [75:02] I can help you with." And then you move [75:04] on. Or you say, "You know, these aren't [75:06] the problems I want to solve anymore." [75:07] And then you move on. Like So, do you [75:09] bill hourly or like based on the scope [75:12] of the work? Like how do you figure out [75:13] how much to charge? Oh. [75:16] Check out that Alan Weiss book. Not you, [75:18] but just everyone. It It's It's really [75:20] interesting. So, I bill like [75:23] project-wise. I've done hourly. [75:26] Um I recently have been really sold on [75:29] this idea of value-based pricing, which [75:31] a lot of people think is just pro- [75:34] project pricing, project-based pricing. [75:36] Um but it's like [75:38] when you're talking to a client [75:40] See, I'm never going to do consulting [75:41] again cuz now everyone's going to watch [75:43] this video and be like, "Oh, you know [75:44] how you sell." [75:45] But But it's basically like [75:48] clients, they want to keep their money a [75:50] lot of time. Well, it depends. Some [75:52] clients I think bigger clients [75:55] you know, if you can solve their [75:56] problem, they'll [75:57] the rationality is like they'll pay [75:59] whatever, but a lot of times people look [76:00] at the cost. They don't look at the [76:02] value. So, value-based pricing, [76:04] value-based selling [76:06] is to they keep They'll keep asking you [76:08] like, "What's the price?" And you [76:10] deflect and you say like, "Well, this is [76:13] the problem [76:14] that you're trying to solve. What's that [76:16] worth to you? What's the value of that [76:17] to you?" And that way, you know, you are [76:21] getting paid for your expertise. You [76:23] know, I'm not getting paid to send sit [76:25] and do this work for all these hours. [76:26] I'm getting paid because there's [76:28] something that I know intrinsically [76:29] through all my Not intrinsically, but [76:31] through all my [76:33] ventures in all these consultancies, [76:34] it's like a compounded interest value [76:37] that you are access to. So, it's like we [76:40] can do hourly, but like like I'm going [76:42] to help you with a solution. So like if [76:44] the result that I deliver is worth $10 [76:48] million, [76:49] is it fair that I get to pay I only get [76:51] paid, you know, $200 an hour? I don't [76:55] know. Like Yeah. [76:56] It's like you might want to try to get a [76:59] portion of that value, you know? [77:02] Yeah, that makes perfect sense. So [77:04] figuring out what like and and generally [77:06] in my humble opinion, not as somebody [77:08] who's done this a whole lot, it's about [77:09] identifying cost centers and reducing [77:11] costs or it's about identifying, you [77:13] know, profit opportunities and like [77:16] realizing those essentially. [77:18] >> Exactly. [77:19] >> help them make money or can you help [77:20] them spend less money? And if you're [77:22] helping them save way more money than [77:23] you're asking for, it's an easy decision [77:26] to bring you on to solve that problem [77:27] assuming that you have the reputation [77:29] and [77:29] the pedigree to to be trusted with such [77:32] a task. [77:33] Uh there is risk associated with it and [77:34] that's going to factor into how much you [77:36] can get from them. But it sounds like if [77:38] you can shift it around value you said [77:41] value-based [77:42] pricing is the term you used. If you can [77:44] shift it around that that [77:46] Yeah. [77:47] Very cool. I don't want to take too much [77:48] of your time. [77:50] But I did want to [77:51] like talk very briefly about Ruby. So [77:55] obviously like a lot of people are [77:56] probably like who even cares anymore? [77:58] It's just [77:59] like LLMs are just going to generate [78:00] Python code or they're going to generate [78:02] JavaScript code or are people even [78:04] looking at the code anymore? Like I see [78:05] a lot of cynicism like that, but you're [78:07] like a die-in-the-die-in-the-wall [78:10] Rubyist. You're a fan of the Ruby [78:12] language, of the Ruby developer [78:14] community. You kind of got your big [78:16] start like and became like a household [78:18] name within the Ruby community through [78:20] like [78:21] extremely popular talks that you've [78:23] given at conferences. And uh I have to [78:26] ask you like [78:27] why Ruby? [78:28] Yeah, I think so the beginning of it was [78:33] I [78:34] you you I was I was looking out [78:37] and I was seeing all these Python [78:39] projects and I was thinking, you know, [78:41] why not Ruby? You know, what is the [78:43] limiting factor? [78:45] And what I found is, you know, the [78:47] Python community, they're more data [78:50] science-y, they're more data focused. [78:51] So, like there was this vibrant [78:53] community around that. I'm like, okay, [78:54] well, why else? [78:56] Well, some of the libraries that you [78:57] need to even do things, to even train [78:59] models aren't aren't in Ruby. And then [79:01] this uh gentleman by the name of Andrew [79:04] Kane just started pouring all these [79:05] Python libraries, like being like, all [79:07] right, Python library, now I'm going to [79:08] build a Ruby one. He just started [79:10] building them all out. And I was like, [79:12] that's really interesting. But nobody [79:13] was like [79:15] like from my side, like not a lot of [79:16] people were using it, like not the [79:18] numbers that you would think would use [79:20] it. [79:21] And so I'm like, well, these people are [79:22] still are not building in Ruby and it's [79:23] like [79:24] um like let me let me try to do [79:27] something cuz I I think it can be done [79:28] and I think if people knew how to do it, [79:30] they would want to do it. So, my first [79:32] task was trying to build like the uh [79:35] learning model that can predict like the [79:37] weather using some data. And I built [79:39] that, I presented it at RailsConf and [79:42] hundreds of developers came in and the [79:44] line was out the door cuz they hadn't [79:45] seen someone do that in Ruby. Like like [79:48] even though the tools were available, [79:49] like someone hadn't sat down and walked [79:52] through every single step of like what [79:54] it takes to [79:55] train a machine learning model. And I [79:57] had just learned that cuz I had paid a [79:58] bunch of money to do a course to do that [80:00] very same thing in Python. I was like, [80:02] just port it port it to Ruby and people [80:04] loved it. And I was like, there's [80:05] something here, like there's motivation. [80:07] So, that gave me like the energy to keep [80:09] working on it. But then [80:10] like what doubled down my thesis because [80:14] at that time I was like, well, I would [80:16] say in the conference side, I was like, [80:17] well, you wouldn't want to like [80:19] you wouldn't want to go tell your CTO, [80:20] let's train all our email models in [80:22] Ruby, right? Like you'd do Python cuz [80:24] that's the standard and like that that [80:26] is what everybody's using. It's like, [80:28] yeah, right. But like but then on the ML [80:30] world, it was totally different. Because [80:32] it's like [80:33] if if all you need to do to start [80:36] building your AI product is make an API [80:39] call to like Anthropic or OpenAI or [80:42] Mistral or whatever your favorite model [80:44] is, [80:45] guess what? None of that is inherently [80:48] Python. None of it's inherently Ruby, [80:50] Elixir. Name your favorite language. And [80:53] so like getting hung up on the language [80:56] just doesn't make sense. And the other [80:58] thing is Ruby and specifically Rails, [81:01] what DHH built, um has a very very [81:04] opinionated way opinionated way of [81:07] standing up a web application, a SaaS [81:10] product application really really fast. [81:13] I would I would I would put all my money [81:15] on a Rubius standing up a a SaaS [81:17] application really fast for their their [81:20] uh [81:20] their um startup faster than a Python [81:23] person, better well-architected, better [81:25] patterned because that's what Rails is [81:27] for. Like people startups are still [81:29] building stuff in Ruby. So if that's the [81:31] thesis, well you build your SaaS [81:33] application and then you do your API [81:34] calls to the LLM and it's great. And so [81:38] I started really believing that you [81:40] could do that with with Ruby and Rails. [81:43] And so I started building out some like [81:45] I built a small library um that was like [81:47] a [81:48] tiny It was called like simple rag. It [81:50] was like a tiny like uh [81:52] retrieval augmented generation library [81:54] to kind of showcase and explain and [81:56] teach like how you can use retrieval [81:59] augmented generation or AI search in an [82:02] application in in Ruby and Rails to say [82:04] like you could do it here too. Um [82:07] now on production scale, you know, [82:10] there's you can't use that, but but but [82:13] but the concepts, you know, trying to [82:16] like get away from just the language and [82:18] and the tools and and like fundamentally [82:19] understand the concepts, you you [82:21] understand that these can be done in in [82:22] multiple language, but is is the support [82:25] there to be able to do this on a [82:26] production level is more of the [82:28] question. That's really interesting. So, [82:31] if I had to try to like summarize to the [82:32] best of my abilities what you just said, [82:35] because the world has kind of moved away [82:37] from doing your own ML to just making [82:39] API calls to these already trained [82:42] models and and letting them do all the [82:43] heavy lifting, uh well, Rails is pretty [82:46] good at just making API calls and [82:48] pulling things into an existing web app. [82:51] Uh and because Rails is so powerful and [82:53] so opinionated and because it's so [82:55] opinionated, it is way faster to build a [82:56] a Rails app than it is to build like a [82:58] Node.js and Express app. Like it like it [83:01] it's [83:02] probably an order of magnitude faster to [83:03] build just a [83:04] >> maintainability and all that stuff. Like [83:06] every time I jump into a Python repo, [83:08] even the Python libraries that like I [83:10] was experimenting porting some stuff [83:12] over to Ruby, I was always like, "This [83:14] code is like it doesn't have it's not [83:16] clean code. It doesn't like these func [83:18] these methods or function are super [83:21] long." And like Rubyists tend to like [83:22] have really small methods. And it's just [83:25] like it was so much [83:27] like I could see how someone can make a [83:30] mess in the Python world of application. [83:33] And you need to hire consult like a lot [83:35] of consultants cuz it's like we got to [83:36] clean that up versus like in Rails, like [83:39] every time I come into a Rails [83:40] application, it's very like the [83:42] structure is always the same. You know, [83:44] like there might be messes, but they [83:46] feel a little bit more contained. [83:48] >> encourage uniformity. Yeah, and it it [83:51] makes a lot of the simple decisions for [83:53] you so you can focus on, "Okay, what [83:54] does this platform actually do?" Like [83:57] not the implementation details of this [83:58] business platform. [83:59] >> the technology, though. It's the people. [84:01] Again, it was back to the people. The [84:02] people in the Ruby and the Rails [84:04] community like have their own ethos [84:06] about like [84:07] a a teaching culture, honestly. Like [84:09] when we when when Rubyists teach other [84:12] Rubyists, they say, "Oh, you know, [84:13] here's the here's the idiomatic way we [84:15] do [84:16] we we write certain things. And you can [84:17] do it multiple ways, but we do it this [84:19] way as a community cuz it makes it [84:21] simpler." And then those patterns [84:23] proliferate everywhere. And so when you [84:25] leave one Ruby Rails corporation, you go [84:27] to another, like you can trust that [84:30] certain patterns are going to be there [84:31] if someone had a like mentorship, you [84:35] know, I I think I think mentorship is [84:37] really strong in in that community. Um [84:39] and I think it's important. [84:41] >> know like Pivotal Labs and a lot of the [84:42] organizations like are kind of I guess [84:44] best practice shops in the sense that [84:47] they bring people in, they kind of steep [84:49] them in the the Rails way of doing [84:50] things and and uh a lot of it's tied to [84:53] you know, Agile methodologies. [84:56] Um Clean Code is a popular book among uh [84:58] and then like the Pragmatic Programmer [85:00] books and stuff like that. So people are [85:02] kind of already steeped in this culture [85:03] of craftsmanship. And so it makes it [85:05] easier to swap in one Rails dev for [85:08] another, for example. Yeah. So So those [85:10] are all huge advantages. And I guess now [85:13] um that thanks to the work that you and [85:17] uh Andrew Kane have been doing and a [85:19] bunch of other people, uh you can [85:21] continue to use Ruby and there's no [85:23] reason to migrate away from it uh [85:25] just because Python may have the [85:26] absolute advantage in terms of like [85:28] tools and libraries, the actual business [85:30] value of those specific those Python [85:32] specific libraries that haven't yet been [85:34] ported to to Ruby are not necessarily so [85:37] great that you need to necessarily use [85:39] Python for it. Does that make sense [85:40] like? Was it Was that a Is that a good [85:42] recapitulation of what you said? Yeah, I [85:45] would say I would say that's true. Yeah. [85:47] Awesome. That's a good [85:48] >> Well, it's [85:49] It's been an absolute pleasure, man. [85:50] I've learned a lot from you and I've [85:53] really enjoyed talking with you about [85:55] like how you're selective about whom you [85:57] apply for [85:58] uh and and like just how important it is [86:00] that there is a good culture fit at uh [86:03] with between you and the organization [86:05] you're applying at, um the role of [86:07] building your reputation through these [86:09] talks. You attend tons of conferences. I [86:12] think you Test Double had like some [86:13] policy where you could basically Test [86:15] Double, the agency you worked at [86:17] previously under uh Justin Searles um um [86:20] had a policy that if you can get a [86:22] conference [86:23] paper submitted that they will like [86:26] essentially pay your way to that [86:27] conference so you can present. That was [86:29] that was that was a great idea. [86:30] >> conferences have you been to? [86:32] Oh, I'd have to check my It's been a [86:34] lot. [86:35] >> dozens? It's [86:37] it might be in the dozen or not. I Maybe [86:39] not quite a dozen, but it's it's been or [86:43] Yeah, I've probably been to almost as [86:44] many conferences that I've talked at cuz [86:46] a lot of the conferences I went to I end [86:48] up speaking, but yeah, it's it's quite a [86:50] handful. Yeah, and like what would be [86:53] your humble like just as a last [86:55] question? Mhm. [86:57] If somebody wants to [86:59] get involved in conferences, like is it [87:01] worth it to go to an in-person [87:03] conference in 2026 when everything's [87:05] online and there are Discords and there [87:07] you know, subreddits and and there like [87:09] so many other places people could go to [87:11] potentially talk about different things. [87:13] Like do you think the in-person [87:15] component of going to a big conference [87:17] has been a big value unlock? Is it [87:19] justifiable? [87:20] Oh, it's 100% justifiable. I think I [87:23] think even more so now, you know, COVID [87:26] forced a lot of people home and so [87:29] just this human interaction, learning [87:31] you know communicating with other people [87:33] is like huge. In fact, so much so that a [87:35] lot of times when when I like I've [87:37] started doing this more. [87:39] I didn't do this in the beginning, but [87:41] it there's this concept called the [87:43] hallway track where instead of going to [87:45] talks and I'm I'm fully about supporting [87:48] talks too cuz [87:49] but but instead of going to talk you [87:50] hang out in the hallway and you talk [87:52] with people. It's like well why would [87:53] anyone spend $900 for a ticket that you [87:57] may have paid for yourself [87:59] you know, just to go into a hallway and [88:00] talk with people. It's like cuz cuz [88:02] having a high density concentration of [88:04] engineers, like-minded people who are [88:07] we're not just talking about software, [88:09] coding, our craft, like [88:11] talking about favorite craft beer, food, [88:13] go out, you know, have a good time, go [88:16] karaoke. Like [88:18] that it's so meaningful. Like it's so [88:20] much fun. And it's like life just can't [88:23] be about sitting at a keyboard, working, [88:25] working, working. I mean, that's fun, [88:26] too. I love my work. You got to get out [88:29] and do other things. Like you got to get [88:30] out and meet people and and build your [88:31] community. And and honestly, my my my [88:34] take is like, yeah, I'm marketing a [88:36] bunch to I'm meeting other people, but [88:39] as as things shift in our in our world, [88:42] you know, with LLMs and just [88:45] you know, work going here and there, [88:47] it's like your defense against change [88:50] becomes your community. Like if you're [88:51] standing by yourself and you're like, [88:53] it's just me, I'm by myself, I'm [88:54] applying to conferences or doing my [88:57] resume, I'm trying to survive on my own, [88:59] it's going to be really hard for you cuz [89:01] there's a whole group of people that are [89:02] talking communicating and helping each [89:04] other out. Like recently, you know, I [89:06] have a friend who's switching jobs and I [89:08] was like, hey, if I can help you in any [89:09] way, either cuz I've had to go through [89:11] the negotiation process before or do [89:14] anything, like look at your resume. I've [89:16] reviewed resumes. A friend like like [89:19] they are not going at a job alone. [89:22] They're going at a job with 20 friends [89:23] who are supporting them and helping them [89:26] land the role both internally in the org [89:28] and outside. Versus like if if if we go [89:30] alone at it, it's much harder cuz you [89:33] don't have all the all the support. And [89:35] so, I think that's huge. Like be [89:37] supportive, support others, have a good [89:40] time, meet other people, you learn so [89:42] much. Like [89:44] I've learned so many different things. [89:45] And people have the most interesting [89:47] career paths. Like they're not even in [89:49] tech and then they're at this [89:50] conference. And you're like, how did you [89:52] go from like music education to to this? [89:55] Or or I think one of my coworkers is [89:57] like music did did I was doing like [89:59] tutoring in music and had a band and [90:01] things like that. And now we're both [90:04] doing AI engineering. And like it's like [90:06] the paths are so interesting and you [90:07] find out you know [90:09] you look at someone's LinkedIn resume [90:11] and you have this idea about how they [90:13] got to where they were, but like the [90:15] truth is like [90:16] everybody's kind of finding their own [90:18] path and it's, you know, it's less [90:20] intimidating that way. Yeah, and [90:23] I will say that like this is one of the [90:24] reasons freeCodeCamp exists. I was a [90:25] 30-year-old teacher who learned [90:27] programming just to try to optimize and [90:29] automate a lot of the back office stuff [90:30] in my school [90:32] before I got involved in freeCodeCamp [90:34] and like started the project and then [90:35] it's grown, but like organizations like [90:37] freeCodeCamp like I meet teachers all [90:39] the time who are in their 30s. I meet [90:41] truck drivers. I meet people who in [90:43] their 40s and 50s who were doing like [90:46] corporate law and wanted to get out and [90:47] do something else who have been able to [90:49] learn programming and the developer [90:51] community is [90:53] it [90:54] almost everybody like not everybody [90:56] there are lots of people who have just [90:57] got a CS degree and got their first [90:58] developer job at 22 and stuff like that, [91:00] but like [91:01] like second career folks career changers [91:04] are very well represented within the [91:06] developer community and if you go to any [91:07] conference you're going to realize that. [91:09] It's not just the people who have been [91:11] doing this [91:12] since their daddy bought them a computer [91:13] when they were 5 years old, right? It's [91:15] going to be people who didn't even have [91:16] computers till they were adults like [91:18] maybe they got a smartphone and they [91:19] went deeper into that and then they got [91:21] a laptop and they went deeper into that [91:22] and then they installed Linux and then [91:23] there's this this whole progression of [91:25] getting more and more technical and it [91:27] sounds like you're you're meeting a lot [91:29] of people like that. Going it alone is [91:32] so much harder. You said I think that's [91:34] a great note to end on. Build your [91:36] community, find your internal champions [91:38] when you're applying for jobs, [91:40] be a chill human being and built like [91:42] get a lot of friends. I imagine Landon [91:44] you have a ton of friends out there in [91:46] the Ruby community, in the consulting [91:48] community, [91:49] just going out there and [91:51] making friends and the those [91:53] relationships are going to help you [91:54] going forward whatever you choose to do. [91:57] That's the secret reason why I go to [91:58] conferences now. It's like I'm I'm going [92:00] to see a specific talk. I'm literally [92:02] just going to see the people I know and [92:04] make new friends a lot of times. Like [92:07] that's like a huge portion. [92:08] Yeah, right on, man. Well, I really [92:10] appreciate you [92:11] giving us so much of your morning and [92:13] putting so much of your words down on [92:17] audio. [92:18] I would encourage everybody to check [92:20] links to your blog, [92:22] which I've linked to in the the video [92:24] description {slash} podcast show notes, [92:27] and to to be on the lookout for you cuz [92:30] I'm really excited to see where you go [92:31] from here. [92:32] Yeah, thanks for having me on the pod [92:34] and of course this has been great. I [92:35] appreciate it. Yeah, well, everybody, [92:38] until next week, happy coding.