Basting_Rootwalla

I agree with the just do something. Idk maybe I'll come up with a ESP32 project and see if I can AI assisted code it or agentic code it or whatever. That's what I loved about working directly woth hardware. It wasn't hard to figure out what the firmware is supposed to do. When it comes to high level software, I hate coming up with an app. I feel like I'll really struggle with giving up more than 20% code control to an LLM and am also much mor curious of applications that aren't just code related.

Thank you for a helpful and insightful reply. This is why I'm finding my way of being interested in AI as doing it at home and understanding the engineering around the LLM. I guess not having used something like enterprise claude, I don't know the results first hand. I just know from the outside, it all sounds like the primary mechanism is pulling a slot machine and finding ways to keep improving the odds.

Makes sense. And going on that line, my question becomes what is "the way we work" now and how would one meaningfully fill in gaps if they haven't been working during the change in how we work.

Thank you. I feel like I'm becoming a FOSS zealot because I feel like we could all shift the power balance back if there is a OS way to do most things. You know how a choice between accessible DIY or buy.

Heya. I was wondering if you could point me to and resources for going full local and open source, particularly with as minimal dependencies as possible (what you're describing.) Any time I try to research, I'm inundated with basically the JavaScript framework equivalent for LLM stuff. 10000 different tools, frameworks, services, etc... where no one really gets what's going underneath it all. Prime example would be how does one set up their own agent locally with just llama-cpp-turboquant and a quantized model? Let's just sat I'm reluctantly switching my study to LLMs because I need to start looking for work again in the high level business and web software world. (I've spent the past 6 months with embedded systems, EE, and low level systems but I gotta shelf that for now even though it's where I really want to push career trajectory.) Maybe it'll all come to fruition and we'll finally make much, much smaller hyper focused LLMs or verticals that then open up the hardware world. Idk what use case there would be for say an LLM embedded into a microcontroller based application, but I'm sure we'll eventually get to more "AI" of some sort integration into specialized systems like tiny ML.

Thanks for the perspective. I love a well written and organized datasheet, but haven't considered what it'd be like to be the person who had to write it.

I feel like part of this can be attributed to going from highschool -> college -> FAANG. I taught myself into programming after well over a decade of working in restaurants and bars. I was always a deep thinker, problem solver, analytical type, (mother always told me "you think too much.") but I learned and developed people skills and team work from working in local bars/restaurants which probably helped round me out, let alone going from making like 35k a year to 80k in my first software job where I could: Work from home, work at a computer, go to the bathroom, make coffee, eat, blah blah ... WHILE getting paid to do something I discovered I love. Perspective really helps develop and maintain humility and you gain perspective either by being incredibly wise and being able to listen and learn from others, or be the fool (me,) and learn by experience and making mistakes.

I think it comes down to finding what particular constraints motivate you, at least that's what I came to after 5 or 6ish years doing typical web and b2b stuff. I dove deep into electronics/hardware/low level because I was so much more fascinated by resource contraints and building self-contained systems rather than systems of scale that operate so far up the levels of abstraction. I've been unemployed for like a year now and kind of stuck for many reasons; lack of childcare/support so not sure how we'd make it work, the turmoil of markets and not knowing whether I should just give in to the current trends and be be "AI first" (gross) or if me going deeper and to more fundamental/lower level stuff will pay off on the long run in setting me up for whatever the direction everything starts to settle in. But I get where you're coming from, because although it's probably so improbable, I'm way more motivated and determined to somehow break into firmware or at least some sort of systems programming. I feel the same dread thinking about going back to web AND having to deal with some varying degree of AI synchofancy.

Alas, the real 10x engineer has been found.

Started with JS and front-end web. Currently building my own power supply and learning everything embedded systems. I would say there is a certain point where the level of abstraction is similar great, it's like it's meta-computing at that point. I went from kind of the top (writing code to run in a web browser) down to C and debugging literal electricity because I saw that divergence point and didn't want to continue in the direction of cloud service providers APIs and "AI." I also had a big change first with writing Go. I absolutely love C and writing firmware. So my shift was realizing the direction I wanted to go was all the way down and how far up I really was and most software is (as in how far away from ever hearing the word coulomb or understanding what a semiconductor actually is and what bits really are.)

I'm not sure why everyone clings to the codegen approach to productivity. (Mostly the mix of non-tech, non-invested tech people/bad devs, and marketing speak) I feel like my productivity has increased a lot with LLMs SOLELY because of the learning assistance. Anyone who has done some software development knows how trivial any documentation examples are for APIs/libraries so much so that they're borderline useless except for getting the basic idea and important functions. Being able to generate a somewhat contextual example is a game changer to me. Not because I expect the code to be immediately usable, but it makes it way easier than Googleing and reading a bunch of articles, docs, SO etc... which I the VERIFY with official sources. I'm not reviewing code generated by LLMs. I'm reviewing knowledge they've synthesized. Imo, the hardest problem is usually figuring out what questions to ask. Until you have enough of a basic conception of what you're trying to achieve, it's hard to know what you need to know. LLMs being able to take an ambiguous question and evaluate it from a linguistically relational sense is the super power. Even if it's just enough to get me to that base understanding that I can then go ask an expert without them guessing what I'm trying to ask about from lacking the same vocabulary or fundamental mental model. It's kind of like a GPS for knowledge.

Au contraire, I have mostly startup experience and am getting call backs for mid-senior roles at all sorts of companies. The experience you get in a startup is not something you can get any other way than working in that sort of environment, in those early stages, and solving problems startups have. Furthermore, it is absolutely transferable up the line because it is largely about autonomy, ownership, and being able constantly rebalance efforts based on a understanding of business, product, and engineering. Those are the same skills you need in technical leadership when owning a project, especially any greenfield or research initiatives.

I'm surprised at how much desire there is to not have a camera on? It seems more odd to me to think about having a virtual meeting and not have some element of presence. I wouldn't be sitting in some screened in booth at a table at an in-person meeting. Seems pretty natural to me that if the medium is a video conference, you would have a camera on. There is a lot of information expressed through facial and body language + if you have your camera off and are muted, how do I know you're even there? I say this as person who hates meetings and is a classic introvert and has had my (then) 2 year old burst into my office and hop on my lap in the middle of a meeting.

I would say it all comes down to tool for the job. For example, I will use something like DBeaver or MySQL Workbench to mess with queries where it's simple to write and execute queries and read their results. I will not use a GUI tool to create or alter schemas or basically anything else. The GUI makes it simple enough to interact with the underlying program and read large amounts of feedback since it's output and formatted in a more visually efficient way to parse. It also gives me near real time feedback with linting etc... but that's just a simple self contained user journey. Write query, execute query, read result without having to change screens or click through menus and buttons. I mean, it's pretty clear in the context of databases why the split exists and most would prefer the CLI method. The amount of work to create the intermediary layer of UI filters/search to composing a query and executing it is so much greater, more complex, and less flexibility than simply knowing SQL and typing it out and executing it in a CLI. The reason I prefer a GUI for crafting or altering  query is because the the visual layout makes it easier to verify the output, not necessarily create or execute the input. Removing the extra abstraction layers between input creation and execution is why it's way easier to automate with scripting. Presenting the output is where a GUI really shines.

Yeah, I would anticipate things reach some sort of new homeostasis consider all the relatively rapid changes in economics and technology over the past few years. Overall, I have always leaned to the side of more software and more security for experienced devs, but we have to get through the current cluster fuck of circumstances. Just sucks knowing it could still be a year or two at the least before there is some new "balance" in the market. Certainly could be worse considering we're still debt free and own a house now being used as an asset. Overall, I'd say we're still doing pretty well comparatively and this past year has really acted as a forcing function to positioning ourselves better for the long term. It's just a bit of a change overload for me as someone who is usually pretty steady, but if it secures a better future for my family, then it's all worth it.

"I’ve had time to process that when the layoff finally comes, it will be a major life change for my family." TL;DR - Laid off a year ago. Wife and I swapped roles with work and stay at home parent. Things were not working out and decided to move our entire lives/family to be near her family several states away. Had 2 interview pipes out of ~70 applications but bombed them due to circumstances. 5+ YoE remote the whole time and will be applying for hybrid/on-siten here, even with possibly long commute. Core Dump of a story: This is me right now. My wife and I have a 3.5 y/o and 2 y/o. I got laid off a year ago now technically (some extraneous details on "technically".) Was working remotely for 5 years as fullstack at the same company. My wife able to be a stay at home mom for a good portion of our kid's lives so far. Then we had to switch roles. She went back to work at the local hospital (RN in L&D) which wound up really not working out over that year. (hospital units can be brutal with work politics aside from the already critical need of nurses at like every hospital making the jobs pretty brutal right now. That's a whole other "market" trend that really started with COVID time and lead to an exodus from health care field.) So I've largely been a stay at home dad for a bit now which is honestly the fucking hardest job I've ever had with having two toddlers. It's the best job a looking back a week ago or even just yesterday, but damn if it isn't hard while you're actually living it in real time. Still grateful we've been able tread water on one income while our kids are so young. It'll be crazy to me once we're both working full time again and how much easier finance matters will become. We wound up making the decision to move several states north to near where my wife grew up for better job opportunities and to have more support since her family is here. We were like 35 min from my family, but due to a lot of life circumstances, we weren't able to get much help. (Like my mother and her sisters rotating taking care of their 96 y/o mother which becomes a pretty restrictive commitment on your life) All that context aside... I started applying some a few months ago and had two interview pipelines start in the middle of us preparing to move our lives in basically 3.5 weeks needless to say, I was exhausted all the time and had no time to do any sort of prep and did extremely mediocre on thr OAs. I spent the first month just wrapping my head around the market and how I should try to be strategic in my job hunt. Spread sheet to keep track of applications and responses as a tweaked my resume. Targeting mid-senior for roles that my experience/resume matches like 80% at least, so basically quality over quantity except the quality is weighted to the employer by minimizing perceived risk in having nearly exactly what they're asking for. Not really being picky about the company on my end because this is an employers market. I had probably sent about 30 applications with my v3 resume (the one that I finally felt decent about after a month) so 2 in 30 isn't too bad I guess, but I still haven't been able to do any studying, prepping, or coding really in 1.5 months now. I also don't have my desktop setup with me up here and will have to get it around Christmas and drive our other car back up with my PC. I'm going to have to apply for hybrid/on-site jobs here for the best chances and I absolutely dread the idea of having to commute again. Especially because it will wind up being like a 1.5 hour commute to the major city. So that's some commentary about my market experience so far but actually sounds better than it feels right now to me since I haven't even sent over 100 apps yet. But the "major life change for my family" part? It's real as fuck. Basicsally our entire lives changed in 1.5 months from the decision to move to finally starting to settle in the place we're renting. We also crash coursed being a landlord and property manager while getting ready to move because we decided to keep our house and rent it out Starting next week, I can finally get time to start the grind again in the job hunt.  I spent a long time thinking about and writing this for some reason. I guess it's some sort of catharsis after these past few months to just lay it out in front of myself reslly.

Right here with you on this. I love that I can do something like ask Claude to "write me a comprehensive example that demonstrates effectively using goroutines in a mock but realistic use case and design pattern." Or something of that nature. I don't expect it to or need it to give me some 100% perfect example. I know enough to make connections or recognize what is wrong or doesn't make sense from the generated code, but being able to have interactive and responsive docs that will produce non-trivial examples or dive deeper on a specific concept really speed up the ground work of wrapping your head around some particular thing.

I'm always blunt and direct. It worries me more when non-ambiguous speech is automatically associated with "being an asshole." In my experience, I was often appreciated for being the one who just cut through the noise, put things back on track, and/or acted as the forcing function of inflection points. I also fall firmly into the "strong opinions, loosely held" camp. I have strong opinions on things I know I have put the time and effort into being adequately knowledgeable or skilled about. If I don't feel that way, I'll premise it and say that I don't have a strong opinion because I don't feel I have the expertise to warrant one, but I will workshop and talk through things as best as I can from a first principles basis. Once I'm presented with a strong case against part or all of a held opinion and I get that "ohhh" moment, I'll immediately admit I was wrong and we should do thst instead. Maybe I'm built a little different or slightly on the spectrum, idk, but if our shared goal is to be as efficient and effective as we can, then I see no reason to weigh personal against conceptual. Delivery is the personality piece, not the content of discussion itself. There is a line between saying "I don't think this is the correct approach" and "you always pick the wrong approach." One is isolated to that instance of discussion and subject at hand. The other is targeted as a critique of the person and not the idea.

Came here to say Wookash. Probably the best inside range podcast, imo. My second is a lot of Modern Vintage Gamer's content. Emulators, old hardware and consoles, etc... I also got a hard on for Inkbox and their Assembly magic.

For me and perhaps more startup or greenfield specific, although useful in product sense period: The balance of minimal, sloppy, fast solutions to drive tighter feedback loops on end user experiences and/or product market fit+PoC because business is first before engineering, but also laying foundations to transition prototyping into sound engineering designed for the typical scalability, maintainability, and observability. We're often tested on engineering aspects and in an established project, there is already a foundation set usually to improve upon and adhere to, but if you've never done early startup or greenfield work, you'll miss out on how important learning and understanding business priority and product sense is. After all, we're building things professionally to for other people to use. Once things operate at larger scale, all of that is offloaded to PMs etc... but in my experience, a good IC who also understands the business aspects can be immensely impactful, just the same way as we can't stand an EM who has no coding experience.

I just wanted to followup and say this comment has been in the back my head and I'm very certain you're right haha I think it's time to start experimenting with some embedded and work upwards from there if that's not what tickles my brain stem.

I think you nailed it. I get the 5% extra pay portion and what it conveys, but I think the real emphasis is we (generally as humans) need to feel purpose. With how deep of a field we have, those who are lacking a sense of purpose, whether in work or outside of work, will find things to apply that same level of consideration to even when it's entirely arbitrary. Like the naming conventions. If there is an agreed upon standard already and it's working well enough, there's no need to bikeshed inconsequential details but there is a desire to channel that energy into something. I remember hitting really, really slow and dull points in my last job/project where I found myself just tweaking and refactoring because I had nothing better to do within the job itself but was pretty invested in the codebase from having a high level of autonomy and ownership. I should have been identifying other problems or brainstorming and prototyping other possible ideas and features. Even if they wouldn't be selected or seen as useful ideas, I would be exercising my product sense. Or I should have been looking outside of my job or field all together and just investing that passion and energy in something else all together that would bring more balance and joy to my life.

Thanks. I needed to read this. It also cements further for me how much I dislike web, even though it's basically what I've done. I find myself reading/watching a lot about computers from the ground up, technologically and historically, to the level of C abstraction. Particularly, a lot of Assembly related content. Web doesn't just feel like an endless wheel of tools and frameworks, but a certain cutoff point of abstraction where it almost makes no sense to spend time on learning anything remotely close to hardware because the shift instead becomes mostly cloud services as the replacement domain knowledge category. It doesn't interest me at all, but I'm not sure if I'd Iike the reality of most embedded roles either. I'm sure programming another coffee pot becomes the equivalent of writing new end points somewhere along the line. Maybe I'd be interested in systems/OS. Not sure what direction I really want to go yet 

The problem, imo, is that "agile" became part of marketing speak. Once it crossed the boundary of the technical side to the business side, it's been over-mangegerial-ized. It became a selling point for business purposes, a key word in job listings, and a tool largely weilded and entirely misconstrued by the management/business side. Ceremony, bureaucracy, and overcomplication are now what comes to mind when I hear "agile" now. But having studied it some in a historical perspective, I feel like it boils down to just a few main concepts: 1. The devs should have the autonomy to "manage" themselves because they'll know how to improve their sprints as a team. 2. Intended for smaller teams as it's meant to be a heuristic driven. e.g. no waisting time making up formulas or large scales for estimating work effort/time. Just talk about the work for the sprint so the team becomes familiar with the requirements and can decide how to break it up amongst themselves. 3. Tighter feedback loops for input from stakeholders because it's hard to think through an entire project from top to bottom from both a business and technical standpoint. The business and product requirements will evolve over time based on many factors that would require changes to the product and therefore the technical plans. So "waterfall" didn"t make sense for increasing complex and more capable tech because it became much harder to minimize unknowns with way more moving parts. Sprints are the engine because it makes a tighter feedback loop for communication for everyone, business and technical. The increased communication leads to more cross-functional understanding of domains which overall better informs everyone, including business->business, business->technical, and technical-> technical, which means better decisions are made.

I feel like they key for me is after 5 years and startup work, I have come to really appreciate writing integration tests. It helps break up the slowness of just building of a feature while doubling up as bug hunting and reducing bugs.

And to piggy back/add, what's important about it and what a junior won't realize is the ongoing maintenance cost as well as complexity added by introducing a whole other system, even if its good, when the rest of the team/org is familiar with the previous system and it won't be realistic or justified in a business sense to go and rip the previous one out and replace it, even if the newer solution is objectively better. It has to be like an order of magnitude better to be worth that just from the technical perspective and then you need to show some favorable numbers of how it positively impacts the business to be worth it.

I've been bitten by unicode byte order marks before when trying to validate a CSV. Always fun learning about hidden characters when parsing strings.

Debugging a deployment on the Digitalocean App Platform.  Everything I tried did not work. I downloaded the app spec yaml from the web platform, through the CLI, etc... remade modifications, tried again with the adjusted spec, no dice. Failed deployments. Made sure I turned off auto formatting in my editor to make sure saving the file wasn't messing up the yaml, switched values, removed values, all sorts of tweaking to see if something was causing a silent error one that didn't give enough information. Started a back and forth with the support ticket process. Even wound up upgrading and paying because the response rate was like once a day unless you pay specifically for more support. They were useless. Finally, I just started with an another example and rewrote the app spec, replacing, removing, and adding values and realized there was a small ordering discrepancy of key:values in the yaml and then it worked. So whatever their side does after ingesting and respitting out the app spec messes something up in their own app spec format that causes the failed deployments and their support couldn't figure it out or were no help. If you were to recopy or download the spec after deploying, it would mess it up again.

My wife stayed at home with our kids (3 and almost 2) until I got laid off end of last year. Now I'm stay at home Dad but looking for a job now since our our oldest (3) started preschool in Sept. and our youngest is almost 2. We wanted to have one of us be the person our kids spent their first 2-3 years with mostly instead of having to give them to someone else. We were fortunate enough to be able to swing it. We live somewhere between MCoL and HCoL. It hasn't been easy. I was making 120k and now, she probably makes somewhere between 90-100k depending on on-calll/Overtime (she's a labor and delivery murse.) But I always figured I'd pay any amount in the future to have had the time we had now with them so young and once we're both bringing income, it should easily be over 200k, so finances will be way, way easier. All that said and reiterating, it has not been easy. I don't think we would have survived without our 3 year old starting preschool. Everything has always been "just enough" under these conditions to get by but not comfortably or even happily tbh. Just enough sleep, money, time, etc... to tread water but not thrive, but our girls have thrived which is whst matters.

I'm not sure why this is getting down voted. The one use i find very beneficial is basically having something of an interactive Google search that can also generate helpful examples. Most documentation has minimal, trivial, and/or overly simplistic examples that isn't always enough to convey concepts. I don't just copy/paste and lots of times there are problems with the generated examples, but being able to have an attempted example based on context is enough for me to get the bigger picture and then do it the right way and tailored to my actual use case. I could probably work on using it more for boilerplate stuff, but since I write a lot of Go, I have my own little library I use anyway with a lot of common utilities or helpers.

I also have a 3 year old and 2 year old. Contrary to your sentiment, I simultaneously do not want to miss any time with them and wish I had more time to dive deeper into computing. So I'm not a virgin and the passion for the craft is actually what makes me wish I had more spare time for it.

Echoing most of the suggestions given here, but more specifically, I would say as someone with 5+ yoe, currently unemployed and looking, I would gladly take this sort of opportunity if the pay was satisfactory. Not out of desperation, but for the reasons stated by mostly everyone else. One person's problem is another's opportunity. I would enjoy the challenge and being able to see real impact I have on a team or product. May be a fool's errand to some, but some of us are fools. I may be wrongly assuming, but are you a bit unsure of your own capabilities or put off by appearing to be the second most experienced? Because this is the situation where you learn how capable you are while still having someone a step ahead of you. It's what I call the "senior junior" job. Some skills you don't get in any way other than navigating the situations themselves. You can always learn more technical things on your own, but you don't learn how to lead a team, understand the business, how to prioritize, plan, delegate, communicate up and down the chain etc...  without being in circumstances where there is necessity for someone to step up or take initiative.

This is good to hear and confirms what I've been anticipating, but I realize I didn't put in the right actions to be ready to capitalize on it. I'd imagine this will be a continually trend and market for a bit, but im realizing now that those who serve to benefit are going to be freelancing/contracting as opposed to being more of a FTE hire. Going to have to consider if I can get myself out there for similar work since I'm currently doing the FTE job search rodeo.

None, like many. I self taught as a career change (no college, bootcamp, etc...) with the internet, a little bit of Udemy, and a good programming discord. First job was an early but chill startup where after a few months, I knew more than the current leads. It was sort of a shit show but fun and I got paid to continue expirmenting and learning really. An essential skill in software is problem solving and learning on your own. The hardest part, imo, is not finding solutions, but figuring out what questions to ask. I believe its rather inherit to what we do that you will have zero to minimal mentorship and part of what decides whether you're someone who actually finds passion in the craft or if its more-or-less just means to an end. There is nothing wrong with doing it for the sake of a good career. There is plenty of work that is (in a reductive sense) task work that needs to be done and we need people who do that sort of work, but like any field, excellence usually comes from a place of passion where you are invested in continual growth and challenge.

I can relate and have been through similar. I started in web frontend and picked up backend stuff along the way until being proficient at "fullstack." (Webserver, database, docker/infra stuff, workers) I picked up Go first and think that was a great language to start branching out further with since you still get more hands on exposure to things like pointers, I/O, static typing and compiled programs, std lib implementations for things that deal with the OS as opposed to a browser being your "system." I've been learning C now, filling knowledge gaps around CS fundamentals like DSA, and expirementing with raylib. So I've basically moved from higher down to lower in phases guided by my own interest and curiousity. At the least, I know for myself I much prefer working with RDBs, internal tools, ETL, distributed systems etc... which is also still applicable as web but also useful for things not meant for browsers.

Gosh this seems so awful as I continually see posts of this theme. I'm on the job hunt right now (unemployed) and actively skipping posts that make a big it a big point as a requirement for "using AI to efficiently blah blah...", let alone skipping AI companies altogether like its crypto. If I were in some sort of scientific domain, I'd be interested in that sort of AI company, but I'm not interested in most of these general purpose or clearly rebranded in the last year as AI companies and shoveled in some GPT wrapper. I'm not sure what will break first; the bubble or my desperation. (I'm not anti-"AI" btw. Its useful for the right tasks in development.)

Fullstack 5+ yoe, no bootcamps, college, etc... worked for a small startup situation the whole time. Learned a whole lot, especially on the business and product side, where everything technical was still just me figuring it out mostly. Unemployed after the project was shuttered. Currently learning some C, DSA, and reading through Designing Data Intensive Applications. Started my job hunt, but still picking up on the "job hunt metagame." I have a wife and 2 young children, so I'd really like to land something more secure, although I'm not sure if that applies the same way beyond startups these days. Yet, I feel as though I'm a bit boxed in to looking for another startup due to my experience. I like challenge, navigating greenfield/product market fit flexibility, learning new domains etc... just worried about expected time commitment and WLB when it comes to finding another startup job. I don't really have the luxury of being that picky though.

Thanks for the insight and I suppose it's my limited perspective, but its odd reading "actually pretty rare" considering how second nature it is to me. I suppose that's the difference in working on something yet to succeed vs. growing and maintaining something that already has succeeded. I guess I'm in a better spot than I think and its just a difference in kind when it comes to earlier and smaller vs mature and bigger.

Any general suggestion of where I should focus on prep pre-apply/interview cycle? Targeting startups/small companies as I believe it to be my best fit and what I enjoy. Context: 5+ years at one and only job so far. Early startup with unique situation of being self funded. Product direction shifted 3 times in a major way. Wound down the past year, resulting in me being only dev until it was shuttered. Self taught (as in no degree/formal learning). I believe I'm a strong generalist in the fullstack web world and enjoy database/data integration sort of ETL/backend stuff the most. It's hard for me to gauge how much time I should spend truly learning things like DSA and practicing leet code because I've never done it before. Also systems design. I worry how much of my skills/experience translate up the company scale since I've tuned much more for the product/value based development, hence why I've never specialized more, but also don't want to play roulette with too small/early of a company and have that lingering thought of instability job wise.

It happens frequently enough, but my IDE is quick to point out a non-existing or depreciated method or type etc... but that's pretty easy to resolve because now I know exactly what I'm looking for if search the web or go to a docs site.

I continue to have a hard time understanding why it seems like nearly all discourse from devs around LLMs is either: A. I don't use it at all B. I try to use it for everything Its been exceptionally helpful for ideation, discovery, etc... since it's kind of like if docs could have a conversation. Can find what I'm looking for even when I'm not sure what I'm looking for, introduce me to tech, patterns, or concepts I may decide to implement, and produce better examples that are more tailored to my specific project or problem. And then I think more about what I'm doing, how I want to do it, and how it works with existing design. The real kicker...? I write the code and iterate from there. Its a huge productivity boost in that it fulfills (for me) the core premise of focusing on and doing the deeper work. Code itself isn't deep, but making sure it works correctly and efficiently while being part of a greater whole is. Basically, it's super charged the researching and planning of something for me which is a non-trivial amount of time and effort and allows me to produce a better mental model while solving a problem, but I still go and solve the problem myself. I guess it's not as sexy or controversial when you frame it as an evolution of search engines even if that's is basically what it does well.

If I may, any tips on how you tailor your resume for finding another startup/small company job or how to go about finding a new job? I've spent my first 5+ years in an early stage stage startup situation and am starting to look for a new job now. The biggest problem I'm facing currently is almost all advice/direction is geared towards enterprise or seeking a FAANG like job (e.g. xyz impact format blah blah) and I don't have a lot of job hunting/interviewing experience since I've pretty much done it once. I'm a strong generalist and my impact is always centered in more business, product, ideation. Basically what you described. We never reached a point where scale up and optimization was priority, so although I'm aware of and care about those things when designing and developing, I haven't had to go deep.

Congrats. Just to echo this for others, I had a similar circumstance, so it is absolutely possible to get hired as completely self taught. For me, it was back in 2020, so of course I had the market in my favor at that time. I quit my bar tending job in May 2019 because of some personal circumstances (met my now wife, was long distance at first, left my job to go live with her for a couple months before she would move states and live with me.) I lived off of savings and took that time to "figure shit out" to try to get into a career. I managed my finances with Google Sheets. Got tired of spreadsheets and what felt limited to me at the time and decided to learn how to build an app. Got into web development and started learning JavaScript in January of 2020 and landed a job in August of 2020. Once I got the first real "ah ha" of programming, I spent like 55+ hours a week learning/building things (emphasis on building things.) So while I was unemployed, my job became learning programming. Now I work fullstack across different kinds of services using different tools/languages, (web frontend, web server, database, building worker services compiled to binaries, some devops, etc...) all still self taught the whole way through. Don't let the modern sentiment of "working hard doesn't pay off" influence you. It's not true. Working hard does pay off, but there is still luck, networking, and all those other things involved. Working hard on it's own doesn't pay, but it still puts you ahead of plenty of others when everything else also lines up.