Maybe not a million, although having a few million, even just in a bank account earning moderate interest (4.5% PA) is easily enough to retire forever.
I do mostly cloud modernization architecture work and use Claude Code daily. But even I’m burning out on it. Not sure if that’s my current job, GenAI, or that I’d rather be playing poker.
Wait, you're upset because they used AI to answer a question they weren't expecting, and couldn't answer? Yet, you used AI as part of your upfront research?
What would you have preferred? They could have just said, "I'm sorry, I don't know the answer to that, please let me get back to you" but instead they tried to get an answer for you.
I think it's perfectly fine for the author to use AI in their research process, and it sounds like they weren't relying on it exclusively.
So they've already asked AI its opinion on the topic. They're explicitly reaching out to an expert because they've exhausted their ability to move forward on their own (even with AI).
If the expert just asking the same question to the AI and returning the answer directly - that's what OP has just done, and it actually is a waste of time. They're looking for insight, not just another quick response from an LLM.
I would imagine that letting the expert know ahead of time so they can research an answer (perhaps with the assistance of AI) would be a good pattern. But it has to be guided by the expert's knowledge - that's the whole point. Using AI is fine, and probably even good if it's being guided by a wise hand, but it isn't sufficient on its own, and bouncing answers directly back from an AI with no refinement is not useful, and dare I say it is somewhat insulting.
These were expert office hours for API modeling. The company has certain requirements and standards - this team is supposed to live and breathe them. The question wasn't that advanced. Performing a basic AI prompt in the authoritative expert office hours is not the answer.
The right approach would be to say: I do not know. Let me discuss / research with my colleagues and get back to you.
And to be clear: I manually read and studied the official guiding documents without AI. Then I used a separate AI setup to more effectively research a larger array of additional internal sources, wikis, etc. I also reviewed the code from other teams / projects to infer any patterns that could apply to my project, so I came prepared with examples for discussion.
> The right approach would be to say: I do not know. Let me discuss / research with my colleagues and get back to you.
I suppose that would be very close to "you've come to the experts for advice and I probably shouldn't be here because I'm not one of them", which nobody wants to admit.
For many, an honest look at themselves would end with "I don't contribute anything". They have the opposite of impostor syndrome - they don't belong, but they feel like they should, and AI helps them pretend.
> No pure managers: Every leader at Coinbase must also be a strong and active individual contributor. Managers should be like player-coaches, getting their hands dirty alongside their teams.
This has always been the case where I work, long before AI.
> This has always been the case where I work, long before AI.
And surely the place you work hired with this in mind. Many places have not, and yet now expect PMs who haven’t coded in years, or in many cases not at all, to contribute to their products’ codebases.
It was the HQ of French company, Alstom's French headquarters is located in Saint-Ouen-sur-Seine, near Paris, at 48 rue Albert Dhalenne, 93400 Saint-Ouen, France. The display shows original props from Luc Besson’s 1997 movie The Fifth Element — including Korben Dallas’s famous yellow flying taxi and the blue-and-silver NYPD police car. Alstom keeps these pieces on display as a fun tribute to their shared focus on city mobility.
This is the KEY difference between people who are willing to adopt this technology and those who aren't.
If you are able to view your job as simply a pursuit of a craft, more power to you.
The reality is likely that over time your employer will realize you are slower than every other engineer, and that your enjoyment of the craft is actually just you being an old slow developer.
The "race" here is the race with every other developer out there. They're getting on bikes, and starting to pull away ... what are YOU going to do?
When are we going to realize these CEOs are just old slow extremely expensive humans? I want to see them replaced with AI as well. I have absolutely no doubt AI can manage a company better than they do.
I love posts like this. AI is easily the most disruptive thing to hit our industry in over a decade and it feels like one of those "this changes everything" moments. Reading how it's impacting others is cathartic and helps shape my own understanding.
Here are some thoughts I have from reading this article:
> The AI can’t “see” the output, so some responsive refinements were just not correct. Within one CSS rule block there were redundant declarations.
This 1,000%.
Vibe coding has its issue and for me personally, frontend polish, responsiveness, and overall quality is the #1 most glaring of them that simply re-prompting often can't solve.
Even with the ability to screen shot your UI that hasn't solved things like glitchy animations. If you want to do anything even remotely above a junior level like scroll animations, page transitions, etc. good luck. AI will certainly try to do it for you, but inevitably it will not work perfectly and you will need to manually refine or even re-write code. When the code base isn't yours, that makes these re-writes a lot less fun.
> The guilty conscience at the same time, like I was cheating. I realized that when I move on like this, my project will never truly feel like my own.
I've wrestled with this over the last year, and still do to some extent. I'm trying to shift my perspective and envision myself as a brand new developer maybe 16 or 17 years of age. Would I think this isn't my work? I doubt it. I'd probably just (correctly) assume that this is the state of the art, this is how you do it.
Unfortunately this doesn't fix a bigger problem... I just don't enjoy vibe coding as a craft. There's something special about sitting down in the morning with your coffee and taking on a difficult programming problem. You start writing some code, the solutions start to formalize in your mind, there's a strong back-and-forth effect where as you code, the concepts crystalize further... small wins fuel a wonderful dopamine hit experience... intellisense completions, compilation completions, page refreshes, etc. are now all replaced with dull moments often waiting for the agent to return its response, which you now read.
> I’m curious (and a little bit scared) to see where we will go from here. I hope that in the end I can be part of a community that values craftsmanship, individuality and honest, high-quality work.
I really hope so too... But speaking honestly, I think this ship is sailing away quite quickly.
Time is money, and it always has been this way. Very few organizations can afford the luxury of time when building, designing, etc. I see no chance for this genie to go back in the bottle, and I believe it has (and will continue) to fundamentally change the nature of our work.
Over time as these models improve, there's a chance it could dramatically reduce the overall need for developers... It will start with low level teams as we're seeing already, but could expand.
I have been saying this to everyone -- what's your exit strategy?
I'm not saying you need to panic, but you need a plan for what happens if / when salaries tank dramatically. I hate to be "that guy" but in life I've found expecting the worst, isn't always a bad thing. Keep your mood up, prepare for the worst possible outcome, and be pleasantly surprised if that's not what happens.
> Unfortunately this doesn't fix a bigger problem... I just don't enjoy vibe coding as a craft. There's something special about sitting down in the morning with your coffee and taking on a difficult programming problem. You start writing some code, the solutions start to formalize in your mind, there's a strong back-and-forth effect where as you code, the concepts crystalize further... small wins fuel a wonderful dopamine hit experience... intellisense completions, compilation completions, page refreshes, etc. are now all replaced with dull moments often waiting for the agent to return its response, which you now read.
Agreed. For me, LLMs don't just reduce the kind of active learning and problem solving that make my job enjoyable; they change and replace it with a "skill" that barely merits the name. "Learning how to use AI" means learning how to use a product. That's worthless. It teaches nothing of durable value.
I'm also not in any way interested in using these tools to learn anything else. They can print out as much information as you want about this or that or that topic, and even if it's correct 100% of the time, you remain a passive consumer of information that the tool is chewing, digesting, regurgitating, and spitting into your mouth.
On the other hand, I'm also profoundly technically and intellectually bored. I can solve all of the problems I encounter in the codebase I work on. I can diagnose issues, refine build, shorten test pipelines, and mentor junior developer -- and I cannot imagine doing this for the rest of my working life. My brain will liquify and dribble out my ears long before I reach retirement age.
If what I were doing were more interesting and technically challenging, maybe I'd feel differently, but if LLMs kill off the kind of programming I do, I'm not sure anybody should grieve, particular if its death sends reasonably smart, curious people to fields where their efforts produce something of greater, or actual, value -- and doesn't wind up lining the pockets of the next generation of insufferable, bs-shoveling "thought leaders."
> I have been saying this to everyone -- what's your exit strategy?
Personally? I'm already preparing to sell our house. I'll keep my current job for as long as I can or for as long as makes sense, and then I'll go back to school for nursing, and become a psychiatric nurse practitioner.
There are also fundamentally different acceptance criteria for a bridge vs a website. Failure modes differ. Consequences of failure are nowhere near the same, so risk tolerance is adjusted accordingly. Perhaps true "engineering" really boils down to risk management... is what you're building so potentially destructive that it requires extremely careful thought and risk management? Engineering. If what you're building can fail, and really cause no harm, that's just building.
> Are these tools necessary to build what we actually need?
I think the entire software industry has reached a saturation point. There's not really anything missing anymore. Existing tools do 99% of what we humans could need, so you're just getting recycled and regurgitated versions of existing tools... slap a different logo and a veneer on it, and its a product.
We still don’t have truly transparent transference in locally-run software. Go anywhere in the world, and your locally running software tags along with precisely preserved state no matter what device you happen to be dragging along with you, with device-appropriate interfacing.
We still don’t have single source documentation with lineage all the way back to the code.
We still don’t treat introspection and observability as two sides of a troubleshooting coin (I think there are more “sides” but want to keep the example simple). We do not have the kind of introspection on modern hardware that Lisp Machines had, and SOTA observability conversations still revolve around sampling enough at the right places to make up for that.
We still don’t have coordination planes, databases, and systems in general capable of absorbing the volume of queries generated by LLM’s. Even if LLM models themselves froze their progress as-is, they’re plenty sophisticated enough when deployed en masse to overwhelm existing data infrastructure.
The list is endless.
IMHO our software world has never been so fertile with possibilities.
The tools are mostly there, but there is a lot of need. Quality can be much better. Quality is UI, reliability, security, and a bunch of other similar things I can't think of offhand.
reply