As a software engineer, I’m convinced “vibe coding” is just a meme. It’s like watching a chaotic system. You need to constantly be wrangling it back on topic, and keep it from bloating the codebase, in order to get anything done. You may be able to vibe a small mockup, but it will inevitably go off and produce garbage that doesn’t make sense.
It is useful as a glorified grep, and a sort of natural language to programming language compiler for simple descriptions. But if you don’t already understand what you expect the LLM to output, you’re gonna have a bad time.
LLMs for coding has improved dramatically over the past year or so. But, I find that its quality varies greatly, depending on the model. I find models like Gemini and GPT to be too overconfident, and it doesn’t communicate well enough. Claude knows when to stop and evaluate the situation for options. I’ve had mixed results with the local models, but I’m still adjusting quantization settings to make it work best with my VRAM.
You still need the skills to understand programming and design engineering, and you frankly need the personality to be meticulous with your reviews, but it’s really nice having something that can code 3-8x faster than what I was doing before.
A group at the web dev team for my company started a Dev Club to meet on Zoom to discuss code info and showcase dev work from them and other dev teams. Being a sysadmin scripter, I wanted to join to get experience in professional dev workflow as a creeper in the shadows. Then genAI code helpers happened and the guy leading it, the senior web dev at my company, started to use Claude and other genAI tools. Now, that’s like all they talk about and showcase. Like 2 or 3 devs decided to outsource their thinking and now that’s all the convo is about. Not even about their stuff, but leading genAI developments.
I stopped attending the monthly Zoom meetings. Not sure how many in there are into genAI, but since it wasn’t my bag, I didn’t want to say anything and just declined the recurring calendar event. Maybe I’ll start my own dev club with blackjack and hookers!
I think this sort of thing happens in software engineering a lot… it doesn’t matter whether genAI tools are the right solution to a problem. Billionaires are throwing an alarming amount of money at this, so they are basically trying to get a slice of that pie by virtue signaling that they love genAI, even if they don’t think it’s that valuable.
My employer wants features delivered, I don’t have time to circlejerk about genAI.
As a software engineer, I’m convinced “vibe coding” is just a meme. It’s like watching a chaotic system. You need to constantly be wrangling it back on topic, and keep it from bloating the codebase, in order to get anything done. You may be able to vibe a small mockup, but it will inevitably go off and produce garbage that doesn’t make sense.
It is useful as a glorified grep, and a sort of natural language to programming language compiler for simple descriptions. But if you don’t already understand what you expect the LLM to output, you’re gonna have a bad time.
LLMs for coding has improved dramatically over the past year or so. But, I find that its quality varies greatly, depending on the model. I find models like Gemini and GPT to be too overconfident, and it doesn’t communicate well enough. Claude knows when to stop and evaluate the situation for options. I’ve had mixed results with the local models, but I’m still adjusting quantization settings to make it work best with my VRAM.
You still need the skills to understand programming and design engineering, and you frankly need the personality to be meticulous with your reviews, but it’s really nice having something that can code 3-8x faster than what I was doing before.
PrimeTime had a good recent video about a senior programmer’s experience with fixing a very hard to find bug.
A group at the web dev team for my company started a Dev Club to meet on Zoom to discuss code info and showcase dev work from them and other dev teams. Being a sysadmin scripter, I wanted to join to get experience in professional dev workflow as a creeper in the shadows. Then genAI code helpers happened and the guy leading it, the senior web dev at my company, started to use Claude and other genAI tools. Now, that’s like all they talk about and showcase. Like 2 or 3 devs decided to outsource their thinking and now that’s all the convo is about. Not even about their stuff, but leading genAI developments.
I stopped attending the monthly Zoom meetings. Not sure how many in there are into genAI, but since it wasn’t my bag, I didn’t want to say anything and just declined the recurring calendar event. Maybe I’ll start my own dev club with blackjack and hookers!
I think this sort of thing happens in software engineering a lot… it doesn’t matter whether genAI tools are the right solution to a problem. Billionaires are throwing an alarming amount of money at this, so they are basically trying to get a slice of that pie by virtue signaling that they love genAI, even if they don’t think it’s that valuable.
My employer wants features delivered, I don’t have time to circlejerk about genAI.