My non-technical client has totally vibe coded a SaaS prototype with lots of features, way bigger product than OP and it sort of works. They spent like 200 hours on it. I wonder what would have been the time needed to clean it up and approve it is secure. I declined to work on it, as I was not sure if it's even possible or if it would be better to rewrite the entire thing from scratch with better prompts. I was not that sure about it given the cost and the fact that they had a product that sort of worked and I let them go to find someone to clean it up. My reasoning is that if the client took 200h to develop this without stopping to check the code, it would take me 2 - 3 x to rewrite it with AI, but the right way, while the cleanup may be so painful it would be way better value for money to rewrite it from scratch.
The 100 hour gap between a vibecoded prototype and a working product
221–230 of 362 posts
Re: The 100 hour gap between a vibecoded prototype and a working product
#222I work as a DevOps/SRE and have been doing it FinTech (bank, hedge funds, startups) and Crypto (L1 chain) for almost 20 years. My thoughts on vibe coding vs production code: - vibe coding can 100% get you to a PoC/MVP probably 10x faster than pre LLMs - This is partly b/c it is good at things I'm not good at (e.g. front end design) - But then I need to go in and double check performance, correctness, information flow…
This is exactly my experience at Lovable. For some parts of the organization, LLMs are incredibly powerful and a productivity multiplier. For the team I am in, Infra, it's many times distraction and a negative multiplier.
I can't say how many times the LLM-proposed solution to a jittery behavior is adding retries. At this point we have to be even more careful with controlling the implementation of things in the hot path.
I have to say though, giving Amp/Claude Code the Grafana MCP + read-only kubectl has saved me days worth of debugging. So there's definitely trade-offs!
Re: The 100 hour gap between a vibecoded prototype and a working product
#223[flagged]
Re: The 100 hour gap between a vibecoded prototype and a working product
#224Earlier quoted context omitted.
I'm building a Java HFT engine and the amount of things AI gets wrong is eye opening. If I didn't benchmark everything I'd end up with much less optimized solution. Examples: AI really wants to use Project Panama (FFM) and while that can be significantly faster than traditional OO approaches it is almost never the best. And I'm not taking about using deprecated Unsafe calls, I'm talking about using primative arrays b…
I think the main issue is treating LLM as a unrestrained black box, there's a reason nobody outside tech trust so blindly on LLMs. The only way to make LLMs useful for now is to restrain their hallucinations as much as possible with evals, and these evals need to be very clear about what are the goal you're optimizing for. See karpathy's work on the autoresearch agent and how it carry experiments, it might be useful…
In my experience, people outside of tech have nearly limitless faith in AI, to the point that when it clashes with traditional sources of truth, people start to question them rather than the LLM.
Re: The 100 hour gap between a vibecoded prototype and a working product
#225Earlier quoted context omitted.
I'm building a Java HFT engine and the amount of things AI gets wrong is eye opening. If I didn't benchmark everything I'd end up with much less optimized solution. Examples: AI really wants to use Project Panama (FFM) and while that can be significantly faster than traditional OO approaches it is almost never the best. And I'm not taking about using deprecated Unsafe calls, I'm talking about using primative arrays b…
> AI really wants to use Project Panama It would help if you briefly specified the AI you are using here. There are wildly different results between using, say, an 8B open-weights LLM and Claude Opus 4.6.
I've been using several. LM Studio and any of the open weight models that can fit my GPU's RAM (24GB) are not great in this area. The Claude models are slightly better but not worth they extra cost most of the time since I typically have to spend almost the same amount of time reworking and re-prompting, plus it's very easy to exhaust credits/tokens. I mostly bounce back and forth between the codex and Gemini models right now and this includes using pro models with high reasoning.
Re: The 100 hour gap between a vibecoded prototype and a working product
#226With sufficiently advanced vibe coding the need for certain type of product just vanishes. I needed it, I quickly build it myself for myself, and for myself only.
Related anecdote: My 12yo son didn't like the speed cubing online timer he was using because it kept crashing the browser and interrupted him with ads. Instead of googling a better alternative we sat down with claude code and put together the version of the website that behaved and looked exactly as he wanted. He got it working all by himself in under an hour with less than 10 prompts, I only helped a bit putting it…
Out of curiosity, did you also implement scramble support? Or just the timing stuff?
Re: The 100 hour gap between a vibecoded prototype and a working product
#227Im an 20 year veteran of application development consulting. Contributor level... not talking head. I do more estimating than anyone you likely know. Consulting is cooked. I just AI native built (not vibe coding...) an application with a buddy, another Principal level engineer and what would cost a client 500-750k and 8-12 weeks, we did for $200 and 1 sprint. Its a passion project but highly complex mapping and navig…
I realize this sounds one sided. Ive also founded companies and worked across the range of startup to faang. Everything has changed... For the better if you ask me.
You are invested in some kind of AI start up, right?
Re: The 100 hour gap between a vibecoded prototype and a working product
#228Earlier quoted context omitted.
> the progress there in the last few years is nothing short of absolutely astounding Yet, all the astounding progress notwithstanding, I don't have a suite of bespoke tools replacing the ones I depend on. I cannot say "hey claude, make me a suite of bespoke software infrastructure monitoring and operational tooling tailored to my specific needs" and expect anything more than a giant headache and wasted time. So maybe…
And what exactly is preventing you from building bespoke software for "infrastructure monitoring and operational tooling tailed to your specific needs"? I could certainly imagine building myself some sort of dashboard. It would seem like a prime use case. You want to hear about a problem solved? Recently I extended a tool that snaps high resolution images to a Pixel art grid, adding a GUI. I added features to remove…
> And what exactly is preventing you from building bespoke software for "infrastructure monitoring and operational tooling tailed to your specific needs"?
Let's say I emit roughly 1TB of telemetry data per day--logs, metrics, etc. That's roughly what you might expect from medium sized tech company or a specific department (say, security) at a large company. There is going to be a significant infrastructure investment to replicate datadog's function in my organization, even if I only use a small subset of their product. It's not just "building a dashboard" it's building all the infrastructure to collect, normalize, store, and retrieve the data to even be able to draw that dashboard.
The dashboard is the trivial part. The hard part is building, operating, and maintaining all the infrastructure. Claude doesn't do a very good job helping with this, and in some sense it actually hinders.
EDIT: I'm not saying you shouldn't take ownership of your telemetry data. I think that's a strategically (and potentially from a user's perspective) better end result. But it is a mistake to trivialize the effort of that undertaking. Claude is not going to vibeslop it for you.
Re: The 100 hour gap between a vibecoded prototype and a working product
#229They're... launching an NFT product in 2026... I know it's not the point of this article, but really?
I have friends (well, friends of friends) who still play the NFT lottery. People love gambling lol
I thought everyone realized by now that a digital image made available via block chain or any other mechanism, can be duplicated indefinitely. The only thing you get is a copyright on some generated image or set of bits. And what are the chances any random digital image is going to be appreciated as art? You can't hang it in a living room or sit it on a coffee table. It's beanie babies, but without even a hill of beans.
Are people just expecting there's going to be enough digital fools to make a market?
Re: The 100 hour gap between a vibecoded prototype and a working product
#230I work as a DevOps/SRE and have been doing it FinTech (bank, hedge funds, startups) and Crypto (L1 chain) for almost 20 years. My thoughts on vibe coding vs production code: - vibe coding can 100% get you to a PoC/MVP probably 10x faster than pre LLMs - This is partly b/c it is good at things I'm not good at (e.g. front end design) - But then I need to go in and double check performance, correctness, information flow…
This is exactly my experience at Lovable. For some parts of the organization, LLMs are incredibly powerful and a productivity multiplier. For the team I am in, Infra, it's many times distraction and a negative multiplier. I can't say how many times the LLM-proposed solution to a jittery behavior is adding retries. At this point we have to be even more careful with controlling the implementation of things in the hot p…
My colleague recently shipped a "bug fix" that addresses a race condition by adding a 200ms delay somewhere, almost completely coded by LLM. LLM even suggests that "if this is not good enough, increase it to 300ms".
That says something about how much some people care about this.