Summarizer

I wish this was true. My experience is co-workers who do lip service as to treating LLM like a baby junior dev, only to near-vibe every feature and entire projects, without spending so much as 10 mins to think on their own first.

The pattern matching and absence or real thinking is still strong. Tried to move some excel generation logic from epplus to closedxml library. ClosedXml has basically the same API so the conversion was successful. Not a one-shot but relatively easy with a few manual edits. But closedxml has no batch operations (like apply style to the entire column): the api is there but internal implementation is on cell after cell basis. So if you have 10k rows and 50 columns every style update is a slow operation. Naturally, told all about this to codex 5.3 max thinking level. The fucker still succumbed to range updates here and there. Told it explicitly to make a style cache and reuse styles on cells on same y axis. 5-6 attempts — fucker still tried ranges here and there. Because that is what is usually done. Not here yet. Maybe in a year. Maybe never.

From the linked project: > The reality: 3 weeks in, ~50 hours of coding, and I'm mass-producing features faster than I can stabilize them. Things break. A lot. But when it works, it works.

As an engineer, I can never actually let a system write code on behalf of me with the level of complacency I've accumulated over the years. I always have opinionated design decisions, variable naming practices. It's memorable, relatable, repeatable across N projects. Sure, you can argue that you can feed all this into the context, but I've found most models to hallucinate and make things unnecessarily opaque and complex. And then, I eventually have to spend time cleaning up all that mess. OP claims they can tell the model over the phone what to do and it does it. Good for OP, but I've never personally had that level of success with my own product development workflow. It sounds too good to be true if this level of autonomy is even possible today without the AI fucking something up.

I liken it to being an author. You want to write a book about people's deepest motivations. Formative experiences, relationships, desires. Society, expectations, disappointment. Characters need to meet and talk at certain times. The plot needs to make sense. You bring it to your editor. He finds you forgot to capitalise a proper noun. You also missed an Oxford comma. You used "their" instead of "they're". He sends you back. You didn't get any feedback about whether it makes sense that the characters did what they did. You are in hell, you won't hear anything about the structure until you fix your commas. Eventually someone invents an automatic editor. It fixes all the little grammar and spelling and punctuation issues for you. Now you can bring the script to an editor who tells you the character needs more development. You are making progress. Your only issue is the Luddites who reckon you aren't a real author, because you tend to fail their LeetGrammar tests, calling you a vibe author.

Weird analogy. This makes sense if you liken this automatic editor to a LSP or compiler of the language you're writing in.

Except that the editor doesn't focus on little things but the structure. It is the job of copy editor to correct all the grammar and bad writing. Copy editor can't be done by AI since it includes fixing logical errors and character names. My understanding is that everybody, including the author, fixes typos when they find them. There is also proofreader at the end to catch typos.

Engineering, to me, is simply "the art of compromise." You can't do that from a high level abstract position. You actually need to stand at the coal face and think about it from time to time. This article encodes an entitled laziness that's destructive to personal skill and quality work.

A lot of more senior coders when they actively try vibe coding a greenfield project find that it does actually work. But only for the first ~10kloc. After that the AI, no matter how well you try to prompt it, will start to destroy existing features accidentally, will add unnecessary convoluted logic to the code, will leave benhind dead code, add random traces "for backwards compatibility", will avoid doing the correct thing as "it is too big of a refactor", doesn't understand that the dev database is not the prod database and avoids migrations. And so forth. I've got 10+ years of coding experience, I am an AI advocate, but not vibe coding. AI is a great tool to help with the boring bits, using it to initialize files, help figure out various approaches, as a first pass code reviewer, helping with configuring, those things all work well. But full-on replacing coders? It's not there yet. Will require an order of magnitude more improvement.

It's fine at adding features on a non-vibecoded 100kloc codebase that you somewhat understand. It's when you're vibecoding from scratch that things tend to spin out at a certain point. I am sure there are ways to get around this sort of wall, but I do think it's currently a thing.

Somewhere on an HN thread I saw someone claiming that they "solved" security problems in their vibe-coded app by adding a "security expert" agent to their workflow. All I could think was, "good luck" and I certainly hope their app never processes anything important...

Found a problem? Slap another agent on top to fix it. It’s hilarious to see how the pendulum’s swung away from “thinking from first principles as a buzzword”. Just engineer, dammit…

That was very vague, but I kinda get where they're coming from. I'm now using pi (the thing openclaw is built on) and within a few days i build a tmux plugin and semaphore plugin^1, and it has automated the way _I_ used to use Claude. The things I disagree with OP is: The usefulness of persistent memory beyond a single line in AGENTS.md "If the user says 'next time' update your AGENTS.md", the use of long-running loops, or the idea that everything can be resolved via chat - might be true for simple projects, but any original work needs me to design the 'right' approach ~5% of the time. That's not a lot, but AI lets you create load-bearing tech-debt within hours, at which point you're stuck with a lot of shit and you dont know how far it got smeared. [1]: https://github.com/offline-ant

Yeah… I'm using Claude Code almost all day every day, but it still 100% requires my judgment. If another AI like OpenClaw was just giving the thumbs up to whatever CC was doing, it would not end well (for my projects anyway).

I played with it extensively for three days. I think there are a few things it does that people are finding interesting: 1. It has a lot of files that it loads into it's context for each conversation, and it consistently updates them. Plus it stores and can reference each conversation. So there's a sense of continuity over time. 2. It connects to messaging services and other accounts of yours, so again it feels continuous. You can use it on your desktop and then pick up your phone and send it an iMessage. 3. It hooks into a lot of things, so it feels like it has more agency. You could send it a voice message over discord and say "hey remember that conversation about birds? Send an email to Steve and ask him what he thinks about it" It feels more like a smart assistant that's always around than an app you open to ask questions to. However, it's worth stressing how terrible the software actually is. Not a single thing I attempted to do worked correctly, important issues (like the discord integration having huge message delays and sometimes dropping messages) get closed because "sorry we have too many issues", and I really got the impression that the whole thing is just a vibe coded pile of garbage. And I don't like to be that critical about an open source project like this, but I think considering the level of hype and the dramatic claims that humans shouldn't be writing code anymore, I think it's worth being clear about. Ended up deleting it and setting up something much simpler. I installed a little discord relay called kimaki, and that lets me interact with instances of opencode over discord when I want to. I also spent some time setting up persistent files and made sure the llm can update them, although only when I ask it to in this case. That's covered enough of what I liked from OpenClaw to satisfy me.

I admire the people that can live happily in the ignorance of what’s under the hood, in this case not even under the layer of claude code because that was too much aparently so people are now putting openclaw+telegram on top of that. And me ruining my day fighting with a million hooks, specs and custom linters micromanaging Claude Code in the pursuit of beautiful code.

I am currently in the process of setting up a local development environment to automate all my programming tasks (dev, test, qa, deploy, debug, etc; for android, ios, mac, windows, linux). It's a serious amount of effort, and a lot of complexity! I could probably move faster if I used AI to set it all up for me rather than setting it up myself. But there's significant danger there in letting an AI "do whatever it wants" on my machine that I'm not willing to accept yet, so the cost of safety is slowness in getting my environment finished. I feel like there's this "secret" hiding behind all these AI tools, that actually it's all very complicated and takes a lot of effort to make work, but the tools we're given hides it all. It's nice that we benefit from its simplicity of use. But hiding complexity leads to unexpected problems, and I'm not sure we've seen any of those yet - other than the massive, gaping security hole.

It is a really impressive tool, but I just can’t trust it to oversee production code. Regardless of how you isolate the OpenClaw instance (Mac Mini, VPS, whatever) - if it’s allowed to browse the web for answers then there’s the very real risk of prompt injection inserting malicious code into the project. If you are personally reviewing every line of code that it generates you can mitigate that, but I’d wager none of these “super manager” users are doing that.

I‘ve done some phone programming over the Xmas holidays with clawdbot. This does work, BUT you absolutely need demand clearly measurable outcomes of the agent, like a closed feedback loop or comparison with a reference implementation, or perfect score in a simulated environment. Without this, the implementation will be incomplete and likely utter crap. Even then, the architecture will be horrible unless you chat _a lot_ about it upfront. At some point, it’s easier to just look in the terminal.