llm/9ad11e16-7acb-4923-bb7e-5d14cd36cf3f/topic-10-410849cf-f910-40a9-8f1e-bc05e1e7401e-input.json
The following is content for you to summarize. Do not respond to the comments—summarize them. <topic> Vibe Coding vs. Engineering # There is a distinction drawn between 'vibe coding'—prompting an LLM to generate scripts or apps—and traditional software engineering. Some users view this workflow as 'slop' or merely prompting, lacking the depth of actual problem-solving, while others find it empowering for quick prototypes or hobby projects. This theme reflects broader tensions regarding the changing nature of software development in the age of generative AI. </topic> <comments_about_topic> 1. I read the Readme. So this is all just stuff you can do with Claude's cli interface? It edits files and runs utilities? And it does this with few enough errors that you can be productive by just chatting with it over ssh? Is Claude the only one that can do this? 2. Possibly Codex, but I've only used Claude Code so far. Worth pointing out I'm not SSHing to a different device. Claude Code installed and running directly in Android terminal on my phone. I've built ASP.NET Core APIs on-device this way. Install dotnet in the terminal and Claude can write code, build, run unit tests, and even run the API on localhost. Then use `git` and `gh` to commit, push and raise a PR. 3. Probably Claude Code and Codex are the currently best ones, Claude Code a bit faster, Codex a lot more precise and "engineering" focused. As long as you figure out how to verify that the built thing actually does what it's supposed to, ideally with automated tests, it's almost fire-and-forget if you're good at explaining what you want and need. 4. The LLM crowd kind of likes hype. 5. By “this scheme” I meant combining these several technologies for vibe coding on an iPhone with Claude Code. It’s been a bit of a viral meme on X this week. 6. >Email is clunky and feedback is not immediate. You're vibe coding on a smartphone into an external computer. You already abandoned "Immediate feedback" and "cohesion". 7. i've seen that work well on existing codebases, but bootstrapping a codebase that way is like pulling teeth in my past experiences. so it really sort of falls back to what you're doing with the llm. code maintenance isn't novel development, which isn't polishing. 8. This pretty much sounds like my dream vibe coding dashboard - basically a personal Github populated by AI agents I can assign tasks to. Does this exist yet? Or can something like gitea be setup to behave this way? 9. You're vibe coding. Clearly what you're working on isn't of enough value to secure anyway. 10. Email is funny - maybe as a backup. Prompting is chatting. 11. I'd rather have an DM interface and each task has its own little icon or face. You still have to set up one of the text servers and also do VPN but if you're already vibe coding that stuff why not make it more pleasant than TUI on your phone? 12. It's amazing to me this is called coding at all. Who knew all project managers and business analysts coming up with business requirements were actually just coding gods sent from the future. 13. In the last 5 years I pretty much fully migrated to my laptop being a terminal for other machines. I more use it like a local machine in HPC: web browsing, word processing, scripting. Anything serious is done remotely. But I also live in the terminal and so realistically what's the difference? 99% of the time the result is that I get to use a "big" computer without having to carry it around. FWIW, I'm not a big fan of AI coding. I use AI (including LLMs) and I am an AI researcher, but the vibe coding just hasn't clicked despite constant efforts. I guess it can make more sense to do it if you're programming from your phone because while normally typing isn't the bottleneck it definitely is on the phone (or at least far less comfortable) 14. > A Claude Pro subscription "Doom Slopping" might be more fitting. 15. I think I'll take any LLM slop code over the written-on-phone-by-"developers" slop code. 16. I assume he means people are too poor to have multiple devices, and if you only have one it's probably a phone. That said I'm dubious anyone who only has a phone is doing meaningful coding 17. How do you avoid doom-coding while learning or experimenting? – Ask HN Lately I have observed this algo in myself while learning something new. I constantly code for very short bursts sometimes on the phone or laptop at night, keep jumping between tools and end up consuming more than creating. It comes off as productive but seldom compounds. A straightforward explanation that has provided me with a helpful point of thought is. Make a mode selection. Did conclusions actually occur? Most doom-coding sessions are loaded with input, no closure. There are 2 small changes that improved it for me. Start sessions with a small, visible output goal (one function, one note, one commit). Time-box input aggressively. I stop scrolling after 15-20 minutes of scrolling. At the conclusion of every session, I would write what I would do next, even if I don’t do it. ~ Wanting to know how others do this. Do you intentionally separate learning sessions and building sessions? Do you have any heuristics to know when you have avoided input? 18. > down time is supposed to be down time. Life doesn't have down time. Should we avoid learning new things because no one is paying us to learn? One of my favorite uses of AI is to quickly make some simple 'hello world' level application that I can run using a given technology. Don't know what an MCP server is? Boot up Kiro and tell it you want to make a sample MCP server and ask it for suggestions on what the MCP server should do. A relatively short while later, with a lot of that time being spent letting AI do it's thing, and you can have an MCP server running on your computer. You have an AI waiting for you to ask questions about why the MCP server does x y or z or how can you get the server to do a, b or c etc As someone who learns a lot better from doing or seeing vs reading specs, this has been monumentally more efficient than searching the web for a good blog post explaining the concept. And when I'm doing these learning exercises, I naturally lean towards the domain my company is in because it's easier to visualize how a concept could be implemented into a workflow when I understand the current pain points of that workflow. I'm not going home and pulling in story's from my board and working on them (generally), I'm teaching myself new concepts in a way that also positions be to contribute better to my employer. 19. I've been using a similar workflow for the past couple of months.Heavily inspired by Simon Willison’s approach of building micro tools, I’ve started building micro-utilities. I do this mostly while I'm commuting or outside or waiting for something at work. Instead of just jotting down an idea in a notes app (and it sitting there for eternity), I’ll open up Jules, describe the tool, and have it scaffold the HTML. I have Cloudflare Pages hooked up to the repo—once Jules submits the PR, the preview branch builds automatically and I can verify the result on my phone immediately. 20. It’s largely asynchronous for me. I'll trigger the generation and come back to the PR whenever I'm free. I'd say the cycle time largely depends on the complexity of the tools you are building. I've built a movie shelf hooking up with trakt.tv under 30 minutes and a mermaidJS diagram editor spanning multiple sessions and couple of days. 21. I don’t get it. How is this different from using the Claude iOS (and I assume Android) native app and use their “Code” option. It fires up a Claude Code session in the cloud and you can vibe code anything while on the toilet. 22. If you need this article to get the idea of using Claude Code from your phone, you won’t build anything substantial anyway. 23. Because, obviously, you should be spending all of your waking time thinking about LLMs, agents, and how you can integrate them into every part of your life. If you have been living properly in the age of impending-AGI, you would have already been desperately seeking more opportunities to interact with these systems. That desperation would have led you to independently discover agents and all the ways you could couple yourself to them even when away from your computer. Are you a parent stuck at home experiencing life with your kids instead of sitting at your desk? Why not escape such a hellscape by whipping out your phone and building a SaaS from your phone while your offspring annoys you with requests for attention and meaningless affection? --- Really, this whole environment of 'coding from my phone with dozens of agents while I'm doing the laundry' feels like satire of the sorts of things we used to laugh at on Linkedin. 24. This makes me worry about the future where I will be unable to hire anyone that actually knows how to solve novel engineering problems via programming with a real keyboard on a real computer with their actual brains. To be honest it is already starting to feel that way. 25. Yeah I feel like I'm missing something here. I'm not sure if people being so dependent on these LLMs generating code is that widespread at this point or if this is some kind of publicity stunt. 26. What does Claude add to this? I've done coding on my phone before by sshing into my home server and just... writing code. Is there a benefit to writing code through a third party instead? 27. Yeah, I use Termux a decent amount, whether it's just updating my todo list on my home server or actually programming on it. I feel like this is just aimed at the people who want to code entire projects with LLMs, cost be damned 28. Programming on a phone is a tough sell for many since typing is slower and you have less screen real estate to view/debug the code. Using an AI agent and typing only prompts makes it more compelling. You input less, and only occasionally have to edit code instead of writing everything from 0. And even with editing, typing a prompt like "separate the X logic from class Y into a new file/class" is much faster on mobile than the equivalent actions. 29. I have been doing this with toad and opencode and it is great for those unprompted ideas that pop up while in the big blue room, but not really useful for large projects. 30. This can be done not just with Claude but also with codex and gemeni cli. Well technically anything that has a cli interface. I run both gemeni (fee) and codex (paid), with tmux thrown in to switch between phone and laptop. Laptop runs vscode with ssh to my server but I could also use the web version of vscode. 31. Pretty cool idea, I'm going to be trying this only using open source Cecli (with DeepSeek API) instead of Claude CLI because I don't have infinite $$ 32. ^^ I probably rely on AI slop than most people on this thread. I've found with the gaps with waiting on Claude Code output match the frequency I'm already checking my phone out of addiction. By no means the healthiest way to spend my time, but if I wanted to spin up a simple website or build out the framework for a project doom coding works for me! Agreed 100% there are healthier uses of my time! 33. I love this! This concept on steroids is one of the main reasons I made https://github.com/knowsuchagency/vibora after trying both happy.engineering and Vibe Kanban for remote coding. There's the claude mobile app, too, but I want to run Claude on my own hardware in a terminal 34. I was expecting this to be about using Termux or similar. Why are LLMs involved here? 35. probably because you just install it, then you log in and youre done. tailscale takes care of the rest. going through any more effort just so you can write some slop code is probably not worth it 36. Account created 16 hours ago posting highly dubious AI hype? This user is almost certainly part of the intense astroturfing campaign likely financed by Anthropic that has been ongoing for days/weeks now. 37. Calling "telling the LLM what to do" coding is dishonest, and I have no respect for any of this. 38. If you don’t write a single line of code that’s not coding. Otherwise my customers are coders to. they to the same. The difference is the recipient of the order 39. Vibe coding is not coding (unless by vibe-coding you meant buttplug.io) 40. Vibe coding is such a bad word. It should be called prompting. Thats all it really is. Its like calling a point and click UI programming 41. I feel it depends whether you inspect and edit the code as part of the workflow, or just test what the AI produced and give feedback without participating in the coding yourself. 42. Most of the slop i witness is the latter. This is evident in huge multi 10K pull requests. The code is just an artifact, while the prompting is the "new" coding. 43. I wouldn’t want to code but I could easily be working on plans with Claude. </comments_about_topic> Write a concise, engaging paragraph (3-5 sentences) summarizing the key points and perspectives in these comments about the topic. Focus on the most interesting viewpoints. Do not use bullet points—write flowing prose.
Vibe Coding vs. Engineering # There is a distinction drawn between 'vibe coding'—prompting an LLM to generate scripts or apps—and traditional software engineering. Some users view this workflow as 'slop' or merely prompting, lacking the depth of actual problem-solving, while others find it empowering for quick prototypes or hobby projects. This theme reflects broader tensions regarding the changing nature of software development in the age of generative AI.
43