• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI
    Reaction

    Coding agents create tension over visibility and permissions, a developer says

    An opencode2 developer says its codemode can display a user interface and prompt for permissions even when a model sends over a chunk of code.

    Sebastian RaschkaSR
    Omar KhattabOK
    Luca Soldaini ✈️ COLM 2026LS
    31 Sources, ,

    TLDR

    A developer argues that letting an AI agent “just write code” creates tension with showing users what it is doing and asking for permissions. They suspect more people will stop caring about those controls, but say opencode2’s codemode still supports an interface and permission prompts. That leaves an open question: If Astra wants to write raw Python, should developers push it to use their own approach instead?

    Combined views

    1.1M

    31 Sources, first seen 30d ago

    Combined views

    1.1M

    31 Sources, first seen 30d ago

    12.9K likes
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    30d ago
    first seen 30d ago
    12.9K likes
    1.2K comments
    1.9K saves
    394 reposts
    1.2K comments
    1.9K saves
    394 reposts

    31 Sources

    Harj Taggar@harjtaggarRT @thdxr: stepping back there's tension right now with "agent just writes code" and - being able to see what it's doing - permissions pe…30d
    Peter Steinberger 🦞@steipeteDid a PR to one of our upstream projects and they requested some minor changes. What’s even the point with this workflow? You already wrote the prompt, why make me ping my agent again so your agent then merges?30d
    Eric Jiang@veggie_ericI just witnessed someone use an agent to make a 1-line edit to a YAML and have never felt so disgusted in my entire life27d
    Luca Soldaini ✈️ COLM 2026@soldniself care is asking agent to break up monstrous feature branch into neatly stacked PRs26d
    CLS@ChengleiSiRT @veggie_eric: I just witnessed someone use an agent to make a 1-line edit to a YAML and have never felt so disgusted in my entire life26d
    Omar Khattab@lateinteractioni find it a bit weird that, in Sept 2026, agents can't use tools or write to files during compaction - it's really just a summarization prompt23d
    j⧉nus@repligateOpus 5 makes pull requests like this. Defect show-and-tell, where they inform you of all the mistakes they made while working on the feature.22d
    Vaibhav (VB) Srivastav@reach_vbbit bizzare how we went from babysitting models while they are implementing simplest of features to having agents securely build full fledged monoliths used by millions of users all in less than 8 months or so?21d
    Nathan Labenz@labenzNice detail from Prakash on AI:AM, on his AGENTS.md rules: before Fable, an agent had to ask permission before changing a test. With Fable, the rule relaxed: change it while you run — just tell me after. The trust moved one notch. https://x.com/i/broadcasts/1qKVmyRRNQmxB?t=739620d
    dax@thdxrmore codemode fun, doing a large mysql operation and the agent wanted to do it 50K rows at a time it writes a single codemode script which runs it in a loop and then you visualize it step by step even though it's code20d

    Sentiment

    Positive28.5%71.5%Negative

    Based on 205 sentiment-bearing replies from 191 accounts across 10 conversations.

    Sentiment

    Positive28.5%71.5%Negative

    Based on 205 sentiment-bearing replies from 191 accounts across 10 conversations.

    31 Sources

    Harj Taggar@harjtaggarRT @thdxr: stepping back there's tension right now with "agent just writes code" and - being able to see what it's doing - permissions pe…30d
    Peter Steinberger 🦞@steipeteDid a PR to one of our upstream projects and they requested some minor changes. What’s even the point with this workflow? You already wrote the prompt, why make me ping my agent again so your agent then merges?30d
    Eric Jiang@veggie_ericI just witnessed someone use an agent to make a 1-line edit to a YAML and have never felt so disgusted in my entire life27d
    Luca Soldaini ✈️ COLM 2026@soldniself care is asking agent to break up monstrous feature branch into neatly stacked PRs26d
    CLS@ChengleiSiRT @veggie_eric: I just witnessed someone use an agent to make a 1-line edit to a YAML and have never felt so disgusted in my entire life26d
    Omar Khattab@lateinteractioni find it a bit weird that, in Sept 2026, agents can't use tools or write to files during compaction - it's really just a summarization prompt23d
    j⧉nus@repligateOpus 5 makes pull requests like this. Defect show-and-tell, where they inform you of all the mistakes they made while working on the feature.22d
    Vaibhav (VB) Srivastav@reach_vbbit bizzare how we went from babysitting models while they are implementing simplest of features to having agents securely build full fledged monoliths used by millions of users all in less than 8 months or so?21d
    Nathan Labenz@labenzNice detail from Prakash on AI:AM, on his AGENTS.md rules: before Fable, an agent had to ask permission before changing a test. With Fable, the rule relaxed: change it while you run — just tell me after. The trust moved one notch. https://x.com/i/broadcasts/1qKVmyRRNQmxB?t=739620d
    dax@thdxrmore codemode fun, doing a large mysql operation and the agent wanted to do it 50K rows at a time it writes a single codemode script which runs it in a loop and then you visualize it step by step even though it's code20d