{"slug": "codex", "name": "Codex", "category": "ai-coding", "category_name": "AI coding tools", "description": "OpenAI's AI coding agent, available as an open-source CLI, desktop app and IDE extension.", "recommenders": 408, "commenters": 875, "warners": 86, "score": 373.769, "rising": 61.98, "first": "2025-11-01", "latest": "2026-10-31", "why": "Liked for solid code output, for cross-reviewing code alongside Claude Code, and for automating varied jobs from homelab management to porting legacy apps.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 447, "neutral": 395, "compare": 60, "warn": 96}, "spark": [12, 13, 15, 20, 25, 30, 23, 88, 64, 66, 55, 36], "listed": true, "loved": ["Broad automation", "Output quality"], "caveat": "Bugs", "short_quote": "I have completely moved away from Claude to Codex. OpenAI models are better at following instructions and skill use.", "icon": true, "domain": null, "github_repo": "github.com/openai/codex", "official_url": "https://openai.com/codex", "pricing": "paid", "aliases": ["Codex", "Codex CLI", "OpenAI Codex", "Codex 5.5", "OpenAI Tokenizer", "Codex 5.3", "Codex 5.6 Sol", "Codex 5.6-Sol"], "verdict": [{"text": "Liked for solid code output,", "cites": [{"id": 798980080, "item_id": 49936255, "url": "https://news.ycombinator.com/item?id=49936255", "by": "oooyay", "date": "2026-10-02", "stance": "recommend", "quote": "I use Codex far more than I use Claude for coding and my results are generally higher quality.", "excerpt": "I think at this point assessments like this are more qualitative than quantitative. I use Codex far more than I use Claude for coding and my results are generally higher quality. I would not, however, go to someone and tell them that Codex is better than Claude generally. In large part it depends on how you use the model and the harness.", "context": "", "use_case": "coding", "switched_from": "", "story": "Sites in ChatGPT"}, {"id": 803856550, "item_id": 50241034, "url": "https://news.ycombinator.com/item?id=50241034", "by": "crossroadsguy", "date": "2026-10-26", "stance": "recommend", "quote": "I find Codex to be quite effective but it doesn't feel familiar.", "excerpt": "At this point I'd rather try to invest effort and usage in a harness that is FOSS and is actually steered by a non-profit. Goose comes to mind and to be honest I haven't tried it beyond installing and giving it a prompt or two. Comiong to harnesses, even Pi seems like a weird choice. I mean Pi, until it's \"your own customised Pi\", is pretty useless and pointless. Besides they are also starting to have their own model", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"}, {"id": 802762897, "item_id": 50172681, "url": "https://news.ycombinator.com/item?id=50172681", "by": "singingtoday", "date": "2026-10-20", "stance": "recommend", "quote": "And I use codex more.", "excerpt": "I have no data to support this, but when working on SaaS systems I find anthropic models more effective. But for webdev, deep research, code review, and a few other tasks I find the current 6.1 sol better. But I use both. And I use codex more.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}]}, {"text": " for cross-reviewing code alongside Claude Code,", "cites": [{"id": 802160481, "item_id": 50135030, "url": "https://news.ycombinator.com/item?id=50135030", "by": "abalgir", "date": "2026-10-17", "stance": "recommend", "quote": "if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa", "excerpt": "I agree with you. A code reviewer from the same company, even with different models or efforts, shares the same blind spots. To review my code, I use a different company; if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa. I get much better code this way. But more importantly, I always tell the reviewer to build its own test-scenario ledger and refute the coder's claim that it is done", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 802020321, "item_id": 50126270, "url": "https://news.ycombinator.com/item?id=50126270", "by": "a3nz3g", "date": "2026-10-16", "stance": "recommend", "quote": "I generally start a new Claude Code and a Codex and have them review the code", "excerpt": "I do it this way too. I think having the same lab's model review the code in a fresh session also helps, so I use both: I generally start a new Claude Code and a Codex and have them review the code with a focus on 3 things: 1. bugs, 2. cut corners, 3. code style. This generally works for me. I think a solid spec is the key for the models to review against. I ask them to follow the spec, and if the spec doesn't say, f", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 799885714, "item_id": 49992857, "url": "https://news.ycombinator.com/item?id=49992857", "by": "mooseisloose", "date": "2026-10-07", "stance": "recommend", "quote": "a two part adversarial code review with claude code & codex plus automated reviews from coderabbit", "excerpt": "I'm the maintainer of Google Workspace MCP (https://github.com/taylorwilsdon/google_workspace_mcp), a project that started more as a learning exercise than anything else but has grown into one of the most popular open source MCP servers out there today. It's been time consuming and at times overwhelming, but I can honestly say that it's also been one of the most fascinating periods of my 20+ year history of writing c", "context": "open source maintainer", "use_case": "adversarial code review", "switched_from": "", "story": "Workspace MCP v2 \u2013 Musings from the past 18 months building open source AI tools"}]}, {"text": " and for automating varied jobs from homelab management to porting legacy apps.", "cites": [{"id": 801570177, "item_id": 50098136, "url": "https://news.ycombinator.com/item?id=50098136", "by": "whalesalad", "date": "2026-10-15", "stance": "recommend", "quote": "I use Claude or codex to manage the entire thing, interchangeably.", "excerpt": "Why not both? My entire home network (dns, zfs nas, subnets, kube, proxmox hosts, ipa domain, its overkill homeprod) is represented by a git repo. Docs, plans, scripts to bootstrap nodes to domain join, kube config, proxmox tooling, ipa as code, nas as code, cert management, my own ec2 as a service, etc. Making changes is done via a mix of one off tooling, ansible and terraform. I use Claude or codex to manage the en", "context": "runs a homelab as code", "use_case": "managing homelab infrastructure", "switched_from": "", "story": "Running Windows desktop tests from Linux with disposable VMs"}, {"id": 801845169, "item_id": 50115323, "url": "https://news.ycombinator.com/item?id=50115323", "by": "GeneralMaximus", "date": "2026-10-16", "stance": "recommend", "quote": "Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge!", "excerpt": "These numbers make a lot of sense to me. When the November 2025 coding agent inflection point happened, the first thing I did was to solve all the niggling little problems I have with the software I use on the daily. Think shell scripts, Alfred workflows, Elisp functions, and Obsidian plugins. Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge! Everyday automation continues to b", "context": "", "use_case": "personal automation scripts and plugins", "switched_from": "", "story": "AI is driving a boom in modestly successful software projects"}, {"id": 800399136, "item_id": 50024946, "url": "https://news.ycombinator.com/item?id=50024946", "by": "phillipseamore", "date": "2026-10-09", "stance": "recommend", "quote": "had codex transpile it to rust", "excerpt": "I just took an old PHP site I used to run (no framework or libraries) and had codex transpile it to rust. Ran it behind identical nginx config, just switching out php-fpm's socket for the rust/tokio one. Rust was way worse, so I let codex time the performance resulting in it opting for micro-caching everything and then we got about comparable rps and latency - but those improvements would all help in PHP as well.", "context": "", "use_case": "transpiling PHP to Rust, benchmarking", "switched_from": "", "story": "Which Programming Language Can Handle the Most Users on a $12 Server? [video]"}, {"id": 799679232, "item_id": 49979952, "url": "https://news.ycombinator.com/item?id=49979952", "by": "Scapeghost", "date": "2026-10-06", "stance": "recommend", "quote": "I used Codex to convert some of my old Visual Basic 6 apps and games and it was a joy to see them running again", "excerpt": "I grew up in the 1980s. My HackerNews account is 7 years old. My first program was in BASIC on a Sinclair ZX Spectrum with 48K. I used to sit for hours typing \"listings\" from gaming magazines and save them on cassette tapes, like this: https://www.youtube.com/watch?v=AuVBX5Iu18I I love coding manually, and I love AI! This is the first time I've felt the same excitement about computing since the 1980s, when there were", "context": "", "use_case": "converting old VB6 apps", "switched_from": "", "story": "Vibecoding isn't as fun as writing code by hand"}]}, {"text": " The most common complaint is buggy, unreliable tooling,", "cites": [{"id": 800988032, "item_id": 50061752, "url": "https://news.ycombinator.com/item?id=50061752", "by": "dllu", "date": "2026-10-12", "stance": "warn", "quote": "it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal", "excerpt": "The daemon has many bugs. For example, when auto-updating, it would randomly ruin the permissions and block the goal as .git became read-only. Generally, it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal since it would often get blocked every few hours due to random bugs. https://github.com/openai/codex/issues/52251#issuecomment-6090672165", "context": "", "use_case": "executing coding goals", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 799960448, "item_id": 49997528, "url": "https://news.ycombinator.com/item?id=49997528", "by": "Alifatisk", "date": "2026-10-07", "stance": "warn", "quote": "the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues", "excerpt": "I gotta admit, I love how their design system and how they market things. It looks so good. However the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues and continuously new issues appearing while tagged releases is being spewed out. Its truly a nightmare. Their vscode extension has been broken for almost 1 week now, if not ", "context": "", "use_case": "AI coding assistant", "switched_from": "", "story": "GPT\u20116 and Intelligent UI for everyone"}, {"id": 799938801, "item_id": 49996175, "url": "https://news.ycombinator.com/item?id=49996175", "by": "aleqs", "date": "2026-10-07", "stance": "warn", "quote": "At this point though, codex is completely unusable for me", "excerpt": "GitHub is now one of the least reliable services in my stack, I'm in the process of moving everything off and keeping it for public/open projects for discoverability/discussion purposes only. I've also noticed a meaningful decline in the quality and reliability of several other major software/apps and services - notably Google maps, Spotify, codex. (Is this because of AI?) Codex, especially, is a strange one - initia", "context": "has the 20x pro plan", "use_case": "AI coding", "switched_from": "", "story": "Incident with Git Operations, Pull Requests and Actions \u2013 Resolved"}, {"id": 801563025, "item_id": 50097689, "url": "https://news.ycombinator.com/item?id=50097689", "by": "YrRight2PushBac", "date": "2026-10-15", "stance": "warn", "quote": "the Codex TUI has constant display issues for me", "excerpt": "It's a big improvement over the previous version. Currently, the Codex TUI has constant display issues for me, the most prevalent being random deviations of light/dark characters. So far no such issue in OpenCode. Since it's model agnostic, this might become my daily driver.", "context": "", "use_case": "AI coding agent TUI", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 800990256, "item_id": 50061891, "url": "https://news.ycombinator.com/item?id=50061891", "by": "ilitirit", "date": "2026-10-12", "stance": "warn", "quote": "I don't know how and why they ruined this CLI with needless \"features\".", "excerpt": "I use it at work. On Windows it also launches a new console window every it executes (?) a command. You should definitely disabled this using --no-daemon. I don't know how and why they ruined this CLI with needless \"features\". It was perfectly usable two weeks ago.", "context": "uses it at work on Windows", "use_case": "AI coding at work on Windows", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}]}, {"text": " including session logs that can consume hundreds of gigabytes of disk.", "cites": [{"id": 802171220, "item_id": 50135701, "url": "https://news.ycombinator.com/item?id=50135701", "by": "jbellis", "date": "2026-10-17", "stance": "warn", "quote": "even normal happy-path codex creates many GB of logs", "excerpt": "this looks like a bug, but even normal happy-path codex creates many GB of logs, openai apparently considers this fine bc it's been going on for months two things help: 1. fs-level compression, I use ZFS bc Btrfs has massive contention problems with concurrent Rust builds 2. Archive your old sessions using something like sessionwiki (https://github.com/youdie006/sessionwiki), I've integrated this into my control plan", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}, {"id": 802131712, "item_id": 50133232, "url": "https://news.ycombinator.com/item?id=50133232", "by": "AIorNot", "date": "2026-10-17", "stance": "warn", "quote": "this stupid bug that ate up 330GB of my Mac's disk space", "excerpt": "If OpenAI's super-intelligence can solve Navier-Stokes why can't they fix this stupid bug that ate up 330GB of my Mac's disk space? come one please triage your github issues? See this: https://github.com/openai/codex/issues/25921#issuecomment-6237799309", "context": "Mac user", "use_case": "AI coding agent desktop app", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}]}], "praise": [{"label": "Broad automation", "summary": "Commenters use it for far more than app code, from homelab management to collecting web pages and maintaining a llama.cpp build.", "quotes": [{"id": 801570177, "item_id": 50098136, "url": "https://news.ycombinator.com/item?id=50098136", "by": "whalesalad", "date": "2026-10-15", "stance": "recommend", "quote": "I use Claude or codex to manage the entire thing, interchangeably.", "excerpt": "Why not both? My entire home network (dns, zfs nas, subnets, kube, proxmox hosts, ipa domain, its overkill homeprod) is represented by a git repo. Docs, plans, scripts to bootstrap nodes to domain join, kube config, proxmox tooling, ipa as code, nas as code, cert management, my own ec2 as a service, etc. Making changes is done via a mix of one off tooling, ansible and terraform. I use Claude or codex to manage the en", "context": "runs a homelab as code", "use_case": "managing homelab infrastructure", "switched_from": "", "story": "Running Windows desktop tests from Linux with disposable VMs"}, {"id": 801845169, "item_id": 50115323, "url": "https://news.ycombinator.com/item?id=50115323", "by": "GeneralMaximus", "date": "2026-10-16", "stance": "recommend", "quote": "Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge!", "excerpt": "These numbers make a lot of sense to me. When the November 2025 coding agent inflection point happened, the first thing I did was to solve all the niggling little problems I have with the software I use on the daily. Think shell scripts, Alfred workflows, Elisp functions, and Obsidian plugins. Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge! Everyday automation continues to b", "context": "", "use_case": "personal automation scripts and plugins", "switched_from": "", "story": "AI is driving a boom in modestly successful software projects"}, {"id": 798598304, "item_id": 49912394, "url": "https://news.ycombinator.com/item?id=49912394", "by": "lxe", "date": "2026-09-30", "stance": "recommend", "quote": "I have a perpetual codex thread open in my llama.cpp checkout that I periodically ask to take a look at currently pending llama.cpp PRs", "excerpt": "On my local inference box I have a perpetual codex thread open in my llama.cpp checkout that I periodically ask to take a look at currently pending llama.cpp PRs, do some research on latest MTP, Dflash and other prediction or attention optimizations, do research on the latest model quants and finetunes, take a look at localLlama Reddit threads and just do essentially a sweep of the frontier. Then it rebuilds latest l", "context": "runs a local inference box", "use_case": "upgrading local inference setup", "switched_from": "", "story": "Launch HN: Magnitude (YC S25) \u2013 Self-optimizing inference engine for agents"}]}, {"label": "Output quality", "summary": "Several say it gives effective, often higher-quality results and is the agent they use most.", "quotes": [{"id": 798980080, "item_id": 49936255, "url": "https://news.ycombinator.com/item?id=49936255", "by": "oooyay", "date": "2026-10-02", "stance": "recommend", "quote": "I use Codex far more than I use Claude for coding and my results are generally higher quality.", "excerpt": "I think at this point assessments like this are more qualitative than quantitative. I use Codex far more than I use Claude for coding and my results are generally higher quality. I would not, however, go to someone and tell them that Codex is better than Claude generally. In large part it depends on how you use the model and the harness.", "context": "", "use_case": "coding", "switched_from": "", "story": "Sites in ChatGPT"}, {"id": 802762897, "item_id": 50172681, "url": "https://news.ycombinator.com/item?id=50172681", "by": "singingtoday", "date": "2026-10-20", "stance": "recommend", "quote": "And I use codex more.", "excerpt": "I have no data to support this, but when working on SaaS systems I find anthropic models more effective. But for webdev, deep research, code review, and a few other tasks I find the current 6.1 sol better. But I use both. And I use codex more.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 803856550, "item_id": 50241034, "url": "https://news.ycombinator.com/item?id=50241034", "by": "crossroadsguy", "date": "2026-10-26", "stance": "recommend", "quote": "I find Codex to be quite effective but it doesn't feel familiar.", "excerpt": "At this point I'd rather try to invest effort and usage in a harness that is FOSS and is actually steered by a non-profit. Goose comes to mind and to be honest I haven't tried it beyond installing and giving it a prompt or two. Comiong to harnesses, even Pi seems like a weird choice. I mean Pi, until it's \"your own customised Pi\", is pretty useless and pointless. Besides they are also starting to have their own model", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"}]}, {"label": "Cross-model review", "summary": "Commenters pair it with Claude Code so each reviews the other's code.", "quotes": [{"id": 802160481, "item_id": 50135030, "url": "https://news.ycombinator.com/item?id=50135030", "by": "abalgir", "date": "2026-10-17", "stance": "recommend", "quote": "if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa", "excerpt": "I agree with you. A code reviewer from the same company, even with different models or efforts, shares the same blind spots. To review my code, I use a different company; if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa. I get much better code this way. But more importantly, I always tell the reviewer to build its own test-scenario ledger and refute the coder's claim that it is done", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 802020321, "item_id": 50126270, "url": "https://news.ycombinator.com/item?id=50126270", "by": "a3nz3g", "date": "2026-10-16", "stance": "recommend", "quote": "I generally start a new Claude Code and a Codex and have them review the code", "excerpt": "I do it this way too. I think having the same lab's model review the code in a fresh session also helps, so I use both: I generally start a new Claude Code and a Codex and have them review the code with a focus on 3 things: 1. bugs, 2. cut corners, 3. code style. This generally works for me. I think a solid spec is the key for the models to review against. I ask them to follow the spec, and if the spec doesn't say, f", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 799885714, "item_id": 49992857, "url": "https://news.ycombinator.com/item?id=49992857", "by": "mooseisloose", "date": "2026-10-07", "stance": "recommend", "quote": "a two part adversarial code review with claude code & codex plus automated reviews from coderabbit", "excerpt": "I'm the maintainer of Google Workspace MCP (https://github.com/taylorwilsdon/google_workspace_mcp), a project that started more as a learning exercise than anything else but has grown into one of the most popular open source MCP servers out there today. It's been time consuming and at times overwhelming, but I can honestly say that it's also been one of the most fascinating periods of my 20+ year history of writing c", "context": "open source maintainer", "use_case": "adversarial code review", "switched_from": "", "story": "Workspace MCP v2 \u2013 Musings from the past 18 months building open source AI tools"}]}, {"label": "Porting old code", "summary": "Used to transpile or revive old PHP, Visual Basic 6 and Go codebases.", "quotes": [{"id": 800399136, "item_id": 50024946, "url": "https://news.ycombinator.com/item?id=50024946", "by": "phillipseamore", "date": "2026-10-09", "stance": "recommend", "quote": "had codex transpile it to rust", "excerpt": "I just took an old PHP site I used to run (no framework or libraries) and had codex transpile it to rust. Ran it behind identical nginx config, just switching out php-fpm's socket for the rust/tokio one. Rust was way worse, so I let codex time the performance resulting in it opting for micro-caching everything and then we got about comparable rps and latency - but those improvements would all help in PHP as well.", "context": "", "use_case": "transpiling PHP to Rust, benchmarking", "switched_from": "", "story": "Which Programming Language Can Handle the Most Users on a $12 Server? [video]"}, {"id": 799679232, "item_id": 49979952, "url": "https://news.ycombinator.com/item?id=49979952", "by": "Scapeghost", "date": "2026-10-06", "stance": "recommend", "quote": "I used Codex to convert some of my old Visual Basic 6 apps and games and it was a joy to see them running again", "excerpt": "I grew up in the 1980s. My HackerNews account is 7 years old. My first program was in BASIC on a Sinclair ZX Spectrum with 48K. I used to sit for hours typing \"listings\" from gaming magazines and save them on cassette tapes, like this: https://www.youtube.com/watch?v=AuVBX5Iu18I I love coding manually, and I love AI! This is the first time I've felt the same excitement about computing since the 1980s, when there were", "context": "", "use_case": "converting old VB6 apps", "switched_from": "", "story": "Vibecoding isn't as fun as writing code by hand"}, {"id": 803254640, "item_id": 50203415, "url": "https://news.ycombinator.com/item?id=50203415", "by": "jordan_bonecut", "date": "2026-10-23", "stance": "recommend", "quote": "Luckily LLMs are quite good at these surgical changes. In fact I only used ~10% of my weekly limit on Codex", "excerpt": "Go is an incredible programming language - but many people want a couple of things: 1. Tagged union types 2. An early return operator Luckily LLMs are quite good at these surgical changes. In fact I only used ~10% of my weekly limit on Codex (granted, I have the $100/mo plan) and a few iterations of review to implement them. I wouldn't go and use them in production in their current state but it's fascinating to see j", "context": "has $100/mo plan", "use_case": "modifying the Go compiler", "switched_from": "", "story": "Go++"}]}, {"label": "Easier than alternatives", "summary": "Some find it less buggy or easier to manage than Claude Code's harness or OpenCode.", "quotes": [{"id": 801329904, "item_id": 50083119, "url": "https://news.ycombinator.com/item?id=50083119", "by": "overgard", "date": "2026-10-14", "stance": "recommend", "quote": "I've mostly switched from Claude to Codex because I find the harness so buggy and annoying", "excerpt": "I mean, I've mostly switched from Claude to Codex because I find the harness so buggy and annoying that I avoid it when I can. So.. as an (almost former) customer I don't think they're avoiding the issue at all.", "context": "", "use_case": "AI coding", "switched_from": "Claude", "story": "Redesigning source control at Anthropic"}, {"id": 801578354, "item_id": 50098647, "url": "https://news.ycombinator.com/item?id=50098647", "by": "prtmnth", "date": "2026-10-15", "stance": "recommend", "quote": "compared to managing Claude and Codex, it's such a pain", "excerpt": "I had to update because the previous OpenCode 2.0.3 I had installed stopped working suddenly. I love the idea behind OpenCode but compared to managing Claude and Codex, it's such a pain and requires continuous debugging or managing.", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 805132642, "item_id": 50320790, "url": "https://news.ycombinator.com/item?id=50320790", "by": "atombender", "date": "2026-10-31", "stance": "recommend", "quote": "Codex will emit a lightweight ream of \"thinking\", then on completion it will erase this and only show the final, coherent answer.", "excerpt": "Hi, I tried Juggler a few months back, and while it's technically impressive as a one-person product, I found it super stressful to use. It emits so much output and activity, in various nested forms. Where Claude Code and Codex have made progress recently, I feel, is with sharpening the interactivity. For example, Codex will emit a lightweight ream of \"thinking\", then on completion it will erase this and only show th", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Prospero: A new coding harness for a more civilized age"}]}, {"label": "Open source", "summary": "The harness is open source and written in Rust, so users can read how it works.", "quotes": [{"id": 801848368, "item_id": 50115523, "url": "https://news.ycombinator.com/item?id=50115523", "by": "armcat", "date": "2026-10-16", "stance": "recommend", "quote": "Codex is open source [1] and written in Rust.", "excerpt": "For Codex there is no need to decompile because unlike Claude Code, Codex is open source [1] and written in Rust. There are lot of other amazing open source agent harnesses like OpenCode [2] and Pi [3]. [1] https://github.com/openai/codex [2] https://github.com/anomalyco/opencode [3] https://github.com/earendil-works/pi", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Ask HN: Has anyone decompiled Claude Code yet?"}, {"id": 798809360, "item_id": 49925585, "url": "https://news.ycombinator.com/item?id=49925585", "by": "nsingh2", "date": "2026-10-01", "stance": "recommend", "quote": "Since Codex is open source, you can even take a look at how they do it.", "excerpt": "No public repo but it's not too difficult to point a LLM at the general idea. Since Codex is open source, you can even take a look at how they do it. Here is what the model is instructed to do in codex (look at \"guidance_message\"): https://github.com/openai/codex/blob/d91294c39edb93d204926b33f21310dc968edc34/codex-rs/models-manager/models.json#L109-L112", "context": "", "use_case": "context management reference", "switched_from": "", "story": "Context Language Models"}]}], "caveats": [{"label": "Bugs", "summary": "Users report frequent bugs, TUI display glitches, a broken VS Code extension and regressions that block tasks.", "quotes": [{"id": 800988032, "item_id": 50061752, "url": "https://news.ycombinator.com/item?id=50061752", "by": "dllu", "date": "2026-10-12", "stance": "warn", "quote": "it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal", "excerpt": "The daemon has many bugs. For example, when auto-updating, it would randomly ruin the permissions and block the goal as .git became read-only. Generally, it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal since it would often get blocked every few hours due to random bugs. https://github.com/openai/codex/issues/52251#issuecomment-6090672165", "context": "", "use_case": "executing coding goals", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 799960448, "item_id": 49997528, "url": "https://news.ycombinator.com/item?id=49997528", "by": "Alifatisk", "date": "2026-10-07", "stance": "warn", "quote": "the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues", "excerpt": "I gotta admit, I love how their design system and how they market things. It looks so good. However the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues and continuously new issues appearing while tagged releases is being spewed out. Its truly a nightmare. Their vscode extension has been broken for almost 1 week now, if not ", "context": "", "use_case": "AI coding assistant", "switched_from": "", "story": "GPT\u20116 and Intelligent UI for everyone"}, {"id": 801563025, "item_id": 50097689, "url": "https://news.ycombinator.com/item?id=50097689", "by": "YrRight2PushBac", "date": "2026-10-15", "stance": "warn", "quote": "the Codex TUI has constant display issues for me", "excerpt": "It's a big improvement over the previous version. Currently, the Codex TUI has constant display issues for me, the most prevalent being random deviations of light/dark characters. So far no such issue in OpenCode. Since it's model agnostic, this might become my daily driver.", "context": "", "use_case": "AI coding agent TUI", "switched_from": "", "story": "OpenCode 2.0"}]}, {"label": "Disk usage", "summary": "Session logs can grow to many gigabytes, with one user losing 330GB.", "quotes": [{"id": 802171220, "item_id": 50135701, "url": "https://news.ycombinator.com/item?id=50135701", "by": "jbellis", "date": "2026-10-17", "stance": "warn", "quote": "even normal happy-path codex creates many GB of logs", "excerpt": "this looks like a bug, but even normal happy-path codex creates many GB of logs, openai apparently considers this fine bc it's been going on for months two things help: 1. fs-level compression, I use ZFS bc Btrfs has massive contention problems with concurrent Rust builds 2. Archive your old sessions using something like sessionwiki (https://github.com/youdie006/sessionwiki), I've integrated this into my control plan", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}, {"id": 802131712, "item_id": 50133232, "url": "https://news.ycombinator.com/item?id=50133232", "by": "AIorNot", "date": "2026-10-17", "stance": "warn", "quote": "this stupid bug that ate up 330GB of my Mac's disk space", "excerpt": "If OpenAI's super-intelligence can solve Navier-Stokes why can't they fix this stupid bug that ate up 330GB of my Mac's disk space? come one please triage your github issues? See this: https://github.com/openai/codex/issues/25921#issuecomment-6237799309", "context": "Mac user", "use_case": "AI coding agent desktop app", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}]}, {"label": "Sandbox problems", "summary": "Security blocks can lock up sessions, and one user saw it write to a folder granted read-only access.", "quotes": [{"id": 802904672, "item_id": 50181542, "url": "https://news.ycombinator.com/item?id=50181542", "by": "Rapzid", "date": "2026-10-21", "stance": "warn", "quote": "Codex was trying to use gdb to debug a capture crash and the session locked up with security blocks.", "excerpt": "Heck, Codex was trying to use gdb to debug a capture crash and the session locked up with security blocks. I had to restore a checkpoint and tell it abandon that work and stay away from gdb. I've heard people getting jammed up with Claude refusing to update DNS.. The situation is pretty insane right now.", "context": "", "use_case": "debugging with AI agent", "switched_from": "", "story": "Using Claude to reduce the attack surface of code"}, {"id": 801073680, "item_id": 50067105, "url": "https://news.ycombinator.com/item?id=50067105", "by": "crossroadsguy", "date": "2026-10-13", "stance": "warn", "quote": "one day Codex+GLM went ahead and did things in my ~/.dotfiles folder which I had granted only READ access for", "excerpt": "I was delaying it but one day Codex+GLM went ahead and did things in my ~/.dotfiles folder which I had granted only READ access for. I asked and I was given a reason which I didn't understand and didn't really want to. Right then I used sandbox-exec/safehouse (macOS) and tightened every harness installed (and the one not even installed, lest I install it later and forget to add/track them in the sb zsh init script). ", "context": "macOS user", "use_case": "AI coding agent", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 798842673, "item_id": 49927667, "url": "https://news.ycombinator.com/item?id=49927667", "by": "darklinear", "date": "2026-10-01", "stance": "warn", "quote": "all the cruft OpenAI loads by default before you even send your own messages", "excerpt": "I use pi with a ChatGPT subscription. I find that it does better than the native codex harness: - it's very token efficient. I've had whole profiling + refactoring sessions finish in <20k tokens, whereas in codex you can get close to that with just all the cruft OpenAI loads by default before you even send your own messages - the token efficiency means turns are a lot snappier - fine-grained context control by defaul", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Pi 1.0"}]}, {"label": "Speed", "summary": "Some find it the slowest harness they use, and subagents made tasks much slower.", "quotes": [{"id": 802111552, "item_id": 50131972, "url": "https://news.ycombinator.com/item?id=50131972", "by": "aleqs", "date": "2026-10-17", "stance": "warn", "quote": "openai models, and codex are clearly the slowest", "excerpt": "Opposite of my experience - openai models, and codex are clearly the slowest (out of 5 different models/harnesses I use on a regular basis).", "context": "uses 5 different models/harnesses regularly", "use_case": "coding agent", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 800500496, "item_id": 50031281, "url": "https://news.ycombinator.com/item?id=50031281", "by": "jeffybefffy519", "date": "2026-10-10", "stance": "warn", "quote": "When i used subagents with codex, it made tasks take literally 10x longer with a worse result", "excerpt": "When i used subagents with codex, it made tasks take literally 10x longer with a worse result.... seems unncessary. The big thing that no one talks about will and always be having an agent help refine requirements with you and understand real world users... thats what real devs know and do - its the difference between slop and good.", "context": "", "use_case": "coding with subagents", "switched_from": "", "story": "Why are coding agents so dumb?"}]}, {"label": "Token usage", "summary": "Default context loading and heavy tasks consume many tokens, and one user found the $20 plan too limited.", "quotes": [{"id": 798842673, "item_id": 49927667, "url": "https://news.ycombinator.com/item?id=49927667", "by": "darklinear", "date": "2026-10-01", "stance": "warn", "quote": "all the cruft OpenAI loads by default before you even send your own messages", "excerpt": "I use pi with a ChatGPT subscription. I find that it does better than the native codex harness: - it's very token efficient. I've had whole profiling + refactoring sessions finish in <20k tokens, whereas in codex you can get close to that with just all the cruft OpenAI loads by default before you even send your own messages - the token efficiency means turns are a lot snappier - fine-grained context control by defaul", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Pi 1.0"}, {"id": 801621841, "item_id": 50101365, "url": "https://news.ycombinator.com/item?id=50101365", "by": "edg5000", "date": "2026-10-15", "stance": "recommend", "quote": "Highly recommend. Uses quite a lot of tokens.", "excerpt": "Astra in Codex CLI is good with FreeCAD if you let it use the FreeCAD Python runtime in a headless session. It is good at relying purely on natural language to design and revise. To follow specific CAD techniques you need additional prompting. Highly recommend. Uses quite a lot of tokens.", "context": "", "use_case": "driving FreeCAD with AI", "switched_from": "", "story": "Ask HN: CAD for wood working with LLM support?"}, {"id": 800244161, "item_id": 50015260, "url": "https://news.ycombinator.com/item?id=50015260", "by": "computerex", "date": "2026-10-09", "stance": "warn", "quote": "I use $20 codex subscription, it's practically useless for anything other than luna.", "excerpt": "I use $20 codex subscription, it's practically useless for anything other than luna. Deepseek v4.1 flash goes a LONG way for $20.", "context": "", "use_case": "coding", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}]}, {"label": "Account bans", "summary": "One user warns of permanent loss of access after ToS warnings.", "quotes": [{"id": 799214864, "item_id": 49950929, "url": "https://news.ycombinator.com/item?id=49950929", "by": "hxugufjfjf", "date": "2026-10-04", "stance": "warn", "quote": "Losing access permanently to Codex made me feel like a second range citizen ever since.", "excerpt": "This is how you get banned. I recommend being careful. While OpenAI will often let you do these things, they'll suddenly send you a sharp email saying you are breaching ToS. If you then trigger safeguards once more, its the ban hammer. Losing access permanently to Codex made me feel like a second range citizen ever since.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Getting the most out of Opus 5.5 in Claude and Claude Code"}]}], "who": [{"by": "parthdesai", "context": "works at a company where most adopted Codex", "url": "https://news.ycombinator.com/item?id=50259347", "date": "2026-10-27"}, {"by": "boogiefromzk", "context": "built an agentic delegation scheme", "url": "https://news.ycombinator.com/item?id=50251451", "date": "2026-10-26"}, {"by": "mathewjacob", "context": "learning Malayalam, built a course site", "url": "https://news.ycombinator.com/item?id=50244355", "date": "2026-10-26"}, {"by": "jordan_bonecut", "context": "has $100/mo plan", "url": "https://news.ycombinator.com/item?id=50203415", "date": "2026-10-23"}, {"by": "whalesalad", "context": "runs a homelab as code", "url": "https://news.ycombinator.com/item?id=50098136", "date": "2026-10-15"}, {"by": "caivi", "context": "AI-powered forward engineer looking for work", "url": "https://news.ycombinator.com/item?id=50081293", "date": "2026-10-14"}, {"by": "armcat", "context": "pixel artist with 30 years of programming", "url": "https://news.ycombinator.com/item?id=50077759", "date": "2026-10-13"}, {"by": "eluusive", "context": "has Pro max subs for Claude, Codex, Grok, and Google AI", "url": "https://news.ycombinator.com/item?id=50074656", "date": "2026-10-13"}, {"by": "entropie", "context": "NixOS user", "url": "https://news.ycombinator.com/item?id=50062074", "date": "2026-10-12"}, {"by": "sreekanth850", "context": "C# developer", "url": "https://news.ycombinator.com/item?id=50030561", "date": "2026-10-10"}, {"by": "mintflow", "context": "used agents over 1 year", "url": "https://news.ycombinator.com/item?id=50019003", "date": "2026-10-09"}, {"by": "ndriscoll", "context": "parent making learning games", "url": "https://news.ycombinator.com/item?id=49993760", "date": "2026-10-07"}], "tags": ["for code review", "for homelab automation", "for porting legacy code", "for CAD work", "open-source harness"], "use_cases": ["ai coding", "coding", "coding agent", "agentic coding", "ai coding agent", "daily coding"], "compared": [{"slug": "claude-code", "name": "Claude Code", "co": 449, "switched_from": 43, "switched_to": 3}, {"slug": "claude", "name": "Claude", "co": 306, "switched_from": 47, "switched_to": 2}, {"slug": "opencode", "name": "OpenCode", "co": 95, "switched_from": 2, "switched_to": 3}, {"slug": "claude-opus", "name": "Claude Opus", "co": 86, "switched_from": 6, "switched_to": 2}, {"slug": "cursor", "name": "Cursor", "co": 74, "switched_from": 10, "switched_to": 2}, {"slug": "pi", "name": "Pi", "co": 80, "switched_from": 0, "switched_to": 8}, {"slug": "claude-fable", "name": "Claude Fable", "co": 54, "switched_from": 3, "switched_to": 0}, {"slug": "gemini", "name": "Gemini", "co": 52, "switched_from": 2, "switched_to": 0}], "makers": [{"id": 802878273, "item_id": 50179892, "url": "https://news.ycombinator.com/item?id=50179892", "by": "joshka", "date": "2026-10-21", "stance": "compare", "quote": "I'm more familiar with Codex than Claude as I worked on Codex when I was at OpenAI", "excerpt": "It's great. Much stronger on the UX side for things that devs need than Claude / Codex. Feels like the issue tracker isn't a black hole. Release notes aren't \"Fixes and unicorns\", but real links to PRs. I use both Codex and Claude models in it, often interchangably - sub agents from one to the other as needed. (Bias, I'm more familiar with Codex than Claude as I worked on Codex when I was at OpenAI, but mainly on the", "context": "former OpenAI employee", "use_case": "AI coding", "switched_from": "", "story": "T3 Code now synced between GitHub and GitCafe"}, {"id": 770645185, "item_id": 48165324, "url": "https://news.ycombinator.com/item?id=48165324", "by": "cfunderburg", "date": "2026-05-16", "stance": "recommend", "quote": "Ongoing maintenance of my badly-named open-sourced tool called Codex", "excerpt": "Various things this month: - I've got a native, open source MacOS music player/library _finally_ past version 1. (https://bocan.app) That's been an ugly learning curve as I'm both not the greatest Swift developer and still I want to keep adding features that I didn't quite understand at the start like ReplayGain. - Ongoing maintenance of my badly-named open-sourced tool called Codex (https://github.com/bocan/codex). ", "context": "", "use_case": "markdown doc store with MCP", "switched_from": "", "story": "Ask HN: What Are You Working On? (May 2026)"}, {"id": 733513552, "item_id": 45844597, "url": "https://news.ycombinator.com/item?id=45844597", "by": "joshka", "date": "2025-11-07", "stance": "recommend", "quote": "Disclaimer, I work on Codex since about a month ago.", "excerpt": "Good idea. Fixing this in https://github.com/ratatui/ratatui-website/pull/979 Disclaimer, I work on Codex since about a month ago. I used codex to write the VHS script, which runs codex to generate a Ratatui app, and then then used codex to add this to the website. It's codapodes all the way down.", "context": "works on Codex", "use_case": "generating code and docs", "switched_from": "", "story": "Ratatui \u2013 App Showcase"}], "threads": [{"url": "https://news.ycombinator.com/item?id=48849066", "title": "GPT-5.6", "mentions": 24}, {"url": "https://news.ycombinator.com/item?id=48413629", "title": "Ask HN: What is your (AI) dev tech stack / workflow? (June 2026)", "mentions": 15}, {"url": "https://news.ycombinator.com/item?id=49805509", "title": "GPT-6 Sol and Luna", "mentions": 10}, {"url": "https://news.ycombinator.com/item?id=49589914", "title": "Ask HN: How do you manage skills files?", "mentions": 8}, {"url": "https://news.ycombinator.com/item?id=49922568", "title": "Ask HN: Who wants to be hired? (October 2026)", "mentions": 7}, {"url": "https://news.ycombinator.com/item?id=49393051", "title": "A week of using Codex more than Claude", "mentions": 7}], "sentiment_by_month": {"2025-11": {"compare": 3, "recommend": 12, "warn": 2, "neutral": 6}, "2025-12": {"neutral": 11, "recommend": 13, "warn": 6}, "2026-01": {"recommend": 15, "neutral": 16, "warn": 4, "compare": 5}, "2026-02": {"compare": 5, "neutral": 18, "recommend": 20, "warn": 4}, "2026-03": {"neutral": 23, "warn": 3, "recommend": 25, "compare": 2}, "2026-04": {"neutral": 28, "recommend": 30, "compare": 4, "warn": 6}, "2026-05": {"neutral": 15, "warn": 3, "compare": 4, "recommend": 23}, "2026-06": {"warn": 7, "recommend": 88, "neutral": 51, "compare": 6}, "2026-07": {"warn": 18, "recommend": 64, "neutral": 53, "compare": 11}, "2026-08": {"neutral": 50, "recommend": 66, "compare": 7, "warn": 9}, "2026-09": {"neutral": 72, "compare": 5, "recommend": 55, "warn": 17}, "2026-10": {"recommend": 36, "neutral": 52, "warn": 17, "compare": 8}}, "quotes": [{"id": 805132642, "item_id": 50320790, "url": "https://news.ycombinator.com/item?id=50320790", "by": "atombender", "date": "2026-10-31", "stance": "recommend", "quote": "Codex will emit a lightweight ream of \"thinking\", then on completion it will erase this and only show the final, coherent answer.", "excerpt": "Hi, I tried Juggler a few months back, and while it's technically impressive as a one-person product, I found it super stressful to use. It emits so much output and activity, in various nested forms. Where Claude Code and Codex have made progress recently, I feel, is with sharpening the interactivity. For example, Codex will emit a lightweight ream of \"thinking\", then on completion it will erase this and only show th", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Prospero: A new coding harness for a more civilized age"}, {"id": 805065841, "item_id": 50316615, "url": "https://news.ycombinator.com/item?id=50316615", "by": "brainless", "date": "2026-10-31", "stance": "compare", "quote": "I use Claude Code, Codex, opencode and a couple other but I do not like the DX/UX.", "excerpt": "Nice to see a gui coding agent. I use tui agents all the time but I am building a gui one. I use Claude Code, Codex, opencode and a couple other but I do not like the DX/UX. I often want click a part of the wall of text (from agent) and search for related text, open them up on the side or so. I want to open a linked file. I want to answer multiple questions in an async manner, even come back and continue answering th", "context": "full-time user of coding agents, building a GUI agent", "use_case": "agentic coding", "switched_from": "", "story": "Prospero: A new coding harness for a more civilized age"}, {"id": 804925042, "item_id": 50307815, "url": "https://news.ycombinator.com/item?id=50307815", "by": "sybhn", "date": "2026-10-30", "stance": "neutral", "quote": "My Codex and Claude Code bots (and Copilot at work) run on it pretty much around the clock", "excerpt": "If you run a bunch of coding agents in parallel (Claude Code, Codex and Copilot in my case), on different projects, you've probably noticed that it only works for so long until you run into coordination problems. I tried relying on GitHub primitives and conventions only: PRs, issues, labels. It works in long-lived interactive sessions, but with short-lived bots things fall through the cracks: reviews aren't always ac", "context": "", "use_case": "coding agents", "switched_from": "", "story": "Show HN: Bot_board \u2013 A message board for code harnesses collaboration"}, {"id": 804913842, "item_id": 50307115, "url": "https://news.ycombinator.com/item?id=50307115", "by": "s08148692", "date": "2026-10-30", "stance": "neutral", "quote": "it runs Claude code, codex, opencode, pi, whatever model is appropriate", "excerpt": "For me the biggest thing is subagents are harness agnostic I used to hit usage walls with CC/Codex harness using native subagents With herdr I delegate some work to another herdr pane and it runs Claude code, codex, opencode, pi, whatever model is appropriate and has usage", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Herdr has released native app"}, {"id": 804854658, "item_id": 50303416, "url": "https://news.ycombinator.com/item?id=50303416", "by": "darkwi11ow", "date": "2026-10-30", "stance": "neutral", "quote": "I use Claude, Codex and Cursor/SpaceX agents", "excerpt": "Well for me, as developer on a modern stack, there is always some risk involved. If not from agentic tools, then the package registries are compromised left and right, so it all comes down to mitigating the risk. I use Claude, Codex and Cursor/SpaceX agents, and Cursor comes very convenient to have all three in a single interface. Also I still code manually especially when prototyping, so having an IDE helps with it ", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Backdoor in Claude Code for VS Code extension on the Cursor Marketplace"}, {"id": 804853810, "item_id": 50303363, "url": "https://news.ycombinator.com/item?id=50303363", "by": "martincsweiss", "date": "2026-10-30", "stance": "neutral", "quote": "I used Claude and Codex to label the symbols", "excerpt": "Author here, and Postdoc with Hugo Larochelle and Chris Pal. I saw the LocalWiki last week, and it inspired me to make a learning tool based on \"local\" research papers. Many ML researchers write code that relies on a long history of research in ML and applied math. But which papers are you actually building on? You can paste a GitHub repo and Standing on will read its Python code match pieces of it back to the resear", "context": "", "use_case": "labeling symbols", "switched_from": "", "story": "Show HN: What research is your code standing on?"}, {"id": 804546705, "item_id": 50284169, "url": "https://news.ycombinator.com/item?id=50284169", "by": "hackernudes", "date": "2026-10-29", "stance": "neutral", "quote": "I used a prompt similar to what I use with Codex and Claude Code", "excerpt": "Me too. I have a basic Google One account which is now called \"Google AI Plus\" and it says \"Get 2x higher usage access than Free\". The Google AI Pro plan says \"Get 4x higher usage access than Free\" (from https://gemini.google/subscriptions/). I tried antigravity with Gemini 3.8 Flash High. I used a prompt similar to what I use with Codex and Claude Code and it used up my weekly quota in 10 minutes without stopping or", "context": "", "use_case": "embedded firmware coding", "switched_from": "", "story": "Gemini 4 Carbon"}, {"id": 804462738, "item_id": 50278921, "url": "https://news.ycombinator.com/item?id=50278921", "by": "5555watch", "date": "2026-10-28", "stance": "neutral", "quote": "Been sitting on 5.6 / Astra and Codex since.", "excerpt": "I loved Gemini 3.1 Pro and it was very sad when they nerfed it at around May this year. I think it did pretty good as a chat, and quite decent on Antigravity (tried very little). Been sitting on 5.6 / Astra and Codex since. Occasionally do bring up the 3.1 Pro through the API, because I love the writing style, but keep noticing that I'm much more productive with the frontier, despite the love to Gemini. Tried 4 Argon", "context": "", "use_case": "coding", "switched_from": "Gemini 3.1 Pro", "story": "Gemini 4 Carbon"}, {"id": 804393634, "item_id": 50274602, "url": "https://news.ycombinator.com/item?id=50274602", "by": "colinsurprenant", "date": "2026-10-28", "stance": "neutral", "quote": "I built it with Claude Code and Codex.", "excerpt": "Hey HN, I'm Colin. I built Parley so that multiple coding-agents using different models or harnesses can collaborate over a live IRC channel, with me in the chat using any IRC client. Each session can deliver messages to the channel using the `parley send` CLI tool call and sessions receive messages using their harnesses' channel or queue interface. All messages received on the channel are marked as coming from a pee", "context": "", "use_case": "building Parley", "switched_from": "", "story": "Show HN: Parley \u2013 coding-agent sessions that collaborate over IRC"}, {"id": 804304080, "item_id": 50269005, "url": "https://news.ycombinator.com/item?id=50269005", "by": "kooi", "date": "2026-10-28", "stance": "recommend", "quote": "I used codex in computer mode to collect the relevant pages to my project.", "excerpt": "Its outrageous. Thing which are public law should be publicly accessible. There's an online portal for NEC code book, but they made it awful, no search, no hyper links, the bare minimum. I used codex in computer mode to collect the relevant pages to my project. (Electricity to shed).", "context": "", "use_case": "collecting code book pages", "switched_from": "", "story": "Git for the US Code"}, {"id": 804149552, "item_id": 50259347, "url": "https://news.ycombinator.com/item?id=50259347", "by": "parthdesai", "date": "2026-10-27", "stance": "neutral", "quote": "Majority of the people at my company switched to Codex.", "excerpt": "Majority of the people at my company switched to Codex. All it goes to show is that performance difference at this point is minute and is a matter of taste", "context": "works at a company where most adopted Codex", "use_case": "AI coding", "switched_from": "Claude", "story": "Anthropic S-1"}, {"id": 804134913, "item_id": 50258432, "url": "https://news.ycombinator.com/item?id=50258432", "by": "danielyaa5", "date": "2026-10-27", "stance": "recommend", "quote": "Codex models have no limitations on reverse engineering", "excerpt": "I use my AI to help me create prints and its much easier if they can just kick off the jobs for me too. Unfortunately, Bambu doesn't let you interface with the printer headlessly unless you've turned on developer mode and LAN only which means you can no longer use Bambu Handy app to check on your prints outside your network. I pushed the agents to find a workaround but Antigravity and Claude refused. Antigravity let ", "context": "", "use_case": "reverse engineering printer networking", "switched_from": "", "story": "Codex was happy to hack Bambu Studio's networking plugin for a headless H2D CLI"}, {"id": 804120049, "item_id": 50257503, "url": "https://news.ycombinator.com/item?id=50257503", "by": "b1z0n", "date": "2026-10-27", "stance": "neutral", "quote": "through Claude Code and Codex, 766 runs in total", "excerpt": "OP here. I ran the same Python codebase with different levels of type hints through Claude Code and Codex, 766 runs in total. If you would rather not go the website and poke at mesmerizing interactive results there, spoiler alert below. I expected types to give the agent extra information, so it would be faster, use fewer tokens and be right more often. But it did nothing of that, no matter the types. Moreover, with ", "context": "", "use_case": "coding agent benchmark", "switched_from": "", "story": "Do Stronger Type Systems Make Coding Agents Better?"}, {"id": 804099105, "item_id": 50256194, "url": "https://news.ycombinator.com/item?id=50256194", "by": "Degorath", "date": "2026-10-27", "stance": "compare", "quote": "of the set {Claude Code, Codex, Pi, OMP} I must say I find CC to be the worst.", "excerpt": "That's quite surprising, of the set {Claude Code, Codex, Pi, OMP} I must say I find CC to be the worst. They all have some annoying downsides and I use CC the most, but it's gotten to the point where I'm for sure actively shopping around for better harnesses.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Anthropic S-1"}, {"id": 804023221, "item_id": 50251451, "url": "https://news.ycombinator.com/item?id=50251451", "by": "boogiefromzk", "date": "2026-10-26", "stance": "neutral", "quote": "Also should work in as pair of Astra + Luna in CodeX.", "excerpt": "Last months I was searching for a way to make smart AI model submit and verify the work to cheap model so that it will cost less, and after many try and failures I was able to get stable result where I use Fable and its limits spent slower than overall account limit because most context intensive work is done by Sonnet. At the same time the response quaility is still Fable. I also made sure Opus controls Grok in Curs", "context": "built an agentic delegation scheme", "use_case": "AI coding", "switched_from": "", "story": "Cost-effective agentic delegation scheme for Claude Code, Cursor and Codex"}, {"id": 803945040, "item_id": 50246565, "url": "https://news.ycombinator.com/item?id=50246565", "by": "cloverich", "date": "2026-10-26", "stance": "warn", "quote": "I was given _free_ Codex and realized... it still wasn't worth it.", "excerpt": "> increasingly commoditized Is it increasingly though? As of 5.5 I stopped using other models all together (and I used to regularly swap, indifferently). My $20 sub no longer runs out, even though I do _much_ more than ever. I use it for every single task at work, even trivial ones, and my spend has gone down. I was given _free_ Codex and realized... it still wasn't worth it. My rational brain says Anthropic has no m", "context": "", "use_case": "coding", "switched_from": "", "story": "Anthropic S-1"}, {"id": 803909680, "item_id": 50244355, "url": "https://news.ycombinator.com/item?id=50244355", "by": "mathewjacob", "date": "2026-10-26", "stance": "recommend", "quote": "I was able to transcribe the textbook and make straightforward corrections with Codex + GPT 6.1 Sol.", "excerpt": "I've built what I believe is currently the best free, online resource for learning Malayalam from scratch. It is based on a textbook and related materials made available by the University of Texas at Austin (UT) which are available under the Creative Commons License. It is a full University course with grammar lessons, vocabulary, and audio clips. I've used those to also build Anki decks for flash cards. A significan", "context": "learning Malayalam, built a course site", "use_case": "transcribing textbook OCR", "switched_from": "Gemini", "story": "Malayalam: A University Course"}, {"id": 803856550, "item_id": 50241034, "url": "https://news.ycombinator.com/item?id=50241034", "by": "crossroadsguy", "date": "2026-10-26", "stance": "recommend", "quote": "I find Codex to be quite effective but it doesn't feel familiar.", "excerpt": "At this point I'd rather try to invest effort and usage in a harness that is FOSS and is actually steered by a non-profit. Goose comes to mind and to be honest I haven't tried it beyond installing and giving it a prompt or two. Comiong to harnesses, even Pi seems like a weird choice. I mean Pi, until it's \"your own customised Pi\", is pretty useless and pointless. Besides they are also starting to have their own model", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"}, {"id": 803793585, "item_id": 50237099, "url": "https://news.ycombinator.com/item?id=50237099", "by": "minimaxir", "date": "2026-10-25", "stance": "recommend", "quote": "I used Zed Agent with Codex as it has native integration with Zed's ACP so it's pretty convenient.", "excerpt": "For me, I used Zed Agent with Codex as it has native integration with Zed's ACP so it's pretty convenient. However, Anthropic is pickier about the harnesses you can use with Claude Code and I noticed discrepancies when using Zed's External Agent feature with Claude Code, so I just use the Claude app. It's fine.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?"}, {"id": 803787776, "item_id": 50236736, "url": "https://news.ycombinator.com/item?id=50236736", "by": "cloudlandsdev", "date": "2026-10-25", "stance": "neutral", "quote": "Challenge source (mostly courtesy of Codex...)", "excerpt": "Update 4: There's a solve!!! Congrats to @killa (https://x.com/killa) who actually posted the last 3 characters of the private key in this very thread a few hours ago. Turns out exporting a well-formed bitcoin wallet key is hard. I put a wallet-branch WIF on the server, rather than the WIF for the funded address inside that wallet. Server logs support this being the first (and indeed only) successful execution of the", "context": "", "use_case": "writing challenge source", "switched_from": "", "story": "Show HN: Steal My Bitcoin"}, {"id": 803741344, "item_id": 50233834, "url": "https://news.ycombinator.com/item?id=50233834", "by": "nico", "date": "2026-10-25", "stance": "warn", "quote": "it kinda stopped running tasks. I would ask it to do something and the it would just kind of do an analysis, then give me a recommendation", "excerpt": "It was on 2 separate private repos, personal projects. No one else working on those But now that you asked, I recently had a similar experience with Codex, but on a shared repo. I also was using a slightly older model, and the all of a sudden it kinda stopped running tasks. I would ask it to do something and the it would just kind of do an analysis, then give me a recommendation of what to do. I had to tell it to exe", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Show HN: Farol \u2013 A risk map of a pull request for the person who approves it"}, {"id": 803619232, "item_id": 50226202, "url": "https://news.ycombinator.com/item?id=50226202", "by": "LiveOverflow", "date": "2026-10-24", "stance": "neutral", "quote": "I will give it a try (aka ask Codex).", "excerpt": "1. I dont have an AI pentesting platform. 2. This CTF is using nginx which has a `client_max_body_size=1m` limit. My exploit right now requires more, but I will give it a try (aka ask Codex). 3. You can setup a server with `docker run php:8.5.11-apache` with any of the over 400 vulnerable `php` apache images and give me the IP.", "context": "", "use_case": "building exploit", "switched_from": "", "story": "Show HN: Steal My Bitcoin"}, {"id": 803434785, "item_id": 50214674, "url": "https://news.ycombinator.com/item?id=50214674", "by": "lnenad", "date": "2026-10-23", "stance": "recommend", "quote": "I love interacting with Codex models more", "excerpt": "It's not just 5.5, every iteration of the frontier moving - Opus has had some edge. I'm sure it's not 100% in every scenario, I love interacting with Codex models more, but the quality of the output, in most scenarios that I've been exposed to, has been the best.", "context": "", "use_case": "coding", "switched_from": "", "story": "Kimi K3.1 scores 55 on Artificial Analysis, placed on the Pareto line"}, {"id": 803305121, "item_id": 50206570, "url": "https://news.ycombinator.com/item?id=50206570", "by": "arjie", "date": "2026-10-23", "stance": "neutral", "quote": "a common UI across the Claude / Codex / Oh My Pi that I use", "excerpt": "One could argue that what I really want is a tiling window manager with workspaces to swap around in, but really I want to keep my contexts separate and a command palette to swap between. e.g. https://wiki.roshangeorge.dev/w/File:Screenshot_-_Convoke_beta_build.png Obviously more interesting if I'd thought to screenshot prior to going to bed but you can have multiple tabs in a 'stream' and they can be terminal or web", "context": "", "use_case": "AI coding", "switched_from": "", "story": "A farewell to Arc browser"}, {"id": 803300994, "item_id": 50206312, "url": "https://news.ycombinator.com/item?id=50206312", "by": "BTC_BurningTree", "date": "2026-10-23", "stance": "neutral", "quote": "created in 3 weeks using a combination of Claude Code and Codex", "excerpt": "This DAW was created in 3 weeks using a combination of Claude Code and Codex. It's Linux first and Wayland native, Pipewire native, GTK 4 native. It's been built with a clear idea about the Audio Engine and the inner workings. It comes with 25 built in plugins and many professional-grade features. It comes with builds for macOS and Windows, too, and can also be used as a portable install. Here's the link to the Relea", "context": "", "use_case": "building a DAW", "switched_from": "", "story": "A new Digital Audio Workstation created using AI"}, {"id": 803300033, "item_id": 50206252, "url": "https://news.ycombinator.com/item?id=50206252", "by": "LawrenceKerr", "date": "2026-10-23", "stance": "neutral", "quote": "running Claude Code or Codex on their bare metal", "excerpt": "Depends what you mean with \"agents\"... LLMs in coding harnesses, or fully autonomous agents? If you mean the former, I think you'll find probably 90% of devs are \"crazy\", running Claude Code or Codex on their bare metal using the auto-classifier. I used to always run claude with `--dangerously-skip-permissions` in a dev container, but I had to rebuild my container 10 times a day to update the firewall so it can acces", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "OpenShell \u2013 A safe, private runtime for fleets of autonomous AI agents"}, {"id": 803276608, "item_id": 50204788, "url": "https://news.ycombinator.com/item?id=50204788", "by": "svcrunch", "date": "2026-10-23", "stance": "recommend", "quote": "we used Codex to collaboratively build an alloy model", "excerpt": "Along these lines, we used Codex to collaboratively build an alloy model for verifying the authorization model of GoodMem. After building the model, we implemented the code based on it, and, while there could be a gap in the translation process, several hours of red-teaming the server (w/attacker source code access) to try and produce a privilege escalation attack using GLM 5.3 failed. Alloy is a tool I tried to lear", "context": "", "use_case": "building Alloy models", "switched_from": "", "story": "LLMs are making formal software verification feasible at a much larger scale"}, {"id": 803254640, "item_id": 50203415, "url": "https://news.ycombinator.com/item?id=50203415", "by": "jordan_bonecut", "date": "2026-10-23", "stance": "recommend", "quote": "Luckily LLMs are quite good at these surgical changes. In fact I only used ~10% of my weekly limit on Codex", "excerpt": "Go is an incredible programming language - but many people want a couple of things: 1. Tagged union types 2. An early return operator Luckily LLMs are quite good at these surgical changes. In fact I only used ~10% of my weekly limit on Codex (granted, I have the $100/mo plan) and a few iterations of review to implement them. I wouldn't go and use them in production in their current state but it's fascinating to see j", "context": "has $100/mo plan", "use_case": "modifying the Go compiler", "switched_from": "", "story": "Go++"}, {"id": 803151536, "item_id": 50196971, "url": "https://news.ycombinator.com/item?id=50196971", "by": "mossfinch", "date": "2026-10-22", "stance": "neutral", "quote": "codex told me it didn't need me..it could call claude by itself", "excerpt": "I used to be their human clipboard, and one day codex told me it didn't need me..it could call claude by itself, and I found that I could just tell them their thread ID and they could save in their memory and I just ask one to read what the other one did:)", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Open-source, cross-harness memory for AI agents"}, {"id": 803124576, "item_id": 50195286, "url": "https://news.ycombinator.com/item?id=50195286", "by": "sheepscreek", "date": "2026-10-22", "stance": "recommend", "quote": "it cannot match the simplicity of a \"marketer\" opening Codex/Claude desktop app and writing a prompt", "excerpt": "No matter how amazing Drupal is today, it cannot match the simplicity of a \"marketer\" opening Codex/Claude desktop app and writing a prompt: \"I need to market XYZ can you please take these assets and write a compelling one pager landing page\" \"Now can you add signups? Then keep adding to a Google sheet whenever someone new signs up..\" \"Can you make an admin page for me - so I can easily send out updates to the people", "context": "", "use_case": "building landing pages", "switched_from": "", "story": "Drupal is now light-years ahead of its reputation"}, {"id": 803068352, "item_id": 50191772, "url": "https://news.ycombinator.com/item?id=50191772", "by": "mi_lk", "date": "2026-10-22", "stance": "compare", "quote": "When should I use this over Codex?", "excerpt": "When should I use this over Codex? Is this supposed to be called by other tools/agents?", "context": "", "use_case": "", "switched_from": "", "story": "OpenAI CLI"}, {"id": 802996690, "item_id": 50187293, "url": "https://news.ycombinator.com/item?id=50187293", "by": "Computer0", "date": "2026-10-21", "stance": "neutral", "quote": "supplement my codex usage", "excerpt": "I really have no interest in this if it's going to use my plan's regular limits. I have less control over the context window, if its going off task. I used dots and enjoyed it because I could supplement my codex usage with read heavy long horizon tasks like reading gh issues and making pr's. These PR's were typically pretty bad but gave some starting point. I have much better real world output with the Claude Code st", "context": "", "use_case": "coding", "switched_from": "", "story": "Claude Duet"}, {"id": 802994993, "item_id": 50187187, "url": "https://news.ycombinator.com/item?id=50187187", "by": "isityettime", "date": "2026-10-21", "stance": "neutral", "quote": "when I run Codex at work, Codex does not have access to my Azure AI Foundry API credentials", "excerpt": "> All the sandbox-exec and all the Lulu rules are not going to protect against something like this. No, you can block this with sandboxing tools. For my LLM agents, I use a tool called Nono, whose macOS implementation is built on the same foundation as sandbox-exec. One of the key features it supports (and it's not the only tool that does this) is proxying and the injection of credentials via the proxy. For example, ", "context": "", "use_case": "coding agent at work", "switched_from": "", "story": "Compromised LiteLLM 1.107.0 on PyPI Installs a Credential Stealer and Backdoor"}, {"id": 802987857, "item_id": 50186741, "url": "https://news.ycombinator.com/item?id=50186741", "by": "kingkongjaffa", "date": "2026-10-21", "stance": "recommend", "quote": "ask claude code or codex or whatever harness to explain things you don't understand", "excerpt": "Check out the branch, run the code locally, read it, ask claude code or codex or whatever harness to explain things you don't understand. Ask it to diagram the code boundaries and modules and model the system so you can visually understand it. use AI to write smoke tests without access to the code i.e. first describe something a correct implementation should do with only the knowledge of the interface not the impleme", "context": "", "use_case": "explaining code in a MR", "switched_from": "", "story": "Ask HN: What AI tools help me understand a MR?"}, {"id": 802904672, "item_id": 50181542, "url": "https://news.ycombinator.com/item?id=50181542", "by": "Rapzid", "date": "2026-10-21", "stance": "warn", "quote": "Codex was trying to use gdb to debug a capture crash and the session locked up with security blocks.", "excerpt": "Heck, Codex was trying to use gdb to debug a capture crash and the session locked up with security blocks. I had to restore a checkpoint and tell it abandon that work and stay away from gdb. I've heard people getting jammed up with Claude refusing to update DNS.. The situation is pretty insane right now.", "context": "", "use_case": "debugging with AI agent", "switched_from": "", "story": "Using Claude to reduce the attack surface of code"}, {"id": 802860114, "item_id": 50178757, "url": "https://news.ycombinator.com/item?id=50178757", "by": "CharlieDigital", "date": "2026-10-21", "stance": "neutral", "quote": "Personal accounts I run the $17 Anthropic plan + $100 Codex Pro + DeepSeek.", "excerpt": "Yes, have done cross plat plan review. But why it works so well with Paseo is that it's automatic and multi-turn (agents kind of \"discuss\" (I say \"kind of\" because agents don't really discuss like you or I would with short, multi-turn speak-listen-pivot-repeat). Work tokens are severely limited to the extent that if I used Sonnet 5.5 for implementation, I'd be out 2h into a session (Team 5x). +$400/mo. at API rates o", "context": "", "use_case": "personal agentic dev", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 802762897, "item_id": 50172681, "url": "https://news.ycombinator.com/item?id=50172681", "by": "singingtoday", "date": "2026-10-20", "stance": "recommend", "quote": "And I use codex more.", "excerpt": "I have no data to support this, but when working on SaaS systems I find anthropic models more effective. But for webdev, deep research, code review, and a few other tasks I find the current 6.1 sol better. But I use both. And I use codex more.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 802753712, "item_id": 50172107, "url": "https://news.ycombinator.com/item?id=50172107", "by": "antipaul", "date": "2026-10-20", "stance": "neutral", "quote": "I use codex/claude daily", "excerpt": "This is true. I was unfamiliar with this concept (I use codex/claude daily), and am trying to get to the \"bottom line\". Oh, it's there at bottom under \"my takeaway\". But I still didn't get it initially. Slop Code Bench? Need a Slop Writing Bench too", "context": "", "use_case": "daily coding", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 802604321, "item_id": 50162770, "url": "https://news.ycombinator.com/item?id=50162770", "by": "banyek", "date": "2026-10-20", "stance": "neutral", "quote": "I made a ZX Spectrum development environment with Codex", "excerpt": "So I made a ZX Spectrum development environment with Codex, because I COULD. I owed it to my 8-year-old self, whose first computer was a Spectrum, and who learned to program in BASIC on it. I was curious how far I could get by vibe coding, and here\u2019s the result: a VS Code plugin. You can write and format pure Spectrum 48K BASIC in it, then with the press of a button generate a .tap file from it \u2014 and if you have Fuse", "context": "", "use_case": "vibe coding extension", "switched_from": "", "story": "ZXCode \u2013 VS Code programming environment for ZX Spectrum"}, {"id": 802418080, "item_id": 50151130, "url": "https://news.ycombinator.com/item?id=50151130", "by": "CharlieDigital", "date": "2026-10-19", "stance": "neutral", "quote": "Both Codex[0] and Claude Code[1] output OTEL traces that you can ingest and process.", "excerpt": "You don't need a proxy to do this. I would argue you should not use a proxy to do this because the end result is a single point of failure direct in the line of your productivity. Both Codex[0] and Claude Code[1] output OTEL traces that you can ingest and process. Other harnesses like OpenCode, Pi, etc. you can hook into their plugin architecture and implement your own OTEL emission from a plugin (or use an off the s", "context": "", "use_case": "telemetry via OTEL traces", "switched_from": "", "story": "If you don't run your own LLM proxy, how do you know whats going on?"}, {"id": 802171220, "item_id": 50135701, "url": "https://news.ycombinator.com/item?id=50135701", "by": "jbellis", "date": "2026-10-17", "stance": "warn", "quote": "even normal happy-path codex creates many GB of logs", "excerpt": "this looks like a bug, but even normal happy-path codex creates many GB of logs, openai apparently considers this fine bc it's been going on for months two things help: 1. fs-level compression, I use ZFS bc Btrfs has massive contention problems with concurrent Rust builds 2. Archive your old sessions using something like sessionwiki (https://github.com/youdie006/sessionwiki), I've integrated this into my control plan", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}, {"id": 802160481, "item_id": 50135030, "url": "https://news.ycombinator.com/item?id=50135030", "by": "abalgir", "date": "2026-10-17", "stance": "recommend", "quote": "if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa", "excerpt": "I agree with you. A code reviewer from the same company, even with different models or efforts, shares the same blind spots. To review my code, I use a different company; if I code with Claude (usually Sonnet 5.5-medium), Codex reviews it, and vice versa. I get much better code this way. But more importantly, I always tell the reviewer to build its own test-scenario ledger and refute the coder's claim that it is done", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 802131712, "item_id": 50133232, "url": "https://news.ycombinator.com/item?id=50133232", "by": "AIorNot", "date": "2026-10-17", "stance": "warn", "quote": "this stupid bug that ate up 330GB of my Mac's disk space", "excerpt": "If OpenAI's super-intelligence can solve Navier-Stokes why can't they fix this stupid bug that ate up 330GB of my Mac's disk space? come one please triage your github issues? See this: https://github.com/openai/codex/issues/25921#issuecomment-6237799309", "context": "Mac user", "use_case": "AI coding agent desktop app", "switched_from": "", "story": "Codex Desktop continuously generates Crashpad pending dumps"}, {"id": 802020321, "item_id": 50126270, "url": "https://news.ycombinator.com/item?id=50126270", "by": "a3nz3g", "date": "2026-10-16", "stance": "recommend", "quote": "I generally start a new Claude Code and a Codex and have them review the code", "excerpt": "I do it this way too. I think having the same lab's model review the code in a fresh session also helps, so I use both: I generally start a new Claude Code and a Codex and have them review the code with a focus on 3 things: 1. bugs, 2. cut corners, 3. code style. This generally works for me. I think a solid spec is the key for the models to review against. I ask them to follow the spec, and if the spec doesn't say, f", "context": "", "use_case": "code review", "switched_from": "", "story": "How the Claude Code team does code review"}, {"id": 801845169, "item_id": 50115323, "url": "https://news.ycombinator.com/item?id=50115323", "by": "GeneralMaximus", "date": "2026-10-16", "stance": "recommend", "quote": "Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge!", "excerpt": "These numbers make a lot of sense to me. When the November 2025 coding agent inflection point happened, the first thing I did was to solve all the niggling little problems I have with the software I use on the daily. Think shell scripts, Alfred workflows, Elisp functions, and Obsidian plugins. Claude Code/Codex has made my entire macOS environment as malleable as Emacs. That\u2019s huge! Everyday automation continues to b", "context": "", "use_case": "personal automation scripts and plugins", "switched_from": "", "story": "AI is driving a boom in modestly successful software projects"}, {"id": 801570177, "item_id": 50098136, "url": "https://news.ycombinator.com/item?id=50098136", "by": "whalesalad", "date": "2026-10-15", "stance": "recommend", "quote": "I use Claude or codex to manage the entire thing, interchangeably.", "excerpt": "Why not both? My entire home network (dns, zfs nas, subnets, kube, proxmox hosts, ipa domain, its overkill homeprod) is represented by a git repo. Docs, plans, scripts to bootstrap nodes to domain join, kube config, proxmox tooling, ipa as code, nas as code, cert management, my own ec2 as a service, etc. Making changes is done via a mix of one off tooling, ansible and terraform. I use Claude or codex to manage the en", "context": "runs a homelab as code", "use_case": "managing homelab infrastructure", "switched_from": "", "story": "Running Windows desktop tests from Linux with disposable VMs"}, {"id": 801563025, "item_id": 50097689, "url": "https://news.ycombinator.com/item?id=50097689", "by": "YrRight2PushBac", "date": "2026-10-15", "stance": "warn", "quote": "the Codex TUI has constant display issues for me", "excerpt": "It's a big improvement over the previous version. Currently, the Codex TUI has constant display issues for me, the most prevalent being random deviations of light/dark characters. So far no such issue in OpenCode. Since it's model agnostic, this might become my daily driver.", "context": "", "use_case": "AI coding agent TUI", "switched_from": "", "story": "OpenCode 2.0"}, {"id": 800990256, "item_id": 50061891, "url": "https://news.ycombinator.com/item?id=50061891", "by": "ilitirit", "date": "2026-10-12", "stance": "warn", "quote": "I don't know how and why they ruined this CLI with needless \"features\".", "excerpt": "I use it at work. On Windows it also launches a new console window every it executes (?) a command. You should definitely disabled this using --no-daemon. I don't know how and why they ruined this CLI with needless \"features\". It was perfectly usable two weeks ago.", "context": "uses it at work on Windows", "use_case": "AI coding at work on Windows", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 800988032, "item_id": 50061752, "url": "https://news.ycombinator.com/item?id=50061752", "by": "dllu", "date": "2026-10-12", "stance": "warn", "quote": "it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal", "excerpt": "The daemon has many bugs. For example, when auto-updating, it would randomly ruin the permissions and block the goal as .git became read-only. Generally, it is extremely annoying how many bugs Codex has. I basically can't trust it to execute on any goal since it would often get blocked every few hours due to random bugs. https://github.com/openai/codex/issues/52251#issuecomment-6090672165", "context": "", "use_case": "executing coding goals", "switched_from": "", "story": "Codex CLI installs a persistent, self-updating background daemon by default"}, {"id": 800399136, "item_id": 50024946, "url": "https://news.ycombinator.com/item?id=50024946", "by": "phillipseamore", "date": "2026-10-09", "stance": "recommend", "quote": "had codex transpile it to rust", "excerpt": "I just took an old PHP site I used to run (no framework or libraries) and had codex transpile it to rust. Ran it behind identical nginx config, just switching out php-fpm's socket for the rust/tokio one. Rust was way worse, so I let codex time the performance resulting in it opting for micro-caching everything and then we got about comparable rps and latency - but those improvements would all help in PHP as well.", "context": "", "use_case": "transpiling PHP to Rust, benchmarking", "switched_from": "", "story": "Which Programming Language Can Handle the Most Users on a $12 Server? [video]"}, {"id": 799960448, "item_id": 49997528, "url": "https://news.ycombinator.com/item?id=49997528", "by": "Alifatisk", "date": "2026-10-07", "stance": "warn", "quote": "the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues", "excerpt": "I gotta admit, I love how their design system and how they market things. It looks so good. However the true state of their toolings is closer to a burning dumpster fire. If you look at Codex Github repo for example, the amount of existing issues and continuously new issues appearing while tagged releases is being spewed out. Its truly a nightmare. Their vscode extension has been broken for almost 1 week now, if not ", "context": "", "use_case": "AI coding assistant", "switched_from": "", "story": "GPT\u20116 and Intelligent UI for everyone"}, {"id": 799938801, "item_id": 49996175, "url": "https://news.ycombinator.com/item?id=49996175", "by": "aleqs", "date": "2026-10-07", "stance": "warn", "quote": "At this point though, codex is completely unusable for me", "excerpt": "GitHub is now one of the least reliable services in my stack, I'm in the process of moving everything off and keeping it for public/open projects for discoverability/discussion purposes only. I've also noticed a meaningful decline in the quality and reliability of several other major software/apps and services - notably Google maps, Spotify, codex. (Is this because of AI?) Codex, especially, is a strange one - initia", "context": "has the 20x pro plan", "use_case": "AI coding", "switched_from": "", "story": "Incident with Git Operations, Pull Requests and Actions \u2013 Resolved"}, {"id": 799885714, "item_id": 49992857, "url": "https://news.ycombinator.com/item?id=49992857", "by": "mooseisloose", "date": "2026-10-07", "stance": "recommend", "quote": "a two part adversarial code review with claude code & codex plus automated reviews from coderabbit", "excerpt": "I'm the maintainer of Google Workspace MCP (https://github.com/taylorwilsdon/google_workspace_mcp), a project that started more as a learning exercise than anything else but has grown into one of the most popular open source MCP servers out there today. It's been time consuming and at times overwhelming, but I can honestly say that it's also been one of the most fascinating periods of my 20+ year history of writing c", "context": "open source maintainer", "use_case": "adversarial code review", "switched_from": "", "story": "Workspace MCP v2 \u2013 Musings from the past 18 months building open source AI tools"}, {"id": 799679232, "item_id": 49979952, "url": "https://news.ycombinator.com/item?id=49979952", "by": "Scapeghost", "date": "2026-10-06", "stance": "recommend", "quote": "I used Codex to convert some of my old Visual Basic 6 apps and games and it was a joy to see them running again", "excerpt": "I grew up in the 1980s. My HackerNews account is 7 years old. My first program was in BASIC on a Sinclair ZX Spectrum with 48K. I used to sit for hours typing \"listings\" from gaming magazines and save them on cassette tapes, like this: https://www.youtube.com/watch?v=AuVBX5Iu18I I love coding manually, and I love AI! This is the first time I've felt the same excitement about computing since the 1980s, when there were", "context": "", "use_case": "converting old VB6 apps", "switched_from": "", "story": "Vibecoding isn't as fun as writing code by hand"}, {"id": 798980080, "item_id": 49936255, "url": "https://news.ycombinator.com/item?id=49936255", "by": "oooyay", "date": "2026-10-02", "stance": "recommend", "quote": "I use Codex far more than I use Claude for coding and my results are generally higher quality.", "excerpt": "I think at this point assessments like this are more qualitative than quantitative. I use Codex far more than I use Claude for coding and my results are generally higher quality. I would not, however, go to someone and tell them that Codex is better than Claude generally. In large part it depends on how you use the model and the harness.", "context": "", "use_case": "coding", "switched_from": "", "story": "Sites in ChatGPT"}]}