... guys, i switched to codex recently because fable is not anymore included in the plans, and honestly gpt 5.6 Sol surprised me by his efficiency. Which one do you prefer, ClaudeCode or Codex ? Comment from Siarhei Hamanovich(@siarheihamanovich): Last 2 months I use Codex more. I switch depending on the job. For a change that touches a real codebase, I care more about whether the agent can follow the existing patterns, run the tests, and leave a small reviewable ... ... conversational loop. So I don t think one tool wins every task. The deciding factor for me is how much verification the task needs after the first answer. Comment from Nika(@busmark_w_nika): I was more used to

... That was basically the problem I wanted to solve. I really like the Codex app especially how clean and focused the interface feels but for harder coding and architecture tasks I still often want to use Claude

... used to end the day by asking ClaudeCode something like: Can you write my daily update based on what I worked on today? The problem was that

... Hello guys, for the context, in Murmell, we run CLI sessions like Claudecode and Codex, and damn this mf is consuming too much RAM around 800mb. For comparaison Codex consumes about 150mb max since openAI rebuilt it in rust. This is forcing us to keep plans a lil high due to cloud costs, we are thinking into rebuilding our own version of it in rust. Does anyone know about other

... everyone! I ve been building StoreHand , an open-source project that turns ClaudeCode into a Shopify store operator. Instead of digging through multiple Shopify admin pages every morning, you can simply ask: How did the store do overnight? StoreHand will summarize new orders, failed payments, low-stock products, cancellations, and more in plain English. Some highlights: Read-only by default No telemetry, no cloud, no account required Built on

... what I could find from benchmarks and my own week of switching between tools, and a few things genuinely surprised me. The efficiency numbers aren't close, and not in the direction I expected. One widely cited test had ClaudeCode completing an identical multi-file feature task using around 33,000 tokens with zero errors , while Cursor Agent burned roughly 188,000 tokens on the same task and hit several errors along ... ... same deliverable. Separately, other community head-to-heads put

... everyone! Launching a new feature of LangWatch today on ProductHunt, developer focussed but not just dev's anymore, so also curious what each of you think. You know what ClaudeCode costs per seat. But do you know which sessions burned the budget, which model did the work, or whether your cache paid off? LangWatch now tracks every

... Yeah, you can't beat Claude! Comment from Chris M. Nzouat(@nzouat): @lucasjpols  Yes, in my opinion, Codex is exceptionally adept at autonomously executing tasks with minimal human intervention . Additionally, Codex comprehends the assignment more effectively than Claude

... been using Codex, ClaudeCode, and OpenClaw for different kinds of work: Codex (now @ChatGPT by OpenAI )is strong at long-context planning and execution. @

... ClaudeCode is amazing. Agentic workflows just... work - at least better than last years options for vibe

... seeing a chat interface (which for some reason all companies are leading towards to), I prefer working on my workspace while utilizing AI on the side. 3. I want more clarity and freedom on the markdown files that ClaudeCode generates. I want to change colors, add notes, highlight certain sections and I want to see all of these in a proper view. Not the weird markdown format. 4. I want to store a version history and the changes ... ... right. Which was pretty good, but then you get the other pain point; ugly markdown files that you can t read and understand. That s when I decided to create Sofos. Why not create a tool that runs

... past few months I've been juggling three AI coding agents Claude

... months ago I started using OpenWolf, a project that gives ClaudeCode a persistent "second brain" by indexing the project, tracking context, and remembering previous decisions. I really liked the idea, but after using it on larger and longer-lived projects I ran into a few practical problems: the .wolf directory kept growing over time token ledgers and logs became unnecessarily large there was no automatic maintenance it only worked well inside

... Solo builder here, using Cursor + ClaudeCode as basically my whole engineering team. Six months in I still don't have a good instinct for this: some days I let it just run with a whole feature and barely glance at the diff, other days I'm rereading every line like it's a junior dev's first PR. Right now my rough rule is billing/auth

... just shipped plugins, MCP, and skills for ToughTongue AI. What this means: Your coding agent (Claude

... days ago Tamamon launched here and you made it #3 Product of the Day. Thank you genuinely. It's a tiny pixel pet that lives on your desktop and grows as you code with Claude

... paid ads. But I have been building automations with ClaudeCode. You describe what you want in plain English. It writes the

... refactor a 14-file React component tree to migrate from class components to hooks. Cursor Composer: - Faster start, immediately surfaced the right files - Stopped twice to ask for clarification, both times useful - Final result needed about 20% manual cleanup ClaudeCode: - Slower to start, took longer to map the codebase - Ran more autonomously, fewer interruptions - Final result needed about 10% manual cleanup but introduced one regression Net: Cursor felt like pairing with a fast junior.

... wanted to show people what a ClaudeCode session actually looks like a real debugging or build session, not a screenshot or a curated demo. The only options were screenshots or giant walls of pasted text, so I built InstaShare. The most honest demo I can give is this: here's the actual session I used to research and build the tool, shared with the tool itself: https://instashare.to/c/zGxvUs-r... That's the entire pitch in one link ... ... small skill to

... Pasting a screenshot into a terminal agent is messier than it should be. On macOS Codex wants Ctrl+V, not Cmd+V. ClaudeCode on Windows wants Alt+V. Some terminals just eat the image. So people end up saving the shot to disk and referencing the path by hand, which kills the flow. What's your real workflow, drag and drop, the --image flag, save-and-reference, something better? Trying to map how people actually do this. Comment from ... ... chore. It's called SlimSnap. Which part would you want most, the auto-send or the annotate step? Comment from Alexander Bickov(@bickov): @rahulbhavsar  That lines up with how I built it, auto-send is the default. The

... Claude/ Chatgpt - helps my research, coding

... Setting IDEs aside for this one - Cursor, Windsurf, VSCode + Copilot live in a different conversation. I'm specifically curious about CLI-and-agent territory. A year ago the answer for most of us was straightforward: open

... actually needed. Second monitor never solved it for me because a stalled terminal looks the same as a working one, you can't notice a silence. I'm solo too, building a few SaaS products in ClaudeCode all day (

... same people spending 8 hours a day in their editor. The context switch to a browser dashboard is exactly where the habit breaks. So I shipped an MCP server for XreplyAI. One config block in .mcp.json and ClaudeCode gets tools to draft, schedule, and publish posts across X, LinkedIn, Instagram, Threads, YouTube, Pinterest, Bluesky, and TikTok without leaving the session. The workflow that clicked for me: finish a PR, ask

... headline: Split Panes Split your window into a live grid and run four different agents from four different vendors at once, each in its own real terminal, next to your dev server and tests. - Different agents, one screen. ClaudeCode planning on the left, Codex implementing on the right, Gemini reviewing below, your npm run dev streaming logs in the corner. No more alt-tab roulette to find which agent is stuck. - Any layout ... ... where