personal-ai-feed

I recently heard about ended its $600,000 annual contract with Salesforce after vibecoding its own tools. That story hit me because I had my own version of that moment… with Evernote. I was paying $249.99/year while using only a fraction of it. Instead of moving to another bloated notes app, I started vibecoding the exact workflow I wanted, with Codex helping me iterate. The first version was rough. But vibecoding is iterative, not magical. I kept loop-engineering it: build → use → notice fricti

O aplicativo deixou de ser algo que você precisa abrir e perguntar. Os agendamentos agora são executados de fato. O daemon cron só era executado em chimera serve --cron antes, o que significava que uma tarefa criada por você nunca era executada enquanto o aplicativo estava aberto — ela simplesmente ficava parada. Agora, ele é executado no mesmo processo, realiza uma verificação real do agente e anexa a resposta a scheduler/cron_results.jsonl , para que nada seja perdido silenciosamente. O cron é

Hi HN, We’re Adeel and Umair, co-founders of Unlayer ( https://unlayer.com/ ). We let you add content creation to your applications without having to build an entire editor, renderer, template, and export stack yourself. Unlayer lets you create emails, web pages, and documents inside your app, in three different ways: in code, visually, or with AI. Here’s a demo: https://www.youtube.com/watch?v=0HsDtNkdMpM . We started with an embeddable email editor because a lot of products eventually need one

https://preview.redd.it/v62fpfrfyseh1.png?width=3456&format=png&auto=webp&s=d060f665e3e3add47f7472025e23436c8b6845fe I made a MacOS application (See the black notch expanded at the top) which allows me to switch between different worktrees VERY quickly (Option + Tab OR Option + Shift + Tab). I basically run 1 worktree/feature dev per desktop, so I switch my desktops like tabs. Maybe something like this already exists I don't know, but I've been using this for months now and it's game changing. I

Yo my fellow Codexians (idk if anyone ever said that lol) A while back I thought "How much would I pay if I used the API instead of just my subscription ?", which can also turn into "What's the actual value of my sub ?". So I built this simple script to generate charts as images, a CSV, a JSON but especially an interactive HTML report where you can see everything at a glance. Link : https://github.com/EDM115/codex-usage-tool Quick feature tour : Grabs data from your local rollouts, databases and

So I just wanted to share that I basically used a day of usage in like an hour. I'm satisfied with the output. The whole session felt smooth, like I was adding features the whole time not bug stomping. Also I like that I get my whole week at once, with no session limit. Very in favor of that. My gut reaction is that this is basically the same as fable tier usage rate. This is not a complaint, I'm still orienting. aiming for 14% a day usage is fine with me, especially if the output is this good.

I linked Codex with my account but I can only choose terra, am I doing something wrong ?

I use GPT 5.6 Sol via my ChatGPT Plus, via Opencode. How long is my data retained by OpenAI? I have turned off “Improve the model for everyone” in ChatGPT’s data controls on the website

Hey everyone, I'm looking for a CLI wrapper or repo to build a massive agent swarm with Codex (running GPT 5.6). My goal is to replicate Claude Code's "Ultrawork" (dynamic workflow spinning up 100+ agents concurrently to blast through tasks). Here's what I've tried so far with no luck: Native "Ultra Mode": Supposed to do exactly this, but it barely launches any sub-agents and feels just as slow. I seem hard-capped at 5 agents. Ruflo : Doesn't seem to activate properly with Codex. Oh-my-openagent

I'm the author of a paper my friends and I wrote after we were curious if a MUD, text games originating in the 1970s, could be used to evaluate LLMs. We've spent the last several months on nights and weekends running this experiment and writing the paper on just our personal computers with about $99 in API credits. Our experiment did have an interesting leaderboard but even more surprising was the measurements of each LLM. We scored each on four behavioral dimensions, two of which lean heavily o

I've been ignoring the whole 5.6 ordeal, and enjoying the free resets (and abusing them frankly), but now I'm wondering if this is a good time to try the new models, but I don't know which are the best bang for buck. I'm specifically wondering for two levels of usage, one that would replace the 5.5 medium, and one that would replace 5.5 extra high, my project is large, I started working on before embracing agents as part of development, has design plans, indexed documentation etc etc, so it's no

So I logged in today to look at the limits while I work with Codex, and I just noticed that there is no 5 hour limits any more! And the new "Weekly limit" goes down like the 5 hour limits! I need to clear one thing, my prompts are short with a codebase that is just about 4 html pages and two .js files with about 100-200 lines in them, I am working with 5.6 Sol at Medium, and the limit are going down FAST! It feels like they just removed the weekly limit and renamed the 5 hour limit to the weekly

Over the past few months, our team has been building more and more slidedecks using web frontend technologies with coding harnesses like Claude Code, but a common complaint is to make even small edits we need to edit the code either manually or via the harness. To avoid this loop, I ended up creating Bento, a single HTML file with everything you need in a slide tool including animations and shared editing. There's no install or cloud login, everything works offline. The default deck is around 56

First off I love 5.6 so this is not about the models but instead about the tools. I've noticed that the interaction between Terminal CLi and Gpt work (which replaced codex app?) is very janky since they emerged Codex and Chatgpt apps. On Codex work I can only select Terra Light or Sol Models. In terminal I get every Luna, Terra, Sol combo. But when I run in Terminal the permission despite what I put in terminal are always defaulting to the App permissions (which isn't always respecting my chatgp

I’ve seen a lot of people say they used a banked reset , only for Tibo to trigger a wider usage reset shortly afterwards. They still got some use from the banked reset, but it felt wasted because everyone’s usage was reset soon after. Now that Codex and ChatGPT Work have reached 10 million users , I had GPT look through Tibo’s recent X posts and replies to answer two questions: Are resets likely to continue after 10M? When could the next one happen? What Tibo has actually said When Codex reached