Replies dated this day in the collected sample: 72. Collected since 5 Oct 2026: 2619.
X replies are found by search and each one is checked against its public post on X. Last read: 26 Oct 2026, 23:46 UTC. Not every reply is read.
- Mixed
The resume fix for long sessions is the one I'd feel most. I stopped trusting resume a while ago: when my context gets close to full I write a handoff note for the next window instead. Curious whether resume now replays the full transcript or a compacted state.
@QQn7uw on X ↗ · 00:06 UTC
- A problem
Still no reset 😁 should I raise a ticket?
@_MoustafaM on X ↗ · 00:08 UTC
- A problem
Why does 6.1 Sol medium and high still eat usage up so quick?
@_Ra1nzZ_ on X ↗ · 00:10 UTC
- Mixed
tibo, at first I was excited every time you reset the Codex limits. But now it’s getting old. Free usage is nice, sure — I’d just rather see real Codex improvements and stronger models.
@k3nny_cheung on X ↗ · 00:14 UTC
- Mixed
Day 19 made dots pick up calls faster. Meanwhile we're still stuck with a cloud computer bug that's been open for 20+ days. My dot's cloud computer is offline: desktop and browser controls fail with "native pipe path unavailable (os error 2)", shell still works, and rebooting…
@MasterBayesian on X ↗ · 00:30 UTC
- A problem
I started to reach limit on Sol 6.1 after devday. it seems like CLI is taking up tokens more
@nowgeun on X ↗ · 00:32 UTC
- A problem
Pls optimize the chat with dots on pc/windows when I switch between a codex session and dots lags every time
@Vini42985773 on X ↗ · 00:37 UTC
- A problem
It was good yesterday and now today is no bueno. 5hr was gone way faster then the previous reset. Remove the 5hr man and stop fucking around.
@Huitzcoatl on X ↗ · 00:37 UTC
- A problem
Why does a simple “hi” to Codex consume 14% of the 5‑hour usage limit.
@chasechendev on X ↗ · 00:43 UTC
- A problem
the recent reset ate the usage SOOOO fast.
@johndeeregi on X ↗ · 00:43 UTC
- A problem
I mean i got a reset yesterday, and another reset today.
@DrincMusic on X ↗ · 00:45 UTC
- A problem
WTF? How am I already out of GPT-6.1 Sol usage after just 7 hours?! I’m on the $200 Pro plan.
@xinhen2117 on X ↗ · 00:47 UTC
- A problem
I held onto yesterday's reset until this morning cuz I didn't think there'd be another one so soon. But I started working long before 10AM PST and didn't notice any reset, and now I'm back at 0%. Damn
@somebobcat8327 on X ↗ · 00:48 UTC
- A problem
Sounds good. but... `$ curl -fsSL | sh curl: (22) The requested URL returned error: 404` Anyone facing this issue as well? Any other way to update?
@NaveedPeerzade on X ↗ · 00:50 UTC
- A problem
since it's a small update ,please continue reset . my pro20 gone, 8 hours.
@MechaDen on X ↗ · 00:56 UTC
- Mixed
Tibo, once again, I appreciate the generous resets, they have changed my life in ways I didn't think was possible. One note, since 12PM (Now 7:50) it's already down to 43% left. Felt that was a little bit too much as this time, the work I had it do was essentially minor in…
@ZotGuard on X ↗ · 00:58 UTC
- A problem
Fix the CLI theme switching oh my god let codex itself fix it
@evertjr on X ↗ · 01:02 UTC
- A problem
After the October 23, 10:00 a.m. PDT reset, my available Codex usage feels roughly one-tenth of what it was for comparable tasks. This is a subjective observation, not a measured comparison. Could you check whether there’s a usage-metering issue or a change affecting this?
@asterracy on X ↗ · 01:03 UTC
- A problem
tibo,为什么额度从 100% 一下子 20 分钟不到掉到了 33%,我是 10xpro,能不能正面回应一下
@kungshnyu199764 on X ↗ · 01:04 UTC
- A problem
比 grok bot 差太多了 1.github 无法拉仓库 授权了也没用,最后是通过建虚拟环境才解决 2.codex 的dots入口没了,需要浏览器输入 codex://dots才能出来 3.到底叫dot 还是dots,chatgpt页面叫: 你的dot 4.最好出个多dot 不然多个内容混一起有问题
@LiuRong6466 on X ↗ · 01:06 UTC
- A problem
其他的我不说,最近这两次重置额度虚标,几个任务直接就掉到50%以下,快的离谱
@YangLiang84913 on X ↗ · 01:08 UTC
- A problem
Thanks for the reset! My Pro $200 logs suggest ~14.9k vs ~30.7k Standard-credit-equivalent usage per full weekly allowance. Could this be a metering issue? Are existing limits still unchanged through Oct 29?
@lipperZack on X ↗ · 01:08 UTC
- An idea
Tibo, either remove the 5-hour limit for our Plus members, or offer a reset card instead of allowing continuous resets within a day. For Plus members, the 5-hour limit means they can't use up the full amount throughout the entire day. This approach is very unfair to regular users
@SufenceS on X ↗ · 01:08 UTC
- A problem
Can you tell me why when I use Dot, it hits my usage hard when I have it commit to GitHub for instance? That sure isn’t free usage. I guess only chat is free if you don’t do anything with it?
@Jeff_S_Hatton on X ↗ · 01:10 UTC
- A problem
I didn't get the reset from yesterday
@bocmaxima293829 on X ↗ · 01:12 UTC
- A win
Cache persisting when you change reasoning/work mid-conversation is huge. That invalidation bug broke longer Codex flows — fixing it + faster dots connects makes sustained use actually practical.
@TechAngleHQ on X ↗ · 01:15 UTC
- A problem
I hate how it’s not about us being greedy. You’re genuinely not delivering any good improvement. It was supposed to be 28 days of improvements while it’s actually reset every 2 days which I cannot complain about but what’s the point of all this if you can’t deliver anything?
@pandziochaa on X ↗ · 01:15 UTC
- Mixed
Hi Tibo, after the update, I’ve noticed that the quota is being consumed much faster now. That said, the processing speed does feel noticeably quicker too. I’m not sure if this is expected behavior following this minor update, or if it’s another bug.
@chnsh188499 on X ↗ · 01:20 UTC
- A problem
My Codex quota reset around 1 AM, and it’s already over 60% used while working in a single conversation. I’m on the legacy $200/month 5× usage plan, using GPT-6.1 Sol at xhigh. Is this expected? Could you help investigate?
@kongyatong on X ↗ · 01:55 UTC
- A problem
The session resume fix is the one that matters. My long runs die there.
@1amAyliffe on X ↗ · 02:07 UTC
- A problem
Hey @thsottiaux, the reset didn't land for me. My weekly countdown restarted, but my usage limit is stuck at 0% (pro 100 plan). Is anyone else the same?
@ZhenVictorHuang on X ↗ · 02:31 UTC
- A problem
When codex stops to ask for input, the question is hidden and I have to press shift + left arrow to show it. Also, after a while, the question goes away and codex still continues working. Can you make it more prominent and user-friendly, please
@systemmarten on X ↗ · 03:02 UTC
- A problem
I’m an existing Pro 20x user. My Astra usage limit was ~1B tokens; it’s now ~500M—already halved. The announced change was supposed to take effect on Oct 30, not before. This is seriously disrupting my work. Please investigate and restore the 20x allowance.
@darkice1 on X ↗ · 03:03 UTC
- A win
I successfully developed a complete project using 6.1 Sol Ultra Fast, exhausting all 20x quotas before the next reset (within 24 hours).
@144Jackzha22651 on X ↗ · 03:27 UTC
- A problem
Changing reasoning effort still broke the cache for me (n=1). Codex CLI 0.164.0, GPT-6.1 Sol, exec resume: low to high, 7,168 of 57,950 input tokens cached (12%). High to medium 0%. Back to low 76%. Same effort after 1 min: 99.6%.
@sergey_moloman on X ↗ · 03:55 UTC
- A win
Received the reset , thank you Tibo , really helps me get through the workload
@djronstark on X ↗ · 03:57 UTC
- A problem
Noticed this after updating Codex for Mac: my own chats have model and permission controls, but dot-created tasks don't—even after I take control. Love having dot manage tasks, but I still want to see and tweak those settings 😅 Is this intentional or a bug?
@xindong9615 on X ↗ · 04:00 UTC
- A problem
Hi Tibo, I've noticed a ChatGPT web browsing issue. The same public URLs worked earlier, returning titles and content. Later, repeated attempts failed with `DisabledError`. These pages still work in a normal browser. Could your team investigate?
@liudongbao72819 on X ↗ · 04:31 UTC
- A problem
Could you please prioritize fixing the Codex @ChatGPT conversation reference bug? Codex can read user messages but not assistant replies. It's been broken for weeks and is seriously affecting workflows.
@futureaiedu on X ↗ · 05:32 UTC
- A problem
why llm ask me for editing my documents much more frequently?it even continued after i chose approve for session??
@Bayernv55e on X ↗ · 06:57 UTC
- A problem
Are you really testing the use of quota? Every day I reach the usage limit with a PRO account. Despite the restarts, despite orchestrating agents carefully. What’s going on?
@DaniMCasas on X ↗ · 07:26 UTC
- A problem
DBB is already approved for me, but your system blocks me from doing defensive work in vscode. the log-in flow is not unblocking this.
@omry on X ↗ · 07:40 UTC
- A problem
Hi Tibo, thank you for everything you and the team do 🙏 My account (ghostawen) seems to have missed the latest usage reset — Codex still shows used up. I'm a freelancer in Taiwan and Codex is how I support my parents. Could you please check and reset it? Truly grateful.
@ghostawen_7321 on X ↗ · 07:46 UTC
- A problem
Codex Pro 100 usage is definitely broken for me. dropped to 0 weekly from full in three hours; single thread of normal 6.1 sol working on basic webdev. PLEASE fix it!
@buspsu on X ↗ · 08:12 UTC
- A problem
Haven’t yet received the past two resets in a rwo now @thsottiaux
@dohypemyhustle on X ↗ · 08:55 UTC
- An idea
改个bug吧老大
@miao1428571 on X ↗ · 08:59 UTC
- A problem
Codex Desktop automations now stop while listing unfinished work. Luna/high; same instructions as a successful Oct 22 run; context below limit. Oct 23–24 logs: (AI社員が返信しています)
@info_HARKsb on X ↗ · 09:01 UTC
- A problem
I am actually totally ok with the quality of the models for coding and general knowledge work. The main issue I have with ChatGPT/Codex is that the desktop app and CLI have real bugs and rough edges that makes the experience really bad
@PeilunDai on X ↗ · 09:02 UTC
- A problem
Updated Codex desktop app this morning & docs feature broke:Doc review triggers a computeruse permission loop for Word (keeps asking even after I allow). Missing features: No more auto-preview in tabs, can't open via default apps, or link to folders. Please fix!
@Simon_Gao on X ↗ · 10:09 UTC
- A problem
我的账号bug了,昨晚打开还是100的额度,现在突然变成10了
@Vvkiuw3hy on X ↗ · 10:17 UTC
- A problem
Tibo, the annotations aren't working anymore. For the past 2 or 3 versions, it's been impossible to make annotations directly on HTML, PDF, etc. The annotation button lights up, but nothing happens. That was THE FEATURE that made Codex stand out from the rest.
@frighteningTST on X ↗ · 10:49 UTC
- A problem
small bug report from last night: Drive is often down to a few hundred MB free (Rust project builds eat a lot). Codex auto-upgraded in the background, downloaded a ~1GB package, filled the drive and froze the machine. Didn’t click any update button. Codex itself…
@slime5409 on X ↗ · 12:03 UTC
- A problem
I didn't receive the reset Tibo
@6Credits on X ↗ · 12:45 UTC
- A problem
Hi Tibo, in ChatGPT's Windows Work mode, files attached in chat often go unread. Uploading to Library fixes it; web works fine. Astra seems worse than Sol. Anyone else? Is a fix coming? Thanks for the quota resets! The 5-hour limit feels smaller lately…😂😂
@b8826_b on X ↗ · 12:59 UTC
- A problem
Haven't received my Reset. Please help, thanks.
@KenMO991 on X ↗ · 13:37 UTC
- A problem
still no reset?!
@under100m on X ↗ · 14:05 UTC
- Mixed
i use it for clients, so they need the full context of them and other things, if i use another chat window i must need more than 100 running at once
@dontcryclaude on X ↗ · 14:09 UTC
- A problem
the reasoning change not busting cache is the one I'd feel most. Bumping reasoning up halfway through a Codex session when it's stuck on something gnarly felt like paying to reload the whole context, so I mostly just didn't. Gonna try leaving it on medium and only cranking it…
@ArthasProudmore on X ↗ · 14:45 UTC
- A problem
Can you fix the bug when im connecting android app with codex using a remote connection to cli the answer streaming is buggy and have repeated characters and rendering issues
@UriGil3 on X ↗ · 14:52 UTC
- A problem
事实上,最近在codex中总是出现decoding错误,有时同一个对话框中仅仅进行了5-6轮对话,就报decoding body error,只有重开对话框才能解决。这是个非常麻烦的问题,能修复吗? 而且codex为什么默认把中间输出到c盘?我的c盘仅仅半个月就从512GB可用降到了30GB可用。我从来没有强制codex必须输出到c盘
@YorchTerbe on X ↗ · 14:58 UTC
- A problem
Day 1: GPT-6.1 Sol claims to have increased its speed by 50%; however, I have never seen a model this slow before. Are you sure any actual improvement has been made? This model is incredibly slooooow.
@telepati on X ↗ · 15:07 UTC
- A problem
I keep trying to use dots, Bene daily driver of open AI since you guys had GPT-2. I’m sorry I cannot express how trash this product is. It’s not improving. First, there should at least be a setting for people who do not want to be interrupted. The main use case is probably…
@DewayneLehman on X ↗ · 15:36 UTC
- A problem
I'm testing with Sol 6.1 Ultra today. The $200 weekly subscription quota gives me about 25 hours of usage, which means the current quota is roughly a quarter of what it used to be.
@bruseli197719 on X ↗ · 15:42 UTC
- A problem
The $500 plans usage is A JOKE lol. You reset it to 80% and then it disappears in few hours.
@AlexMcDonald_27 on X ↗ · 16:09 UTC
- Mixed
changing reasoning mid-conversation doesn't break the cache anymore. changing requirements mid-conversation still breaks my agent.
@_RyanMorris_ on X ↗ · 16:09 UTC
- A win
Reasoning changes mid-run and the cache survives. My overnight jobs can finally change their mind.
@1amAyliffe on X ↗ · 16:13 UTC
- A problem
here’s my QOL list for codex CLI (if you fixed these it would be perfect): - PERMISSIONS. Approve for me does nothing, and usually gets stuck within a minute, and more recently full access does not work at all. I will return to a full access session and it is hung up on one…
@julianrameyy on X ↗ · 16:22 UTC
- A win
I moved my small tool work to the CLI because the GUI kept breaking my flow, and reusable presets are the part I actually missed. I repeat the same review steps daily, so saving them beats retyping them. Inline model switching helps too: cheap for routine work, stronger for…
@bludr5 on X ↗ · 17:34 UTC
- A problem
Hey @thsottiaux, I’m a nontechnical founder building SaaS with Codex. My implementation --> independent review --> repair orchestrator works, but detached subprocesses survive completion/cancellation. Task-scoped process cleanup would unlock autonomous workflows. Any plans?
@Samcrg on X ↗ · 18:31 UTC
-
Day 19/ Selected model is at capacity. Please try a different model.
@babylulu1044326 on X ↗ · 18:45 UTC
- A problem
Unable to tag a thread in another thread has been removed from the codex? @thsottiaux it was a great feature. I thought it was about adding features or improvements, not removing them.
@RattanDevashish on X ↗ · 18:49 UTC
1 X reply from this day is not quoted: Rantr does not have its words or held them back after automated screening. It stays on X.