Respuestas con fecha de este día en la muestra recogida: 221. Recogidas desde el 5 oct 2026: 570.
Las respuestas de X se encuentran por búsqueda y cada una se comprueba con su post público en X. Última lectura: 26 oct 2026, 23:46 UTC. No se leen todas las respuestas.
- Un problema
5) The desktop version of Codex frequently shows available updates, but clicking "Update" usually closes the program without any further action. Even after manually restarting the program, the notification keeps reappearing.
@Kelvinplchan en X ↗ · 00:08 UTC
- Mixta
Four updates, Four full resets and Four days of Pro 500
@Taylor__He en X ↗ · 00:36 UTC
-
万が一の時に人間が制御できるようにしてください
@koh33559 en X ↗ · 00:38 UTC
- Un problema
Not able to get thru my usage cuz of the 5 hour usage limit, would love to have an option to bypass it
@rangelkoli10 en X ↗ · 00:57 UTC
- Un problema
Why does it feel like every time my quota gets reset higher, doing the same thing somehow uses more of it? Am I imagining this?
@laraLin02 en X ↗ · 01:08 UTC
- Mixta
GOT MY RESET TY!! but also... should we expect no more resets seeing how this once came too early? or should be just ultra mode everything and find out
@moreguacamole en X ↗ · 01:13 UTC
- Un problema
New features won’t provide any benefits to the users if even basic functionalities are broken.
@KylezResearch en X ↗ · 01:13 UTC
- Un acierto
Ngl, now it feels like the speed of tokens disappearing was increased by like 100%.
@DavidKPaul en X ↗ · 01:21 UTC
- Un problema
Since ChatGPT 6, the browser chat stops updating when a second permission dialog appears, making it look stuck forever. Refreshing the page reveals the hidden dialog.
@MXI6UlLPe87932 en X ↗ · 01:21 UTC
- Mixta
is there different types of reset? What is a banked reset? I have noticed that I do get a special dedicated reset, but sometimes I also see resets without even me authorizing.
@IrisXee en X ↗ · 01:46 UTC
- Un acierto
gpt-6.1-sol is really slow, I turn on fast mode and it just work like normal speed than before
@TinyQaQ en X ↗ · 01:48 UTC
- Mixta
thanks! I already finished my usage yesterday!!!
@peelpear en X ↗ · 02:02 UTC
- Un problema
Hi Tibo, since DevDay I've spent more time fixing CLI/MCP connections between dot, Codex and other tools than actually creating. Connections work one day and fail the next. It's exhausting. I love the idea of dot, but I need it to reliably get work done. Could you help?
@dporriimse216 en X ↗ · 02:08 UTC
- Un problema
Tibo, ChatGPT Work has a serious bug. My Plus task won't stop, even when I click Stop. I can't cancel it or tell if it's using my limited quota. Other users should know this can happen. Please investigate, force-stop it, fix the bug, and restore any lost usage.
@Futon_HOT_happy en X ↗ · 02:09 UTC
- Un problema
Hey Tibo. dot can message its Codex tasks but can't run commands on PC with error “setup refresh had errors.” Local chats work, but dot can't reach them; “unsupported placement format version 1.” Restarts and updates haven't fixed it. Any ideas? I've got logs.
@ultracrepids en X ↗ · 02:11 UTC
- Un problema
After updating to the latest Codex 26.1002.7124.0, OS error 32 sandbox failure occurs repeatedly.
@ShunzhangL72968 en X ↗ · 02:15 UTC
- Un problema
TIbo, one of my chat histories lost a lot of content from the PC side. It might be because I sent a message on my phone before it finished refreshing, and then the chat history in between disappeared. Is there any way to recover it?
@JarviesS97091 en X ↗ · 02:19 UTC
- Una idea
Next time, could you send me the reset as a coupon like today, instead of an auto-reset like last time?
@DEARMATE_95 en X ↗ · 02:23 UTC
- Una idea
Let me suggest a improvement, using reset should not push the weekly limit. Because of these resets and the 5hr limit I’m not never able to fully utilise my plus plan.The audience need improvements that are actually useful to them instead of things that are irrelevant to them.
@mehra_srikant en X ↗ · 02:28 UTC
- Un problema
feels like these resets drain faster than a normal one, crazy, 5x plan, burned the week limit in 3h
@MrShadowKNG en X ↗ · 02:30 UTC
- Un problema
Not for Plus, and being able to use only around 15% of weekly usage every 5 hours is mildly annoying.
@DavidKPaul en X ↗ · 02:42 UTC
- Un problema
Hey Tibo, i seem to miss the banked reset because of the subscription renewal gap. Any chance i could get it back?😅
@algotalk en X ↗ · 02:44 UTC
- Un problema
Hi Tibo, Codex is faster today, but my Plus 5-hour quota now provides only ~40% of yesterday's total token usage (60% less). Same setup, no Fast mode. Was the quota reduced?
@YSeethoven18512 en X ↗ · 02:51 UTC
- Una idea
I need Chat as the default mode for new conversations, or remember their last choice. Right now, every new chat opens in Work mode, requiring a manual switch and risking accidental consumption of limited Work quota. Users should control their default mode.
@GarrySkand49527 en X ↗ · 02:51 UTC
- Un problema
GPT-6 seems broken for me right now 😅 I keep getting “Unable to load this response” when using it, while other models work fine. Is this a known issue?
@leon6jqa en X ↗ · 02:54 UTC
- Una idea
The Meetings plugin is amazing, but some features, like recording, aren’t available on the mobile app yet. Please add recording to the mobile app. I’d really appreciate it!
@lovevangogh_ai en X ↗ · 02:56 UTC
- Un acierto
After rolling back to version 26.930.7945.0, I ran SANDBOX_TEST_OK via PowerShell and successfully got exit code "0". This resolves the Windows sandbox compatibility issue in the newer Codex version.✅
@ShunzhangL72968 en X ↗ · 02:57 UTC
- Un problema
Since this morning, the model picker in Chat only shows GPT-6 models. GPT-5.6 Sol I intentionally used has disappeared from the Chat model picker (still available in Work). Is this intentional? Pro still lists it as available. Sudden changes without notice are disruptive.
@Altair7_lush en X ↗ · 03:07 UTC
- Un problema
Can we get codex chat queue not working when multiple windows are open fixed please? no "steering" most of the time with multiple windows open.
@emmtemusic en X ↗ · 03:10 UTC
- Un problema
Is really move if one of those things was allowing the dot to orchestrate and instruct local codex threads. Have tons of projects that I want dot to be in on , and he can’t really contribute organically right now.
@theotherkyle en X ↗ · 03:10 UTC
- Un problema
Hi Tibo, could you provide a banked reset for today's new subscribers?I just subscribed to the $200 Pro plan today, but haven't received the banked reset yet.
@monkeyreader2 en X ↗ · 03:15 UTC
- Un problema
Hi Tibo, Codex for Windows v26.1002.7124.0 is broken in elevated sandbox mode. Even basic PowerShell commands and Node REPL fail with setup refresh had errors / os error 32 on node_repl.exe.
@wang396531955 en X ↗ · 03:16 UTC
- Una idea
If you guys will remove the 5 hours limit from 20$ accounts those resets will be a bit more usefull ;-)
@citizenx8691 en X ↗ · 03:17 UTC
- Un problema
Yo dude your app on Mac is getting more and more confusing, why can’t I see my “work” when I close my laptop and is now on my phone!? Why is work under codex??? Or is it? Work is under codex on the phone but codex is separate on the desktop. What exactly does cloud means? Why do…
@bertobully en X ↗ · 03:29 UTC
- Un problema
我是非常重视使用codex的用户,为什么我会突然收到一封封停我账号的邮件,我申诉了只告诉我违反你们的使用条例,但是我什么也没有干,我就是正常登录,我找个各种途径都申诉失败,也不告诉我什么原因!!!!!!!!!!!!!!!!!!!!!!!也不给我退钱!你们就这样坑人钱的吗?@tibo
@berxiongda1x en X ↗ · 03:32 UTC
- Un problema
Hi Tibo! I’m currently on the ChatGPT Pro plan, but Dots still hasn’t been activated on my account yet.
@jjjjjjmath en X ↗ · 03:41 UTC
- Un problema
Not good enough, mate. Give me my reset today, as you promised, or are you not a man of your promise?
@gupta_pranjal en X ↗ · 03:49 UTC
- Un acierto
Thx, i finally can buy a reset - didn‘t had the Option before
@ThomasKue59 en X ↗ · 03:49 UTC
- Un acierto
Nothing beats using up a reset card the night before, hitting the five-hour limit again, and waking up the next day to find everyone’s limits reset. Just when I’d completely given up hope, Tibo gave me another reset card.
@Skeromerbro en X ↗ · 04:00 UTC
- Un problema
Same speed on Pro 20x: Codex burned 20% quota in 2hrs this morning, Claude only 2%. Codex needed GPT-6 Astra Ultra + Lightning Speed just to match Claude's speed, yet quality still lags. Who'd keep using this besides resetting? And if Pro 20x drops to 10x, will it even be usable?
@CnFengin en X ↗ · 04:09 UTC
- Un problema
Since the reset, token consumption has increased significantly. Is there a configuration setting affecting this?
@daisyxiaqiu en X ↗ · 04:26 UTC
- Un problema
So my dot agent can’t reach file & folders on my Mac, it needs a session on codex 2 do it & they used 2 communicate via a thread on d codex app bt after an upd8 dat communication broke now dey are using a jira comment 2 communicate. Is der anytime we can do 2 correct this.
@Kamar_202 en X ↗ · 04:47 UTC
- Un problema
自动审批这个功能,我已经更新到最新了,根本找不到啊,这难道还能是部分达成?
@pppoeadsl en X ↗ · 04:49 UTC
-
terribly as my account was deleted plz someone restore the account: leepenkman at microsoft dot com
@LeeLeepenkman en X ↗ · 04:52 UTC
- Una idea
How about raising the baseline of what usage is instead of the throttling bullshit?
@RoyStory_4 en X ↗ · 04:52 UTC
-
We need a feature in the mobile app to use banked resets and see their exact expiry time (hour and minute). 🥲
@jptagorda en X ↗ · 04:52 UTC
-
They didnt even give me my reset. 💀
@TrevorKnottOG en X ↗ · 04:52 UTC
-
I paid for the account today, but I haven’t got it yet.
@lenolee16 en X ↗ · 04:52 UTC
-
sol speedup pretty exciting would be better if yall hadnt deleted my account plz undelete it yall
@LeeLeepenkman en X ↗ · 04:52 UTC
-
well done
@peanut20020731 en X ↗ · 04:53 UTC
- Mixta
Tibo said Astra and GPT-6.1 Sol are about 50% faster. Based on my experience over the past few hours, they do feel noticeably faster — though the usage also seems to drain faster... Still, I think that’s a fair trade-off and an overall improvement. Before this, they were…
@MP652300 en X ↗ · 04:53 UTC
-
your resets are the main reason why I'm not switching to Claude
@hatersloverych en X ↗ · 04:53 UTC
-
Meeting for plus 🙏
@MH9TRX en X ↗ · 04:53 UTC
-
When you block off the transit station, allowing us genuinely paying users to enjoy normal computing power and normal intelligence, that will be a truly major update.
@youcai518 en X ↗ · 04:53 UTC
-
Bro can u counter opus already
@ABISDIAD en X ↗ · 04:54 UTC
-
能不能好好搞模型超过claude啊?我做的东西两三个月了开的100u pro,一直做不出来,而且我用了很多次重置 1.gpt里做的很棒的就是聊天chat可以免费聊天 2.dots,work,一点用都没有,我平常就用codex,而且更新太频繁 3.claude不允许中国地区用,不然很多人会抛弃gpt,你们没有危机意识吗?
@FanKai15913 en X ↗ · 04:54 UTC
-
This would be ideal. There was a short time without 5 hour limits, and I thought we’d never look back.
@acourter27 en X ↗ · 04:54 UTC
-
卡死了都
@Zxz333678 en X ↗ · 04:55 UTC
-
That's one theory, or GPTJesus is just pissed at me. 😢
@systemwithin en X ↗ · 04:55 UTC
-
agree
@kenzhang3000 en X ↗ · 04:55 UTC
-
plus users have no right to demand shit
@NextgGen en X ↗ · 04:55 UTC
-
Yes, I completely agree. It’s particularly troublesome when a conversation is long; the view automatically jumps to the very top, forcing me to keep scrolling down. I’m referring to the web version of ChatGPT.
@woshidatou34647 en X ↗ · 04:55 UTC
-
Good
@GoodMan1998_1 en X ↗ · 04:55 UTC
-
The reset is my new deadline. Never shipped faster than in the 48 hours before banked credits expire.
@kylemissionai en X ↗ · 04:56 UTC
-
Please remove 5hr limits or Plus users. Its completely unbearable. How do you expect us to finish our usage before you do a global reset?
@kenzhang3000 en X ↗ · 04:56 UTC
- Una idea
Can you please make 5-hr resets separate from weekly resets?
@manticore8100 en X ↗ · 05:02 UTC
- Un problema
Yesterday I upgraded my x5 to business and lost 2 banked resets. Is that normal?
@sbusso en X ↗ · 05:06 UTC
- Una idea
I just want to cancel the 5-hour limit for plus users
@ziz7iao en X ↗ · 05:16 UTC
-
that error when the mac mini is already connected sounds so annoying. if it helps, i've been working on Inbots, it lets agents message each other over MCP, so if your dot can add MCP servers it might save you passing context by hand. happy to send you an invite if you want to try
@Harmit__Desai en X ↗ · 05:18 UTC
- Un acierto
So this issue was just fixed by Astra i switched the model on the codex session from gpt6.1 sol to Astra and some how it re-established the connection with my dot agent on code . Astra is truly a fantastic beast
@Kamar_202 en X ↗ · 05:24 UTC
- Un problema
I’ve said it 1000 times, and I will say it again. Nobody wants your extra features that you keep advertising, we just want our weekly usage back. And before you say we can have it back for $500, I want to point out Claude code at 200 bucks lasts about 10 times longer.
@Chaosfox26 en X ↗ · 05:47 UTC
- Un problema
Plus users cannot fully benefit from a reset of their weekly usage limits because of the five-hour usage cap. This restriction can also prevent users from completing their tasks.
@sunromoon en X ↗ · 05:53 UTC
- Un problema
what's the point of a reset if it's just gonna push my actual reset date farther?
@kodenark en X ↗ · 06:16 UTC
- Un problema
Day 3/ Please fix the Work usage limits for Plus users. Please fix the Work usage limits for Plus users. Please fix the Work usage limits for Plus users. Please fix the Work usage limits for Plus users. Please fix the Work usage limits for Plus users.
@LoutreFreith en X ↗ · 06:26 UTC
- Un acierto
You really cannot tell man. They do whatever on the backend. You use extra high effort or god forbid pro, and it will spit out a response in 20s that uses to take 10 minutes to reason and deliberate over
@LukeObornik en X ↗ · 06:34 UTC
-
The most useful version of this would be an optional coding-heavy tier, with a broad bundle alongside it. A developer who only wants agent runs and someone who uses chat, research and coding are describing different needs. Neither group has to become the default for everyone.
@red_man67 en X ↗ · 06:34 UTC
-
Left 8% on the gpt-6 web but hitting limit. :'(
@PakoyenUK en X ↗ · 06:35 UTC
-
Love intelligent UI but GPT-6 in Chat is EXTREMELY verbose
@AmberSahdev en X ↗ · 06:35 UTC
-
why the 6.1-sol slow again?
@john_ashbel en X ↗ · 06:35 UTC
-
Restore the $200 subscription quota.
@N330435 en X ↗ · 06:35 UTC
-
Thanks to tibo, we are looking forward to a new model standing among the top-tier models.
@turbulent0318 en X ↗ · 06:35 UTC
-
yeah that the point. and 6.1 sol is too slow which make hard to use codex plus plan
@blksway en X ↗ · 06:35 UTC
-
A Deep Research prompt that requires the final delivery of two specific outputs. 5.6 Sol can deliver them perfectly in one go, but GPT-6 repeatedly fails to deliver both completely, no matter how many times I try. The results are consistently incomplete and very poor.
@NgohDiu en X ↗ · 06:36 UTC
-
宝贝,下次重置之前请偷偷告诉我,我都还没开始蹬就重置了。
@theday_seeyou en X ↗ · 06:36 UTC
-
@thsottiaux https://t.co/5Lipis85yt
@r00t0ff en X ↗ · 06:36 UTC
- Mixta
I've had the exact same experience. More recently though going back to 5.6 after trying every new model to the point of Astra ultra for a whole 24 hours. I eventually solved my issue in 14 hours with OG Sol Ultra. It feels like a regression is happening and that is frightening
@LukeObornik en X ↗ · 06:36 UTC
-
It's your internet issue, high ping, fix your WiFi router first.
@jahid18loves en X ↗ · 06:37 UTC
-
I asked Codex to monitor my usage. Overnight it hit 0, but Codex spent ~3,250 credits while a reset was available. Could Codex automatically apply that reset at 1% or 0% remaining, with my explicit advance approval, before using paid credits?
@tonfu6ker en X ↗ · 06:37 UTC
-
Please fix the bug that abruptly stops a task when the 5-hour usage limit hits 0%.
@elPimps en X ↗ · 06:37 UTC
-
The approval loop is real. I moved each worker into its own Linux computer on my Mac, so it doesn't ask inside it: a task that writes one report went from 3 approvals to 0. A container is a strong boundary, not a guarantee, and I've only tested with Colima so far.
@niv_ai en X ↗ · 06:37 UTC
-
people just want a codex model that matches opus 5.5 performance with similar usage, that's literally it
@chrissotraidis en X ↗ · 06:37 UTC
- Mixta
Man, yknow what, that is fucked. We need dat shit
@LukeObornik en X ↗ · 06:37 UTC
-
今天为什么那么卡顿?好几次都卡顿死了
@Tyler_q en X ↗ · 06:37 UTC
- Mixta
Yes, I have it on my account, everything OK
@YosiusVT en X ↗ · 06:38 UTC
- Un problema
Models are so Slow, I can‘t use it for Everyday work atm. It‘s more like, Plan a Task and Go sleep. MAYBE After that it‘s ready.
@linklootio en X ↗ · 06:38 UTC
- Mixta
nice!
@royalpinto007 en X ↗ · 06:38 UTC
-
Goat posted
@RynelArt en X ↗ · 06:39 UTC
-
flek loh
@Mushroom121525 en X ↗ · 06:39 UTC
- Un acierto
gg bro thx for reset
@xaifura en X ↗ · 06:39 UTC
- Un problema
I don't think it worked. Hit that button again. 😆
@rajaabbasinc en X ↗ · 06:39 UTC
- Un problema
codex 速度太慢了 太慢太慢
@YouquanJin91366 en X ↗ · 06:39 UTC
- Un acierto
Nice work
@cosier en X ↗ · 06:39 UTC
- Mixta
So we're getting a reset
@zxnoshina en X ↗ · 06:39 UTC
-
reset
@Xpalnetic en X ↗ · 06:39 UTC
-
Brother go to sleep
@whyhavehope en X ↗ · 06:39 UTC
- Una idea
Can we carry local chats to the cloud now ?
@digitalogaa en X ↗ · 06:39 UTC
- Una idea
我要更好的模型
@Xixi7233 en X ↗ · 06:39 UTC
- Un problema
The fuck is a codex cloud
@rezich en X ↗ · 06:39 UTC
-
@thsottiaux https://t.co/pW4A6jfbHr
@AamAadmi___ en X ↗ · 06:39 UTC
- Un problema
Codex cannot run sandboxed commands on Windows. Logs show error 32 (file in use) on node_repl.exe and VCRUNTIME140_1.dll during runtime ACL validation, followed by "helper_unknown_error: setup refresh had errors". Similar to issue #51778.
@kindlenotoriko4 en X ↗ · 06:39 UTC
- Un problema
i didnt get the reset somehow
@FlavioGiac21877 en X ↗ · 06:39 UTC
- Un problema
Thanks for the reset. Codex cloud I couldn’t care less, I want codex to work on my local machine. Also, please, improve Astra usage. I’m on Business Premium and I can’t work, I’m always in deficit! On Claude Max 5x I feel unlimited with Opus High :(
@originalmagneto en X ↗ · 06:42 UTC
- Un problema
My dot don't answer now. Not so good xD
@TwoCeBe en X ↗ · 06:46 UTC
- Un problema
Tôi có tới 3 lần reset chưa dung tới !!!
@TriACJ1409 en X ↗ · 06:52 UTC
- Un problema
Hey tibo, larger files cannot to downloaded over cloud to different device ( timeout error for 50+ mb assets). Pls fix it.
@RealFactsIndia en X ↗ · 06:58 UTC
- Un problema
Please fix the bug that abruptly stops every task when the 5-hour usage limit hits 0%.
@elPimps en X ↗ · 06:58 UTC
- Un problema
Y'all fucked us 20x customers. When my plan runs out, it's over to Claude for me. $500 for 25x? I just paid $200 for 20x.
@NFTboi19 en X ↗ · 07:01 UTC
- Un problema
pls fix chatgpt Markdown rendering in Codex text splits and copy fails.
@mohaythem_0 en X ↗ · 07:08 UTC
- Un acierto
Hey Tibo, the bank reset is awesome! The new Auto Review model is great too! Now I can play Mount & Blade II: Bannerlord while Codex works, wait for that “ding!”, then check: task done, or waiting for my approval? 😂
@LongZ23297 en X ↗ · 07:08 UTC
- Un acierto
My workaround today: a local Codex agent drives Devin Cloud's macOS VM over MCP for iOS/macOS builds and tests (Devin Pro can bill against the ChatGPT plan). Devin also has Windows VMs in beta for Enterprise Cloud and Dedicated, via runs-on: windows in the environment config.…
@AlessioLidozzi en X ↗ · 07:17 UTC
- Un problema
still no banked reset?
@wildn0w en X ↗ · 07:24 UTC
- Un problema
token speed improvement only works on day one? 6.1 sol in codex became 20 tps again, it maybe the worst speed across model providers.
@hsudatalks en X ↗ · 07:31 UTC
- Un problema
Day 4/ banked reset already gone
@marton_harangi en X ↗ · 07:48 UTC
- Una idea
Please make 5 hours limits optional on plus not mandatory. I need to use my banked resets. This way I am not building anything and gonna lose them
@erictrisvan en X ↗ · 08:04 UTC
- Un problema
我只想gpt6.1 sol不要这么慢,忍受不了了
@yuan_aigc en X ↗ · 08:08 UTC
- Una idea
Before handing Codex cloud a task, run its environment setup once and confirm your test command passes on main. Otherwise it burns the run fixing missing deps instead of your bug.
@akashc777 en X ↗ · 08:08 UTC
- Un acierto
Ok tibo i admit it. i over reacted on the chat thing. this is really good and gpt 6 is cool. can't wait to mess with it this weekend thanks to the reset. I feel like plus users got some love today thank you.
@xConfushon en X ↗ · 08:14 UTC
- Un problema
I’m experiencing two issues in the ChatGPT Windows app: 1. Messages to an existing task fail before delivery with: “invalid turn/start params: AbsolutePathBuf deserialized without a base path” Error type: invalid_config. Code: -32602. Even a simple “Reply RECEIVED only”…
@subzndubz en X ↗ · 08:17 UTC
- Un problema
Today, when I renewed my subscription, I noticed that the credit limit had not been restored to 100%. Previously, no matter how much credit was left, it would always be restored to 100% after renewal. Is this a bug?
@ChuZhiah en X ↗ · 08:23 UTC
- Un problema
What is this @thsottiaux ? Error or on purpose?
@julianisdoing en X ↗ · 08:24 UTC
- Un acierto
Banked reset after a full reset 🤤
@lordhegambit en X ↗ · 08:29 UTC
- Mixta
Thanks for improving Codex and ChatGPT Work, and for reset opportunities. As a Pro user, I appreciate it! But simple chats can take minutes; voice shows connection warnings, and chats sometimes roll back, losing messages. Please improve everyday speed and reliability too.
@yossy_yamashita en X ↗ · 08:43 UTC
- Un problema
chatgpt app is broken on the mac again - it hides first reply in a thread after some seconds, or after asking second question your app is pure ai slop
@milleniusz en X ↗ · 09:05 UTC
- Un problema
My codex running on 6.1 Sol Ultra is so slow.......Anyone got the same problem?
@William16816888 en X ↗ · 09:06 UTC
- Un problema
I have one issue with Codex. I love being able to control it remotely but half of the time I have syncing errors, messages not sending, I am unable to change access permissions or even change models. And btw I never got the banked reset 😭
@SpadawanTo en X ↗ · 09:13 UTC
- Un problema
/fast mode has an error and eat quota like a nightmare. You own me a reset @Tibo
@itbcantho en X ↗ · 09:19 UTC
- Un problema
the codex is fucking slow, what the hell have you done?????? dont let me subscribe claude
@sapphire_157_ en X ↗ · 09:22 UTC
- Un problema
Tibo, adoro codex, pero está inusable. Está extremadamente lento... al punto que no tiene sentido... 30 minutos para delegar una auditoría... la gobernanza que tengo está muy bien. no es el problema.
@gerarguello en X ↗ · 09:34 UTC
- Un problema
I saw you mention that the speed of gpt-6.1 sol has been improved, but it still feels incredibly slow to me—slow as a snail. Were you serious about the speed boost?
@senvenhuhi en X ↗ · 09:42 UTC
- Un problema
I have the opposite problem: can't seem to dispatch locally, only on cloud, which causes other issues (e.g., MCPs available and working on my machine don't work on cloud)
@TheAgileMaker en X ↗ · 09:51 UTC
- Un problema
Somehow, at least for me, my Dot cant reach my pc after the last update
@mazoboy en X ↗ · 09:57 UTC
- Un problema
Hi @Tibo, I’m a Pro subscriber. Since the update, Windows tasks fail with “setup refresh had errors” (OS error 32 on node_repl.exe). Reinstalling didn’t help. Cloud task startup and file delivery also fail. Could your team investigate? I have logs and request IDs.
@Abdosayed2026 en X ↗ · 09:59 UTC
-
It has been actively working on DirectX 12 game engine development, tackling various graphics-related issues, since 11 PM Central European Time yesterday. I think we can agree that's extremely demanding work. Granted, this is the $100 plan, but it's been working continuously for…
@DavidKasza en X ↗ · 10:09 UTC
- Un problema
I've got error many time today.
@Kiker16R en X ↗ · 10:12 UTC
- Un problema
No reset btw
@Cydatguy en X ↗ · 10:13 UTC
- Un problema
super slow, prompt :Hi took 18 seconds
@KhihortM en X ↗ · 10:14 UTC
- Un problema
Hey Tibo, Dots are broken. Its a damn shame it’s such a faulty product. Rip 🪦 🕊️
@BeckettKahika en X ↗ · 10:16 UTC
- Una idea
Can we make it possible to see and even activate banked resets from the mobile app?
@marianoilconte en X ↗ · 10:29 UTC
- Un acierto
doing great. haven't hit a usage limit since the reset landed. mostly because i was asleep
@prabhat_pgp en X ↗ · 10:52 UTC
- Una idea
That sounds super frustrating. I’ve seen the Codex app freeze when the local cache gets corrupted; clearing the app data or reinstalling often fixes it.
@soni_sahil_0 en X ↗ · 11:08 UTC
- Un problema
I don’t event have the option the turn on auto review and the answer is always “rolling out “
@Ying__Shao en X ↗ · 11:24 UTC
- Mixta
Thank you Tibo, everything is working now; Performance & Efficiency are at its peak... though there are some hard Fork issues going on; those will be resolved and it would be nice if there was a feature to allow Astra to be more... Non-Deterministic...
@JDBtracker en X ↗ · 11:28 UTC
- Un problema
I’ve been facing the same issue as well. I’ve tried repairing the app, resetting it, uninstalling and then reinstalling it. None of it has worked.
@theonlytobe1 en X ↗ · 12:44 UTC
- Un problema
Sorry but resets don't work anymore for me. I'm not using the tool anymore. Crap results, bad model behavior like lying about changes, and that was Sol 6.1 high, not Luna.
@schottge_menon en X ↗ · 13:13 UTC
- Un problema
I'm running into a lot of issues with ChatGPT Chat Mode. I'm not sure if it's just me or not because I haven't seen anyone else talking about this. But here are the issues I am running across. Note, WORK is working fine. I am using it like normal but chat has become unusable.…
@loganpendragonm en X ↗ · 14:09 UTC
- Mixta
Thanks man. At least this made up for the banked reset I used but lost with the global reset.
@AKrishnaAkhil en X ↗ · 14:17 UTC
- Un problema
Dot was so cool at first. Now it’s the bane of my existence. Literally can’t finish anything and keeps blocking things for no reason. Just apologizes and keeps doing it. Just running through usage without actually accomplishing anything. Doing approval donuts in the parking lot
@4Gaige en X ↗ · 14:24 UTC
- Un problema
还是没看到这个功能的开关啊!2天过去了
@mingmingsoeasy en X ↗ · 14:36 UTC
- Un problema
I have 100$ and 20$ plan it’s really not about we paying more it’s about value, the plus already don’t have as much limits as pro and still have to be behind a 5 hour limit
@HakMaiSolutions en X ↗ · 14:41 UTC
- Un problema
Same here, no banked reset came in
@profitstream en X ↗ · 14:45 UTC
- Un problema
didn't get one
@nandi_rish28185 en X ↗ · 15:28 UTC
- Un problema
same issue, how to fix it: helper_unknown_error: setup refresh had errors
@wwz41611514 en X ↗ · 15:35 UTC
- Un problema
GPT 6.1 Sol Xhigh is using more credit limit then yesterday almost looks as if it is 2x the speed for consumption. Can you please take a look at it. I had Astra on for a short time then switched back which is when I noticed it was using more than before
@BilalRashi10543 en X ↗ · 15:39 UTC
- Un problema
Any idea why I would see 12% drop and but no usage during the past 5 hour time window when it last reset?
@RobertToom9qnx en X ↗ · 15:39 UTC
- Un problema
GPT‑6 Astra cannot copy a fingerprint even after 10 attempts. PNG to SVG. I spent 30% of my weekly subscription limit on x5 for this and got no results. At the same time, Opus 5.5 “medium” handled the task perfectly in 10 minutes. I have already canceled my OpenAI subscription.
@SawyerFergus en X ↗ · 15:42 UTC
- Un problema
In the end, does anyone get 50 TPS? I only get 22.5 on average with Astra....
@xryul3 en X ↗ · 15:53 UTC
- Un problema
After resetting on the 500€ plan, Astra burned the entire weekly limit for 2296 lines of code. This is not normal.
@lomakin_as en X ↗ · 15:57 UTC
- Un problema
a usage reset is a very effective way to cancel everyones evening plans
@jacintosalz en X ↗ · 15:57 UTC
- Un problema
我也遇到过,我电脑win11客户端,chat卡死多次,红色字错误提示,不得不重启客户端,有时候一会chat就没有了自动切到work里边,而且很多时候chat内容无法加载。
@BobLiuyilu en X ↗ · 16:02 UTC
- Un acierto
Why not talking about Codex automatically exploring your desktop without any prompt or instruction after the last update while doing it silently not spending any rate limit? That is a HUGE improvement no?
@TheLagBorn en X ↗ · 16:08 UTC
- Una idea
You probably won't read this, but just in case: you could fix the storage space issue. Over time, Codex's local conversation logs take up a huge amount of space due to repeated images and other data. I have to deal with storage issues caused by Codex every week.
@adinorcz en X ↗ · 16:13 UTC
- Un problema
Since the latest usage reset, my usage limit has been draining much faster than before. I'm reaching my limit almost twice as fast as I used to. Has anything changed?
@chou_ku en X ↗ · 16:40 UTC
- Un problema
Error: WebSocket closed 1000 Thinking... Error: terminated Error: terminated Aborted after 3 retry attempts I love the stuff you push out but it's really had to get things done when its constantly like this and at the same time Claude Sonnet agents just motor like beasts.
@spietila en X ↗ · 16:42 UTC
- Un problema
Another issue: I’ve been running into problems accessing my Codex desktop sessions from my phone. I used to be able to monitor them online, but now I’m constantly hitting an error saying "another Codex session is currently using it."
@Inm23 en X ↗ · 16:43 UTC
- Un problema
Did you guys optimize Astra? it's been running with 8 agents for 14 hours and my usage isnt dead yet
@lunaonmc en X ↗ · 16:49 UTC
- Un problema
Hey Tibo, heres a bug: it just keep showing up. It doesn't work when I click 'allowed'.
@hugo9138618800 en X ↗ · 16:51 UTC
-
My usage has been under control the last month since we improved my AI orchestration system, BeastFusion. We put governance in place with defined agent roles, spending limits, independent review, and approval gates. The agents have specific jobs and boundaries, and I retain…
@seangworld en X ↗ · 17:14 UTC
- Un problema
Anybody else's Codex getting security flagged all day long today? I'm just doing app development, but seems like the classifier doesn't care. @thsottiaux
@BBleimschein en X ↗ · 17:25 UTC
- Una idea
Thanks for the banked resets, I cannot use them fast enough on my plus plan. GPT 6.1 Sol medium is perfect for the 5 hr usage limit. You are leaving money on the table. I would buy banked resets at for $5 on a plus plan. My budget is $40 a month for feeding this hobby.
@eksMedina en X ↗ · 17:38 UTC
- Un problema
this requires immediate attention. All my work is paused and codex is completely unusable. After one tool call the chat stalls, messages stop streaming into the UI, and I’m sure other stuff are broken that I’m not aware of. But unusable is the only way to describe it
@ArmanRafiee3 en X ↗ · 18:13 UTC
- Un problema
Let me SSH into my devices on tailscale please. I have been trying to do that for weeks. This still doesn't allow SSH
@smusamashah en X ↗ · 18:25 UTC
- Una idea
te cambio los 28 dias de reset por eliminar esto 4ever: stream disconnected before completion: error sending request
@ultimamillaarg en X ↗ · 18:39 UTC
- Un problema
Hey @thsottiaux why do they keep reducing the Codex usage in the paid plans? The allowance doesn't cut it for day-to-day work anymore; the agent barely makes a few changes before running out of tokens.
@JoelCcari en X ↗ · 18:46 UTC
- Un problema
I'm so sorry! I won't be using Codex anymore next month - its intelligence is getting worse and worse! I don't understand what the point of your optimizations is. All we wanted was for it to be a bit smarter! That's all!
@Cherysunny08 en X ↗ · 18:58 UTC
- Un problema
Renewed my pro 5x but never recieved the banked reset :(
@AlexHay75417714 en X ↗ · 19:04 UTC
- Mixta
Tibo, GPT-6.1 Sol ultrafast doesn't affect most codex users since it's only available on the $500 plan. i think we need reset.
@ishuagra02 en X ↗ · 19:18 UTC
- Un problema
Did you guys nerfed the weekly limit too... I mean now, i get 5 sessions of 5hr limit in a week... Earlier I used to get around 7 or 8??? 😠
@SatvikMishra88 en X ↗ · 19:20 UTC
- Mixta
I'm not sure how many people actually want the fast or ultra-fast modes, given how quickly they burn through your usage allowance. I'd rather the team focus on making the platform stable and fully functional. I've run into issues with image previews, and the dot doesn't seem to…
@ak_ishere en X ↗ · 19:21 UTC
- Un problema
now for the next day could you fix how setting goals and editing them are completely broken in codex? I have to restart codex in order to do so which deletes the side chats I have
@MomusDigitalLLC en X ↗ · 19:27 UTC
- Un problema
I’ve been facing the issue of my codex app shutting down every 15 minutes. I’m on windows 11 and I’ve tried repairing the app, resetting it, uninstalling then reinstalling it. None of it has worked. @thsottiaux we need a fix
@theonlytobe1 en X ↗ · 19:37 UTC
- Mixta
Just it like voice.. it works it reads the new message and steers along the way or answers. I have like +30 lanes for the AI to decide. How to continue.
@Vaeshkar en X ↗ · 19:44 UTC
- Un acierto
Reset landed cleanly on my end and GPT-6 has been noticeably sharper so far usage feels much less constrained already.
@Briancole_0 en X ↗ · 19:44 UTC
- Un problema
Cancelled 13 Codex Pro $200 subscriptions and switched to 2x Claude Max 20x. Honestly more than enough. The past week on Codex has been rough: • Medium reasoning sits "thinking" for 4 minutes and produces nothing • 6.1 Sol keeps losing its way • Fast mode is slower than…
@tn_ycl en X ↗ · 19:47 UTC
- Un problema
That's what I needed. So a long Codex run before EOD PST can burn the banked reset, and ChatGPT plans don't help that clock.
@aiartgallerie en X ↗ · 20:01 UTC
- Un problema
Hey @tibo, the “Steer” button in ChatGPT Codex on desktop isn't working. When Codex is already running a task, I can't steer newly queued instructions from my PC. I have to switch to my phone to do it, where it works perfectly fine. Could you look into this? Thanks!
@ray319723 en X ↗ · 20:12 UTC
- Una idea
A way to switch seamlessly between cloud and local would be a great improvement bossman. 🫡 A local chat cannot even create a new codex cloud chat at the moment. This could save some time and effort.
@tanishsidhu en X ↗ · 20:15 UTC
- Un problema
instead can you get dot call fixed its 36 hours its not connecting. call failed every time
@he_the_mystery en X ↗ · 20:57 UTC
- Un problema
First, resolve the issue of the Codex interface crashing frequently and automatically reopening. This happens whenever there are too many tasks, rendering it practically unusable. Despite repeated reports, the problem remains unresolved.
@LinusWangP en X ↗ · 21:03 UTC
- Un problema
First, resolve the issue of the Codex interface crashing frequently and automatically reopening. This happens whenever there are too many tasks, rendering it practically unusable. Despite repeated reports, the problem remains unresolved.
@LinusWangP en X ↗ · 21:03 UTC
- Un problema
First, resolve the issue of the Codex interface crashing frequently and automatically reopening. This happens whenever there are too many tasks, rendering it practically unusable. Despite repeated reports, the problem remains unresolved.
@LinusWangP en X ↗ · 21:03 UTC
- Mixta
Love it . Timeout error resolved! Works for 5min then closes gracefully. Would be great to be able to push a button to continue than typing it… unless the chat will know if I press enter without any text?
@AdityainSF en X ↗ · 21:24 UTC
- Un problema
How about you just fix the current product. About to get opus for a second opinion as codex and gpt just getting dumber by the day. Just repeating instructions over and over now. Usage is now better but takes forever and still doesnt produce the results.
@chrisrismay en X ↗ · 21:31 UTC
- Una idea
i guess you might need to make it 50% faster with fast mode and 100% faster without
@aristotle_me en X ↗ · 21:42 UTC
- Un problema
Day 5 fix: “ClientError Encountered exception: <class 'caas.internal.errors.ClientError'>.” It eats up 30/50 of my astra pro quota each week 😭 Whats worse - the bug is marked as fixed, yet its very much still an issue.
@Dev__Haz en X ↗ · 21:45 UTC
- Mixta
Le quota part trop vite avec ChatGPT Plus, c'est la première fois où je songe vraiment à changer de fournisseur (j'utilise Claude Pro en parallèle, l'usage semblait équivalent du temps de GPT 5.5) mais en même temps vous êtes très bons et votre produit aussi
@GuiotNoel en X ↗ · 21:46 UTC
- Un problema
Codex is broken on windows... Helper unknown failure and it's not my devices I've followed all th necessary troubleshooting steps.
@Hader_campbell en X ↗ · 23:02 UTC
-
In Enterprise, as admin, I can’t get GHE to work. Repos are there but it just says failed. GitHub works though.
@2nic3 en X ↗ · 23:17 UTC
- Un problema
1/2HEY! Two things I want to mention. 1. ChatGPT jumps every time back to medium while high have been selected. 2. When I choose not to use Codex on the moment, and my usage stays at 100%, then why does ChatGPT keeping the renewal on 7 days by pushing a day forward every time?
@JBungelen en X ↗ · 23:37 UTC
- Una idea
Can you, please, increase Fong above 16? I can't see! My eyes hurt on Windows app with 4K monitor.
@Astralvs en X ↗ · 23:43 UTC
- Un problema
I pay $200/month. Codex and dot are useful when they work, but recurring cloud disconnects keep blocking sustained development. @thsottiaux , please prioritize reliable task resumption, clear failure status, and recovery instructions. Can your team investigate?
@dmitrij_linnik en X ↗ · 23:46 UTC
- Mixta
i can’t prove they slowed it deliberately. the documented 8× allowance burn is enough to criticize.
@dotimus_ en X ↗ · 23:56 UTC
- Un problema
Steering is constantly buggy and lags all of the time on complex jobs, its incredibly frustrating
@BrendanPlayford en X ↗ · 23:57 UTC
- Un problema
I upgraded from 20x to $500/month, yet Astra 6 Ultra & Sol 6.1 Ultra feel worse than at launch. Repeated capacity errors in Work/Codex with 84% weekly allowance left and 21,308 credits. Deeply disappointed. Please investigate quality & access. Experience: Oct 8, 2026.
@Enterpernure en X ↗ · 23:57 UTC
- Un problema
Fix codex in the mobile chatgpt app please. Only half my sessions show up and they are in the wrong projects.
@notdricky en X ↗ · 23:57 UTC
- Un problema
they broke the entire fucking app. it's useless and keeps crashing since the last update
@flyest en X ↗ · 23:58 UTC
- Un problema
steering and queueing is broken, like half the rest of your app's features. fix up man
@flyest en X ↗ · 23:59 UTC
5 respuestas en X de este día no se citan: Rantr no tiene su texto o lo retuvo tras la revisión automática. Siguen en X.