Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- The author says the optimization has rolled out and concerns how ChatGPT and Codex desktop apps save state on disk. The 60% figure is his expectation for that stored state, not a result measured for this article.
- The post supplies no supported-OS list, minimum version, per-client rollout completion time or schedule for reclaiming old data. It does not establish an update requirement or cleanup command.
- Compare like with like: installation size, state data and total disk usage are different measurements. Without an earlier baseline, a current reading cannot establish the reduction.
- This is a storage change. The post does not announce a new usage allowance, a reset or delivery of the separate plan for more usage. Do not delete app data simply to chase the percentage.
What actually changed?
The current returned version of Tibo’s post is marked last edited on October 13, 2026, at 23:32 UTC. It calls this progress “Day 9” and says optimizations have rolled out for how the ChatGPT and Codex desktop apps store state on disk. That edit time is October 14 at 07:32 in Asia/Shanghai; this is a time-zone conversion, not confirmation of the original publication time.
Think of state here as data an app writes to disk to retain its working state. The post does not list the affected files or directories, so it does not establish that chat history, project files, caches or a particular database make up the entire affected set.
The author says the change should reduce the amount of state stored on disk by 60%, citing feedback from power users running out of space. This is his description of the expected effect. The post provides no per-device measurements and does not promise that every user will save the same number of bytes.
Why might the whole app shrink by much less than 60%?
The denominator matters. The post refers to state stored on disk, not the installation package, the entire app directory, every project and download, or all used space on a computer. It also does not say memory use or execution speed will change by the same percentage.
As a hypothetical illustration, if the affected state previously occupied 1 GB under a particular measurement, a 60% reduction would mean about 0.6 GB less. This is an explanation of the ratio, not a test result. It does not imply that a computer with 100 GB of used disk space will gain 60 GB.
How can you observe the change without comparing different numbers?
First note the desktop app, operating system and version you actually use, then consult the storage statistics already available in your operating system. If you have an earlier record for the same location and measurement, compare it after normal use. Switching measurement categories prevents you from attributing the full difference to this optimization.
Without an earlier record, today’s reading is only a new baseline. Workload, saved data and other files on the computer may all change. A single total-disk reading cannot verify the author’s 60% figure; an absence of an obvious decline alone also does not prove the change is ineffective.
Which version is required, and should you clear data manually?
The post gives no operating-system list, minimum version, client-by-client rollout progress or timing for reclaiming old state. Saying the optimization has rolled out does not establish that every computer has completed an update. Check your own version information and subsequent official instructions for the client you use.
The post does not require reinstalling the app, deleting a database or clearing an app directory. Without an explicit file scope and recovery instructions, do not treat manual deletion as a necessary activation step—especially where it could remove work you need to keep.
Disk savings and more usage are separate developments
This Day 9 announcement describes a storage change. It does not announce a new usage multiplier, reset time or plan eligibility. Tibo’s earlier preview of more usage must still be followed on its own evidence; shipping a disk optimization does not establish delivery of that separate plan.
Mentioning feedback from power users explains the motivation. It does not create an “only power users are eligible” rule. Some users in the returned public replies made their own inferences; those are not additional promises from the author. If space remains tight, identify what is actually occupying it before following explicit instructions for the relevant client.
Sources
Frequently Asked Questions
Does the announcement also cover mobile apps, the web or the API?
It names only the ChatGPT and Codex desktop apps. It does not extend the same disk-state optimization to mobile, web or API use, so the desktop statement cannot establish a guarantee for those entry points.
Do I have to uninstall and reinstall to save space?
The post gives no such requirement and provides no reinstall or manual-cleanup procedure. Do not delete app data just to reach a percentage without explicit backup and recovery instructions.
Can I calculate how many gigabytes I will gain from the 60% figure?
A proportional estimate requires the earlier amount of affected state measured on the same basis. Total disk usage or installation-package size is not that baseline, and the actual change still needs observation.
Did the author call this Day 9?
Yes. The original post explicitly says “Day 9.” It identifies progress in his series; it does not mean there are two updates each day or that this announcement also grants a usage reset.