OPENCODEX / Ice hockey / 24 Please call me JUN btw
Seoul
Joined February 2025
- Tweets1.8K
- Following447
- Followers1.3K
- Likes1.8K
True so headache what to build
We will all be learning in coming weeks & months, that most human-made software:
▫️ Was really slow
▫️ Had many security vulnerabilities
▫️ Was under-specified and incorrect
AI will rewrite it all. Many times. It will make it fast, cost-efficient, and secure. It will think about failure modes we couldn't wrap our heads around.
It will prove, formally verify, load test, harden. It will turn memory-unsafe foundations into robust, optimized ones. If you ask, it'll thoroughly document it, teach you about it, make diagrams, animations, and slide decks.
In many ways, we are witnessing the true birth of software engineering.
Got rejected by YC today.
Made it to the interview, didn't make the batch.
Fair call. We're a Korean team with Korean customers, and I still pitch in broken English.
So here's the plan:
moving to SF in December anyway
shipping every single day until then
posting every day. Day 1 of 40.
Day 1 of 40/ GOAT runs your coding agents (Claude Code, Codex, Cursor) on every machine you own. Laptop, home desktop, SSH box, cloud VM. One app, open source.
I also built OpenCodex (18K stars).
which setup your all models configs and account in one click setup
What should I ship first?
quick quote
multi account also violate TOS
and cli proxy does it
btw opencodex gives more native flow to claude code
than cli proxy api
if you use account pool via claude code
you could use it natively
Im just want to make models and harness layer
freedom
Change company name codexgepete
Bad news for ChatGPT users: starting in December, Plus, Pro and Business limits will be shared across ChatGPT and Codex.
OpenAI quietly updated its help page. Usage will now be measured in tokens, so bigger models and harder tasks burn through your allowance faster. Only Instant chat stays free of the limit. Hit the cap and you wait, upgrade or buy credits. Heavy Codex users may feel this one.
Kimi membership giveaway winners 🎉
@minpeter: K3.1 Pro + SolidCraft for the keyboard case you're designing
@Cold_brew_T: Kimi as the security reviewer on your read-only Toss Securities dashboard in pi
@kernelKain: Backboard (Codex + VS Code) for hackathon builds, starting with Flip the Switch on Nov 4
Codes are on the way in your DMs.
Thanks to everyone who replied, 97 answers and a lot of great harness setups in there.
As a Kimi Ambassador, I'm giving away 3 Kimi memberships ($39 each).
Reply with which harness you'd run it in, or what you'd build with it. I'll pick 3 in 48 hours.
My current setup added Kimi K3.1 Pro in opencodex,
I use kimi to review code for security issues, p
lans the review and hands the pieces out to sub-agents.
It feels like running a model with Daybreak access. It digs in instead of hedging.
Thanks for the welcome kit kimi!
#BuildWithKimi #kimi
The best part about opencodex is
that you can run every harness
|on the exact same model setup.
Codex, Claude Code, OpenCode, Pi: a native experience with any model, set up in one click. Easy enough that my mom could do it.
For example, I really love Codex's code mode,
and I enjoy watching Opus fire off parallel tool calls in there while using Codex's goated Mac computer use.
can someone tell me that
what is the result of 28 day challenge
🤥
we just want reasonable codex ux improvement
but my codex connection getting worse
and dot is ngl far behind than grok bot
Actually I dont serve local models at all
Btw to run 100+ sessions
In “codex”
I need 512+ ram in my linux workstation
you have many device
and want to use a lot of accounts
You dont need to login each of them
and just use opencodex remote
you can manage pull in once and
auto config same to all devices
Use opencodex and try it automatic
Day 26/ You can now easily switch between multiple accounts on mobile, and coming soon to desktop too. Next two days will be resets… banked… do with this what you may 👀.
help.openai.com/en/articles/…