HeyDust

A designer，Share commercially viable AI-generated artwork and Notion templates

[HeyDust @retned]

This week's Codex quota is running out really fast—just a few casual conversations used up 10% of it. Have you noticed?

[HeyDust @retned]

@HuggingModels YES!

[HeyDust @retned]

@thsottiaux Complete Step 10B—Faults and Fault Attribution

[HeyDust @retned]

@Gino_miaoyan 招行外币卡可以直接充

[HeyDust @retned]

If your usage quota was resetting soon and you still had a lot left, what would you put on autopilot?
Looking for tasks where I can give the model a goal, let it work for a while, and only check the final output.
What are your best low-supervision AI workflows?

[HeyDust @retned]

I just reserved my Cloudflare Wallet tag: https://t.co/gxgkmJ8jpA.

Reserve yours now at https://t.co/BasffuXHci

[HeyDust @retned]

@edwardartist Good job! 🥳

[HeyDust @retned]

@thsottiaux The God of Reset

[HeyDust @retned]

Yes, Luna Max doesn't follow the rules and breaks through the limitations set by the developers. I tried it once and gave up because reworking it requires more tokens.

[Haz @diegohaz]

GPT 5.6 Luna Max vs. Sol Medium

I spent the last 2 days running several sessions with both, and I think the official benchmarks are misleading in suggesting they're equivalent.

Luna is obviously much cheaper. Despite using many more tokens, it cost an average of $1.20/session, while Sol averaged $29. That's a big difference, and it also affects subscription rate limits! So if you're prioritizing cost over other factors, Luna is probably the way to go.

However, I had to redo most of Luna's work. In fact, Luna wasn't able to produce any shippable code for my use cases, so the huge cost reduction was just an illusion.

The bigger issue with Luna was that it often ignored initial instructions, especially as the conversation grew longer. Toward the end of tasks, it got lost on very basic stuff, which made me switch to Sol for most of them.

I had Opus 5 review all their work, and Luna generally introduced more P1 regressions.

After reviewing the sessions, both Opus 5 and Sol xhigh concluded that Sol Medium produced better results despite the price, which matches my own assessment.

[HeyDust @retned]

We often talk about the “barrel effect.” The key is simple: maximize your strengths, let Codex fill the gaps, and together build the perfect barrel.

[HeyDust @retned]

You need to keep your Codex at the “almost empty” level at all times; otherwise, when Tibo resets, it'll feel like you've lost hundreds of millions!

[HeyDust @retned]

Compared to using Luna Max, switching to manual review mode can also save a significant amount of tokens.

[HeyDust @retned]

To make the reset more rewarding, I’ve already used up all my allotted credits in advance—it’s easy to get hooked that way.

[HeyDust @retned]

@thsottiaux I can't create a dialog in the newly created project directory either.

[HeyDust @retned]

I encountered a bug in Codex. When I click “Create Dialog” next to the project directory, the message cannot be sent properly. If I use “New Task,” it works, but I can't link it to the project folder. I'm not sure if anyone else has encountered this issue.@thsottiaux

[HeyDust @retned]

@thsottiaux Could you add a “jump to response” / answer navigation feature like ChatGPT web?

When Codex generates long responses (code explanations, debugging steps, file changes), it becomes difficult to quickly locate previous answers.

[HeyDust @retned]

@jxnlco I use conversations with ChatGPT to guide Codex’s operations; ChatGPT is the brain, and Codex is a pair of skilled hands.

[HeyDust @retned]

I plan to keep it for use in complex projects.

[Tibo @thsottiaux]

Now that you can bank usage resets in Codex. Are you a hoarder or do you use them without breaking a sweat? How do you think about them?

[HeyDust @retned]

@elonmusk Good~

[Claude @claudeai]

Claude now connects to the tools creative professionals already use.

With the new Blender connector, you can debug a scene, build new tools, or batch-apply changes across every object, directly from Claude.