PixemWeb

Fullstack Developer, Blogger & YouTuber at and

[PixemWeb @PixemWebDesign]

👉 WordPress Theme Development combining old school and new school methods and tools.

https://t.co/Sk7aLYQ1Ty 👈

#WordPress #wordpresswebsite

[PixemWeb @PixemWebDesign]

@theo I was looking into the Mac Studio when the 512gb memory version is released. I wonder which would perform better. My MacBook Pro M4 Max chip with 128gb of memory is decent but not performant enough. Any thoughts on?

[PixemWeb @PixemWebDesign]

@bridgemindai I’ve been saying that for a minute now. Google is baking something big. They have a long horizon and lots of runway. While other ai labs have to make pitches to VC’s, Google is letting their secret source simmer for a long while.

[PixemWeb @PixemWebDesign]

@thsottiaux What happened to the reset? Make it a banked reset please 🙏🏼

[PixemWeb @PixemWebDesign]

I’m on the $200 plan for both and I can barely make it through the week. I’m seriously considering tagging in Grok into the mix at either the $100 or $300 plan. I figure it’s best to mix models and plans vs paying for extra usage credits which should be a last resort. Orchestration is what’s important at this point.

[PixemWeb @PixemWebDesign]

@williamsba There are still ignorant people out there 🤦🏻‍♂️

[PixemWeb @PixemWebDesign]

@thsottiaux Note to self. Check @thsottiaux before hitting the reset button. I guess I wasted a reset. That’s obviously my fault 🤦🏻‍♂️

[PixemWeb @PixemWebDesign]

I think plugins have a bright future. Here’s my logic. People will say something to the effect that anyone can build with AI right now, and that’s true. The issue is that the true cost of AI is orders of magnitude higher than the subscriptions offered today and with OpenAI stating that the $200 plan is no longer 20x and is now just 10x and also introduced the $500 plan at 25x, I think we can all see where this is going. Anything other than the absolute basic plugin will eventually be cost prohibitive. Note I have the ChatGPT Pro plan and the Claude Max plan and a few months ago I could build multiple plugins simultaneously. Now I can barely make it through the week before my usage is exhausted and I’m begging for a reset like the kid in that movie saying “Please sir, can I have some more” bowl in hand. So only the developers and agencies that currently exist (or will exist) and have the pockets to pay will be able to build complex solutions. That being said, I do see workarounds where people can follow a project that’s similar to their idea and that they know is built with a frontier model, and maybe use a lesser plan or lesser models or even open weights models to dissect the project and recreate something closer to their original idea (low budget distillation). Or an even better idea, they might just buy the plugin and keep it moving. Just my two cents.

[PixemWeb @PixemWebDesign]

I spent the past 6 months developing an alternative to Quickbooks Advanced Online for this very reason. I have a beta version I’m testing internally for the next 2 months before releasing it. It’s a WordPress Plugin with all the features that Quickbooks offers with even more CRM capabilities. The price increases have been insane.

[PixemWeb @PixemWebDesign]

@thsottiaux It seems to have corrected itself. All's good now.

[PixemWeb @PixemWebDesign]

@thsottiaux Astra is a great model, I'm really enjoying it, but I noticed that for some reason, the usage went from 77% available to 0% in minutes. There must be a glitch since it also previously had my re-up for usage on 9-15 and now it says 9-12. Did something happen?

[PixemWeb @PixemWebDesign]

@claudeai Can we get a reset?

[PixemWeb @PixemWebDesign]

@thsottiaux Could use a reset right about now 🙏🏼

[PixemWeb @PixemWebDesign]

@Eng_china5 @three_cube @grok is this real?

[PixemWeb @PixemWebDesign]

@claudeai @DanielMiessler Reset limits please 🙏🏼

[PixemWeb @PixemWebDesign]

Working hard to exhaust my Fable Usage before the session ends. Tic Toc

[PixemWeb @PixemWebDesign]

Tip of the day. If you run out of usage tokens with either Claude Code or Codex CLI:
1. Download Ollama onto a dedicated user account (optional. You can install Ollama on your main account if you want)
2. Download a local model
3. Then in terminal use the command “ollama serve”
4. Then switch to your main dev account on your computer
5. Open terminal
6. Then use “ollama launch claude” or “ollama launch codex”
7. select your local model of choice that you downloaded (like qwen3 coder)

Then you can continue working with either claude or codex as the agent shell, and your local model for all token usage.

Your skills continue to work and you don’t need to worry about exhausting your tokens from Codex or Claude.

[PixemWeb @PixemWebDesign]

@DataChaz That was me 😂
Now they extended it and I'm out of Fable.
So I'm like that kid in the movie Oliver saying
"Please sir, can I have some more"

[PixemWeb @PixemWebDesign]

@claudeai Reset limits please.

[PixemWeb @PixemWebDesign]

@claudeai Reset limits please

[PixemWeb @PixemWebDesign]

@claudeai Reset weekly limits please