All Activity

I have to agree with @Cyberdevs. I've tried several AI's but Claude Code is the best for my needs. It made my life much easier when I was setting up my server. It even fixed few issues that I had with my Hackintosh.

For me Claude is the go to source, it really does the job for me. I've tested ChatGPT, DeepSeek, Gemini and Grok and Claude by far is the best one for my needs. I've tested some local AIs but never get the results I was looking for. Besides my GPU (RX6800 XT) seems to be not that suitable for large projects. On macOS it's nearly impossible to do anything heavy but in Windows I used Zluda and Stable Diffusion Forge for AMD and it generates images with 720x1080 between 11 to 21 seconds based on the Loras and the Models used. For Text generation I've tested Ollama and LM Studio which can utilize AMD GPUs if you have ROCm installed and the result is much better than macOS. Stability Matrix has an application for Windows and macOS (Apple Silicon only) which can download and install various Generative AI tools and it works pretty nice for Image and Text Generation. I've recently discovered another app called LynxHub which works on Intel Macs as well and has plenty of repos to install but haven't used it much.

It's a pity AI is not genius and sometime or always it can't resolve a problem. 1. First I see https://www.insanelymac.com/forum/topic/362634-nootedgreenkext-is-on-air-its-a-long-long-road-to-complete/ The last message "The end of the road". Very pity! 2. I was asking Qwen.ai How to compile Wine 11 to have an 32-bit support in macOS? After some attempts he said it is impossible. But no. I made it with my own brain! https://github.com/SergeySlice/Wine-Sherry/releases/tag/SherryWoW-11.1-23 3. VoodooHDA is still not finished. It will be separate story. Have I any success? Only in small parts...

DisplayPort should be asymmetric.

@Green206 Take a look at the WhateverGreen.kext boot-arg -wegnoegpu. Oh... you're welcome.

Extract kext from patcher, works excellent with BCM4352. One minor symptom, ater kext in OC 1.0.8 on Sequoia, my sleep/wake schedule which i cancell all and lock, made my HP wakes to strange power events which does not exists before. That is the same symptom as in your patched BlueToolFixup kext and -btlfxwakecrash boot arg.

Interesting! Let me think and test it.

Tell us your story what you asked from AI. Tell us non-standard approach to the dialog. Ask a question how to use AI. Tell us what AI do you prefer and why. Show images created with AI. Tell us funny story related to AI.

I use it a lot for cover design and to know if the response is accurate or imagined.

I don't know if it would be interesting or not, but it would be good to know the monitor connections, an example done with ChapGPT as we discussed in the new IA channel.

@chris1111 Got them Thank you!!!

If I use an HDMI switch and connect one cable to the dedicated card (the RX 6400) for Windows and the other to the integrated graphics for Mac, will that work, or will I have to manually switch the video source for the Mac?

Sometimes happens to me, but its randomly, even with a eGPU, maybe drivers?

I send you the Disk GOldenGate in PM

@LockDown there are multiple such apps for Windows. You should be able to find one with a search. I'm not aware of similar apps for macOS, but they may exist, too. If you don't want to find the app, there are published test results that you can find with a search. EDIT: I stand corrected. I did measure Radeon dGPU power in macOS. See here.

@ird It's going to take me a little longer to test a 65W EliteDesk 800 G4 Mini with i5-8600, because the only such unit that I have includes a motherboard that I tinkered with while attempting to produce a 65W motherboard with RX 560x dGPU. I de-soldered / re-soldered components on this particular motherboard and while it "works", its multi-core performance is lousy. It appears that I have caused some kind of thermal problem in the CPU power MOSFETS. They get insanely hot during GB6 multi-core tests, so the motherboard (fortunately) throttles CPU speed and never achieves full multi-core performance. If someone else can test the EliteDesk 800 G4 Mini 65W with i5-8600 / 230W power supply, that would save me some trouble of piecing together a new test system. By the way, if you're curious about my attempts to create a 65W G4 Mini motherboard with RX 560x dGPU, the short answer is that I failed after multiple attempts. The long answer involves destroying one 65W motherboard, destroying one 35W motherboard, crippling another 65W motherboard (the one I was hoping to use for this GB6 test) and burning up a m.2 NVMe SSD. The benefit of this exercise is that I now own a soldering heat gun, learned how to de-solder and solder components with a heat gun and I created my own soldering hot plate. It was a rather time-consuming, expensive experiment.

OT What app to know how much power drawn?

macOS Tahoe 26.5.2 is just around the corner

Initial work to update the app to Golden Gate. macOS app and disks icons courtesy of @chris1111 EDIT: server error trying to upload the pics, sorry.

@kgp About This Hack needs to be updated to Golden Gate. New icon and the rest. Working on it.

@fantomas As @Slice said “Where user may exchange information about using different AI: Claude, ChatGPT, DeepSeek etc.” or something like this.

This time a sequential OTA update of just 10 GB! With macOS 27 beta 2 (26A5368g), Apple appears to have fixed the following issues on my MacBook Pro M1: Keyboard and trackpad now work immediately after waking from sleep or opening the lid. No Wi-Fi issues after waking from sleep or opening the lid. Secondary click now works perfectly when configured as "Click in Bottom Right Corner". Interestingly, Apple released Kernel Debug Kit macOS 26.7 build 26A5368g, apparently confusing macOS 27 beta 2 with macOS 26.7. The very bad part of this funny confusion at first glance: Any future Tahoe 26.7 KDK release is unlikely to carry a build number higher than 26A. Consequently, Kernel Debug Kit macOS 26.7 build 26A5368g - which actually belongs to macOS 27 beta 2 (Golden Gate) — would not be compatible with macOS Tahoe. If Apple keeps this naming scheme, OCLP's automatic KDK matching could become problematic beyond macOS Tahoe 26.6, as it may no longer be possible to reliably distinguish Tahoe 26.7 KDKs from Golden Gate KDKs based on build numbering alone. This assumes, of course, that macOS 26.7 will be the final major Tahoe release. 🤕 Whether this is an actual mistake or an intentional change in Apple's KDK naming scheme remains unclear. If intentional, OCLP and similar tools may eventually need to adjust their KDK matching logic accordingly. I have already submitted a Feedback Assistant report to Apple regarding this apparent KDK naming issue.

Kernel Debug Kit 26.6 build 25G5043d is also available on Dortania's KDK mirror, and OCLP correctly selects the direct KDK match for macOS 26.6 beta 2. Thanks to Apple for choosing the correct KDK nomenclature this time. 😄 Modern Audio, modern Wi-Fi, AirDrop, AirPlay, Screen Mirroring, Personal Hotspot, Continuity Camera, and sleep/wake continue working flawlessly with my BCM943602CDP. Interestingly, both Apple and Dortania's KDK mirror also expose Kernel Debug Kit 26.7 build 26A5368g. However, 26A5368g is in fact the build number of macOS 27 beta 2, which means Apple has apparently introduced yet another KDK nomenclature flaw by confusing macOS 27 with macOS 26.7. The very bad part of this funny confusion at first glance: Any future Tahoe 26.7 KDK release is unlikely to carry a build number higher than 26A. Consequently, Kernel Debug Kit macOS 26.7 build 26A5368g - which actually belongs to macOS 27 beta 2 (Golden Gate) - would not be compatible with macOS Tahoe. If Apple keeps this naming scheme, OCLP's automatic KDK matching could become problematic beyond macOS Tahoe 26.6, as it may no longer be possible to reliably distinguish Tahoe 26.7 KDKs from Golden Gate KDKs based on build numbering alone. This assumes, of course, that macOS 26.7 will be the final major Tahoe release. 🤕 Whether this is an actual mistake or an intentional change in Apple's KDK naming scheme remains unclear. If intentional, OCLP and similar tools may eventually need to adjust their KDK matching logic accordingly. I have already submitted a Feedback Assistant report to Apple regarding this apparent KDK naming issue.

Does Apple like spoilers?

https://swdist.apple.com/content/downloads/08/22/140-31668-A_UOOEA5WMR8/yx13s8l4tygi9n1ar7rxhrhpn9t04wvk1n/InstallAssistant.pkg <dict> <key>BUILD</key> <string>26A5368g</string> <key>VERSION</key> <string>27.0</string> [ 35 ] 2026-06-22 17:16:36 .. 140-31668 .. 27.0.0.26A5368g macOS Golden Gate Beta