I think I need to find a way to manage AI skills and code memory. I'm always intrigued by the idea of having a coding machine separate from my desktop, but then the browser is missing, all the memories are missing, and MCPs are missing.
I would also love to experiment with stuff like Hermes or OpenCode, but I have the same issue.
I just found out that you can deduplicate files on a #Mac with #APFS. #jdupes --recurse --dedupe [your coding workspace] Saved me 25% of disk space without deleting a single file. https://github.com/jdlorimer/jdupes
I wasn't aware that we have so many interesting #EU#LLM hosting options now. #Berget Ai: seems to offer an unlimited flat rate with #GLM 5.2 and #Kimi K3 (will need to check out) #Melious: Also has an unlimited offer (although it seems to only cover GLM 5.2) #Mistral also offers GLM now and, if you are ok with API pricing, you can use Cortecs as an open router alternative with a good selection of providers.
TIL Batch jobs on the MistralAI api can randomly be ignored. It's probably better to set a 1 hour timeout and just resubmit the batch whenever it fails.
I started a gigantic #TYPO3 upgrade/refactoring task with #Fable as an orchestrator. It has been going for more than 12 hours. It is looking almost perfect. Over 170 #Opus sub-agents. Zero input after the start. The site looks identical despite using completely different tech under the hood. That would have taken ~ maybe 4 weeks before?
I think I need to find a way to manage AI skills and code memory. I'm always intrigued by the idea of having a coding machine separate from my desktop, but then the browser is missing, all the memories are missing, and MCPs are missing.
I would also love to experiment with stuff like Hermes or OpenCode, but I have the same issue.
I wasn't aware that we have so many interesting #EU#LLM hosting options now. #Berget Ai: seems to offer an unlimited flat rate with #GLM 5.2 and #Kimi K3 (will need to check out) #Melious: Also has an unlimited offer (although it seems to only cover GLM 5.2) #Mistral also offers GLM now and, if you are ok with API pricing, you can use Cortecs as an open router alternative with a good selection of providers.
I started a gigantic #TYPO3 upgrade/refactoring task with #Fable as an orchestrator. It has been going for more than 12 hours. It is looking almost perfect. Over 170 #Opus sub-agents. Zero input after the start. The site looks identical despite using completely different tech under the hood. That would have taken ~ maybe 4 weeks before?
I just found out that you can deduplicate files on a #Mac with #APFS. #jdupes --recurse --dedupe [your coding workspace] Saved me 25% of disk space without deleting a single file. https://github.com/jdlorimer/jdupes
TIL Batch jobs on the MistralAI api can randomly be ignored. It's probably better to set a 1 hour timeout and just resubmit the batch whenever it fails.
