Ah
mike_wooskey
joined 2 years ago
I don't understand the joke.
Yeah, my thinking was definitely biased to my environment. I selfhost llama.cpp so even if Hivekeep doesn't require significant resources, whatever LLMs it runs will use my hardware.
agents are activated serially per message, not all firing at once, and the persistent memory is exactly what keeps each context small
It sounds like I need to try it!
I laughed out loud (well giggled out loud) at this one, twice.
This looks interesting, especially the persistent memory. I want to try it out but it seems likely to me that multiple simultaneous agents would require significant hardware. Even if they were serially activated, reloading contexts with each switch would take time. I have a pretty beefy GPU and experience significant (almost ridiculous) slowdown when opencode runs 2 subagents simultaneously.
But perhaps the memory storage/lookup keeps contexts very small?
Anyway, I can't find any mention in the repo or docs what the suggested minimum hardware is.
I'm so sorry. I'm glad you and he had each other.
Tangent: I've stopped using "google" as a verb as part of my philosophy of removing Big Data from my life (and hopefully having a tiny, tiny influence on others to do the same). It's still possible and likely that Adobe Photoshop was used for this pic, but it's quickly becoming even more likely that AI was the tool. Maybe "this pic was manipulated"?
I'm glad this conversation is happening.
LuisCore is a X, X X for X X at scale: an X for X, X, and X X across X. The X X is luiscore.com, with X at X and X.
I often feel like that's what I'm reading, but I assume that enough people understand it that the trend goes on. Like maybe I should just not read articles about tech any more.
I really like the quote "If you can’t explain it to a six-year-old, you don’t really understand it yourself" (possibly by Feynman). I'd like this to be how the works works, but it would mean that most people don't understand what they're talking about or they're intentionally being pretentious and exclusive. But technical things do get made, so it seems more likely that I just don't understand.
Yes. You can use opencode, the agenetic coding tool, with just about any llm runner or model. But Opencode Go is their cloud-llm suscription plan, with limited/slightly-dated llm models.
Good thinking to mention open-webui - I was only thinking agenetic coding, but I use open-webui for llm chatting. I think it's fantastic.
FYI, I think opencode go is kind of a subscription model, not a direct credits-to-tokens model. In terms of value it's nowhere near as good as Claude's subscriptions, but it seems way more valuable that paying for tokens directly. However they only offer a few models - decent ones, but not many and a little behind the times.
The Purina for Professionals website's "cookie management" (to opt out of cookies) is heavily dark-patterned. Not only does the popup at the bottom of the page not have cookie management (it has a link that takes you to a page that has a link to take you to cookie management), but the wording is intentionally counter-intuitive to opt out:
The option is called "Do Not Sell or Share My Personal Information", which suggests that by selecting the option, you want them to not sell or share your information. But in parentheses it adds "slide left to opt out of sale/share", which suggests that disabling the option means for them to not sell or share your information.
Which of the 3 buttons does which of the 4 commands listed next to them?
Of the 3 buttons in the red-outlined-box on this Brother printer control panel, which does what function? Grrr!
view more: next ›
Only minor performance degradation with compression, but does the compression also mean the MOE model can run on a GPU with less VRAM? E.g., I don't think I can run GLM 5.2 on my W7900 (48G) even though the experts only take 40G, but would this compression allow it to run?