segmondy

13 ms·

391.
by segmondy 8mo ago
Great news! I have always worried about ggml and long term prospect for them and wished for them to be rewarded for their effort.
392.
by segmondy 8mo ago
there's a difference between a book and data or music and data. that is their data. if you have a painting and i take a picture of it and store it on my drive. it's my data, i don't own the copyright to it tho, but it&#x27
393.
by segmondy 8mo ago
where were the alternatives before tailscale? we could only read bout BeyondCore with envy before tailscale. i'm going to keep supporting them unless they do something naughty.
394.
by segmondy 8mo ago
Thanks, any one with kids experiences this. It's so frustrating. For the same kid you could literally have 5 different sizes that are the same. So you have to keep track of sizes by brand. Trying a new brand is often an adventure.
395.
by segmondy 8mo ago
It's fairly well known to some of us. It's one of the few shows I own DVDs of.
396.
by segmondy 8mo ago
depends on what you mean by impractical. but some of us are trodding quite along.
397.
by segmondy 8mo ago
It is, but Anthropic is deluded if they think they can gain a moat from that. They gladly copy/borrow ideas from other projects that's why they are paranoid that folks are doing the same.
398.
by segmondy 8mo ago
you are comparing to DeepSeek's old OCR, there's DeepSeek-OCR2 which btw is amazing from my experimentations. https://huggingface.co/deepseek-ai/DeepSeek-OCR-2
399.
by segmondy 8mo ago
RLVR. Reinforcement Learning with Verifiable Rewards. Prior to this it was RLHF, reinforcement learning with human feedback. The models can now be trained without human in the loop for coding problems, you give them code to solve. yo
400.
by segmondy 8mo ago
They want to know your accounts, they will force you to link it so they can reach there and claw back money for disputes etc.
401.
by segmondy 8mo ago
For those claiming they rigged it. Do you have any concrete evidence? What if the models have just gotten really good? I just asked Gemini pro to generate an SVG of an octopus dunking a basketball and it did a great job. Not even Deep
402.
by segmondy 8mo ago
you have to choose to keep vibing or learn to code. there's a possibility that programming might not matter and being good at vibing would be what matters. no one knows yet ... for now it does pay to know how to.
403.
by segmondy 8mo ago
if you audit it, then you're not vibing.
404.
by segmondy 8mo ago
It was obvious to me that they were top contender 2 years ago ... https://www.reddit.com/r/LocalLLaMA/comments/1c0je6h/google_...
405.
by segmondy 8mo ago
Wow, Great story. The title was boring, but I'm procrastinating and decide to read on. It's amazing how LLM is supercharging our capabilities. I was just telling my wife that we need to record more videos of the kids. I use
406.
by segmondy 9mo ago
This is actually a good benchmark, I use to roll my eyes at it. Then I decided to apply the same idea and ask the models to generate SVG image of "something" not going to put it out there. There was a strong correlation between
407.
by segmondy 9mo ago
No, you're just getting used to them.
408.
by segmondy 9mo ago
Local models are not behind. There are many specialized local models on huggingface that can do things that none of the closed/commercial models can do. The only way to get that edge is to run locally. When I say many, I mean in the
409.
by segmondy 9mo ago
let the crew paint a new flag for the country they want to dock at.
410.
by segmondy 9mo ago
LLM thingz
411.
by segmondy 9mo ago
It's already difficult getting your Tesla repaired, it's going to be a nightmare when Melon pivots away completely from EV which he will once it's obvious that he's getting crushed all around, not just in sales but desig
412.
by segmondy 9mo ago
This is nothing new, folks have been doing this for since 2023. Lots of paper on arxiv and lots of code in github with implementation of multiagents. ... the "limit" were agents were not as smart then, context window was much smal
413.
by segmondy 9mo ago
I cancelled my digital ocean server of almost a decade late last year and replaced it with a raspberry pi 3 that was doing nothing. We can do it, we should do it.
414.
by segmondy 9mo ago
eventually we will have smarter smaller models, but as of now, larger models are smarter by far. time and experience has already answered that.
415.
by segmondy 9mo ago
Nah, their "moat" is CC, they are afraid that as other folks build effective coding agent, they are are going lose market share.
416.
by segmondy 9mo ago
you do realize claude opus/gpt5 are probably like 1000B-2000B models? So trying to have a model that's < 60B offer the same level of performance will be a miracle...
417.
by segmondy 9mo ago
The green/yellow/red indicators are based on what you set for your hardware on huggingface.
418.
by segmondy 9mo ago
Not applicable... the models just process whatever context you provide to them, context management happens outside of the model and depends on your inference tool/coding agent.
419.
by segmondy 9mo ago
No, at Q2 you are looking at a size of about 26gb-30gb. Q3 exceeds it, you might run it, but the result might vary. Best to run a smaller model like qwen3-32b/30b at Q6
420.
by segmondy 9mo ago
1 6000 should be fine, Q6_K_XL gguf will be almost on par with the raw weights and should let you have 128k-256k context.

More ›