segmondy

11 ms·

421.
by segmondy 9mo ago
Try KimiK2.5 and DeepSeekv3.2-Speciale
422.
by segmondy 9mo ago
children? how about adults? After USA destabilized Libya, it turned horrible. In Libya there are open slave markets. Adults. Africans trying to who travel a great deal trying to get to Europe are often kidnapped and kept as slaves in Lib
423.
by segmondy 9mo ago
AGI is already here and arrived without a bang, AGI arrived last year. To each his own and their own reality.
424.
by segmondy 9mo ago
this is where design patterns come in handy even tho folks roll their eyes at it.
425.
by segmondy 9mo ago
you can add 1 more GPU so you can take advantage of tensor parallel. I get the same speed with 5 3090's with most of the model on 2400mhz ddr4 ram, 8.5tk almost constant. I don't really do agents but chat, and it holds up to 6
426.
by segmondy 9mo ago
You can run it on a mac studio with 512gb ram, that's the easiest way. I run it at home on a multi rig GPU with partial offload to ram.
427.
by segmondy 9mo ago
The old Kimi K2 is better than GLM4.7
428.
by segmondy 9mo ago
read the tech report
429.
by segmondy 9mo ago
I'm so sorry to hear that you are falling behind. With this attitude you are going to keep falling behind. Good luck.
430.
by segmondy 9mo ago
grifters gonna grift
431.
by segmondy 9mo ago
4 3090s, epyc MB with 8 channel memory, 7352 cpu, slow 2400mhz ddr4 rams.
432.
by segmondy 9mo ago
There's been so many moments that folks not really heavy into LLM have missed, DeepSeekR1 was great, but so was all the "incremental" improvements, v3-0324, v3.1, v3.1-terminus, and now v3.2-speciale. With that this is the
433.
by segmondy 9mo ago
I run KimiK2 at home, Most of it on system ram with a few layers offloaded to old 3090s. This is a cheap budget build. Kimi-K2-Thinking-UD-Q3_K_XL-00001-of-00010.gguf Generation - 5,231 tokens 604.63s 8.65 tokens/s
434.
by segmondy 9mo ago
... and what does this prove? what can you decide to use one LLM to solve over another based on this tetrisbench besides play tetris?
435.
by segmondy 9mo ago
If you give LLM a hammer everything looks like a nail, you give it a saw everything looks like wood. You ask LLM to find issues, it will find "issues" At the end of the day, you will have to fix those issues, if you decide to ha
436.
by segmondy 9mo ago
"The browser could be a sandbox" but the browser is definitely not a sandbox. The browser is an environment.
437.
by segmondy 9mo ago
Your ranking is way off, Deepseek crushes Minimax and GLM. It's not even a competition.
438.
by segmondy 9mo ago
You are not familiar, yet you claim a bias. Bias based on what? I use pretty much just open-source models for the last 2 years. I occasionally give OpenAI and Anthropic a try to see how good they are. But I stopped supporting them when
439.
by segmondy 9mo ago
Seriously? Does Netflix or Spotify cost the same everywhere around the world? They earn less and their buying power is less.
440.
by segmondy 9mo ago
Folks in tech built that for them knowing the purpose.
441.
by segmondy 9mo ago
if you want to do math proofs use AI built for proof https://huggingface.co/deepseek-ai/DeepSeek-Math-V2 https://huggingface.co/deepseek-ai/DeepSeek-Prover-V2-671B
442.
by segmondy 9mo ago
right, but the stolen phones get sold in other countries where the carriers don't care if the phone was stolen but care that someone is spending money on their service.
443.
by segmondy 9mo ago
sure, but let's look at the name of most folks leading their AI efforts, 80-90% have degrees.
444.
by segmondy 9mo ago
farmer carpenter auto mechanic dog sitter nurse teacher chef homeless etc
445.
by segmondy 9mo ago
If you have to ask, then the answer is no.
446.
by segmondy 9mo ago
> The reason this matters is that LLMs are incredibly nifty often useful tools that are not AGI and also seem to be hitting a scaling wall I don't know who needs to hear this, but the real break through in AI that we have had is no
447.
by segmondy 9mo ago
without searching the internet. imagine you have no internet. do you have any computer with a modem? do you know anyone that does? can you get access to one? do you have dialup software? do you have a dial up international intern
448.
by segmondy 10mo ago
Yup, Anthropic has constant performance problems (not enough GPU), OpenAI is too messy with their politics and Altman.
449.
by segmondy 10mo ago
10 years worth of cash? So all these Chinese labs that came out and did it for less than $1 billion must have 3 heads per developer, right?
450.
by segmondy 10mo ago
no, there's no such agreeable thing. everyone has their own idea. but if i was to recommend such today, i would say, go on a self discovery method and find your idea books for algorithms/algorithm analysis & data structure,

More ›