lambda

13 ms·

121.
by lambda 6mo ago
You're going to need to provide a source for that outrageous claim. Also, which sites that are impacted by the age verification laws are involved in grooming in any way? Please be specific.
122.
by lambda 6mo ago
You do have to set the kernel parameters once to set the max GPU allocation, I have it set to 110 GiB, and you have to set a BIOS setting to set the minimum GPU allocation, I have it set to 512 MiB. Once you've set those up, it's
123.
by lambda 6mo ago
> Only Apple has the unique dynamic allocation though. What do you mean? On Linux I can dynamically allocate memory between CPU and GPU. Just have to set a few kernel parameters to set the max allowable allocation to the GPU, and set the
124.
by lambda 6mo ago
I've got a 128 GiB unified memory Ryzen Ai Max+ 395 (aka Strix Halo) laptop. Trying to run LLM models somehow makes 128 GiB of memory feel incredibly tight. I'm frequently getting OOMs when I'm running models that are pushing
125.
by lambda 6mo ago
So, this is all true, but this calculation isn't that nuanced. It's trying to get you into a ballpark range, and based on my usage on my real hardware (if I put in my specs, since it's not in their hardware list), the results
126.
by lambda 6mo ago
It discusses it, and they have data showing that they know the number of active parameters on an MoE model, but they don't seem to use that in their calculation. It gives me answers far lower than my real-world usage on my setup; its c
127.
by lambda 6mo ago
Yeah, I looked up some models I have actually run locally on my Strix Halo laptop, and its saying I should have much lower performance than I actually have on models I've tested. For MoE models, it should be using the active parameters
128.
by lambda 6mo ago
It may be prettier looking. I've seen many Git GUIs that are prettier than Magit. But none of them that I've tried have ever come close to the workflow. I can stage and unstage individual hunks, do complex interactive rebases, squ
129.
by lambda 6mo ago
Magit is absolutely the best Git GUI ever. Unfortunately, for most people the fact that it's part of Emacs is a blocker. And because most people use worse Git tools, they tend to use workflows that are easier with more cumbersome tools
130.
by lambda 7mo ago
Yeah. I've been finding a scary number of people saying that they never write code by hand any more, and I'm having a hard time seeing how they can keep in practice enough to properly supervise. Sure, for a few weeks it will be OK
131.
by lambda 7mo ago
Please relearn to use your brain. I cannot imagine how you can properly supervise an LLM agent if you can't effectively do the work yourself, maybe slightly slower. If the agent is going a significant amount faster than you could do it
132.
by lambda 7mo ago
I mean, there's also the whole GitHub free tier. It used to only be for public repos, so mostly OSS plus "shared source", but now they allow it for private as well. But it still costs them money to host your code and provide
133.
by lambda 7mo ago
What do you mean it's unfit for training? It's a form of reinforcement learning; the end result has been selected based on whether it actually solved the need. You need to be careful of the amount of reinforcement learning vs cont
134.
by lambda 7mo ago
Oh, I do frequently have biscuits and gravy topped with fried eggs, though the biscuits and gravy would definitely pull it further towards the flour end of the spectrum, maybe not quite in the dark region. Also, hollandaise is pretty integr
135.
by lambda 7mo ago
Are we counting butter as milk? Because hollandaise sauce is an emulsion of egg yolk, butter, and lemon juice, so no milk directly unless we count butter.
136.
by lambda 7mo ago
Qwen3 Coder Next in llama.cpp on my own machine. I'm an AI hater, but I need to experiment with it occasionally, I'm not going to pay someone rent for something they trained on my own GitHub, Stack overflow, and Reddit posts.
137.
by lambda 7mo ago
https://bitinfocharts.com/comparison/bitcoin-transactionfees...
138.
by lambda 7mo ago
They've since added some hacks to enable it to handle more transactions and bring the price down. Effectively, the network had hit its limit on the number of transactions it could fit in a block, so you had to pay high fees to get acce
139.
by lambda 7mo ago
What? I was replying to someone who explicitly referenced the Bitcoin whitepaper, they were clearly talking about BTC. And the protocol from the whitepaper was actually pretty bad, from a cost and transaction time point of view. It's g
140.
by lambda 7mo ago
What? "Instantly and for a very low fee"? Fees have historically gone up above $100 per transaction. They've since added hacks on top of the original Bitcoin protocol to get the price back down again, but the original design
141.
by lambda 7mo ago
The thing is, in our current legal system, property rights are fairly fundamental; they own certain things, and that gives them a legal right to control it. And the money they derive from that has become more and more influential in our pol
142.
by lambda 7mo ago
But part of the faith is faith that God can communicate through imperfect mortal vessels. > Moses said to the Lord, “Pardon your servant, Lord. I have never been eloquent, neither in the past nor since you have spoken to your servant. I
143.
by lambda 7mo ago
There is not enough arable land in the US for every person in the US to be able to live on subsistence farming. Subsistence farming is a lot more work, and much less efficient, than large scale industrialized agriculture.
144.
by lambda 7mo ago
Yeah, wow. Definitely setting off my AI summary alarm.
145.
by lambda 7mo ago
After some more testing, yikes, MiniMax M2.5 can get painfully slow on this setup. Haven't tried different things like switching between Vulkan and ROCm yet. But anyhow, that 17 tokens per second was on almost empty context. By the tim
146.
by lambda 7mo ago
A piece of software that you write, in code, unless you use random numbers or multiple threads without synchronization, will operate in a deterministic way. You know that for a given input, you'll get a given output; and you can reason
147.
by lambda 7mo ago
Sorry, I shouldn't have been that harsh about it; I apologize. Doesn't look like I can delet the comment, though. Anyhow, I'm just a bit bitter about the onslaught of obviously vibe coded projects that very little effort has
148.
by lambda 7mo ago
OK, with MiniMax M2.5 UD-Q3_K_XL (101 GiB), I can't really seem to fit the full context in even at smaller quants. Going up much above 64k tokens, I start to get OOM errors when running Firefox and Zed alongside the model, or just fail
149.
by lambda 7mo ago
The example you gave is just so buggy; why do you think that this is worth sharing with the world yet? The very first question is full of obvious bugs. You have 'find . -name "notes.txt"' selected, it then says 'You
150.
by lambda 7mo ago
Huh? Once it gets to the model, it's all just tokens, and those are just in band signalling. A model just takes in a pile of tokens, and spits out some more, and it doesn't have any kind of "color" for user instructions

More ›