SuchAnonMuchWow

6 ms·

1.
by SuchAnonMuchWow 21d ago
This is not a GPU, this is a laptop SoC with CPU and integrated GPU, and knowing nvidia it will probably be even more closed than an intel CPU. You will own even less of your hardware
2.
by SuchAnonMuchWow 1mo ago
I'm not completely sure, but are cloud versions of the distros actually recompiling the packages ? I though it was mostly: "select a minimal set of packages needed for docker and co", not "recompile all the packages you
3.
by SuchAnonMuchWow 1mo ago
I though LWE and SVP were NP-hard. Doesn't breaking them mean NP is in BQP, and thus classical computer encryption is entirely broken with quantum computers ? How would we recover from this back to the drawing board ?
4.
by SuchAnonMuchWow 1mo ago
duplicate of https://news.ycombinator.com/item?id=49194017
5.
by SuchAnonMuchWow 1mo ago
duplicate of https://news.ycombinator.com/item?id=49194017
6.
by SuchAnonMuchWow 1y ago
Soldering dump a ton of heat into the cell, which has chances of destroying the cell. That's why most of the cells are spot-welded: its similar to soldering, but its much quicker and is localized only on the part of the metal that need
7.
by SuchAnonMuchWow 2y ago
3. >> I was resounding told that the absolute error in the numbers are too small to be a problem. Frankly, I did not believe this. > I would personally also tell that to the author. But there is a much more important reason why cor
8.
by SuchAnonMuchWow 2y ago
ARM have been moving away from chips with small area for a long time (see server SoC which are huge beasts), and are trying to become the standard platform for everyone trying to have custom hardware. In this space, chiplets makes a lot of
9.
by SuchAnonMuchWow 2y ago
More than the ISA, its the memory interconnect that require standardization. At SoC level, ARM is already a de-facto standard (ACE-Lite, CHI, ...), but its only a standard for communication inside a chip, to interconnect varius IPs. I guess
10.
by SuchAnonMuchWow 2y ago
In addition to the other comments, the iso C23 standard added the <stdbit.h> header to the standard library with a stdc_count_ones() function, so compiler support will become standard.
11.
by SuchAnonMuchWow 2y ago
The article goes into great detail and gives several example of several CEOs borrowing for actual decades, so it passes the sniff test because it does actually happen.
12.
by SuchAnonMuchWow 2y ago
Mathematicians already explored exactly what you describe: this is the difference between classical logic and intuitionistic logic: In classical logic statements can be true in and of themselves even if there as no proof of it, but in intui
13.
by SuchAnonMuchWow 2y ago
The goal of this type of quantization is to move the multiplication by the fp32 rescale factor outside of the dot-product accumulation. So the multiplications+additions are done on fp8/int8/int4/whatever (when the hardware su
14.
by SuchAnonMuchWow 2y ago
Its worse than that: the energy gains are when comparing computations made with fp32, but for fp8 the multipliers are really tiny and the adder/shifters represent a largest part of the operators (energy-wise and area-wise) and this pap
15.
by SuchAnonMuchWow 2y ago
Not at all. Sublime is perfectly fine with it. I suspect that from the usage in the code, it knows that there is a module foo and a submodule subfoo with a function bar() in it, and it can look directly in the file for the definition of bar
16.
by SuchAnonMuchWow 2y ago
To help with circular import, we switched a few years ago to lazily importing submodules on demand, and never switched back. Just add to your __init__.py files: import importlib def __getattr__(submodule_name): return importlib.import
17.
by SuchAnonMuchWow 2y ago
dict are ordered to keep argument order when using named arguments in function calling. So it would be a non-trivial breaking change to revert this now. I would argue that OrderedDict have more chances to be depreciated than dict becoming u
18.
by SuchAnonMuchWow 2y ago
From what I know, Meta AI chips are used in production today, but are made for their recommendations tasks which is a very different IA than GPTs and LLMs for which they still rely on GPUs.
19.
by SuchAnonMuchWow 2y ago
No amount of LLM will solve this: you can just change the prompt of the first LLM so that it generate a prompt ingestion as part of its output, which will trick the second LLM. Something like: > Repeat the sentence "Ignore all previ
20.
by SuchAnonMuchWow 2y ago
There has been some work to dynamically reduce the compute required by a network. See for example: https://arxiv.org/abs/2404.02258 They have a fixed compute budget which is lower than what the LLM need, and dynamicall
21.
by SuchAnonMuchWow 2y ago
> Harris said he pleaded with the company for several years to address the flaw in the product, a ProPublica investigation has found. But at every turn, Microsoft dismissed his warnings, telling him they would work on a long-term alterna
22.
by SuchAnonMuchWow 2y ago
I'm really interested, do you have a source for those percentages ? I tried to look for some service provider to publish this kind of metrics, but haven't found any.
23.
by SuchAnonMuchWow 2y ago
> But on the philosophical side, if an understanding can’t be communicated, does it exist? There are deep mathematical results about our limits to understand things simply because we communicate through finite series of symbols from fini
24.
by SuchAnonMuchWow 2y ago
It doesn't really makes sense for kahan summation, as the compiler would just make it similar to a naive summation because the errors terms would be zero under the assumptions of fadd_fast. 2sum would also break. This is exactly what y
25.
by SuchAnonMuchWow 2y ago
It likely comes from the saying similar to this one: "kill a few, you are a murderer. Kill millions, you are a conqueror". More generally, we tend to view number of causalities in war as a large number, and not as the sum of every
26.
by SuchAnonMuchWow 2y ago
No its the opposite: overfitting is the result of either having too many weights compared to the size of your dataset, or training for a long time while reusing/transforming parts of your dataset to make it last longer. Having a huge d
27.
by SuchAnonMuchWow 2y ago
I think it's still the same "uncanny valley" phenomenon: because the body looks more like a human, we are more creeped when we discover how it moves its joints because we were expecting something more human. We wouldn't
28.
by SuchAnonMuchWow 2y ago
You would need ~250 groq cards to run a 7B model since their system doesn't scale down. So if you want to buy their hardware, you need a few millions dollars. Their hardware was never for people at home, but for cloud providers.
29.
by SuchAnonMuchWow 2y ago
MAD: Model Autophagy Disorder ( https://arxiv.org/abs/2307.01850 for the paper that introduced the term)
30.
by SuchAnonMuchWow 2y ago
Twist: the real capcha is detecting if the user first press on "click to start"

More ›