functional_dev

5 ms·

31.
by functional_dev 5mo ago
The article shows WASM being 1.2-3.7x slower, and your experience confirms it. Do you have any idea which operations regress the most?
32.
by functional_dev 5mo ago
Chain of thought happens in the <think> tags, not the visible output. Caveman only strips filler from what you see... the reasoning depth stays the same. I found this visualisation pretty interesting - https://vectree.io&#x
33.
by functional_dev 5mo ago
Great find! It appears uppercase tokens are completely unknonw to the tokenizer. But the character still comes through in response :)
34.
by functional_dev 6mo ago
The wiki diagram helped me too, thanks! One thing I did not understand before: why SNAT must happen at POSTROUTING? Because the exit interface is only known after the routing decision... before that kernel does not know which source IP to w
35.
by functional_dev 6mo ago
Good point, full self driving probably felt like marketing win back then... but now it is a liability. Rebranding now would just highlight all missed promises
36.
by functional_dev 6mo ago
Looked at their source code out of curiosity. They use Elasticsearch as a geo backend, tile server that renders PNG images server side for each map tile request. - low zoom -> server aggregated grid - high zoom -> switcher to point ti
37.
by functional_dev 6mo ago
yes, this part I find to be most interesting.. how many losses before the picture changes for the public?
38.
by functional_dev 6mo ago
maybe it keeps getting rediscovered, becaues a good tooling around it is still missing?
39.
by functional_dev 6mo ago
so the atmosphere acts as giant lamp lit from behind by Moon? never thought of it that way
40.
by functional_dev 6mo ago
exactly, embeddings destroy information.. exact keywords, acronyms, etc. They get squashed into floats. That is why grep still beats it for code. I generated visual schematic of every stage of the pipeline - https://vectree.io&
41.
by functional_dev 6mo ago
I didn't know how old this actually is. I was surprised to learn Marcus Aurelius wrote his Meditations as a private journal... never meant to be published. found this if anyone wants to explore: https://vectree.io/c&#x2
42.
by functional_dev 6mo ago
had the same question and after reading about it, I found there are multiple layers on each other. Existing plans are built to run 40-60 years.. retiring creates "stranded assets". pension funds fight hard to avoid that. The renew
43.
by functional_dev 6mo ago
That would be great
44.
by functional_dev 6mo ago
Isn't the point of mirrors to concentrate heat for mining, not to get more total power?
45.
by functional_dev 6mo ago
30 years and nobody (pharma, govts, etc.) have not found it worth to study this properly. If it worked, would not at least some researched pick this up?
46.
by functional_dev 6mo ago
yeah, encrypting the bits before embedding seems like a valid first step... even if someone suspects steganography, they still can't read it
47.
by functional_dev 6mo ago
this made me look into how cloud hypervisors actually work on HW level.. they all offload it to custom HW (smart nic, fpga, dpu, etc..). cpu does almost nothing except for tenant work. AWS -> Nitro, Azure -> FPGA, NVIDIA sells DPUs. H
48.
by functional_dev 6mo ago
what surprised me is how Proton works under the hood... no emulation at all! wine translates win API -> Linux. Then DXVK converts DirectX calls into Vulkan in real time, and VKD3D-Proton for DX 12. so it always native Vulkan.. no wonder
49.
by functional_dev 6mo ago
you are right.. separating this 2 questions is big safety advantage. exactly like in software we separate logic from security (like AWS IAM)
50.
by functional_dev 6mo ago
agreed, attackers can use these AI tools to scan open source code and find bugs very fast... if project maintainers do not have access to such tools, it because an ufair fight
51.
by functional_dev 6mo ago
yeah, it appears to support audio and image input.. and runs on mobile devices with 256K context window!
52.
by functional_dev 6mo ago
I was wondering about this too! I did not know how they can aim a laser from so far at a moving spaceship. I generated this visual map about to help me understand it - https://vectree.io/c/aiming-space-lasers-gimbals-an
53.
by functional_dev 6mo ago
your point about gas pockets making the shield blow out made me curious I found this visual schematic of the spalling helpful - https://vectree.io/c/the-physics-of-ablative-spalling-in-ori... It finally makes sense why
54.
by functional_dev 6mo ago
the list is funny :) So it counts how many times I was angry?
55.
by functional_dev 6mo ago
is it that bad? maybe that is a secret for a long life. I want a job that never disappears :)
56.
by functional_dev 6mo ago
exactly, makes me think... if person who makes the bread does not feed his own family, something is wrong
57.
by functional_dev 6mo ago
I did not know, that NVFP4 was handled at the silicon level... until I dug deeper here - https://vectree.io/c/llm-quantization-from-weights-to-bits-g...
58.
by functional_dev 6mo ago
Agreed, Spring Framework is quite simple... just look here - https://vectree.io/c/spring-framework
59.
by functional_dev 6mo ago
BTW magpies have passed the mirror test
60.
by functional_dev 6mo ago
valid point, it reminds me of video games. GPUs got faster, devs pushed higher resolutions, more complex lighting instead of saving power :)

More ›