eating tide pods and doing nothing useful all day long

> USB4 host interface device. See Links to ACPI-Related Documents under the heading “USB4 Host Interface Specification”

where's that document?

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

remote fTPM service to pass Epic Games attestation - /dev/tpm0 over the internet

Release day RTX Spark N1x windows driver looks a bit rushed

NVDLA firmware inside was compiled on a dev machine and presumably pushed to the repo

/home/akumaranand/l4t-dev-main-generic_int-debug-eng-20251125T095455/cv/dla/firmware-t35x/engine_sched/scheduler.c

noooooooooooooo

ugh. Chase sent me this today...

> the idea that it's unethical to distill from AI models, but totally ethical to train on copyrighted source material

Best plan would probably be a USB-SPI bridge (something like the MCP2210) with the standard TPM 2.0 SPI header and let people supply their own TPMs in order to make it hard to simply block them

Epic's remote-attestation plan for Linux is actually a cunning plot to produce a market for USB TPM dongles that can be reset at runtime and programmed with arbitrary PCR values

> in-kernel implementation of Vulkan built from scratch

CUDA 14.0 will move to Armv8.2-A as the minimum supported architecture for ARM64-SBSA.

it'll get worse don't worry 🎵

https://www.telegraph.co.uk/news/2026/10/13/suicide-rates-highest-ever-among-middle-aged-men/

a lot more likely to be just an interchangeable cog in a wheel with no purpose, and death is better than that outcome as a leech on society with nothing of value to add whatsoever

LLMs as a cherry on top of this story, enjoy!

qualcomm ripple

https://github.com/qualcomm/ripple/blob/c9b3ccf800eb9282c89964baa68faa3ca9a1eebb/llvm/docs/RippleRef.rst

> Ripple is an IR-level mechanism that lets frontends describe block shapes (logical SIMD/tensor extents per processing element) and lets the backend (through the Ripple pass) specialize and vectorize scalar code based on those shapes.

From Chapter 11 filing

> Character serves approximately 12 million active monthly users located around the world and, while usage varies widely, on average, active users spend approximately 120 minutes a day on Character.Al

the godly blessing of migraines

Hexagon v85 adds non-LL/SC atomics that perform a word or double-word read-modify-write atomically

*However*

V85 ERRATA: Atomicity may be lost if a competing instruction is not an atomic. To guarantee atomicity, only atomic instructions should be used for the memory location

Hexagon v83 adds hardware CFI

Hexagon Element Accelerator User Manual
formally the
Hexagon V85 HLX Programmer’s Reference Manual

https://docs.qualcomm.com/doc/80-N3040-23/topic/index_80-n2040-68_v85_hlx_prm.html

the neural accelerator will help AI give you a warm hug

This is disheartening to read. I wasn't aware the OpenAI math paper would be so impactful...

https://proofsandprompts.com/2026/10/08/100-reactions-to-100-solutions/