Tootfinder

Sources: Nvidia plans to unveil a new AI inference chip at its GTC conference in March; the system will have a Groq-designed chip and OpenAI is a customer (Wall Street Journal)
https://www.wsj.com/tech/ai/nvidia-pla…

i implemented Ascon-Hash256 in Amaranth! (as well as laid the groundwork for the rest of Ascon family functions, including Ascon-AEAD128)
it comes at 1268 LUTs for iCE40 in the "compact" configuration, processing an average of 5.3 bit/cycle. a little expensive, but modern and cheaper than many better known alternatives

US Citizenship and Immigration Services (USCIS) said that it has stopped processing all immigration requests relating to Afghan nationals.
The Department of Homeland Security (DHS) said it was expanding that directive to include a review of all asylum cases approved under the Biden administration.
DHS did not clarify whether it is reviewing all asylum cases from only Afghanistan or from other countries, as well.
The USCIS director, Joseph Edlow, later added he was also dire…

from my link log —
Ordered dithering with arbitrary or irregular colour palettes.
https://matejlou.blog/2023/12/06/ordered-dithering-for-arbitrary-or-irregular-palettes/
saved 2026-02-27

Ordered Dithering with Arbitrary or Irregular Colour Palettes
Introduction Hello. I have recently been spending a lot of time dithering. In image processing, dithering is the act of applying intentional perturbations or noise in order to compensate for the in…

⏳Seems like forever since I pulled out the 90mm/2.8 lens that I used to shoot this African daisy last summer... and yeah, I will reboot my photo processing pipeline one of these days
#photo #photography #flowers

#LaVCa: LLM-assisted visual cortex captioning https://arxiv.org/abs/2502.13606 using "large language models (LLMs) to generate natural-language captions for images to which voxels are selective"; to be presented a…

LaVCa: LLM-assisted Visual Cortex Captioning
Understanding the property of neural populations (or voxels) in the human brain can advance our comprehension of human perceptual and cognitive processing capabilities and contribute to developing brain-inspired computer models. Recent encoding models using deep neural networks (DNNs) have successfully predicted voxel-wise activity. However, interpreting the properties that explain voxel responses remains challenging because of the black-box nature of DNNs. As a solution, we propose LLM-assiste…

Source: Meta has signed a multiyear deal to rent Google's TPUs to develop new models and has also been in talks to buy TPUs for its data centers as soon as 2027 (The Information)
https://www.theinformation.com/articles/go

Google Strikes Multibillion-Dollar AI Chip Deal With Meta, Sharpening Nvidia Rivalry
Meta Platforms has signed a deal to rentGoogle’s AI chips, known as tensor processing units, to develop new AI models, according to a person involved in the talks. The multi-year deal is worth billions of dollars, said a person who was briefed about it.Meta has also been talking to Google about ...

🥵 Reservoir thermal energy storage offers efficient cooling for data centers
https://techxplore.com/news/2025-12-reservoir-thermal-energy-storage-efficient.html

Reservoir thermal energy storage offers efficient cooling for data centers
The rise of artificial intelligence, cloud platforms, and data processing is driving a steady increase in global data center electricity consumption. While running computer servers accounts for the largest share of data center energy use, cooling systems come in second—but a new study by researchers at the National Laboratory of the Rockies (NLR), formerly known as NREL, offers a potential solution to reduce peak energy consumption.

🍳 Light yet strong material inspired by egg whites can cool devices faster
#materials

Light yet strong material inspired by egg whites can cool devices faster
A research team at the Korea Institute of Materials Science (KIMS) has developed a high-performance heat-dissipating composite material that achieves both eco-friendliness and low-cost processing.

Qwen releases Qwen3-Max-Thinking, its flagship reasoning model that it says demonstrates performance comparable to models such as GPT-5.2 Thinking and Opus 4.5 (Qwen)
https://qwen.ai/blog?id=qwen3-max-thinking