Ultrathin business card runs a fluid simulation
github.com
HorizonDB, a geocoding engine in Rust that replaces Elasticsearch
radar.com
Getting Good Results from Claude Code
dzombak.com
The Rise of Ritual Features: Why Platforms Are Adding Daily Puzzle Games
productpickle.online
Linear sent me down a local-first rabbit hole
bytemash.net
Food, housing, & health care costs are a source of major stress for many people
apnorc.org
Astronomy Photographer of the Year 2025 shortlist
rmg.co.uk
Show HN: Trayce – “Burp Suite for developers”
trayce.dev
Telefon Hírmondó: Listen to news and music electronically, in 1893
en.wikipedia.org
Show HN: Synchrotron, a real-time DSP engine in pure Python
synchrotron.thatother.dev
How Attention Sinks Keep Language Models Stable
hanlab.mit.edu
Flipper Zero dark web firmware bypasses rolling code security
rtl-sdr.com
Show HN: Aha Domain Search
ahadomainsearch.com
Complex Iterators Are Slow
caolan.uk
Exit Tax: Leave Germany before your business gets big
eidel.io
The BLS Can't Be Replaced by the Private Sector
bloomberg.com
FLUX.1-Krea and the Rise of Opinionated Models
dbreunig.com
GPT-5: Key characteristics, pricing and system card
simonwillison.net
OpenAI's new open-source model is basically Phi-5
seangoedecke.com
> Researchers had observed similar patterns in BERT, where "a surprisingly large amount of attention focuses on the delimiter token [SEP] and periods," which they argued was used by the model as a sort of no-op. The same summer at Meta, researchers studying vision transformers found similar behavior, observing that models would repurpose uninformative background patches as computational scratchpads.
This seems to go beyond just transformers. For example, I recall reading a paper a while ago that showed a similar effect in an image to image model with a GAN/U-Net architecture [1].
[1] https://arxiv.org/abs/1712.02950