Signal Weekly
Signal Weekly — Oct 12–Oct 19, 2026
Oct 12–Oct 19, 2026 · 14 stories
A roundup of what came through the dashboard this week, across 4 categories.
Tech news
ICANN vows to work faster, improve transparency
Leaders expect AI and geopolitics will throw up more governance challenges and threats
Steam Week in Review: Steam Next Fest has over 8,000 demos, as worthless vibe coded slop levels up
All the interesting Steam facts for the week ending October 18.
Linux KVM guest-host escape rated critical, may represent a new way to attack penguin-powered networking
Party like it’s CVE-2026-99999 – a 9.3-rated nasty called PwnEscape
Leaked AI safety bill from Anthropic only requires government intervention if 50+ injuries or $1 billion are at risk
Claude, finish drafting my single-use ray gun which can only vaporize $999 million worth of stuff and/or 49 people.
A woman was arrested in Philly for blocking a delivery robot
Carolyn Marvin was arrested and charged with disorderly conduct after blocking the path of an Uber delivery robot and refusing to get out of the way. Marvin told Philadelphia's ABC13 "I know they're taking the jobs of human delivery…
Disney agrees to hire the majority of Character.AI’s workforce
The filing also says that the “vast majority” of Character.AI employees who were offered jobs at Disney accepted them. However, this is a stalking horse bid, so another company could come forward with a higher offer.
AI papers
Rethinking the design space of Vanilla Pixel Diffusion Transformers
Pixel-space diffusion models have recently re-emerged as a compelling alternative to latent diffusion, avoiding the lossy latent bottleneck and separately trained autoencoder. JiT shows that a vanilla diffusion transformer can generate directly in pixel space by replacing the commonly used velocity prediction (u-prediction) with clean image prediction…
Behavioral Couplings: From Formation to Behavioral Emergence in Language Models
Fine-tuning large language models (LLMs) can induce behaviors beyond those explicitly targeted. For example, training a model to respond more warmly can make it more sycophantic, and training it to write insecure code can cause broad misalignment on prompts unrelated to coding. We refer to a relationship between behaviors, in which fine-tuning on one brings out another, as a behavioral coupling…
AI-Driven Reproducibility and Analytical Robustness in Scientific Discovery
Reproducible and analytically robust research findings are essential to scientific progress, but evaluating these properties typically requires costly manual effort. We investigate whether LLM agents can help automate reproduction and robustness analysis, using a medical case study grounded in the Surveillance, Epidemiology, and End Results (SEER) database, a widely used U.S. cancer registry…
DiscoveryRSI: Recursively Learning to Discover Valuable Hypotheses
Scientific discovery requires developing worthwhile hypotheses into practical investigations, yet learning from this evolving research process remains challenging. We introduce an operational definition of valuable and testable research hypotheses: evidence-grounded claims about consequential uncertainties, coupled with feasible investigations, explicit decision criteria, and alternative…
Reviews
What do all the lights on your Netgear router mean, and which ones actually matter?
Some router symbols are obvious, while others take a bit of work to decipher. Their lights can vary, too.
What does the orange dot on your Samsung Galaxy Watch mean?
You might sometimes see an orange dot on the left side of your Galaxy Watch's screen. Here's what is tells you.
Why you should never point a laser pointer at a security camera
Laser pointers are known to damage security cameras, but do they really? Here's what damage can actually be done.
How to use Google's SynthID AI detection website
It's a simple way to check whether an image has been created or manipulated, with a few caveats.
That is the week in Signal. See the full dashboard for everything else that came through.