Hacker News. Daily summary. Top 20 stories.
Anonymous GitHub account bikini publishes untouched 0-days, offering PoCs for public credit.
LLM-driven exploit hunting shifts open-source security debate; SSN exposure is systemic design flaw.
DeepSeek's DSpark speculative decoding speeds LLM inference in production.
DeepSeek's openness contrasted with US labs' secrecy and capital spending.
OpenRA releases a new playtest with random map generators for Red Alert, Tiberian Dawn, and Dune 2000.
Users debate AI difficulty and cheating, while praising OpenRA's balance and community improvements.
Meta keeps escalating its legal war against whistleblower Sarah Wynn-Williams, even after she fell silent onstage.
General condemnation of Zuckerberg, with one commenter disputing the claim about China access.
A handbook on software engineering patterns for systems where money is the primary focus.
Widespread endorsement of the handbook but a split on integer minor-units vs. string representation for money.
Digital purchases are often revocable licenses, not ownership; physical media offers lasting access.
Piracy dominates as the workaround, with pirated copies often better than purchased ones.
Lived experience matters more than AI's knowledge, using a Good Will Hunting monologue to argue human uniqueness.
Two camps: those who dismiss AI's impact and those who cite real harm like therapy and coding reliance.
A website catalogs 13,911 open webcams accessible on the public internet.
Split between condemning the site as invasion of privacy and defending it as exposing user negligence.
Marfa Public Radio's fall membership drive podcast reads boring work documents to lull you to sleep.
Most find the concept clever; some recommend similar podcasts they prefer.
California makes loud streaming ads illegal starting July 1, extending CALM Act rules.
Commenters broadly welcome the law, with debate over technical feasibility of volume normalization.
Open-source Town Square adds anonymous stick-figure visitors to your website to create a sense of co-presence.
The crowd is split on whether anonymity or persistent handles better captures the old-web feeling.
Asian AI startups launch models comparable to Anthropic's banned Mythos.
Split: ban backfires on U.S. AI leadership vs. models unproven as Mythos rivals.
Sharp policy thresholds create perverse incentives, like losing money to keep healthcare subsidies.
Broad consensus that benefit cliffs are harmful, with many citing UK, Slovenian, and other examples.
OpenTTD 16.0-beta1 adds backwards-driving trains, improved map generation, and a title game competition.
Commenters appreciate OpenTTD's longevity, while some find signaling confusing and wish for easier NewGRF setup.
Ford rehired hundreds of veteran engineers after its AI automation strategy backfired.
Commenters predict more such stories, noting short-term executive incentives and AI's lack of nuanced judgment.
An interactive guide compares 29 public DNS resolvers on privacy, filtering, speed, and jurisdiction.
Quad9 praised but criticized for false positives; some run self-hosted DNS.
Former Meta director sues, says company surveilled her for a year to enforce gag order.
Many commenters say Meta's legal tactics are driving people to buy the book.
A guide for setting up a two-node AMD Strix Halo cluster with RDMA for vLLM inference.
Enthusiasm for Strix Halo clusters tempered by steep price increases and comparisons to alternatives.
Michigan spent $1.8 billion in incentives but created only 602 jobs.
Outrage over $2.5M-per-job cost; many call it corruption, not job creation.
GLP-1 drugs like Ozempic may boost mood by altering gut microbes.
Many users report mood and anti-inflammatory benefits, not just weight loss.