Auto-download daily
Issue · June 28, 2026

Top of Hacker News

24 stories of the day: AI, technology, business and simply interesting discussions — recapped and unpacked.

GitHub Trending: Today — Plane, Directus, Svelte, MONAI, Summer2026-Internships (github.com)

In today's GitHub Trending: Plane (a self-hosted Jira and Linear alternative for teams allergic to per-seat pricing), Directus (turns any SQL database into an instant REST and GraphQL API), Svelte (the compiler-first web framework), and MONAI (a PyTorch toolkit for medical imaging).

The loudest entry isn't code: SimplifyJobs' internship board, a job listing that lives entirely inside a Git repo and gets edited by opening an issue. And two near-identical "replace your SaaS" tools draw opposite verdicts from the people who run them.

Not just a headline — the discussion unpacked

The top HN threads: facts from the discussions, both sides, insider voices. Email, Telegram, or RSS.

THE FINALS снова лезет вверх в чартах, но отзывы полны усталости (store.steampowered.com)

THE FINALS — бесплатный шутер от Embark Studios про полностью разрушаемые арены: стены, полы и целые здания рушатся прямо по ходу боя. На неделе игра прибавила в чартах и получила обновление магазина 10.13.0.

Парадокс в реакции: вместо радости от косметики в отзывах копится раздражение из-за читеров и подбора соперников.

An anonymous account dumped a pile of "0-days" on GitHub. The experts who read them shrugged (github.com)

An anonymous repo dropped dozens of exploits at once, neatly documented, daring readers to report them for credit. The framing — a flood of fresh 0-days — set off the usual alarm.

Then people who reverse-engineer for a living read the code: most aren't 0-days at all, several aren't even vulnerabilities. One reviewer found a "7-Zip exploit" that only proves the parser is reachable. The unsettling part isn't the bugs — it's who, or what, is generating them at scale.

DeepSeek says its decoding trick makes generation 57–78% faster. The thread asks who's left to sell tokens to (github.com)

DeepSeek dropped a paper on DSpark, a speculative-decoding module bolted onto its V4 models, claiming 57–78% faster per-user generation at the same hardware capacity — and shipped the weights to Hugging Face the same day.

The thread's sharpest voice wasn't about the math. It was the warning that token-sellers eyeing an IPO have a market that's 'almost solely developers', open models a step behind, and no customer boasting higher profits from AI. Also noted: the headline isn't even the paper's title, and speculative decoding dates to 2022.

Meta's arbitrator fined a whistleblower $50,000 per criticism. She responded by sitting silent on stage for an hour (pluralistic.net)

After her memoir, a Meta-paid arbitrator ordered Wynn-Williams to stop promoting it at $50,000 per criticism — over $11 million, more than she and her husband have earned in their lives. So at a UK festival she sat on stage for an hour and said nothing.

Cory Doctorow frames it as a Streisand-effect self-own. The thread's skeptics aren't defending Meta so much as noting his essay never once mentions Meta's side — that she signed an NDA, took severance, and that he overstates the China claim.

Headlines said 'Asian startups clone Mythos.' One is Japanese, Khosla-backed, and the 'clone' claim is unproven (techcrunch.com)

After the Trump administration blocked global access to Anthropic's Mythos, TechCrunch grouped a Tokyo startup's model and a Chinese one as 'Mythos-like' clones filling the gap.

A CTO in the thread pulled it apart: 'Asian' lumps a Khosla-backed Japanese ally in with China, and the 'distilled from Anthropic' line is unproven — these are different multi-model architectures. A skeptic added the obvious tell: you don't ship a frontier model out of nowhere with nothing released before it.

🚀Tech ★ 158 · 💬 137 · Jun 28 · #AI#policy#startups#360#Anthropic#Fable 5+2

A Kindle owner reopened a novel he'd bought. It had quietly become the German translation (dervis.de)

An essay argues physical media is the only copy that 'remains under the buyer's control' — citing Sony pulling purchased films, Disney+ erasing 50+ titles, and 87% of pre-2010 games now unavailable.

The thread's killer detail is smaller: a paid Kindle novel silently turned into its German translation on redownload. The contrarians don't disagree about ownership — they just point past it. The only copy nobody can revoke or rewrite, several conclude, is the pirated one.

A fintech handbook hit the front page. Then the author showed up — and so did engineers who'd built the real thing (w.pitula.me)

A solo-posted Fintech Engineering Handbook — money as integers, ledgers, idempotency, webhooks — climbed the front page, and the author turned up in the replies, baffled by his own reach.

What followed was a working seminar. Ex-fintech CTOs validated the core; others flagged the 'minor-units' integer trick as a trap with stablecoins, called event-sourcing-everywhere overkill, and argued the PII-deletion advice would get a compliance team a very bad day with regulators.

How many elementary particles are there? Quanta says 17 to 995.5. A physicist in the thread says the question is broken (quantamagazine.org)

Quanta lines up the counts: 17 in the textbook, 61 once you split quark colors and boson handedness, 995.5 if you count quantum degrees of freedom. David Tong's verdict: 'the true answer is not an integer.'

A physicist in the thread says the article breaks where it counts '8 gluons' as if they were discrete like the W and Z. Gluon color is a continuous space; the basis is arbitrary. Particles are just excitations of fields — count the fields, and the number collapses.

A site turned the internet's exposed webcams into a browsable live atlas. Half of them are bedrooms (ipcrawl.com)

IP Crawl collects webcams left open to the public internet and presents them as a living atlas you can browse — pools, shops, churches, and, unnervingly often, bedrooms.

An NVIDIA engineer said he couldn't open it in good faith and it should come down. Others noted the problem is a decade old — 'if you think nobody would connect that to the internet, at least 1000 people did' — and that the useful version of this site would warn the owners, not the voyeurs.

🚀Tech ★ 252 · 💬 128 · Jun 28 · #privacy#security#Insecam#IP Crawl#Shodan

Meta sent staff to photograph a former exec at a literary festival. Now there's a lawsuit and a docket to read (fortune.com)

Wynn-Williams' new federal suit alleges Meta surveilled her for over a year, sending people to her public appearances to photograph her, while seeking $50,000 per claimed breach of a non-disparagement deal.

Meta says she's using the courts to sell books. The thread's skeptics agree she signed an NDA and took severance a decade ago. The counter is harder to dismiss: a company that pays staff to photograph a critic who says nothing is doing exactly what she accuses it of.

California makes obnoxiously loud streaming ads illegal on July 1. The industry's excuse got demolished in the thread (arstechnica.com)

A federal law quieted loud broadcast-TV ads back in 2012, but it never touched streaming — so streaming ads stayed loud. California's new law extends the same standard to streamers on July 1.

Industry groups argued that server-side ad insertion and a 'broad range of devices' make consistent loudness hard. The thread's top reply took that apart: loudness is a measured, solved problem at the source. 'Stop trying and fix it. It's your own damn system.'

🌍Other ★ 254 · 💬 79 · Jun 28 · #policy#web#California#CALM Act#Instagram+1

Michigan spent $1.8B chasing 20,595 jobs. It got 602 — about $3 million each (reason.com)

Eight marquee Michigan subsidy deals promised 20,595 jobs for $2.7B in incentives. Actual spend: $1.8B. Actual jobs: 602 — roughly $3 million apiece. Ford's $210M deal has created zero so far; a $261M SanDisk site is 'a big empty field.'

A Xoogler argued the headline cheats: tax abatements aren't 'spending,' since no business means no revenue either way. A VC fired back that this case was actual cash transfers — which makes it worse, not better.

Lennar's CEO said it 'value-engineers every component.' Buyers say that's a polite word for the leaks (hntrbrk.com)

An investigation pins shoddy new-build quality on Lennar and D.R. Horton 'value-engineering' to hit cost targets, then offloading the defects via warranties and arbitration. The CEOs said as much to investors.

A reader flagged the catch printed on the site itself: the publisher makes money from litigation and a short fund, so read it as positioned PR. A builder's relative went further — any tract builder is this bad; it's the incentives, not the logo.

A volunteer remake of Red Alert didn't just preserve the game. It quietly out-balanced the original (openra.net)

OpenRA rebuilds Red Alert, the original Command & Conquer and Dune 2000 with modern RTS features — attack-move, veterancy, fog of war, online play — on Windows, Mac and Linux.

A 25-year veteran's verdict cuts deeper than nostalgia: it's better balanced than the originals. The one wall remains — you still need the original game's assets — but the thread mostly turned into love letters, including one about a 50-something farm neighbor who wrecked everyone at Red Alert 2 over dial-up.

Why marathon finish times spike at 3:30 and 4:00, and what that says about every threshold you've trusted (danluu.com)

Dan Luu's 2020 essay maps the spikes that appear wherever a hard threshold sits: 9.8M marathon finishes clump at round times, Polish exams bunch at the pass mark, drug charges pile up exactly at the sentencing cutoff.

The thread adds fresh cases — chess ratings, AWS gaming its own P90 latency target — and one self-correction: a commenter called the lead example bogus, realized it predated a policy change, and withdrew it. The marathon spikes turn out to be social, not just vanity: pace-runners with flags.

🌍Other ★ 216 · 💬 65 · Jun 28 · #data#economy#science#AWS#Dan Luu#Lichess

A novelist explains why he still writes books in a 1978 word processor. The thread says it's not nostalgia — it's defense (sfwriter.com)

Sci-fi novelist Robert J. Sawyer keeps writing in WordStar because its commands sit under a touch typist's fingers and it never interrupts — 'I'm in control, not the program.' George R.R. Martin and Anne Rice agreed; Rice called Word 'pure madness' beside it.

The thread reframes the appeal as distraction-defense, not nostalgia. One writer switched to a typewriter to stop staring at a screen. Another nailed it: the longing isn't for the software, it's for the quiet.

A microbiome researcher thinks Ozempic lifts mood by getting your gut bacteria to make cannabis-like compounds (psychologytoday.com)

A microbiome writer lays out a mechanism for why GLP-1 drugs lift mood: they grow a gut bacterium that makes cannabis-like endocannabinoids. In mice, removing the microbes erased the effect; a fecal transplant restored it.

Then the humans answered. One reports the opposite — anhedonia, a known side effect. Another lost insurance and now pays $450 a month, the returning weight 'making me more depressed.' Mice, as one reply noted, are not people.

🌍Other ★ 91 · 💬 196 · Jun 28 · #health#science#GLP-1#Ozempic#Zepbound

'Arrested for going 5 seconds over' went viral. The bodycam shows a longer story — and a worse one underneath (gadgetreview.com)

A man was arrested at an Oklahoma council meeting on a proposed data center, the headline saying he ran five seconds past a three-minute limit. Newly released bodycam captured an officer's 'Arrest him.'

Commenters who watched it push back: he stood at the dais shuffling papers for nearly a minute and refused repeated orders to leave. But the buried facts hold up — water and power decided before residents spoke, and the city first quoted $1,750 for the footage.

A 'landmark' study says screens damage babies. Its own authors say it found no such causal link (theguardian.com)

A 'landmark' review by an advocacy team across four UK universities is reported as proof that screens harm babies' development.

The thread's top comment puts the headline next to the study's own words: it 'did not establish causal links.' Critics call it advocacy dressed as research. The one signal that survives isn't the headline — it's a separate finding that babies under a parent's phone show the same stress as the 'still-face' experiment.

Vendors say Mythos changes cybersecurity forever. A practitioner says the real change is who Washington lets use it (cephalosec.com)

A security blogger's 'keep calm and carry on' post pushes back on the Mythos hysteria — most breaches are still misconfigurations and bad practice, and vendors hawking 'Mythos solutions' the day it was announced gave the game away.

A higher-karma reply moves the target: the danger isn't the model's marginal edge, it's Washington choosing who gets it, creating defenders who can break everyone else's systems. The release-ban-IPO timeline, one reader notes, reads like a stock pump.

A Show HN keeps your Mac awake while AI agents work. The thread: that's a built-in command and a 15k-line reinvention (github.com)

Adrafinil, a Show HN named after a modafinil derivative, keeps a lid-closed Mac awake only while AI agents are working, then sleeps. Useful — until the thread points at the wheel it reinvents.

macOS ships 'caffeinate -dimsu.' Amphetamine has done conditional triggers for years. One commenter compressed the whole thing to a three-line shell loop; another asked why a sleep toggle needs 15,000 lines of code; a third noted the coding agent itself just shipped the feature.

🚀Tech ★ 93 · 💬 56 · Jun 28 · #AI#dev tools#Adrafinil#Amphetamine#Codex+1

A widget shows you the strangers reading the same page — no accounts, no history. The old web fans loved it; the veterans winced (cauenapier.com)

Town Square is a tiny widget that shows other live visitors as stick figures on your site, lets them chat in real time, and keeps nothing — no accounts, no profiles, no history. The goal: make the web feel like there are people on the other side of the screen again.

The indie-web crowd adored it. Veterans pushed back on the two things it waves away: moderation (one recalled PlayStation Home, accosted within minutes) and whether ephemerality is even what made the old web human — the permanence was.

A writer answered AI slop with Robin Williams' bench speech. The thread noted film's best 'I've lived' monologue was written for a machine (jayacunzo.com)

A marketer answers AI slop with Robin Williams' Good Will Hunting monologue — knowing about war isn't living it. His line: 'AI has read the internet. It can't read the room. It hasn't lived a life.'

The thread agreed, then complicated it: cinema's greatest lived-experience speech, Blade Runner's 'tears in rain,' was written for an artificial being. And the bench scene itself is a script, performed — the authenticity it sells was manufactured by a writer and an actor.