Jul 2Friday, July 3, 2026 · all days
1.Virginia bans sale of geolocation data(hunton.com)
841 points by toomuchtodo 16 hours ago | 130 comments | permalink
tl;dr: Virginia Governor Spanberger signed S.B. 388, amending the VCDPA to ban the sale of geolocation data effective July 1, 2026. However, Virginia's ban is narrower than similar laws in Maryland and Oregon because the VCDPA defines "sale" as only monetary exchanges, excluding transfers for other valuable consideration. Similar bans are pending in California, Massachusetts, Vermont, and Washington.
HN Discussion:
  • Law has loopholes allowing fuzzy or anonymized geolocation data to still be sold
  • ~Concerns about enforcement scope and jurisdiction across state/company lines
  • Ban should be broader, covering all sharing not just monetary sale
  • Supports the law as bare minimum but wants stronger punishments for violators
  • Neutral questions or added context about the geolocation data market
2.CarPlay Is Additive(caseyliss.com)
372 points by sprawl_ 12 hours ago | 492 comments | permalink
tl;dr: Rivian's CSO defended the lack of CarPlay support by claiming screen mirroring "takes over every single pixel," but the author refutes this by noting standard CarPlay only uses part of the screen (as seen in Volvos) and is entirely optional for users. The author argues CarPlay is additive—drivers who prefer Rivian's native UI simply wouldn't use it—and says he won't buy a Rivian R2 or R3X despite wanting one until the company adds support.
HN Discussion:
  • CarPlay's consistency across vehicles and personal device integration makes it invaluable
  • CarPlay is a must-have feature confirmed by market data and personal buying decisions
  • Rivian's native infotainment is genuinely inferior, undermining their anti-CarPlay stance
  • CarPlay is actually inferior to well-designed native systems like Tesla's for navigation
  • CarPlay is unnecessary since phone mounts or basic systems work fine
3.Right to Local Intelligence(righttointelligence.org)
325 points by thoughtpeddler 13 hours ago | 107 comments | permalink
tl;dr: Summary not available
HN Discussion:
  • Requests more specifics about what laws or actions the article references
  • Skeptical of the campaign's purpose since local model rights already exist practically
  • Strongly supports proactive defense of local AI against corporate regulatory capture
  • Agrees local models threaten hyperscalers who push regulation to protect valuations
  • ~Warns the article's safety carve-outs will be weaponized to effectively ban local models
4.crustc: entirety of `rustc`, translated to C(github.com)
310 points by Philpax 14 hours ago | 61 comments | permalink
tl;dr: A developer translated rustc (Rust nightly) into 46 million lines of C that compile with GCC into a functional Rust compiler, as a demo for their unreleased "cilly" Rust-to-C toolchain. Cilly aims to bring Rust to obscure/legacy platforms that have C compilers but no LLVM/GCC support, using "witness programs" to probe compiler features and even supporting network-transparent compilation over TCP to remote C compilers. The full toolchain isn't publicly available yet due to the author's job, thesis, and a blender incident.
HN Discussion:
  • Admiration for the dedication and originality of the niche transpilation project
  • Suggests practical applications like Diverse Double-Compiling to verify compiler trustworthiness
  • Questions whether extending LLVM/GCC would be more practical than building a transpiler
  • ~Notes existing alternatives like LLVM's C backend that could serve the same purpose
  • Humorous quips about Rust-to-C being ironic given the 'rewrite in Rust' trend
5.Since Linux 6.9, LUKS suspend stopped wiping disk-encryption keys from memory(mathstodon.xyz)
497 points by IngoBlechschmid 22 hours ago | 209 comments | permalink
tl;dr: Summary not available
HN Discussion:
  • Title is clickbait since luksSuspend is a Debian-specific extension, not officially supported kernel feature
  • Explains the technical context of why keys remain in RAM during suspend but not hibernate
  • This regression matters because silent security bugs don't announce themselves
  • ~Hibernating to disk is a better protection than relying on LUKS suspend key wiping
  • Recurring critical C bugs suggest systemic issues with large open source C codebases
6.Reality has a surprising amount of detail (2017)(johnsalvatier.org)
292 points by vinhnx 5 days ago | 109 comments | permalink
tl;dr: Reality contains vastly more relevant detail than is apparent from a distance—simple tasks like building stairs or boiling water fracture into countless nuanced sub-problems once you actually engage with them. This matters because critical details are invisible until noticed, then become transparent once internalized, which makes it dangerously easy to get intellectually stuck without realizing evidence is right in front of you. The remedy is deliberately seeking out details you'd normally overlook, especially those that others (particularly people you disagree with) find important.
HN Discussion:
  • Personal anecdotes confirm that real-world details defeat naive assumptions and planning
  • Contrarian view: reality actually requires surprisingly few details to be useful
  • ~Programming is appealing precisely because its fiddly details are tractable and fixable
  • Learning happens when assumptions break, validating the article's point about hidden details
  • Praise for the writing quality and memorable framing of the article
7.Podman v6.0.0(blog.podman.io)
566 points by soheilpro 23 hours ago | 225 comments | permalink
tl;dr: Podman 6.0.0 modernizes its networking stack by transitioning from slirp4netns and iptables to Netavark, Pasta, and nftables, and adds experimental Pesto rootless port forwarding that preserves source IPs. The release also overhauls Quadlets with REST API support, improves `podman machine` with multi-provider support and a new `os update` command, revises config file handling for multi-user environments, and further tightens Docker API compatibility.
HN Discussion:
  • Podman is a great Docker replacement with easy migration and no daemon needed
  • Quadlets and rootless containers make Podman excellent for homelab/server hosting
  • Podman's incomplete Docker compatibility causes subtle issues for users
  • Poor distribution support (e.g., Ubuntu) limits Podman adoption vs Docker
  • Questions about how Podman compares to alternatives like OrbStack or native macOS containers
8.Exapunks (2018)(zachtronics.com)
303 points by yu3zhou4 18 hours ago | 103 comments | permalink
tl;dr: EXAPUNKS is a 2018 Zachtronics programming puzzle game where players write assembly-like code to control EXAs (Execution Agents) that hack into networks, banks, game consoles, and other systems in a cyberpunk setting. The game includes in-fiction zines ("TRASH WORLD NEWS") with tutorials, a custom puzzle creation tool (Axiom VirtualNetwork+) using JavaScript, and mini-games like HACK*MATCH. Physical zine sets remain available via print-on-demand through Lulu.
HN Discussion:
  • EXAPUNKS captures the essence of programming and makes it fun
  • Zachtronics games inspired career growth and reduced fear of assembly
  • Recommending related Zachtronics catalog and similar hacking games
  • Physical zines and detailed manuals are a wonderful throwback
  • ~EXAPUNKS language didn't click, other Zachtronics titles are better
9.Immich 3.0(github.com)
474 points by hashier 23 hours ago | 232 comments | permalink
tl;dr: Immich v3.0.0 introduces Workflows (a drag-and-drop automation builder with triggers, filters, and actions), HLS real-time video transcoding, a custom web video player, integrity checks that detect untracked/missing files and checksum mismatches, and a "Recently Added" page. Mobile gets non-destructive editing parity with web, OCR text selection, slideshow support, direct-to-album uploads, and significantly more reliable Android background backups via a new periodic task scheduler. The release contains numerous breaking API changes (mostly affecting third-party integrations), drops pgvecto.rs support, and requires updating IMMICH_VERSION to v3 in the .env file.
HN Discussion:
  • Enthusiastic users praising Immich as a high-quality free replacement for Google/Apple Photos
  • Users sharing self-hosting setups and deployment tips for securing Immich
  • ~Debate over the lack of end-to-end encryption, with some defending its absence and others preferring alternatives like Ente
  • Questions and concerns about mobile app sync reliability and background backup behavior
  • Criticism of poor import tooling from Google Photos/iCloud and buggy Live Photo handling
10.An American Privacy Emergency(scottaaronson.blog)
336 points by flowercalled 13 hours ago | 100 comments | permalink
tl;dr: A June 2026 Commerce Department directive (DAO 216-26) bans differential privacy, noise infusion, and swapping from Census Bureau and BEA publications, restricting confidentiality protection to 1970s-era techniques like coarsening and suppression. Cynthia Dwork and other leading researchers argue this politically motivated order—tied to Project 2025 and aimed at enabling citizenship data extraction—will force a lose-lose tradeoff between useless statistics and violations of the Census Act's confidentiality requirements, since coarsening alone can be trivially reversed via basic algebra when multiple statistics interact. They call on scientists to pressure Congress to rescind the directive.
HN Discussion:
  • Quoting and highlighting the article's key claims about the directive's bans
  • Providing action links to contact legislators as the article urges
  • Asking what the political motivation behind targeting these techniques is
  • Article lacks technical detail and relies on contrived scenarios to favor its preferred solution
  • Author is being bombastic and the piece isn't really about privacy
11.Android Developer Verification: Threat masquerading as protection(f-droid.org)
1641 points by drewfax 1 day ago | 706 comments | permalink
tl;dr: F-Droid frames Google's mandatory Android Developer Verification (ADV) — rolling out September 30 in Brazil, Indonesia, Singapore, and Thailand — as a trojan-like system service that will block installation of apps from developers not centrally registered with Google. They argue the program does little to actually prevent malware distribution, forces developers to hand over ID and agree to vague terms where "malware" is undefined (and could be weaponized against ad-blockers or other disfavored software), and threatens F-Droid's open-source distribution model despite widespread opposition from EFF, FSF, ACLU, and over 70 organizations.
HN Discussion:
  • Users should switch to Linux-based mobile OSes to escape Google's control
  • F-Droid's tone is childish and undermines their credibility with mainstream audiences
  • Users have the right to install whatever they want on devices they own
  • Google itself is the real trojan/malware through pervasive data harvesting
  • Mandatory signing/verification threatens freedom, democracy, and tinkerer/developer autonomy
12.PeerTube is a free, decentralized and federated video platform(github.com)
627 points by doener 1 day ago | 306 comments | permalink
tl;dr: PeerTube is an AGPL-licensed, federated video platform from Framasoft that uses ActivityPub to interoperate with the Fediverse (Mastodon, Pleroma, etc.) and WebRTC for peer-to-peer load sharing between viewers. It supports both on-demand and live streaming, lets instances cache each other's videos for redundancy, and offers creator support via donation links rather than ads or recommendation algorithms. Anyone can self-host an instance or join an existing one, with no vendor lock-in.
HN Discussion:
  • Lack of monetization makes PeerTube unviable for professional content creators
  • Platform lacks critical mass of content and audience to be useful
  • Poor UX and discovery make it inaccessible to average users
  • Genuinely useful for hosting FOSS/tutorial content and a welcome alternative to YouTube
  • ~PeerTube only solves distribution, not the harder problems of discovery and monetization
13.Why Switzerland has 25 gbit internet and America doesn't(stefan.schueller.net)
500 points by talonx 9 hours ago | 380 comments | permalink
tl;dr: Switzerland treats fiber as a natural monopoly, mandating point-to-point architecture with four dedicated fiber strands per home and open Layer 1 access, allowing multiple ISPs to compete over shared physical infrastructure — enforced by regulators who fined Swisscom 18M francs in 2024 for trying to switch to a shared P2MP model. In contrast, the US allows territorial monopolies with shared connections and no meaningful competition, while Germany wastes resources on redundant parallel builds ("overbuild"). The lesson: infrastructure should be built once as a neutral asset, with competition happening at the service layer.
HN Discussion:
  • Article is clickbait ignoring geographic scale and coverage nuances of the US
  • US ISPs provide humiliatingly poor service, reinforcing need for better infrastructure model
  • Multi-gigabit speeds are unnecessary in practice; 1Gbit is more than enough
  • Public investment and infrastructure regulation succeed elsewhere (Italy, Australia's failed NBN)
  • Raw speed numbers matter less than peering quality and real-world performance
14.Spain Orders Blacklist of Palantir from Public and Private Companies(clashreport.com)
677 points by mgh2 22 hours ago | 273 comments | permalink
tl;dr: Spain's government is directing SEPI-controlled companies—including Telefónica, Indra, and Navantia—to stop contracting with Palantir over national security and sovereignty concerns, mirroring similar moves in France and Germany. Palantir retains a €16.5M contract with Spain's military intelligence (CIFAS) expiring in November, which military leadership wants renewed despite Moncloa's resistance. The blacklist reflects geopolitical tensions between PM Sánchez and the Trump administration (given Thiel and Karp's ties), and Spain is funding domestic alternatives like Openchip via a €5B gigafactory initiative.
HN Discussion:
  • Spain is setting a positive example other European countries should follow
  • The ban is hypocritical since Spain uses other foreign firms like Huawei, suggesting corruption not security
  • ~Skepticism that the policy will endure through future political changes
  • Questioning why the ban targets only Palantir rather than being a general law
  • Requesting more specifics about the actual national security concerns cited
15.The fall of the theorem economy(davidbessis.substack.com)
268 points by varjag 1 day ago | 115 comments | permalink
tl;dr: A former mathematician argues that AI's rapid progress on theorem-proving benchmarks (like the recent "First Proof" project, where labs solved 6-8 of 10 research-level problems) threatens mathematics as a profession, but exposes a deeper misunderstanding: math's real value lies in human comprehension and concept-building, not proofs themselves. He warns that AI-generated proofs are often unintelligible "Mathslop" that doesn't accrete to the corpus, and that the mathematical community's honor code—which rewards only theorem-proving—leaves it dangerously vulnerable to being declared "solved" by AI while its actual cognitive contributions go unrecognized.
HN Discussion:
  • Mathematics' real value is intuition and insight, echoing Egan's 'truth mining' vision
  • Proving theorems is a historical accident; math like software can rely on testing and usage
  • AI dominance may end open science as actors hoard AI-generated knowledge
  • Article correctly identifies understanding as core, and AI-generated proofs won't be understood by humans
  • ~Article misses that math's insularity and failure to communicate is the real problem
16.AI can't be listed as inventor on patent applications, Japan's top court rules(japannews.yomiuri.co.jp)
381 points by mushstory 23 hours ago | 199 comments | permalink
tl;dr: Summary not available
HN Discussion:
  • Patents themselves don't demonstrably boost innovation, so restricting them is welcome
  • AI lacks accountability so should not own patents or benefits
  • ~Ruling is inadequate since humans can just list themselves as inventors of AI-generated work
  • Obvious/common-sense ruling consistent with other jurisdictions like the US
  • Denying AI inventorship raises concerns about future AI personhood and rights
17.How to ask for help from people who don't know you(pradyuprasad.com)
490 points by FigurativeVoid 1 day ago | 72 comments | permalink
tl;dr: When asking strangers for help, put yourself in their mind: establish credibility (ideally through proof of work rather than institutional affiliation), provide concise context that connects to their interests, and make the request specific, bounded, and low-friction. Critically, make it easy to say no—a pressured yes poisons the relationship, while help freely given builds one. And never lie, since any whiff of dishonesty kills the request.
HN Discussion:
  • ~Concise, minimal-effort requests can outperform elaborate ones showing high effort
  • Proof of work must be substantive and deep, not superficial demonstrations
  • ~Showing self-effort to solve the problem matters more than request formulation
  • Offering to pay or citing the person's work signals seriousness effectively
  • People overestimate how often experts are asked, skewing outreach decisions
18.The Egg Bandits Made a Thousand Times the Fine They Just Paid for Price Fixing(thebignewsletter.com)
472 points by toomuchtodo 1 day ago | 239 comments | permalink
tl;dr: Summary not available
HN Discussion:
  • Consumer boycott: sharing lists of Cal-Maine brands to avoid buying
  • Surprise that the egg price spike was price fixing, not just inflation/avian flu
  • Market concentration and lack of competition enable this kind of price fixing abuse
  • Fines are inadequate; need harsher penalties like corporal punishment or executive licensing
  • Need a stronger FTC with real antitrust enforcement teeth
19.The primary purpose of code review is to find code that will be hard to maintain(mathstodon.xyz)
360 points by ColinWright 1 day ago | 174 comments | permalink
tl;dr: Summary not available
HN Discussion:
  • Code review has multiple purposes; maintainability is just one of many important ones
  • Knowledge transfer and shared ownership are the most valuable aspects of code review
  • The article's claim promotes lazy reviewing by narrowing review's legitimate scope
  • Bugs actually can be found through code review via code smells and patterns
  • Clarifying the author's mathematical framing about bug-finding limitations
20.Bring back crappy forums(tedium.co)
559 points by pentagrama 1 day ago | 343 comments | permalink
tl;dr: The author reminisces about early web forums (WWWBoard, phpBB, vBulletin, UBB) and their quirky tech like BBCode, arguing they fostered better communities than modern algorithm-driven social media. Forums lost out to Web 2.0 platforms like Reddit and Digg largely due to novelty-seeking and the burden of self-hosting, not because they were worse. The piece suggests users may eventually return to smaller, forum-style communities as they tire of scale-driven engagement and "context collapse."
HN Discussion:
  • Reddit-style threaded comments are a genuine improvement, not just novelty-seeking
  • Nostalgic agreement that forums fostered deeper communities and information than modern platforms
  • ~Forums declined because they were genuinely crappy and hard to maintain, not just due to novelty
  • ~Forums still exist and thrive in niche communities of practice; no need to bring them back
  • Social media won due to network effects and dark patterns, not user preference