StreetComplete: Fixing OpenStreetMap, one tiny quest at a time
537 points · streetcomplete.app
StreetComplete is an Android app that gamifies contributing to OpenStreetMap. Instead of requiring users to learn JOSM or iD editor, it finds missing map data near your location and presents it as simple quests — “What’s the surface of this road?” or “Does this restaurant have outdoor seating?” You answer on-site, and the data gets submitted to OSM under your account.
It’s a clever on-ramp for a project that has always struggled with contributor growth. The barrier to editing OSM has historically been high — you need to understand tagging schemas, aerial imagery interpretation, and editor tooling. StreetComplete sidesteps all of that by reducing contributions to binary or multiple-choice questions. The tradeoff is that you can only add data this way, not correct existing errors or add complex geometry. Still, for filling in the long tail of metadata that makes OSM actually useful for routing and search, it’s an effective approach. Open source, available on F-Droid, no accounts required beyond your OSM login.
Chat Control 1.0 and 2.0 Explained
206 points · fightchatcontrol.eu
There aren’t one but two parallel “Chat Control” laws moving through EU institutions, which is why the reporting seems contradictory. Chat Control 1.0 was a temporary derogation from the ePrivacy Directive that let providers voluntarily scan private messages for CSAM. It expired in April 2026 after Parliament refused to extend it. Chat Control 2.0 is a proposed permanent regulation that would make detection mandatory — and it’s been stuck in trilogue negotiations for months.
The core fight: the Council wants “voluntary” suspicionless scanning plus broad risk-mitigation duties (effectively mandating scanning through the back door), while Parliament insists scanning should only target specific suspects with judicial authorization. The Council’s own legal service recently flagged that the proposal constitutes generalized scanning incompatible with the EU Charter. Meanwhile, Google, Meta, Microsoft, and Snap have stated they’ll continue scanning regardless of the legal basis.
A better way to tie gym shorts (or any drawstring)
371 points · youtube.com
A 15-second video from “FIRST CLASS AMATEUR” showing a drawstring knot technique. 3.1 million views, 85K likes. This is peak Hacker News: a trivially simple practical tip that somehow nobody knows, posted by a channel with 278K subscribers. The video is four years old and still climbing the front page. Sometimes the most upvoted content is just a genuinely useful thing explained well, with zero venture capital involved.
Chat Control passed first round in EU Parliament
412 points · heise.de
In a maneuver opponents are calling unprecedented, the EU Parliament voted 331–304 to adopt an urgency procedure that lets them vote on reviving Chat Control 1.0 on Thursday — the last session day before summer break. Parliament President Roberta Metsola put it on the agenda at the behest of member states and the EPP group. The tactical calculus is grim: since it’s second reading, rejecting the law requires an absolute majority (361 votes), but passing it only needs a simple majority of MEPs present. With many parliamentarians already departed for summer, the re-enactment is considered nearly inevitable.
Four EU Commissioners pressured MEPs with a letter warning about a “regulatory gap.” Pirate MEP Markéta Gregorová called it a farce; rapporteur Birgit Sippel (SPD) refused support but her group caved anyway. The irony: IT security researchers keep warning about unacceptably high false-positive rates of AI scanning tools, and a Society for Informatics board member has filed an urgent application with Germany’s Federal Constitutional Court. The EU is about to resurrect mass surveillance infrastructure that its own lawyers say is incompatible with fundamental rights — and they’re doing it by timing the vote for when most opponents have left the building.
Microsoft fires idTech team at id Software
379 points · gamefromscratch.com
As part of Xbox’s restructuring under new CEO Asha Sharma — 3,200 roles eliminated across FY27 — most or all developers working on idTech at id Software have been laid off. The idTech engine is arguably one of the most consequential pieces of game technology ever built, ranking 4th in GameFromScratch’s all-time engine list. It powers not just id’s own games but a sprawling family tree of engines and titles descended from the original Quake codebase.
The loss of institutional knowledge here is staggering. id Software’s engine team had decades of expertise in rendering, optimization, and platform-specific tuning that you can’t replace by hiring fresh graduates. Microsoft bought Bethesda for $7.5 billion and is now gutting the technical foundation that made one of its studios legendary. The pattern — acquire, restructure, discard the talent — is depressingly familiar in tech M&A. id’s games will presumably move to Unreal or some internal Microsoft engine, but the specialized knowledge that made idTech uniquely performant is walking out the door.
98% isn’t much
405 points · whynothugo.nl
Hugo Barrera makes a deceptively simple point: 98% sounds high until you think about what it means in context. A restaurant where 98% of customers don’t get food poisoning is getting people sick weekly. A website that works for 98% of visitors is broken for ~150 million people globally. And crucially, 98% of the general population doesn’t mean 98% of your audience — the author found that a CSS feature with “wide support” (98% globally) only worked for 70% of one client’s actual visitors.
This is a direct shot at the browser compatibility mindset that treats “98% support” as green-lit for production. The post argues that truly robust engineering means graceful degradation, not just optimizing for the majority. It’s a short, sharp piece that makes you reconsider what “supported” actually means when your denominator is real humans trying to use your product. Worth reading for anyone who ships web software and has ever said “it works in Chrome.”
Amazon without the knockoffs
262 points · knockoff.shopping
Knockoff is a Chrome extension that filters trademark-squat pseudo-brands out of Amazon search results. You know the ones — SZHLUX, HORUSDY, EHEYCIGA, and other ALL-CAPS keyboard smashes that exist only to flood Amazon listings with white-label products. The extension checks every listing against a bundled database of 5,000+ known real brands and scores unknown names using linguistic heuristics (ALL-CAPS, unpronounceable consonant runs, vanishing vowels).
It’s open source (Fair Source, converting to MIT after two years), runs entirely locally with no accounts or tracking, and works across all Amazon marketplaces. You can set strictness levels — relaxed catches notorious offenders, strict only shows recognized brands. Community reporting feeds misclassification fixes back into the shared brand list within a day. The problem it solves is real: Amazon’s marketplace has become nearly unusable for finding quality products amid the flood of algorithmically-named dropship garbage. Whether Amazon will tolerate a browser extension that hides their sponsored listings (an optional feature) is another question entirely.
Throughline
Two dominant themes today: the erosion of digital privacy and the fragility of institutional knowledge.
The Chat Control stories are the big one. The EU is on the verge of resurrecting mass surveillance of private messages using procedural tricks timed to when opposition is thinnest. The fightchatcontrol.eu explainer and the heise.de report together paint a picture of a legislative process being gamed — the Council pushing “voluntary” scanning that its own lawyers say violates fundamental rights, while tech companies have already stated they’ll scan regardless of legal authority. This isn’t child protection; it’s the normalization of infrastructure that can be repurposed for anything.
The Microsoft/idTech layoffs and the “98% isn’t much” piece both speak to how institutions discard hard-won knowledge. Microsoft is destroying decades of engine expertise to hit quarterly restructuring targets. Web developers treat 98% browser support as “good enough” while excluding millions. The knockoff-shopping extension is a band-aid on Amazon’s failure to maintain marketplace quality. And StreetComplete is the bright spot — a tool that makes it easy for ordinary people to contribute to a genuinely open, community-maintained resource rather than depending on a corporation to get it right.