Hacker News Digest — 2026-02-13-PM
PM HN summary for February 13, 2026, focusing on the highest-scoring front-page stories and what the comments fixated on.
Themes
- Basics are regressing: typing/text selection, payment UX, and other foundational interactions feel worse (and people are mad).
- Dark patterns are moving from annoyance to enforcement: the EU is framing infinite scroll/autoplay/recs as addictive design under the DSA.
- Tools for expressing ideas in text remain popular (ASCII/Unicode diagrams, CSS-based generative patterns).
- AI is both a productivity accelerator (pattern finding in math/physics) and a governance/safety lightning rod.
Fix the iOS keyboard before the timer hits zero or I’m switching back to Android (https://ios-countdown.win/)
Summary: A one-page ultimatum with a WWDC countdown arguing Apples iOS keyboard has become unreliable (autocorrect, swipe typing, and text selection) and demanding a fix or public commitment.
- Lots of me too on missed taps, laggier typing in long fields, and selection behavior becoming unpredictable.
- Mini how-to (and complaints) about how Select All now appears only in certain states.
- Some technical speculation that heavy UI transitions/input-queue handling is causing double-taps and dropped input.
MonoSketch (https://monosketch.io/)
Summary: An open-source browser app for building box-drawing/ASCII-ish diagrams from simple shapes, aimed at code/docs/presentations.
- Comparisons to Monodraw (Mac app), with requests for cross-platform portability and small UX fixes.
- ASCII pedantry: Unicode box-drawing isnt ASCII, but its what people actually want.
- Debate: terminal/text diagrams vs Mermaid/Graphviz vs images (diffability, searchability, speed).
Skip the Tips: A game to select “No Tip” but dark patterns try to stop you (https://skipthe.tips/)
Summary: A short web game that makes No Tip hard to click, highlighting how UI nudges can steer people into paying more.
- Thread shifts to payment-terminal dark patterns like dynamic currency conversion (often a 1215% markup).
- Multiple anecdotes of merchants hitting accept conversion for customers.
- Complaints about tap-to-pay devices hiding the price vs older transparent POS displays.
GPT-5.2 derives a new result in theoretical physics (https://openai.com/index/new-result-theoretical-physics/)
Summary: OpenAI and collaborators claim single-minus gluon tree amplitudes can be nonzero in a special kinematic regime, with GPT-5.2 helping simplify cases and conjecture a general formula later verified by humans.
- Pushback on the hype framing; heavy focus on whats actually novel vs known MHV/ParkeTaylor results.
- Clarifications that the claimed novelty is about a special regime where a vanishing theorem has a loophole.
- Meta: LLMs shine on problems with strong verification constraints; spec discovery remains the bottleneck.
Lena by qntm (2021) (https://qntm.org/mmacevedo)
Summary: A fictional encyclopedia entry about the first runnable brain upload, emphasizing legal/economic incentives to copy and exploit a person-as-software.
- Recommendations for qntm and adjacent sci-fi; people relate it to current AI trajectories.
- A long detour into the classic Lenna test image controversy (consent, licensing, cultural value).
- Some read the story primarily as a labor/rights caution: abstracting people behind an API enables abuse.
Zed editor switching graphics lib from blade to wgpu (https://github.com/zed-industries/zed/pull/46758)
Summary: Zed merged a PR removing Blade and moving the Linux renderer to wgpu, reflecting consolidation around pragmatic Rust graphics layers.
- Rust GUI ecosystem discourse: lots of churn and under-maintained deps; GPUI gets credit for being for a real app .
- Mixed Zed user experiences: fast and stable vs recent updates caused hangs/crashes .
- Editor vs IDE expectations: deep refactors are typically LSP-provided, not editor-native.
OpenAI has deleted the word ‘safely’ from its mission (https://theconversation.com/openai-has-deleted-the-word-safely-from-its-mission-and-its-new-structure-is-a-test-for-whether-ai-serves-society-or-shareholders-274467)
Summary: A governance/accountability article noting OpenAI removed safely from IRS-filed mission language amid restructuring, capital raises, and ongoing safety controversies.
- Tension between safety guardrails and usefulness for sensitive conversations; sycophancy called out as a real risk.
- A side-argument about archive.is/archive.today ethics and paywall workarounds.
- Disagreement on legal liability and what responsibility AI companies should bear.
The EU moves to kill infinite scrolling (https://www.politico.eu/article/tiktok-meta-facebook-instagram-brussels-kill-infinite-scrolling/)
Summary: The EU is treating infinite scroll/autoplay/push notifications and personalized feeds as addictive design , preliminarily finding TikTok in breach of the Digital Services Act.
- Key link is an EC press release (not just a generic ban on infinite scroll): https://ec.europa.eu/commission/presscorner/detail/en/ip_26_312
- People argue about vague vibes-based enforcement vs the difficulty of writing precise anti-addiction rules.
- Some claim the true root cause is advertising incentives; others emphasize user agency and legitimate infinite-scroll use cases.
Sandwich Bill of Materials (https://nesbitt.io/2026/02/08/sandwich-bill-of-materials.html)
Summary: A parody SBOM spec for sandwiches mapping supply-chain jargon (CVE scanning, provenance, lockfiles, licenses) onto lunch.
- Mostly riffs: semver bread, hashing mayo, and the General Pickle License.
- A few practical notes on where real SBOMs show up (procurement and due diligence).
CSS-Doodle (https://css-doodle.com/)
Summary: A web component that generates generative-art patterns by applying CSS-like rules (plus custom functions) over a grid of DOM elements.
- Amazement at how much variety comes from small rule sets.
- Debate whether its really CSS or more like a declarative JS drawing engine; canvas vs DOM tradeoffs.