Hacker News, Distilled

Project behavior and purpose

• Tool configures heartbeat-based checks in external alerting systems (currently PagerDuty).
• Jobs/services “check in” at expected times; missed check-ins cause alerts regardless of the monitored infra’s state.
• Unlike simple intervals, supports arbitrary schedules (e.g., specific times on weekdays) and time-window deltas.
• Runs as a self-hosted service that only calculates and updates snooze/heartbeat configurations; the actual alerting is done by the vendor.

Comparison to other tools and IaC

• Some suggest this logic could live in Terraform or use existing Terraform providers (e.g., for PagerDuty).
• Compared to services like Cronitor, BetterStack, Checkly, OpsGenie, etc., this is positioned as:
  • “Missing heartbeat → alert” wiring specifically into PagerDuty.
  • More expressive scheduling than some interval-only heartbeat monitors.
• Question raised: if you already depend on SaaS (PagerDuty), why not use a SaaS that directly hosts heartbeat endpoints?

Infra, reliability, and dependencies

• Supporters argue: monitoring should not rely solely on your own infra; using a robust external provider increases fault tolerance.
• Counterpoint: dead-man switches feel too important to rely on third-party services you don’t control.
• Retort: everything ultimately depends on someone else’s infra (power, ISP, cloud); prepaying a cloud service is often simpler and more reliable.
• Some already use multi-region Alertmanager, OpsGenie, or similar to monitor the monitors.

Naming and concept (“dead man’s switch” vs heartbeat)

• Several note this is essentially a watchdog/heartbeat/cron-job monitor, not a literal “dead man’s switch.”
• Others argue the definition fits: if the operator/system fails to act (send heartbeat), an automatic action (alert) triggers.

Use cases beyond infra & after-death actions

• Thread veers into “what happens after I die”:
  • Interest in digital estate handling, debt/estate law nuances, and ensuring survivors can access assets.
  • Ideas include “death README” documents, attorneys as executors, or specialized “dead man service” companies.
  • Concerns about robustness (e.g., hospitalization, coercion scenarios, immutable tasks).

Integrations and future directions

• Interest in adding more alerting vendors (Squadcast, telco systems, messaging apps with “send later,” blockchain-based control contracts).
• Project maintainer open to contributions and treating core logic as a library.

Manifest V3 and ad blocking capabilities

• MV3 removes the powerful blocking variant of webRequest and replaces it with declarative rules (declarativeNetRequest).
• Rule limits have been increased over time (now up to ~330k static + 30k dynamic shared across extensions, with per‑extension minimums), but are still seen by some as arbitrary and constraining.
• Critics say you cannot fully express uBlock Origin’s dynamic/heuristic behavior as static rules, and features like “element zapper” and deep per‑site tweaking are degraded.
• Several users report that uBlock Origin Lite and AdGuard MV3 extensions block “all” or “most” ads in practice, including YouTube for some, while others see more leaks and breakage, especially on complex sites.
• There is concern advertisers will adapt once most Chrome users are constrained to static lists.

Security, privacy, and user‑control debate

• Pro‑MV3 side: webRequest‑based extensions are a major malware/spyware vector; declarative filtering lets adblockers work without seeing raw requests, improving privacy and reviewability.
• Critics argue Chrome could have kept MV2 for vetted extensions or gated APIs behind stricter review instead of removing capabilities, and that MV3 reduces user agency over their “user agent.”
• There is disagreement whether blocking fewer APIs truly improves security, given other powerful capabilities (e.g., script injection) remain.

Firefox and other browser options

• Firefox explicitly states it will keep MV2 “for the foreseeable future,” so full uBlock Origin and similar extensions continue to work.
• Firefox features discussed: Multi‑Account Containers (per‑tab cookie isolation), existing but clunky profile management (an overhaul is in progress), vertical tabs via forks (Floorp, Zen) or add‑ons, but loss of PWAs on desktop.
• Some criticize Mozilla for funding dependency on Google and for mishandling extension reviews (e.g., friction over uBO Lite), but many still see Firefox as the main escape hatch.
• Chromium‑based alternatives: Brave and Vivaldi plan limited MV2 support tied to upstream; Brave and Vivaldi both rely more on built‑in blockers. Long‑term maintenance of MV2 forks is seen as costly/unclear.

Enterprise and policy workarounds

• Chrome offers an enterprise policy (ExtensionManifestV2Availability=2) that keeps MV2 working until June 2025; users share registry and policy snippets for Windows, macOS, and Linux.
• This “managed browser” mode can disable Chrome’s DoH resolver unless additional policies are set.

Network‑ and OS‑level blocking

• Pi‑hole, NextDNS, and MITM proxies are suggested as MV3‑proof layers, but:
  • DNS‑level blocking is coarse (domain‑wide, weak exceptions), can be bypassed via DoH, and struggles with YouTube ads.
  • MITM proxies can be effective but run into certificate pinning and setup complexity.

Broader ecosystem concerns

• Many see this as part of a wider pattern of platforms locking down user control (DRM, hardware attestation, secure boot) and of Google leveraging its browser monopoly to weaken adblocking.
• Others think MV3’s impact is being overstated and expect most users to stay with Chrome, with only a technical minority migrating to Firefox or niche browsers.

Markdown vs. Rich Text and Data Longevity

• Some are fatigued by “yet another Markdown editor” and want true WYSIWYG rich-text with Markdown stored only behind the scenes.
• Others see Markdown as essential because notes should outlive any single app; plain-text files make migration and multi-app use easy.
• Concerns raised that Markdown isn’t well standardized; every app has its own dialect and feature set.
• Several users praise hybrid approaches (WYSIWYG + Markdown view) in apps like Joplin, Obsidian, Bear, etc.

Platform, Filesystem, and iOS

• Debate over whether iOS “has a filesystem.”
  • Some say iOS’s Files app and provider system now act like a real filesystem, including network shares and third‑party file providers.
  • Others still see it as constrained versus traditional desktop filesystems.
• Desire for Linux support appears repeatedly; web/PWA is appreciated but seen by some as weaker than a native app.

Comparison with Other Note Apps

• Obsidian is the main comparison point:
  • NotesHub praised for low one‑time price, app‑store availability, Git sync, and web version.
  • Obsidian favored for plugins, powerful search (with plugins), and ecosystem; its sync and Git integrations have mixed reviews.
  • Licensing differences matter: Obsidian is free for personal use but commercial use requires payment; NotesHub’s free tier reportedly allows work notes.
• Other comparisons: Apple Notes (great UX but Apple‑only, flaky for some), Joplin, Zettlr (FOSS), iA Writer, Note‑taking via VS Code/Neovim, etc.
• Some worry about closed‑source lock‑in; others argue Markdown storage mitigates this risk.

Architecture & Cross-Platform Tech

• Windows app uses Electron but is reported to be relatively light; macOS app is non‑Electron.
• Mobile apps are hybrid (webview-based) using a React web app without frameworks like Flutter/Capacitor; some dispute calling this “native.”
• Web version is a PWA with offline capabilities and is positioned as the Linux solution.
• Marketing copy is critiqued for heavy use of passive voice.

Sync, GitHub, and Privacy/Self-Hosting

• Git integration is a major draw; users like editing notes in a Git repo and committing from the app.
• Some are uneasy that the GitHub OAuth flow requests broad permissions; fine‑grained PAT + “generic Git” provider is suggested but reportedly doesn’t always work (unclear if user error or bug).
• Support for self-hosted GitLab via generic Git is mentioned; full self-hosting of the app is not available.
• Requests for end‑to‑end encryption and self‑hosted versions are common.

Features, Missing Pieces, and UX Requests

• Users appreciate Kanban, LaTeX, music notation, Mermaid diagrams, whiteboarding, and Git-backed notebooks.
• Requests include:
  • Browser clipping/extension (like Evernote/Notion/Joplin).
  • Plugin system (seen as a key Obsidian advantage).
  • Better drawing tools (pen types, zoom behavior), mind‑mapping, themes, HTML export, and custom CSS.
  • LFS support and stronger image/file handling.
• Overall sentiment: app is impressively polished and featureful for a solo dev, but some hesitate due to closed source, lack of plugins, limited Linux story, and missing crypto/self-hosting.

Nature of “reasoning” in LLMs

• Many commenters see current models as doing “associative” or pattern-based reasoning: they match to seen templates and local patterns, not robust logical calculus.
• Others argue that with enough training, models learn compressed internal procedures that function as reasoning, even if implemented via statistics over tokens.
• There is debate over definitions: some insist “reasoning” should mean formal, verifiable deduction; others say if next-token prediction reliably yields correct chains of thought, the label is mostly philosophical.

Benchmarks, perturbations, and overfitting

• The discussed paper and related work show large drops in accuracy when:
  • Irrelevant details are added to word problems.
  • Numbers or surface phrasing are changed but underlying structure is the same.
• Some see this as evidence of overfitting and benchmark gaming (Goodhart’s Law, possible training contamination, RLHF targeted at famous puzzles).
• Others note newer models (e.g., reasoning-optimized ones) degrade less, and claim the paper overemphasizes small or older models to argue for fundamental limits.

Concrete successes and failures

• Examples where models succeed: classic “Alice’s siblings” puzzle, some nontrivial algebra, and advanced functional analysis guidance that a math PhD student finds genuinely helpful.
• Examples where they fail:
  • Slightly modified river-crossing puzzles and family riddles.
  • Misinterpreting a doctor riddle when pronouns change.
  • Simple arithmetic/algebra (e.g., decimal comparisons, binomial expansion) in some settings.
  • Struggling with irrelevant clauses in GSM-like problems and arbitrary-length exact computation.

Comparison with human reasoning

• Several note strong similarities to average students: performance collapses when multiple steps, distracting details, or unfamiliar phrasings are introduced.
• Others stress key differences:
  • Humans can notice distraction, withhold answers, and update their understanding across a conversation.
  • LLMs remain confidently wrong and often revert to earlier misinterpretations.
• One prominent mathematician is quoted as likening a top model to a “mediocre but not incompetent” grad student that still needs heavy hinting.

Architecture, data, and tools

• Some blame tokenization of numbers, finite depth/attention, and fixed compute per token for reasoning brittleness; others say these can be mitigated with better schemes or external tools.
• Synthetic math data is widely discussed: easy to generate for formal math, harder for genuine quantitative reasoning; current gains on benchmarks are incremental, not saturating.
• Many argue real reliability will come from LLMs embedded in larger systems (tools, solvers, formal languages), not from raw next-token prediction alone.

Implications and outlook

• Optimists: scaling, improved data, and “slow thinking” variants (explicit chains-of-thought, tool use) will push LLM reasoning past most humans in many domains.
• Skeptics: core limitations (fragility to irrelevant info, lack of verifiable guarantees, error rates on simple tasks) make them untrustworthy for correctness-critical math and serious agents.
• Several tie this to the broader AI investment bubble and warn that headline math/olympiad claims may overstate true, general reasoning ability.

State of Fundamental / Particle Physics

• Many agree particle physics is in a rut: the Standard Model keeps being confirmed, but no “sexy” new physics appears.
• Others stress there is still substantial progress (tetra/pentaquarks, precision tests, g-2, W mass updates, gravitational waves), just not paradigm-shifting.
• A particle physicist argues the field is not “dying”; advances are slower because experiments are harder and costlier, but work continues steadily.

Loop Quantum Gravity (LQG) and Lorentz Invariance

• The video’s main technical claim: LQG’s discrete areas conflict with Lorentz invariance unless the minimum area goes to zero or invariance is broken.
• Experimental bounds from astrophysical observations reportedly rule out the Lorentz-violating version.
• A prominent former LQG researcher agrees the situation is at least as bad as described.
• Some question whether the argument is as decisive as presented; others say the math really does force a harsh either/or.

String Theory and Testability

• Strong criticism that string theory has produced no clear, falsifiable prediction over decades and can be tuned to fit anything.
• Counterpoints:
  • It is practically hard to test quantum gravity at all, not just string theory.
  • Some string-inspired ideas (e.g. supersymmetry at accessible scales) have effectively been ruled out, but this doesn’t kill the whole framework.
• Debate over whether string theory has drifted into unfalsifiable “mathematical physics” rather than empirical physics.

Experiments, Colliders, and Funding

• Disagreement over future large colliders:
  • Critics: next collider has no targeted, likely discovery comparable to the Higgs; money might be better spent on other avenues.
  • Supporters: without new high-energy data the field will stagnate; also, if you stop building, you may lose the capability entirely.
• Some accuse big labs of overselling benefits; others call that a “false dichotomy” and say multiple approaches can be funded if plans are concrete.

Sociology, Ego, and Epistemology

• Several comments focus on ego, cognitive dissonance, and sunk-cost fallacy:
  • Senior researchers may resist abandoning decades of work.
  • Groups can become self-reinforcing and hostile to criticism.
• Comparisons are made to other domains where confident experts systematically overpredict.
• Disagreement over whether formal philosophy of science helps; some argue it’s mostly irrelevant, others think understanding limits of knowledge is essential.

Broader Speculation and Offshoots

• Some propose that physics may have followed a wrong fork long ago and needs deep backtracking, revisiting discarded ideas (aether, many-worlds, extra time dimensions, Kaluza–Klein).
• One elaborate conspiracy theory suggests fundamental physics was tacitly “soft-suppressed” after nuclear weapons to avoid dangerous discoveries; most find this implausible given geopolitical competition.
• Tangents arise about climate science, AI/ML winning physics-related Nobels, and the difference between testable science and unfalsifiable speculation.

Views on the Video and Communicator

• Many find the critique valuable, especially the technical part on LQG; endorsement by established experts in the field is noted.
• Others dislike the clickbait style, personal tone, and lack of detailed sourcing in a short “rant” format, saying it makes it hard to separate strong points from overreach.
• Some worry that painting whole subfields as “not even wrong” undermines public trust and oversimplifies a complex mixture of genuine exploration, dead ends, and institutional incentives.

Reaction to the Nobel Peace Prize choice

• Many see awarding Nihon Hidankyo as an “excellent” and timely reminder of nuclear dangers, especially amid renewed nuclear rhetoric and weakened arms‑control treaties.
• Others are disappointed the committee did not highlight actors in ongoing conflicts, reading the choice as an expression of concern rather than a celebration of concrete progress.
• Some praise it as a more “legitimate” award than past controversial laureates; others question whether the organization has real impact versus symbolic value.

Survivor testimony and memory

• Multiple commenters reference visiting Hiroshima/Nagasaki museums and meeting survivors; these experiences are described as profoundly affecting.
• Graphic details are recalled: skin sloughing off, slow radiation deaths, permanent “shadows” on stone, lifelong disabilities.
• There is concern that “time witnesses” will soon all be gone, making organizations like Nihon Hidankyo more important as custodians of memory.

Physical effects of nuclear weapons

• Extended discussion contrasts “vaporization” vs “disintegration,” with clarifications that most victims were burned and blasted rather than literally turned to gas.
• The “human shadow” on steps is explained as differential stone discoloration, not bodily residue.
• Later comments detail fireballs, blast waves, thermal pulses, and radiation, emphasizing that many deaths would be delayed and horrific, not instantaneous.

Deterrence, risk, and international law

• Some argue nuclear weapons have prevented major wars via mutually assured destruction; others highlight near‑misses and call MAD itself “madness.”
• Debate over whether moral taboo, fear of nuclear winter, or cold strategic logic is the main restraint.
• Customary international law on non‑use is discussed; critics say it is weak, unenforceable, and shaped by nuclear states.

Hiroshima/Nagasaki justification

• One strand claims the bombings weren’t necessary to end the war and were driven by signaling to a rival and avoiding shared occupation.
• Others argue conventional invasion would have caused far higher casualties and that the bombings did force surrender.
• Some suggest the demonstrated horror of actual use helped create the later nuclear taboo; others reject this as a moral justification.

Japan’s wartime atrocities and historical framing

• Several commenters worry global focus on Hiroshima/Nagasaki eclipses Japan’s mass atrocities in Asia.
• They criticize Japanese historiography and political symbolism (e.g., shrines) as downplaying aggression compared to post‑war Germany.
• Counterpoints stress that civilian victims of the bombings can be mourned independently of state guilt.

Modern geopolitics and escalation fears

• Strong concern about current nuclear tensions involving Russia, Ukraine, parts of the Middle East, and other nuclear powers.
• Disagreement over whether firm resistance or compromise (including territorial concessions) better reduces nuclear risk.
• Some fear that avoiding all confrontation with nuclear states effectively grants them a license to invade non‑nuclear neighbors.

Civilizational survival and space colonization

• A subthread argues humanity is “a button click away” from catastrophe and should establish self‑sustaining off‑world colonies.
• Others dismiss this as unrealistic and vastly harder than stabilizing Earth, insisting that solving climate and biodiversity crises is the real survival priority.

Robotaxi concept & vehicle design

• Two-seat “Cybercab” design draws mixed reactions: some praise the sleek, retro‑futuristic look and big rear trunk; others say a taxi that only seats two is inherently limited and not inclusive (families, disabled, elderly).
• Lack of steering wheel or pedals is seen as the first “all‑in” bet on autonomy, but also alarming: no way for passengers to intervene if software fails.
• Questions about practicality: butterfly doors, no rear window, inductive charging, and heavy automation of cleaning/charging seen as either clever optimization or showy concept‑car theater unlikely to reach mass production.

Autonomy readiness & FSD performance

• Strong divide between owners who say FSD (v12.x) is “stress‑free” and already better than many human drivers and those who report frequent dangerous mistakes (bad merges, wrong‑lane turns, near‑misses with trucks, bike lanes, or poles).
• Numerical claims vary wildly: some say ~1 intervention per 200–1,000 miles; others cite independent tests around 1 per 13 miles and argue even 1 per hour is far from taxi‑grade reliability.
• Many highlight Tesla’s own labeling of FSD as SAE Level 2 “supervised” and that autonomy without a driver requires orders of magnitude higher reliability.

Demo credibility & teleoperation

• Multiple commenters believe the Hollywood‑lot Cybercab rides were highly geofenced, low‑speed, pre‑mapped demos, not representative of real streets.
• Tesla “Optimus” robots serving drinks are widely assumed to be teleoperated; several say nothing shown exceeds 20‑year‑old robots like Honda’s Asimo.

Regulation, safety & liability

• Debate over “recalls” vs OTA updates and whether media use of “recall” is technically correct or intentionally anti‑Tesla.
• People note Tesla has not pursued California’s formal autonomous testing/deployment permits and avoids public disengagement stats.
• Some expect big regulatory and labor pushback similar to Uber; others point to existing driverless Waymo service as evidence regulators will eventually allow it.

Business model & competition

• Skepticism that individual owners will profitably run personal robotaxis; concerns about bans or opaque corporate control, analogized to being locked out of Uber.
• Some argue the real value is Tesla’s future taxi platform and AI, not car sales; others see this as stock‑price “hype” after years of missed FSD/robotaxi timelines.
• Waymo is repeatedly cited as 5–10 years ahead in real robotaxis; Chinese players (e.g., BYD, Baidu Apollo) are mentioned as rising EV/AV competition.

Public transit vs car‑centric future

• Many argue robotaxis are a distraction from proven solutions (trains, trams, buses, bikes, walkable cities) and will worsen congestion.
• Others see robovans and small taxis plus fewer parked cars as a path to reclaiming urban space, if autonomy actually works.

Overall sentiment

• Enthusiasm centers on the long‑term vision and visible progress in FSD for some users.
• Skepticism focuses on safety, repeated slipped promises since ~2016, lack of hard data, and whether this is another “2 years away” announcement.

Title & pricing confusion

• Many readers initially misread “$2 H100s” as a purchase price, not rental, and called the original title click‑baity.
• Once clarified as ~$2/hour rental, most agreed this is still a sharp drop from ~$8/hour and notable for the market.

Economics of $2/hr (and below)

• Back‑of‑the‑envelope math: at ~$50k per H100, $2/hr implies ~3–3.5 years to break even assuming near‑full utilization, not counting power, networking, rack space, labor, or financing.
• Some argue $2/hr (or promotional ~$1–1.5/hr) must be loss‑making unless hardware or space/power is effectively “free,” or the provider is burning VC money.
• Others counter that once hardware is bought, sunk cost is sunk: renting at or near marginal cost (power, ops) is rational to minimize losses versus idling or fire‑selling.

Capacity, clustering, and real constraints

• Key distinction: single PCIe H100s vs tightly networked SXM/Infiniband clusters. The latter (for large‑scale training) still command higher prices.
• Comments stress that “$2/hr H100” often means constraints: PCIe cards, weaker networking, spot‑style capacity, beta‑quality uptime, or limited user access.
• Some say the real money is in renting whole, reliable multi‑node clusters; cheap single‑GPU time is a side effect of over‑reserved capacity.

Bubble, AI winter, and Nvidia

• Opinions split on whether this signals an AI infra bubble bursting or just a normal boom‑bust supply cycle like oil or crypto mining.
• Some are short or cautious on Nvidia at its current valuation; others argue demand will rebound as cheaper compute spurs more fine‑tuning and inference.
• Several note that next‑gen GPUs (Blackwell, MI300X, etc.) and ongoing foundational‑model races will keep aggregate compute demand high.

Open models, training vs. inference

• Broad agreement that open‑weights models (e.g., Llama families) reduce the need for most companies to train from scratch.
• Thread consensus: only a small number of teams globally need very large H100 clusters for new foundation models; most just fine‑tune and serve inference.
• This shift depresses demand for top‑end training clusters while expanding use of smaller, cheaper GPU setups.

Infrastructure, reliability, and cloud vs. bare metal

• Debate over whether low‑cost providers are inherently unreliable versus “good enough” for checkpointed training workloads.
• Some emphasize that hyperscaler‑style uptime and 45kW+ racks are extremely expensive; others say cloud reliability is oversold and many on‑prem setups could match it.
• Power cost per H100 is relatively small compared to capex and depreciation; rack density, networking, and financing dominate economics.

PHP CMS alternatives (ProcessWire, Kirby, ClassicPress, etc.)

• ProcessWire is heavily praised: fun to develop with, flexible, suitable for both simple sites and web apps, and easy to update by replacing a single directory.
• Critiques: less “batteries-included” than WordPress; better suited to developers than non-technical users and has weaker “site profile”/theme marketplaces.
• Kirby gets positive mentions from long-time users, but some criticize its outdated admin stack (Vue 2).
• ClassicPress appears as a more “classic” WordPress fork with performance comparisons linked.
• Other PHP options noted: Textpattern, Dotclear, GetGrav, PluXml, Chyrp Lite, Typecho, Movable Type, BloKi.

Static Site Generators & Static Hosting

• Many advocate SSGs (Hugo, Jekyll, Astro, Lektor, Middleman, Publii, Quartz, Logseq, Soupault, Pandoc).
• Pros: security (hard to hack), cheap or free static hosting (GitHub Pages, Netlify, Deno Deploy), version control workflows, performance.
• Cons: most are poor fits for non-technical users; collaboration and image handling are recurring pain points; commenting is seen as inferior to WordPress’s built-in system.
• Publii is highlighted as a GUI middle-ground SSG, though collaboration and Git integration are viewed as weak.

Drupal, Wagtail, and heavier frameworks

• Drupal supporters argue it’s a powerful, extensible, PHP-based alternative with modern initiatives (browser-based “Drupal CMS”, AI tools, commerce/CRM, no-code ECA).
• Others say Drupal has grown more complex and less friendly for “ordinary” deployments, with many sites stuck on Drupal 7.
• Wagtail (Django-based) is praised for UX, flexibility, and stability but criticized for opacity, learning curve, and need for substantial developer setup. Some see it more as a CMS framework than a drop-in WordPress replacement.

Hosted platforms, headless, and niche tools

• Hosted platforms like Wix, Squarespace, Shopify, Ghost (hosted), Hashnode, Bear, Framer, and Sanity (headless) are mentioned as realistic options for non-technical users.
• Specialized or visual tools: Frappe Builder, Netlify + JAMstack, Astro as a flexible “engine,” magecdn for static image handling.

What people want in a “true” WordPress alternative

• Desired traits: self-hosted, open-source, easy install, long-lived, rich plugin/theme ecosystem, ACF-like custom fields, Gutenberg-style visual editing, REST API, and workable e-commerce.
• Many argue that for the vast majority of non-technical users, nothing matches WordPress’s combination of simplicity, ubiquity, plugins, and support—alternatives either lack features or require more technical skill.

Ecosystem & politics

• Some criticize WordPress’s current ecosystem (naggy freemium plugins, hosting pain, leadership drama), yet others note GPL and forking as safeguards.
• Political stances by CMS creators (on social media or endorsements) spark debate: some want politics kept out of recommendations, others see ethical stances as necessary.

Population density vs other explanations

• Many reject density as the main cause. If it were, politics would be easy in dense NYC and hard in sparse Wyoming, which doesn’t match people’s experience.
• Some note density correlates with party preference, but see that as insufficient to explain conversational taboos.
• Others argue broader US culture (individualism, mobility, fragile social ties) matters more than physical proximity.

Norms, risk, and “psychological safety”

• Long‑standing etiquette (“don’t talk politics or religion”) reduces practice, so conversations are brittle.
• People fear job loss, HR trouble, social ostracism, or even violence (especially in a heavily armed country).
• In workplaces, politics is seen as “bad for business,” so many deliberately avoid it.
• Reduced social capital and weaker neighborhood ties make rupture cheaper: it’s easy to drop friends or change circles.

Two‑party system, identity, and tribalism

• The US two‑party, first‑past‑the‑post system forces all-or-nothing bundles of issues; mixed views (e.g., pro‑choice + low taxes) are hard to represent.
• In multi‑party and coalition systems (Europe, parts of India), people report it’s easier to separate issues from overall identity and still cooperate.
• In the US, party and leader attachment becomes identity; politicians are “worshipped,” rallies feel cultish, and criticism is taken as an attack on the self.
• Politics often functions as tribal affiliation rather than policy debate; sound bites and culture‑war framing dominate.

High stakes vs “catastrophizing”

• Many say current issues are existential (abortion, trans rights, immigration, racial equality). For affected groups, opposing views feel like threats to survival or basic personhood, not abstract disagreements.
• Others argue “life or death” language is overused to shut down debate and is sometimes out of proportion to actual risk.
• Abortion, trans rights, and immigration generate especially sharp conflict: some focus on bodily autonomy and civil rights; others on fetal life, social order, or border enforcement.
• Disagreement over whether mass deportation or anti‑minority rhetoric is comparable to historical genocidal trajectories.

Culture, media, and online dynamics

• Comparisons with India and Europe: India has intense political violence yet open discussion; some attribute that to cynicism about efficacy (politics as “sports”) and strong in‑group bonds.
• US politics is seen as increasingly fear‑driven; media and social platforms amplify outrage, personalize issues, and reward division.
• Online debate norms (bad faith, “debate bros,” dogpiling) bleed into offline life and raise the perceived cost of engagement.

Suggestions for better conversations

• Focus on shared problems and underlying values before party labels or “solutions.”
• Set boundaries (“I’m done for today”) without demonizing; recognize opportunity costs, especially for marginalized people constantly asked to justify their existence.
• Some argue shunning bigots is morally necessary; others see engagement and continued relationships as more effective but acknowledge this is easier for the less directly targeted.

FBI Sting Website and Execution

• Commenters identify what appears to be the FBI’s fake project site (“NexFundAI” / “NextFundAI”), noting:
  • Low-effort, mostly AI-generated feel with stock robot images and over-annotated HTML.
  • Despite looking cheesy, it still successfully attracted targets, which some see as evidence of how low the bar is in crypto scams.
• A banner now explicitly confirms it as a government sting.
• Some jokingly ask where they can buy the coin, underscoring how blurred the line is between parody and real tokens.

Legal, Entrapment, and Tactics Debate

• Major thread on whether this is entrapment:
  • One side argues it’s not entrapment to create a token and wait for scammers; it’s analogous to setting up a honeypot server or posing as a customer of a criminal service.
  • Entrapment is described as inducing crime in someone not predisposed (e.g., threats, long-term cajoling).
  • Bait-car and undercover-prostitution analogies are used to argue the FBI’s behavior is lawful, if morally gray.
• Some worry law enforcement prefers “easy” sting operations over harder investigations of existing crimes, potentially misallocating public resources.

Market Manipulation vs. Legit Market Making

• Discussion on “market makers” who allegedly advertise control over pumps, dumps, and insider-like trading.
• Comparisons drawn to large crypto trading firms that do market making and arbitrage:
  • Some see a continuum from sophisticated market making to outright manipulation.
  • Others note ongoing investigations into big players and suggest it’s too early to equate them with blatant fraud.

Crypto Value, Use Cases, and “Pump and Dump” Claims

• Strong skepticism that most crypto is effectively pump-and-dump:
  • No intrinsic yield; value depends on selling to someone else at a higher price.
  • Still marginal as a payment method; overwhelmingly used for speculation.
• Counterpoints:
  • Use cases cited: cross-border payments, remittances, escaping inflation or capital controls, and “disciplining” central banks via stablecoins.
  • Some liken Bitcoin to “digital gold,” while others argue gold and Bitcoin differ significantly in stability and utility.
  • Debate over whether value is “intrinsic” vs. subjective, with references to economic theories of value and collateral use.

AI Workloads as “Proof of Work”

• Brief exploration of tying crypto mining to useful AI computation:
  • Supporters like the idea of “useful work.”
  • Critics argue technical and economic mismatches: AI workloads are finite and don’t scale in the same way as difficulty-adjusted mining.

Overall reception of “The Copenhagen Book”

• Many commenters find it clear, practical, and a welcome contrast to “don’t roll your own auth” warnings.
• It’s praised for being concise, actionable, and focused on how to use crypto/auth correctly rather than deep math.
• Some skepticism about treating it as authoritative, partly due to concerns about some design recommendations (e.g., session extension).

Passwords, magic links, and alternative auth flows

• Strong disagreement over “don’t support passwords in 2024”:
  • One side: passwords are universal, work well with managers and TOTP/security keys, and shouldn’t be removed.
  • Other side: email magic links / OTPs are more secure in practice for typical users than reused passwords.
• Magic links draw heavy criticism for UX: email delays, spam filters, device mismatch (phone vs desktop), and corporate email gateways “clicking” links.
• Some argue magic links should only log you in on the device that clicked them to reduce phishing, accepting worse UX on public computers.

Passkeys, OAuth, and centralization

• Predictions that integrated password managers or passkeys will dominate; others are wary of centralization (Apple/Google lock-in, account bans).
• Concerns that OAuth and passkeys leak usage patterns to big providers and create brittle dependencies, so passwords should remain an option.

Security vs UX and library design

• Repeated theme: many real security failures are UX and API-design problems, not cryptographic ones.
• Crypto APIs are criticized for “character soup” names and unsafe defaults; praised examples favor high-level, purpose-driven functions with safe defaults.
• Good UX examples: simple password hashing APIs, well-integrated hardware keys; bad UX: cumbersome security key flows and PGP-style complexity.

Sessions, tokens, and lifetimes

• Debate around extending session lifetimes vs regenerating tokens:
  • Some worry about effectively “immortal” tokens.
  • Others argue rolling sessions with DB lookups and user-visible session management are a reasonable tradeoff.
• Long-lived sessions are desired by users frustrated with frequent expirations and 2FA prompts; others note risk and provider-imposed caps.

Email handling and identifiers

• Disagreement about treating email addresses as case-insensitive:
  • Standards say local parts are case-sensitive, but most real systems are not.
  • Common compromise: store original casing for display, but compare case-insensitively.
• Notes that many transactional email providers only support ASCII and that exotic email formats are often painful in practice.

Other gaps and wishes

• Requests for coverage of JWT, SAML/SSO, ID tokens, and more language examples (Python/Node).
• Some see the “how to build” approach as preferable to ever-more-complex auth libraries.

Title wording and tone

• Several dislike “better than you think” / “you” headlines as condescending or click‑baity.
• Others argue readers understand it’s generic, find the meta‑complaints unproductive, or see it as normal copywriting.
• Some note irony: provocative titles drive upvotes, but also provoke guideline‑violating meta‑threads.

---

TypedDict vs dataclasses / slots / Pydantic

• Pro‑dataclass arguments: faster attribute access, lower memory with slots=True, nicer .attr syntax, and runtime attribute‑name errors.
• Counterpoints: Python’s general slowness can dwarf L1/cache details; algorithmic choices matter more; __slots__ especially pays off with many tiny objects.
• Some strongly recommend Pydantic (or msgspec) as “transformative” for modeling and validating data, but others note:
  • TypedDicts (especially total=False) are more ergonomic for modeling partial/patch data, including function kwargs (PEP 692).
  • Pydantic is heavier and slower; good at boundaries, not necessarily for everything.

---

Patch semantics, None, and modeling optionality

• Discussion around representing HTTP PATCH‑style updates where a field can be:
  • Left unchanged
  • Set to a value
  • Explicitly set to None / “unset”
• None alone can’t encode both “leave as is” and “remove”, so people propose:
  • Using sentinels or tagged unions (e.g. set/unset/ignore actions).
  • TypedDicts with optional keys to model “field omitted” vs “present and null”.

---

Static typing, enforcement, and tooling

• One side: annotations “do nothing” without a runtime enforcer like Pydantic or decorators; Python happily ignores them.
• Other side: static type checkers (mypy, Pyright) plus CI can give strong guarantees inside the codebase once external inputs are validated.
• Disagreement over whether optional typing is “worst of both worlds” or a sweet spot:
  • Critics: no global guarantees, types can lie, external tools aren’t exhaustive.
  • Supporters: any static checking is better than none; hints improve editor support, docs, refactoring.

---

Static vs dynamic languages and Python’s type system

• Many argue static types improve readability, correctness, and refactoring even in small scripts; others find them verbose, ugly, and slowing down experimentation.
• Comparisons arise to Go, TypeScript, Java, C#, Rust, Scala, Haskell, etc., with mixed views:
  • Some think Python’s typing is weaker and fragmented (multiple checkers, subtle differences).
  • Others value its gradual, opt‑in nature and use hints primarily as lightweight metadata.

---

Real‑world TypedDict use

• Several recount messy “dict‑everywhere” Python codebases where TypedDicts made it possible to reason about data flow without large refactors.
• TypedDicts are seen as a low‑risk, drop‑in way to retrofit structure on string‑keyed dicts, whereas dataclasses or full refactors feel too invasive.

Overall reception

• Many commenters praise the project as an approachable, humane intro to HTML, especially for non‑professionals and complete beginners.
• Several see it as aligned with the original, open, “small web” ethos and as a counterweight to complex JS-heavy stacks and corporate platforms.
• Some plan to use it in teaching (high school classes, spouses, non‑technical friends) or as a “next step” after simple blogging/CMS tools.

Teaching approach & scope

• Strong approval for starting with:
  • plain text editors (Notepad/TextEdit),
  • a single index.html,
  • immediate deployment (e.g., Neocities),
  • then gradually adding structure and better tools.
• Many like the decision to lean on a simple external CSS (simplecss.org) instead of teaching full CSS early.
• Some argue sandboxes (online editors) are better than local files for reducing friction and “small mistakes,” others insist basic file handling is an essential skill.
• Debate on whether to add more CSS, templates, or keep the focus tightly on HTML.

HTML semantics & minimal documents

• Discussion about what constitutes a “proper” HTML5 document:
  • <!DOCTYPE html> is required to avoid quirks mode.
  • Only doctype and a non-empty <title> are strictly necessary; <html>, <head>, <body> are optional.
  • Many tags can be left unclosed and still be valid HTML5; some dislike treating HTML like XHTML.
• Some advocate for semantic structure (sections, header/main/footer, avoiding “sea of divs”) and minimal/no JS.

Accessibility, defaults, and styling

• Some lament that browser default styles are unattractive; speculate that nicer defaults might have encouraged more plain-HTML sites.
• Mentions of CSS resets and tiny baseline styles to improve readability (margins, line-height, fonts).
• One accessibility-oriented critique of another popular tutorial (low text contrast) used as a contrast to this site’s clarity.

Computer literacy & filesystem issues

• Repeated concern: many modern users, especially younger/mobile-first ones, struggle with files, folders, downloads, and encodings.
• Some think this is a serious barrier to following “create a folder and a file” instructions; others find this overblown or a symptom of weak education.
• Suggestions include linking to basic OS/file tutorials or short YouTube primers.

Meta: who HTML is “for”

• Strong agreement that HTML should be for “everyone who wants it,” not just professionals.
• Counterpoint: most people neither need nor want markup; WYSIWYG and higher-level tools are better for them.
• Some argue LLM/code tools now reduce the incentive to learn HTML; others still see lasting value in understanding the basics.

Recall integration with Explorer

• Thread centers on reports that Recall is now an Explorer dependency; some note it’s not a hard DLL dependency but more of a “dark pattern” integration that’s hard to fully disable.
• Others say disabling Recall via DISM currently works and preserves the modern Explorer UI, though people fear it could be silently re‑enabled by future updates.
• Comparisons are drawn to past tying of IE to the OS and to other hard‑to‑remove Windows components.

Privacy, surveillance, and security concerns

• Many see Recall (periodic screenshots plus AI search) as “spyware‑like,” especially since it records sensitive screens (PHI, legal work, trade secrets).
• Even if data is local, users object to CPU usage and attack surface; people recall earlier versions storing screenshots in an unencrypted DB.
• One comment cites Microsoft documentation that Recall uses a local model in a virtualized secure enclave with TPM‑encrypted storage; others immediately question how long that will remain unbreakable, given past security issues.

Debloating and mitigation strategies

• Numerous tools/scripts are recommended: O&O ShutUp10++, various Win11 “debloat” scripts, privacy.sexy, Blackbird, winutil, Ameliorated builds, and switching to LTSC/IoT Enterprise SKUs.
• Some warn these scripts often break Windows, especially feature updates and Windows Update itself.
• There’s irony noted: users worried about Microsoft telemetry still run unvetted privileged scripts from random repos.
• Others argue consent matters: they’d rather risk self‑inflicted damage than accept unasked‑for features like Recall and Copilot.

Linux as alternative

• Large subthread: “just switch to Linux.” Some report successful full‑time moves (often with Steam/Proton and Lutris), saying Windows now feels hostile and ad‑ridden.
• Others push back: Linux still lacks key commercial apps (Adobe suite, high‑end CAD like SolidWorks/Revit, some photography tools, vendor device software, robust VR support). For these users, Windows remains mandatory.
• Several advocate dual‑boot, GPU‑passthrough VMs, or “Windows To Go” external SSD installs so Windows is used only when needed.

Gaming, VR, and hardware support

• Proton is praised for near‑complete game support, but anticheat systems and niche peripherals (steering wheels, VR headsets like Valve Index) remain pain points.
• Experiences conflict: some report Rocket League and various VR titles working fine under Proton/SteamVR; others describe broken compositors, calibration issues, and missing drivers/TrueForce features.

Developer experience and user attitudes

• Some developers say Linux is the best environment for server‑side/web work; others find Windows (especially with Visual Studio/.NET or WSL) equally or more productive.
• Several note that most ordinary Windows users neither know nor care about these issues; power users feel “trapped” by app/driver lock‑in and increasingly see Windows as indistinguishable from malware.

Pre‑movie featurettes and spoilers

• Many object to spoiler-heavy interviews/recaps shown immediately before classic films (e.g., Alien, The Matrix anniversaries).
• Even fans who know the films well feel it diminishes tension and rediscovery.
• Some note this is especially frustrating when bringing kids or first‑timers to see a classic on the big screen.
• A minority argue that for 25–45‑year‑old blockbusters, most viewers already know key moments via cultural osmosis, so impact is limited.

Re‑releases and who they’re for

• One view: re‑release events are primarily for existing fans who don’t want to wait through credits for extras, so putting them up front makes sense.
• Counterview: fans often bring newcomers; theaters can’t know who’s seen it, so they should avoid spoiling “the movie we are seated to see.”
• Suggestions: move featurettes after the credits; or clearly separate pre‑feature and main feature times so people can skip.

Trailers and modern marketing

• Broad agreement that trailers now routinely spoil major plot points; some people avoid them entirely and enjoy films more.
• Examples include trailers and toy/LEGO tie‑ins that reveal twists or entire plots.
• Others note trailers are cut by separate marketing teams whose sole goal is selling tickets, not preserving surprises.

Cinema vs home viewing

• Many complain about phones, talking, and long ad blocks, leading some to abandon theaters for home setups.
• Others report better experiences at indie cinemas or strict chains (no phones, fewer trailers), though food service can be distracting.
• Reserved seating and predictable pre‑roll timing let some arrive late to skip previews.

Attitudes toward spoilers

• Strong divide: some say spoilers ruin anticipation and prime you to wait for specific beats; others claim great works remain great even when you know the ending.
• Several mention books and DVD/BD menus whose introductions or menu loops spoil key scenes or themes.
• A few see spoiler‑heavy extras as catering to audiences who “need things spelled out,” reflecting broader changes in media and attention.

Scope of the Problem

• Many see the hurricane conspiracies as a symptom of broader societal decay and mainstreaming of fringe beliefs, especially in the US but also globally (UK, Eastern Europe, India, Brazil, Poland, etc.).
• Others argue “nutcases” have always existed; what’s new is visibility and amplification via the internet and 24/7 media.

US Politics and Conspiracy Mainstreaming

• Strong focus on US right‑wing politics: a sitting Representative openly alleging government weather control is seen as legitimizing conspiracy thinking.
• Some argue this reflects a long trend of anti‑intellectualism and “doing your own research,” supercharged by Trump‑era rhetoric.
• A minority push back, framing Russia‑related narratives and Trump investigations themselves as conspiracies or “political theater.”

Distrust of Institutions vs. Education

• One camp blames underfunded or misdesigned education and lack of critical‑thinking training for people’s inability to distinguish fact from fiction.
• Another camp argues the primary driver is accumulated scandals and lies from governments, corporations, and media, which erode trust even in genuine science.
• Several note that distrust now extends beyond government to scientists, journalists, doctors, and even other countries’ governments.

Role of Social Media and Algorithms

• Widely cited as a core driver: engagement‑optimized feeds amplify extreme, divisive, and conspiratorial content, turning a small minority into a loud majority.
• Some describe social platforms as “digital virus” ecosystems where corrections are socially punished and fringe communities self‑reinforce.

Psychology and Sociology of Belief

• Explanations include: need for simple narratives and scapegoats; loss of purpose in an abstract, spreadsheet‑driven economy; “intellectual revenge” by people who feel looked down on; and a desire for belonging and status.
• Several stress that conspiracism is not purely an IQ or education issue; well‑educated people, including professors, can be true believers.

COVID, Cognition, and Generational Change

• A side thread suggests COVID infections and long COVID may mildly reduce cognitive function, possibly worsening trends; others say conspiracy culture predates COVID by decades.
• Multiple comments worry about younger students’ declining reading focus and older generations’ susceptibility to online misinformation.

Weather, Climate, and “Weather Tech”

• Some conflate legitimate weather modification (e.g., cloud seeding patents) with hurricane creation/control; others emphasize the enormous energy scales involved and demand evidence of actual capability.
• Climate‑change denial and minimization appear alongside respect for individual meteorologists’ lifesaving work, illustrating fragmented trust.

Overall reception

• Strongly positive response, especially from users with ultrawide or 4K monitors and frequent screen‑sharing.
• Seen as a “genius” quality-of-life tool that solves a long-standing annoyance, especially with tools like MS Teams that lack “share region” features.
• Several plan to adopt it immediately for work, teaching, and pairing.

Primary use cases

• Sharing a readable portion of a very large or high‑DPI display without constantly changing system resolution.
• Keeping a dedicated “sandbox” area for meetings while preserving the main desktop layout.
• Avoiding accidental exposure of notifications or unrelated windows.
• Using a lower‑resolution virtual screen (e.g., 1080p) so viewers on laptops can see code, spreadsheets, or SAP screens clearly.

Alternatives and cross‑platform solutions

• Windows: RegionToShare, FancyZones in PowerToys, virtual display drivers, VMs/remote desktop, OBS with virtual camera or projector windows.
• Linux/X11: Xephyr/Xnest, xrandr --setmonitor, xvfb + VNC, VLC + slop, nested compositors; on Wayland: headless outputs plus mirroring (sway, Hyprland), pipewire screencast, xdg-desktop-portal hooks.
• macOS: BetterDisplay (paid, more features), Screegle, OBS projectors, system accessibility zoom, RDM for resolutions.
• Some note that OBS is very powerful but can be overkill or confusing; DeskPad is praised for simplicity and being free/open.

Screen-sharing UX & readability

• Common complaint: viewers on small screens can’t read content from large/HiDPI displays, especially in Teams.
• Suggestions include:
  • Using a low-res virtual monitor via DeskPad and sharing only that.
  • Not maximizing shared windows; keeping them smaller for better scaling.
  • Letting presenters control viewport instead of asking viewers to zoom.
• Debate around window-sharing UX: some find it manageable; others say it’s clumsy and slow to switch between multiple windows.

Implementation, limitations & concerns

• Uses private macOS APIs for virtual displays; questions about how these APIs were discovered and Apple’s stance on their use outside the App Store.
• macOS shows a persistent “screen recording” indicator, which some find annoying but accept as unavoidable.
• macOS-only nature is noted; multiple people discuss how Linux has had similar capabilities for years, though often with more complex tooling.
• Some Intel/legacy Mac users worry about performance and OS support longevity.

Changing Definitions and Diagnosis of Hypertension

• Guidelines have repeatedly lowered BP thresholds (e.g., to 130 mmHg in 2017), dramatically increasing the number of people labeled hypertensive; some see this as evidence-based, others suspect profit incentives.
• Several comments stress that proper diagnostic protocols (multiple office readings, 24‑hour ambulatory monitoring) are rarely followed in routine practice.
• White-coat and “labile” hypertension are common themes; many report high office readings but normal home or 24‑hour values.

Challenges of Wrist and Cuffless BP Measurement

• Wrist cuffs are described as highly position‑sensitive (must be at heart level, supported); small deviations cause large errors.
• Cuffless, PPG-based approaches are seen as promising but technically fragile: waveform strongly depends on sensor pressure, orientation, and movement, and may only proxy true BP.
• Medical-device engineers note that FDA clearance for purely PPG wrist devices is still missing and likely difficult.

Existing Devices and Calibration

• Aktiia and Samsung Galaxy watches are cited as working, but require periodic calibration with an arm cuff and are often approved only in certain regions.
• Calibration is seen as both essential and annoying; measurements calibrated to one person fail badly on others.
• Cheap Chinese smartwatches exist; some users find trends roughly plausible, others assume they are close to random without validation.

Continuous Data: Promise and Interpretation Problems

• Many expect BP data to mirror CGM patterns: large intra‑day variability, spikes with exercise, cold, caffeine, stress, etc.
• Debate on usefulness: some say only standardized resting readings matter for current risk models; others argue daily averages, patterns, and variance may be more predictive.
• Concern that clinicians currently lack frameworks and guidelines for acting on continuous BP streams.

Clinical Phenomena and Causes

• Multiple reports of asymptomatic high BP, post‑Covid onset, white‑coat anxiety, and strong sensitivity to caffeine.
• One claim attributes most hypertension to high sugar/fructose intake; others push back, citing low‑carb diets yet persistent high BP and likely multifactorial causes.

Behavior Change, Privacy, and System Issues

• Continuous or easy BP tracking can motivate lifestyle changes and improve control for some users.
• Worries include cloud‑dependent devices, poor after‑sales support, unclear device accuracy, and future insurance or workplace use of stress/BP data.

Overall reaction & technical achievement

• Commenters are impressed that fully reverse‑engineered, open GPU and x86 layers on Apple Silicon can now run many AAA Windows games on Linux.
• The stack combines x86 translation (e.g., FEX), DirectX→Vulkan translation, and a VM to cope with 4K vs 16K page sizes, yet still reaches playable performance.
• Some highlight this as evidence of how far Linux gaming and translation layers (Wine/Proton‑style tech) have come.

Performance, overhead, and limits

• Newer AAA titles often don’t hit 60 fps; emulation and virtualization add CPU, GPU, and RAM overhead.
• 16 GB of RAM is described as a practical minimum because games run inside a VM to fix page‑size assumptions.
• Emulation overhead and page‑size workarounds can cause noticeable memory pressure on lower‑RAM machines.
• Some games run better on Asahi than via macOS tools in specific anecdotes; others report the opposite.

Asahi vs macOS (Game Porting Toolkit, Crossover, Whisky)

• Several argue macOS with Apple’s Game Porting Toolkit, Crossover, and Whisky still offers broader game compatibility, more features (e.g., ray tracing), and less overhead today.
• Others report poor experiences on macOS (visual glitches, low fps) and significantly better results on Asahi for particular titles.
• Consensus: Asahi’s achievement is big, but macOS + existing tools generally remains ahead for Apple‑hardware gaming right now.

Apple GPU architecture, Vulkan, and Metal

• Discussion of missing hardware tessellation and geometry shaders on Apple GPUs; these are emulated with compute or mesh shaders.
• This complicates Vulkan and certain emulators; some see this as a reason Apple prefers Metal and avoids official Vulkan.
• Opinions split on whether this is mostly technical (API design, die area, developer experience) or mainly about platform lock‑in.

Hardware support and project priorities

• Users lament slow progress on M3/M4 support, DP‑alt‑mode, Thunderbolt, and microphones compared to the fast graphics advances.
• Others note different contributors work on different subsystems; lack of progress in some areas reflects who shows up, not simple priority trade‑offs.
• M3/M4 GPUs differ significantly (new instruction encodings, ray tracing, “dynamic caching”), so support will take time; lack of an M3 Mac mini also hurts CI.

Distros, ecosystem, and philosophy

• Fedora Asahi Remix is the “first‑class” target; NixOS and Arch can work but require more manual effort, especially for the newest gaming stack.
• Broader discussion touches on Proton’s role in making Linux gaming viable vs native ports, and on why Linux users still want access to Windows‑only games.
• Some debate ARM vs x86 efficiency; claims conflict and are left unresolved in the thread.

Referrer and HN‑specific behavior

• Side discussion about the project site previously serving an anti‑HN page based on HTTP referrer.
• This leads to explanations of the Referer/Referrer‑Policy mechanisms and privacy trade‑offs; some consider referrer‑based blocking childish but technically straightforward.