HN Brief: 2026-11-02
Today's front page was a quiet revolt against solutions that replace judgment with a procedure. The biggest thread, HackerBooks, showed a list of books mentioned on HN and immediately faced the problem of Mein Kampf sitting in the top spots. The Lenovo encrypted drive was just an SD card, the perfect metaphor for checkbox security. The Apollo economist argued AI has delivered no payoff for most companies. The mood found its counterbalance in a police chief physically unplugging Flock's cameras—a rare story of a human winning a fight against an opaque system.
The threads most worth clicking are "HackerBooks," for the split between purists who want raw conversational volume and skeptics who insist parsing praise versus trashing is essential; "Lenovo's encrypted fingerprint USB," for the devastating look at checkbox security culture; "Police chief unplugs cameras after Flock ignores contract cancellation," for the argument over whether defending the distinction between Flock and Ring just helps the surveillance vendor; "AI delivers zero financial payoff to most companies," for the insistence that judgment is the real bottleneck once production is commoditized; and "Mathematics has changed. Let's act like it," for the existential identity crisis over funding human mathematicians if machines can do the grunt work.
Show HN: HackerBooks – HackerTube, but for books mentioned on HN [comments]
1288 points · 223 comments · hackerbooks.org · 16h ago
Someone used Claude to build HackerBooks, a site that logs every book mentioned by at least two different commenters across HN history, linking each book back to the actual discussions and making the data downloadable. The thread’s biggest blowback hit the original “recommended” label—*Mein Kampf* and *Das Kapital* sat in the top spots—and the creator swapped the wording to “mentioned” within hours, earning real goodwill for the quick fix. Beyond the semantics, everyone was genuinely impressed by the speed of the static JSON site, and the thread immediately devolved into a brainstorming session demanding “HackerMovies” and other meta-HN spin-offs. There’s a clean split between purists who want raw, unsentimented conversational volume and skeptics who argue a proper recommendation engine has to parse whether people were praising or trashing a book to be useful.
Restrooms along the Norwegian Scenic Routes [comments]
687 points · 406 comments · www.visitnorway.com · 21h ago
The linked article wasn't available to this summarizer; from the discussion, the submission spotlights the official guide to the famous architect-designed restrooms along Norway's scenic routes. The whole thread immediately detoured into a sprawling political fight over Donald Trump calling Norway a 'shithouse country' after the Nobel snub. A solid consensus formed around correcting his apparent assumption that the Norwegian government controls the Peace Prize—the Committee is formally independent—though a contrarian thread argued the conflation is understandable. This widened into a direct split on whether Trump is genuinely clueless or deliberately indifferent to accuracy, completely consuming the discussion. Completely outside the politics, those trying to engage with the submission universally griped that the Visit Norway website scrolls horribly on mobile because of its video backgrounds.
Lenovo's 'encrypted' fingerprint USB is just an unencrypted SD card in a case [comments]
466 points · 158 comments · www.reddit.com · 17h ago
The thread covers the exposure of Lenovo's "encrypted" fingerprint USB drive as nothing more than an SD card in a case—the fingerprint sensor only controls power to the USB port, with zero encryption actually applied to the data itself. The overwhelming takeaway was that this isn't a bug but a feature of checkbox compliance culture, where these devices exist to let enterprises and government agencies tick an "encrypted" box on a procurement form without anyone ever bothering to verify the claim. That kicked off a genuine split between the crowd arguing you should never trust a hardware black box and should just use LUKS or VeraCrypt, versus the counterargument that software encryption is vulnerable to a compromised host, making a properly designed hardware solution with independent key entry necessary for real security. Beyond the specific Lenovo debacle, the real weight of the discussion dug into the broader class of these devices—the follow-up finding hardcoded keys in other "secure" drives and the fundamental insistence that biometrics like fingerprints are irrevocable public identifiers that completely break the cryptographic model. The whole thread gets capped with the perfect damning anecdote about a government agency shipping an encrypted drive with the password written on a sticky note taped to the box, because the requirement was that the data *be* encrypted, not that it be secure.
Police chief unplugs cameras after Flock ignores contract cancellation [comments]
444 points · 149 comments · www.thedrive.com · 17h ago
The linked article in *The Drive* covers the Plymouth, Michigan police chief who physically unplugged Flock license plate cameras after the company ignored the city’s contract cancellation. The thread largely treated this as a stark example of a surveillance company operating without accountability, drawing direct comparisons to Uber’s old regulatory playbook. The real split, though, was over Flock vs. Ring cameras: one camp insisted Flock’s dragnet tracking across jurisdictions is a unique civil liberties threat, while another pushed back that normalizing Ring data sharing makes the two functionally equivalent and defending the distinction just helps Flock. Beyond the surveillance taxonomy debate, the discussion dug into civic despair versus action—arguing over whether the chief’s unplugging proves local democracy can still force a confrontation, or if direct intimidation and property destruction are the only things these companies ultimately respect.
Patrick Stewart taught us what good men do with power [comments]
416 points · 294 comments · www.vox.com · 14h ago
The Vox piece argues Jean-Luc Picard was the idealized post-Cold War leader, wielding hegemonic power justly from within trusted institutions. The thread immediately undercut the premise by debating whether mining fiction for "thought leadership" has any value, splitting the conversation wide open. Detractors argued that TV shows compress messy reality into satisfying, crowd-pleasing resolutions, making them poor real-world guides, while defenders relied on the old guard of storytelling—Homer, Shakespeare—to argue that allegory is the very foundation of cultural ethics. A deeper correction routed around the article's tidy politics to point out that Roddenberry's original vision was space communism, not liberalism, and that *Deep Space Nine*'s exploration of the bloody costs of maintaining utopia is where the show actually grappled with the burden of just power. A final needle was the observation that if we're skeptical of fictional characters as role models, the US has elected actual actors to the presidency three times, blurring the line between performance and leadership entirely.
Ask HN: Which TV do you recommend for privacy?
312 points · 460 comments · news.ycombinator.com · 12h ago
This Ask HN asked for a privacy-respecting TV, specifically a large 4K panel to act as a dumb monitor for an Apple TV, and the dominant answer was to buy any modern TV and never touch the network settings. That advice immediately collided with a civil war in the comments: one camp insists disconnected smart TVs secretly beacon over neighbor Wi-Fi or LTE, a claim their opponents treat as an urban legend begging for a single documented example. A quieter split formed over whether an Apple TV is a genuine privacy solution or just a trade for a less visibly creepy one, with Apple defenders pointing to deliberate engineering trade-offs against critics highlighting recent ad pushes in the App Store. The thread’s real consensus settled around hotel-grade commercial displays or accepting the friction of a fully offline setup.
Mathematics has changed. Let's act like it [comments]
287 points · 306 comments · terrytao.wordpress.com · 13h ago
The article is a guest post on Terry Tao's blog by Frank Calegari, urging mathematicians to stop grieving the AI disruption and adapt their hiring, publishing, and training culture toward broader exploration instead of just racing to solve old problems. The thread largely bypassed his specific reform ideas and went straight for the jugular of an existential debate over whether any human mathematicians are needed at all. The split was sharp and simple: one camp sees AI as an inevitably superior proof-finder making the human game irrelevant, while the other insists that mathematical *understanding*, paradigm invention, and the social act of asking the right questions are the irreplaceable core of the field. A huge chunk of the discussion got sucked into the weeds of Lean verification, debating whether bugs in the software or mis-specified statements actually render the celebrated OpenAI proofs invalid. Running underneath the whole thing was the uncomfortable practical question of why anyone should fund human math as a profession if a machine can do the grunt work, which made Calegari's optimistic prescriptions for the profession feel like a hard sell.
U.S. workers took home their smallest share of capital since at least 1947 [comments]
285 points · 152 comments · fortune.com · 8h ago
The Fortune piece reports that U.S. workers' share of economic output just hit its lowest level since the government started tracking it in 1947, even as corporate profits and GDP soar. The thread largely saw this as a feature, not a bug — multiple people pointed out that decades of tax cuts and deregulation were explicitly designed to shift wealth to capital, and automation is just accelerating that, not causing it from scratch. The UBI argument got picked apart from both sides: one camp insisted it would just inflate away, while others argued that controlled prices or negative income taxes avoid that problem, and that the real issue is political will, not economic feasibility. A darker strain ran through the comments about what happens when labor has no leverage — several people drew parallels to the Great Depression and warned that the social contract only holds as long as workers have something to bargain with.
Show HN: OmniDiff – Opensource, super fast, content aware Git difftool, in Rust [comments]
225 points · 55 comments · github.com · 19h ago
OmniDiff is a new open-source Rust diff tool that isn't just for code—it handles images, PDFs, archives, and more with different algorithms per type. The HN thread went straight to comparisons with delta and difftastic, and the author had benchmarks ready: claims of 4ms vs 35ms speed and major wins on moved-code detection. A split emerged over the AGPL license, which the author defended as a deliberate political choice against Big Tech, while others argued permissive licensing would win and even questioned whether copyleft holds when AI assists in the writing. There was practical pushback on missing 3-way merge support and the 99.95% metric, but the creator of `git difftool` offered to help get OmniDiff upstreamed—a strong nod to the engineering quality.
AI delivers 'zero financial payoff' to most companies, Apollo economist says [comments]
220 points · 217 comments · techcrunch.com · 9h ago
Apollo’s chief economist argues that outside the top seven tech giants, AI has delivered zero payoff to corporate America, with the rest of the S&P 500 seeing flat profits and productivity. HN treated this as premature at first—comparing it to 2001 skeptics of the web—but the debate sharpened into a fight between developers who swear they’re shipping faster and a camp that says cheaper output from a commodity calculator doesn’t create a durable business advantage. The thread converged on the idea that AI is a cost-cutting tool, not a growth engine: if every competitor gets the same model lift, the only P&L signal is headcount reduction, which looks exactly like the “zero payoff” Slok points to. The economist’s emphasis on judgment over calculation carried the day, with most agreeing that accountability and unwritten rules are the real bottleneck now that production has been commoditized.
OccupySaaS [comments]
210 points · 140 comments · occupysaas.com · 7h ago
The OccupySaaS site pitches a movement of laid-off engineers using AI to build open-source drop-in replacements for SaaS platforms, arguing the cloud is a rent-seeking toll booth. HN immediately seized on the irony of using AI—the very technology behind the layoffs that spawned the movement—as its primary weapon, with a strong current of opinion calling the whole project tone-deaf, naive about the depth of enterprise software, and a prime example of an "AI fever dream." Beyond the ideological split, the technical feasibility took a real beating: critics pointed out that the chosen targets already have open-source alternatives, that enterprise support contracts are the actual moat, and that building a business on generalized vengeance isn't a viable product strategy. There was pushback on the framing too—the idea that well-paid tech workers are the "99%" rankled—but some defended the project's spirit, arguing that turning the tech industry's weaponized AI back on entrenched incumbents is the only fitting response to mass layoffs and vendor lock-in.
Frontier Airlines Likely Breached [comments]
203 points · 77 comments · www.reddit.com · 15h ago
The submission is a Reddit PSA warning that Frontier Airlines customers are receiving phishing emails containing their real flight details and confirmation numbers, pointing to a likely data breach at the airline. The Hacker News thread used this as a springboard to debate whether corporate security is finally facing a genuine reckoning, with one side arguing that breaches will continue until they carry real business consequences and the other pointing to attacks like the one on Jaguar Land Rover as proof that existential damage is already happening. Several people with cybersecurity backgrounds stressed how unusually convincing the scam was and called out Frontier for not issuing a direct customer warning, and the conversation spiraled into how AI tools are making both attacks more potent and compliance theater easier to fake. A clear split emerged between engineers who blame corporate indifference and those who think the legal system’s hostility toward responsible disclosure is the deeper rot.
Nurses Sound Alarm as 'Uber for Nursing' Apps Push to Deregulate Healthcare [comments]
175 points · 157 comments · nurse.org · 13h ago
The article flags a new report warning that gig nursing platforms are using algorithmic pricing to set different rates for the same shift and lobbying states to deregulate themselves out of traditional staffing rules. HN largely bypassed the specific report to debate whether provider salaries are the real driver of healthcare costs—a claim that was immediately met with the stat that physician and nurse pay accounts for roughly 15% of total spending—putting the focus instead on insurance profits, administrative bloat, and systemic inefficiency. The crowd split between people convinced the AMA is a protectionist racket deliberately limiting doctor supply and those who blame the administrative/insurance machinery, but a dense tangent cut in with a precise correction that the actual bottleneck is ACGME residency funding rather than the professional association itself. The loose consensus was that the gig model isn't a cause of the dysfunction but a perfectly rational adaptation to a deeply irrational system.
AI agent swarms in Minecraft fail to build civilization, drown in bureaucracy [comments]
166 points · 79 comments · www.tomshardware.com · 18h ago
The piece reports on Y Combinator's experiment where GPT and Claude agent swarms in Minecraft were tasked with building a civilization, only to completely fumble it by generating absurd bureaucratic overhead—like four agents spending eleven minutes appointing a single "fertilizer operator." The HN crowd immediately seized on this as a dead-on satire of human organizational systems, with a lively split forming between those who see it as a perfect indictment of EU-style regulation and those arguing that "bureaucracy" is really just an emergent property of synchronization that people conveniently moralize. The conversation seriously diverged from the article itself, as much of the thread instead devolved into venting about the sheer awfulness of the Tom's Hardware website (aggressive anti-adblock, impossible popups), while a separate, more substantive camp questioned the scientific value of the entire exercise, asking whether this was "machine psychology" or just a headline-grabbing demo with no control for prompt quality. A practical firsthand account of trying to get a single Claude agent to play Minecraft well—involving hours of careful corrections—added weight to the suspicion that the swarm's failure might simply reflect that the individual base models aren't good enough at the game yet to exhibit any interesting emergent social dynamics.
Seoul Viosys begins production of rare-earth-free white light emitter [comments]
166 points · 48 comments · www.semiconductor-today.com · 15h ago
Seoul Viosys claims it has cracked mass production of a white micro-LED that requires none of the rare-earth materials traditional white LEDs depend on, generating white light directly from gallium nitride in chips as tiny as 10µm. The discussion took that boast apart pretty quickly over color quality, arguing that hitting the CIE white region on a chart doesn't mean the light is anything but ugly if the spectral distribution is spiky, making it great on paper but potentially disappointing in a room. The real stickiness, though, was the gallium problem—everyone pointed out that swapping rare earths for gallium just swaps one heavily China-dominated critical mineral for another, and since gallium prices have already skyrocketed the claim of a cheaper, more ethical product looks shaky. The AI data transmission tie-in was treated as naked buzzword-bait, derailing an otherwise interesting look at the tech's genuine potential for backlighting and micro-displays, while the mods even had to swap the article source after the community flagged it as an AI-generated rewrite.
Landline phones will switch off across Britland in 100 days, no extensions [comments]
153 points · 124 comments · www.theregister.com · 16h ago
The Register reports that the UK's analog landline network is being shut off in 100 days with no extensions. The thread immediately derailed over the headline's use of "Britland," which apparently isn't a term recognized in the UK—some commenters defended it as typical Register piss-taking, while others admitted they clicked only to figure out where Britland was. The substantive split was about the loss of legally guaranteed quality of service: the old POTS network ran on exchange batteries that kept phones working for days during power outages, while VoIP and mobile are "best effort" services with no such requirements, and commenters from Australia, Finland, and California chimed in with similar stories of degraded reliability after their own transitions. A few commenters pointed out that the underlying copper is often the same cable; the phone just runs through the router now, and the new complexity of VoLTE and SIP stacks in telcos introduces failure modes the old circuit-switched network never had.
It wasn't AI that flipped Proton's charge [comments]
152 points · 86 comments · blog.raed.dev · 13h ago
A Proton user and longtime recommender reflects on the CEO calling a security test request “loser thinking” and argues the real story isn’t about AI code quality but about a leadership culture that treats Proton’s core promise of careful security as the main obstacle to growth. The discussion homed in on that exact contradiction: if velocity and ARR targets now dictate priorities, what distinguishes Proton from the services its users are paying to escape? A strong tangent debated whether AI specifically amplifies authoritarian “founder mode” or just gives bad bosses a fresh excuse to steamroll prudent engineering. A smaller faction tried to spin the engineer as a Luddite blind to AI’s testing power, but the thread largely dismissed that as missing the point about governance and professional respect. The CEO’s clunky apology and the suspiciously uniform “dissenters” defending the company on Reddit cemented the impression that Proton’s leadership has stopped believing in the constraints that made the product worth paying for.
Proving and improving the Platonic Representation Hypothesis with JEPAs [comments]
132 points · 67 comments · arxiv.org · 18h ago
This paper formally attempts to prove the Platonic Representation Hypothesis, showing that Joint-Embedding Predictive Architectures trained on paired "good views" converge to shared representations across modalities radically more efficiently than next-token prediction does. The reaction immediately split between those buying the formalization and a strong camp arguing it’s basically a tautology about optimal data compression, a point backed by a physics argument that the premise of a single "true" representation contradicts the standard practice of switching coordinate systems. The deepest divergence went straight into philosophy—one side insisted the math guarantees convergence to a single reality, while others countered that the world is full of incompatible, value-laden worldviews and that this framework has no way to resolve whose subjective definition of "better" wins out. Ignoring the metaphysics, the concrete result cut through cleanly: a tiny JEPA model matching or beating the biggest LLMs on vision alignment is a massive engineering finding, and the thread largely parked on that efficiency win as the real story regardless of the Platonic framing.
Novellas with Groff [comments]
120 points · 51 comments · movq.de · 18h ago
The article is a developer's diary of writing two novellas—40,000 words in ten days—using groff, the old Unix typesetter, and laying out the minimal macro package he built for it. The discussion quickly became a showcase of writers-who-code sharing their own idiosyncratic pipelines, from Google Docs to custom markup to people who still use troff for university papers, all circling the familiar tension between optimizing the tool and actually getting words on the page. A persistent tangent dragged in the writing routines of famous authors—GRRM's WordStar 4.0, King's longhand—and the irony that most of them are notorious for not finishing things, which dovetailed into a meta-complaint about how coders obsess over their typesetting stack rather than writing. Historical context surfaced about O'Reilly's original books being set in troff, along with some technical griping about groff's page break quirks. The thread landed somewhere between genuine admiration for the author's output and a group self-diagnosis of pipeline-bikeshedding disorder.
A tiny Pacific country rolls out universal basic income [comments]
98 points · 137 comments · www.economist.com · 21h ago
The Economist piece is about the Marshall Islands funding a universal basic income for all citizens using payments from the US under the Compact of Free Association. HN immediately read this as a geopolitical tool rather than a pure social experiment, arguing it's designed to stop emigration to the mainland and keep the islands populated by US allies to counter China. The thread diverged into detailed arguments over the math of a Negative Income Tax versus a pure UBI, dark speculation about universal income as a prerequisite for an AI-driven labor collapse, and whether the US welfare system is already a comparable or better mechanism. The central split was between people who see this as a pragmatic evolution of the social contract and those who dismiss it as an unsustainable dependency that only works for a tiny nation being subsidized by a superpower for strategic reasons.
F/OSS Comics #13: BSD Unix, Part 3 – The Beginning of BSD Unix [comments]
92 points · 9 comments · fosscomics.com · 6h ago
This F/OSS Comics installment covers the moment BSD became a formal distribution—Bill Joy bundling the Pascal system, vi, and the C shell onto tapes and getting Unix running on the VAX with virtual memory. The thread immediately got snagged on a seemingly minor detail in the Unix genealogy, specifically the panel that traces the 'ex' editor back to the obscure 'em' rather than 'ed'. That sparked a surprisingly fierce split over whether Ken Thompson's dismissal of visual editors—he didn't want to see the file state while editing—was a reasonable stance for the hardware of the day or a monumentally stubborn opinion from someone building an OS kernel in C without syntax highlighting. The discussion also tightened up a few of the comic's historical labels, arguing over whether early WWII computers were electronic or electro-mechanical. There was also a strong appetite for the series to dive further into the political context—like the free speech and antiwar movements—that the author hints at in the BSD origin story.
AI fan wikis riddled with errors are muddying the waters of games culture [comments]
79 points · 67 comments · www.eurogamer.net · 6h ago
The article takes aim at AI-generated "fan wikis" that flood search results for indie games, drowning out legitimate volunteer-run resources with pages full of hallucinations and fake information. The discussion immediately treated this as a perfect microcosm of the wider web's collapse, far outrunning the article's tentative optimism about community resilience and landing squarely on the idea that the internet's "dark forest" era is fully upon us with no credible off-ramp. Most participants saw any technical fix as futile, arguing the incentives to produce slop are too strong, though a vocal minority pushed for retreating into verified human-only spaces, invoking the "low-background steel" analogy to describe pre-AI content as a precious, exhaustible resource. A long, fierce side-argument erupted over whether the same rot is infecting software engineering, splitting between those who think LLMs raise the floor of code quality and those who insist the new ambiguity of "did a human even intend this logic?" makes debugging fundamentally harder than a messy but meaningful codebase ever was. The overwhelming verdict was that the article captured just one skirmish in a lost war—the basic assumption that a piece of online content reflects human effort is breaking down, and nobody has a good plan for what replaces it.
Domain hacks with unusual Unicode characters (2018) [comments]
75 points · 51 comments · shkspr.mobi · 21h ago
The article demonstrates how Unicode characters that normalize to TLDs (like ℡ for .tel or № for .no) can be used to shorten or stylize URLs. The thread's dominant take wasn't the party trick itself, but the immediate security backlash—the argument that homograph attacks render the URL bar untrustworthy, and that browsers are already right to aggressively block these domains. The point was heavily reinforced by the discussion itself, where a significant chunk of people reported their browsers mangling or refusing to open the examples at all. The broader consensus pinned the blame on Unicode's complexity as fundamentally incompatible with the stringent security needs of the DNS, a theme that spun off into war stories about early domain registration loopholes and the specific quirks of zero-width characters.
Show HN: Wikihack – HackerTube, but for Wikipedia articles mentioned on HN [comments]
73 points · 16 comments · wikihack.netlify.app · 2h ago
Wikihack scrapes HN for Wikipedia article mentions and packages them into a scrollable feed, directly riffing off Karpathy's HackerTube framework. The dominant take in the thread was a heavy skepticism, framing it as the definitive signal that the 'HackerX' clone trend has exhausted its runway, moving from genuine intellectual curiosity into what critics bluntly called low-effort fast food slop. A clean split emerged: one camp insisted the whole genre relies purely on Karpathy's influencer clout for visibility, while the other argued this specific implementation is actually a better product because it surfaces the HN comment context where the article was linked, which is almost always more valuable than the Wikipedia page itself. Beyond the meta-critique of Show HN's culture, practical feedback touched on the mobile-only desktop layout and the reliance on a temporary free domain, but the heart of the discussion was a referendum on whether imitation has hollowed out the original promise of the format.
Trigger.dev customer environment variables exposed [comments]
71 points · 28 comments · trigger.dev · 6h ago
Trigger.dev disclosed a security incident where an attacker stole customer environment variables, claiming the exploit is blocked. The real fight was over their decision to mark the incident "Resolved" on their status page—half the thread saw it as a pragmatic update, the other half as a massive red flag that torpedoes any trust the disclosure was meant to build. A chunk of the discussion also criticized the lack of technical detail, noting there’s no way to assess the real damage or vector without a proper postmortem. This led naturally into a broader argument against outsourcing core infrastructure like queues and secrets to third parties at all, framing the whole affair as the inevitable cost of convenience.
So I guess everyone uses LLMs to think now [comments]
68 points · 109 comments · eva.ac · 10h ago
The linked article is a blog post from xyzeva arguing that they don't use LLMs for their security research because the tinkering process itself is the fun part, and they push back hard against AI marketing hype in the security industry. The thread got almost completely derailed by the author's choice to write entirely in lowercase, devolving immediately into a bitter split over whether that's just a legitimate informal stylistic preference or an annoying act of "authenticity signaling" that makes it harder to read. On the core argument, the response was split too—the analogy of people still playing chess for fun despite engines dominating the game came up as a strong defense of the author's position, while others countered that you don't get the luxury of refusing efficiency gains when your job performance is measured on output. A few domain experts jumped in with factual corrections on the specific security examples cited (like the eMudhra key being an intermediate cert, not a root in Chrome), and there was general pushback on the claim that people don't complain about confident wrongness, which many saw as the most common LLM complaint on the internet.
Apple captures 50% of consumer revenue of the top half of the S&P 500 [comments]
68 points · 50 comments · asymco.com · 10h ago
The article argues Apple is the last true consumer company in the top tier of the S&P 500, capturing half of all consumer revenue while the rest of big tech shifted to selling infrastructure to other businesses. The thread quickly tore into the methodology, specifically how the author credited Amazon and Walmart for only the thin margin on their retail sales rather than total revenue, which makes their consumer exposure look negligible. The definitional weirdness extended to classifying Meta and Nvidia as zero-percent consumer companies, ignoring that advertising is entirely consumer-driven and Nvidia still ships massive volumes of gaming GPUs. There’s some sympathy for the broader point about Apple’s unique leverage over the end user, but the overwhelming takeaway is that the statistical sleight of hand turns the whole analysis into a textbook case of lying with data.
Enlightenment 0.28.0 Release [comments]
66 points · 30 comments · www.enlightenment.org · 16h ago
The linked article is the full changelog for Enlightenment 0.28.0, a release that adds NetBSD support, Network Manager integration, and the undeniably wonderful restoration of the Pants module with skirts, kilts, and lederhosen. The HN thread immediately split straight down the middle between people flooded with nostalgia for the e16/e17 days of being blown away by compositing on barely-there hardware, and a reality-check camp who abandoned it long ago because it was crash-ridden and impossible to daily drive compared to modern KDE. A meta-argument over the sheer density of the changelog led someone to actually run it through DeepSeek and post a reorganized version, which the maintainer promptly asked for, and the conversation kept circling back to the bizarre longevity of a window manager stuck at version zero for 26 years. The consensus is that the sheer stubborn dedication keeping it alive is the thread's real story, even if its practical relevance is slimmer than ever.
The Megadonors Who Have Poured $1.2B into the Midterms [comments]
64 points · 53 comments · www.nytimes.com · 8h ago
The New York Times piece lays out how the top 20 megadonors have pumped $1.2 billion into the midterms, with Republicans soaking up most of it from a new wave of crypto and tech billionaires. The discussion barely engaged with the breakdown of who gave what and instead dove straight into the dysfunctional doctrine of *Citizens United*, arguing the numbers show the whole system is fundamentally broken. A sharp divide opened up between the court-packing crowd and a pragmatic subset pushing a novel tax strategy—hitting political donations with a punishing overhead tax to make massive spending counterproductive without triggering a free-speech fight. The conversation also served as a timely fact-check, with several people noting the article was already outdated and that Elon Musk has personally blown past the $90 million cited here to the tune of nearly half a billion. The prevailing sentiment was a resigned acknowledgment that no matter how you slice the law, the scale of money from this handful of people has made a joke of the idea that elections belong to the voters.
Sam Altman says DRAM capacity should go to data centers, not personal computers [comments]
61 points · 77 comments · www.tomshardware.com · 21h ago
Sam Altman makes the claim that it's a waste of global resources for personal computers to have DRAM when it could be in his data centers running AI, essentially arguing for centralized compute over local hardware. HN immediately called hypocrisy, with many pointing out that Altman's own multiple homes sit empty most of the time and that data center GPUs often go unused, making his utilization argument deeply convenient. Others dug into the central-planning implications, comparing this to HOA rules or a landlord demanding tenants give up their spare bedrooms, while a few defenders noted that modern OSes rarely leave RAM idle anyway. The thread landed hard on the idea that this is less about efficiency and more about protecting a subscription business model where you'll own nothing and rent your computing from a cloud oligopoly.
Generated 2026-11-02 08:07 UTC
Generated by Sauron from Hacker News discussions and linked articles.