Hacker News, Distilled

Practical Effects and Model Work

• Commenters love the breakdown of ILM’s miniature work, especially the large Red October model mounted on a hidden pylon.
• Some feel the added explosions and depth-charge effects are visually “intrusive,” but consensus is they’re analog elements composited in, not early CGI.
• The recently purchased large-scale Red October is noted; some argue it’s more than a “prop” given its engineering complexity.

Underwater Filming Techniques

• Discussion centers on ILM’s choice of a smoke-filled “underwater” stage instead of real water.
• Advantages cited: no need for underwater cameras/pools, easier motion-control, and avoidance of scale issues like unrealistic splashes and bubbles.
• Some find these shots occasionally look like “a model in smoke,” and compare them (often unfavorably) to miniatures in Das Boot.
• Other films are mentioned that faked underwater scenes on land for practical or actor-related reasons.

Submarine Design and Life Aboard

• The real Typhoon-class swimming pool sparks surprise and interest; linked videos and photos confirm it existed.
• Explanation: Typhoons’ twin pressure hulls and missile arrangement created unusual internal volume, allowing luxuries like a pool.
• Multiple anecdotes from US and allied navies emphasize how important food, small comforts, and “playing hard” are for morale on subs and ships.

Accuracy, Realism, and Veteran Perspectives

• A former US submariner describes taking cast/crew to sea pre-production and praises the film’s jargon and operational details as largely accurate.
• Some note that, like Apollo 13, the film amps up interpersonal drama compared to the more restrained behavior of real crews.

Film Craft, Performances, and Longevity

• Many say the movie “weirdly” still holds up—visually, tonally, and structurally—despite its Cold War setting.
• The linguistic trick that shifts from Russian to English mid-scene is widely admired.
• Scott Glenn’s captain is especially praised as the definitive on-screen version; Connery’s accent as a Soviet officer is seen as both iconic and implausible.

Source Material, Inspirations, and Related Works

• The real mutiny on the frigate Storozhevoy is referenced as Clancy’s inspiration.
• Early Clancy novels, especially Red Storm Rising, are praised for insight into Soviet/Russian doctrine, spawning a tangent on Cold War vs. modern Russian capabilities.
• Recommended adjacent works include Das Boot, Crimson Tide, The Wolf’s Call, a submarine-effects YouTube series, and a podcast on ’90s political thrillers.

Historical and fascism analogies

• Many see the “Amazon Prime, but with human beings” remark as chilling, drawing explicit parallels to industrialized deportations and the early stages of 1930s–40s European atrocities.
• Commenters argue that past genocidal regimes also began with “deportation schemes” that were normalized and bureaucratized before escalating.
• Some reject the idea that invoking these parallels is hyperbole, asserting that openly fascistic figures and movements are already visible and in power.

Dehumanization and business framing

• The core outrage centers on treating deportation as a logistics problem and people as packages or SKUs.
• Several argue that “evil begins when you treat people as things,” linking this to broader corporate practices where workers are reduced to IDs and productivity metrics.
• The context that the remark was made at a “Border Security Expo” for vendors is seen as making it worse, not better, because it highlights the commercialization of state violence.

Prisons, incentives, and quotas

• Discussion connects ICE to the US prison system: private detention, perverse incentives to keep beds full, and de facto deportation quotas leading to arbitrary or excessive detentions (including tourists).
• Others note that most US prisoners are in public facilities, but forced labor and lack of rehabilitation make the public system little better than private.
• Counting prisoners and undocumented immigrants for political representation while denying them the vote is compared to earlier systems that treated marginalized people as population but not citizens.

Economic and labor impacts

• Some point out that mass deportation would remove workers, consumers, and taxpayers, harming the economy except for contractors who profit from enforcement.
• Others argue that migrant labor is used to suppress wages, and popular anger reflects an economy skewed toward business over workers.
• There is frustration that aging societies objectively need immigrants but still mobilize against them.

International reactions and border culture

• Multiple anecdotes depict US border and visa processes as uniquely hostile and performatively authoritarian compared to Europe or even China.
• Non‑Americans describe rapidly hardening anti‑American sentiment, driven by perceived US hubris, threats toward allies, and erratic leadership.
• Smaller countries feel deeply affected by US internal politics yet powerless to influence them.

Politics and global migration rhetoric

• Commenters note that weaponized fear of migrants is not unique to the US: examples include Brexit, German “remigration” debates, and Australian “stop the boats” politics.
• One view is that US “immigration problems” are largely self‑inflicted: one party makes legal immigration hard; the other blunts enforcement, producing large undocumented populations living and working in limbo.

Root causes and proportionality

• Some argue the US should stop bombing, destabilizing, and exploiting other countries if it wants fewer migrants.
• There is concern that a system supposedly targeting “worst abusers” in practice sweeps up ordinary visitors and residents, making weeks-long detention for visa issues routine and normalized.

Translation tangent (TGV & idioms)

• Several comments debate literal vs idiomatic translation: “train à grande vitesse” maps cleanly to “high‑speed train”; “train of/at great speed” is seen as misleading.
• French speakers note that the preposition “à” often marks a defining property (as in “avion à hélice”), not something easily mapped word‑for‑word into English.
• Similar complaints are made about over‑literal explanations like “al dente = to the tooth”; some argue such glosses confuse rather than teach.

Cost and convenience: trains vs cars (and planes)

• One side argues US rail is mainly a “novelty” and more expensive than driving; they emphasize flexibility of cars and poor US last‑mile transit.
• Others counter that people drastically underestimate total car costs (fuel, insurance, maintenance, parking, depreciation, externalities).
• European examples show wide variation: in some places annual public‑transport passes are clearly cheaper than car ownership; in others (e.g., Switzerland with a full “GA” pass) cars can be cheaper depending on usage.
• Time and “freedom” are framed differently: trains allow work, rest, and comfort; cars require constant attention but offer door‑to‑door travel.

Geography, density, and where rail works

• Debate over whether US size and low density truly preclude HSR.
  • Critics say this is exaggerated “cope”: there are obvious dense corridors (NE corridor, West Coast, Texas triangle, Midwest routes) where rail would be viable but isn’t built.
  • Others stress that outside major metros, both in the US and Europe, public transit is thin and car dependence is high.
• For distances <600–1000 km, many argue high‑speed rail can be as fast or faster door‑to‑door than flying, but this depends heavily on station placement and local networks.

US political, institutional, and cost barriers

• Repeated theme: the core obstacle is not technology but US governance, regulation, and fragmentation.
  • Multi‑layer permitting, environmental review, land acquisition, and litigation drive costs and delays.
  • Federal–state–county fragmentation makes it hard to fund and align on interstate projects; localities often oppose lines that don’t directly benefit them.
• Some point to entrenched car/oil interests, decades of highway subsidies, and cultural car preference as reinforcing factors.
• A few highlight how “hyperloop” and similar ideas were used politically to undermine California HSR support.

Comparisons: Europe, Japan, China, and within Europe

• Commenters stress that even Europe’s HSR network is patchy, often expensive for passengers, and strongly centered on major cities; many regions still lack fast connections.
• Examples like France, Japan, and China show transformative effects where true high‑speed lines exist (city pairs under ~600 km), but also illustrate huge up‑front investment and political commitment.
• Some European countries (e.g., Austria, Ireland, Denmark) are used to show how geography, mountains, or historical underinvestment limit speeds despite rail being culturally accepted.

Specific US and other projects

• Amtrak Acela is noted as TGV‑derived but hamstrung by lack of dedicated high‑speed track.
• Brightline in Florida is cited as the most successful recent US intercity rail, though speed, safety at crossings, and cost are criticized.
• California HSR is intensely debated: some call it the best US rail project under construction; others dismiss it as stalled or misrouted, with route compromises made to appease regional politicians.

Scope of the NIH Freeze and Targeting of Columbia

• NIH is halting both new grants and payments on existing ones; work technically may continue but requires prior NIH approval to access funds.
• Commenters note Brown and Princeton are also affected, and Harvard is under review; one link lists ~60 universities under federal civil-rights investigation.
• People question why Columbia is “first,” suggesting symbolism: elite, New York–based, heavily Jewish student body, and prominent in recent protests. A decades-old Trump–Columbia land dispute is mentioned as a possible personal grudge, but remains speculative.

Impact on Research, Medicine, and Industry

• Many stress that “everyone loses” when public research is disrupted: harm to clinical trial participants, brain drain, and long‑term damage to US industrial competitiveness.
• Some see this as part of a broader pattern: firing or sidelining key personnel at FDA/CDC/USDA (e.g., bird flu vaccination staff), undermining health agencies, and then blaming resulting crises.

Academic Freedom, DEI, and University Culture

• One camp argues US academic freedom was already “dead” due to DEI pledges, culture-war politics, and campus protest dynamics; others see this as overstated or cherry‑picked.
• Non‑US commentators say what really degraded US universities’ global standing is extreme tuition and debt, not campus politics.

Antisemitism vs. Anti-Israel Protest

• Large subthread debates whether anti‑Israel protests are inherently antisemitic.
• Some distinguish clearly: a state vs. a people; criticism of Israeli policy ≠ hatred of Jews. Others argue that certain chants, pro‑Hamas rhetoric, and calls to dismantle Israel cross into antisemitism.
• Several note historical patterns where classic antisemitic tropes (“global control,” “lobbies”) are simply relabeled as attacks on “Israel” or “the Israeli lobby.”
• Others counter that antisemitism is being “weaponized” by this administration to crush dissent and shield Israeli policy from legitimate criticism.

Political Motives and Authoritarian Drift

• Many see the move as punishment for “woke” universities and a warning shot: federal funding is contingent on suppressing certain student speech.
• Commenters frame this as a broader authoritarian project: attacking universities, redefining DEI as “racism,” reshaping language (Orwell/Chomsky “newspeak”), and creating a climate of fear and self‑censorship.
• Several highlight the irony of an administration with documented flirtations with Nazi symbolism and antisemitic figures using antisemitism charges as its main justification.

Revenue Sharing, Equity, and Risk

• Several stories describe swapping unpaid invoices for revenue share or equity, sometimes leading to very large upside, but commenters stress these are rare exceptions.
• Others with long consulting careers say almost no client ever had meaningful profit; in their experience, only immediate cash payment is reliable.
• Some see payment timing as a negotiating lever: earlier, safer cash for less total money vs delayed, riskier payouts (rev share, ownership) for potentially more.

Client Trust, Nonpayment, and Fraud

• Many recount clients stretching terms (NET 30 → NET 60/90) or simply refusing to pay, forcing court action or making lawsuits uneconomical.
• Tactics mentioned include “phoenix” bankruptcies, offloading IP to shells, firing contractors before suing makes sense, and running startups on unpaid invoices.
• A minority report genuinely good-faith clients who renegotiate fairly when cash runs short—but others argue this is unusually lucky.

Startups, Cash Flow, and Loyalty

• Founders and early employees describe going unpaid or underpaid in exchange for partnership, equity, or later backpay; sometimes it worked out very well, sometimes they never got paid.
• Several emphasize that this is only viable if you can afford the risk (e.g., partner income, savings); early-career workers with no buffer simply cannot “work for free.”
• “Loyalty” is frequently used as pressure: some see staying through no-pay as an investment in relationships; others see it as emotional manipulation and a red flag.

Legal Protections and Enforcement

• In some EU countries, a single unpaid employee can trigger insolvency proceedings; wages rank first in bankruptcy. Contractors have weaker tools but can affect credit scores or sell debts.
• In US states like California and Massachusetts, late final paychecks can be very costly for employers (penalties, attorney fees, personal liability), but enforcement can be slow and underfunded.

Unpaid Internships and Early-Career Exploitation

• Unpaid internships are widely criticized as exploitative and primarily benefiting the rich; in tech, they’re almost universally advised against.
• Legal status varies by country and context, with some loopholes (university placements, “volunteering”) and inconsistent enforcement.

Setting Boundaries: No Pay, No Work?

• Many argue for a hard line: once an invoice is overdue or a paycheck is missed, stop working.
• Others advocate situational flexibility—trading short-term pay for long-term relationships, reputation, or equity—but insist this must be an informed, voluntary choice, never an obligation.

What Rapid Storage / Colossus-Based Buckets Are

• New Cloud Storage zonal bucket type colocated with GPUs/TPUs for much higher random-read throughput (claimed up to 20x vs regional buckets).
• Built directly on Colossus’ stateful protocol; a gRPC client is planned that is essentially a thin wrapper over Colossus.
• Targeted at AI/ML workloads and analytics that need very high random-read bandwidth (e.g., large Parquet datasets, LLM training).

Zonal vs Regional / Multi-Region Semantics

• “Zonal” = tied to a single availability zone; may still be replicated but replicas can share failure domains.
• In Google’s terminology, “regional” usually implies transparent multi‑zone replication; “zonal” does not.
• Rapid Storage complements existing regional and dual‑region buckets; users can choose latency/durability/cost tradeoffs via the same GCS API.

Comparison to AWS S3 Express One Zone and Other Providers

• Several comments frame Rapid Storage as GCP’s answer to S3 Express One Zone (low-latency, single‑AZ object storage).
• S3 Express offers much lower latency but is significantly more expensive than standard S3; naming is criticized as misleading.
• Some argue GCP now uniquely offers low‑latency zonal, standard regional, and transparent dual‑region object storage under one consistent API; others counter that S3 has overlapping but not identical multi-region features.

Performance, AI Branding, and Hypercomputer Marketing

• Mixed reaction to marketing: some praise finally exposing Colossus-like capabilities; others see “AI infrastructure” and hypercomputer FLOPS comparisons as heavy spin.
• Confusion over claims that a TPU pod exceeds the world’s largest supercomputer; clarified that Google is comparing 8‑bit AI FLOPS, not traditional 64‑bit supercomputing FLOPS.

Cost, Elasticity, and DIY Alternatives

• Detailed back-of-the-envelope comparisons argue Google’s “HDD prices” rhetoric is overstated versus self-built storage and cheaper cloud storage (e.g., Backblaze, Hetzner).
• Counterpoints emphasize elasticity and operational convenience: instant bucket creation, scaling to TBs then deleting, avoiding hardware deployment/maintenance, and fine‑grained isolation.

Colossus Semantics and Tradeoffs

• Colossus objects are append-only with a single writer; objects can be “finalized” to disallow further writes; no random writes.
• Advocates: dropping POSIX features like multi-writer atomic updates enables much higher performance, cost efficiency, and strong multi‑tenant isolation at scale.
• Skeptics note that such semantics can be hard to retrofit into existing POSIX-based applications, which likely delayed a direct Colossus offering.

Anywhere Cache vs Rapid Storage

• Anywhere Cache = SSD cache in front of normal (often multi‑regional) buckets; improves latency and avoids egress on cache hits.
• Rapid Storage = new bucket type with all data locally stored and fast, including writes, plus fast durable appends—semantics not available in standard buckets.

Adoption and Product Stability Concerns

• Some excitement from users in scientific computing and analytics who expect major speedups from data locality.
• Others caution against early adoption due to Google’s history of killing or reshaping products; recommendation to wait and see if Rapid Storage “sticks.”

Older British systems and wiring oddities

• Several comments reminisce about the older round‑pin BS546 system (2A/5A/15A), still used for lighting in the UK and widely in India.
• Some homes had all three generations at once: small and large round‑pin plus modern BS1363, with very old rubber/cloth insulation and scary fuseboards.
• People report bizarre legacy receptacles in old buildings (mixed standards, ambiguous voltages, abandoned knob‑and‑tube, mystery fuse boxes).

Everyday experience & “stepping on plugs”

• The trope that UK plugs are agony to step on is widely referenced; some say it’s mostly a meme, others have vivid childhood memories of doing exactly that.
• Arguments arise over how often this actually happens and whether “only a careless person” would leave a plug on the floor.

Mechanical design & usability

• Supporters praise BS1363 as solid, hard to dislodge, and suitable for high‑power loads like 3 kW kettles and heaters. The perpendicular orientation lets furniture sit close to walls.
• Critics find it huge, heavy, unfriendly for travel, and dislike that the cord must exit sideways and that grounding is effectively mandatory even for double‑insulated devices.
• Multiple people complain that fuses can overheat or even melt cheap plugs/sockets if loaded near the 13A limit or misused.

Safety features: shutters, fuses, ring circuits

• Shutters keyed by the longer earth pin are widely viewed as a strong child‑safety feature, and less finicky than US “tamper‑resistant” outlets.
• The plug fuse is explained as protection for thin appliance flex on high‑capacity 32A ring circuits, originally a WWII copper‑saving measure.
• Some argue fuses could instead live inside appliances; others note that then damaged cords wouldn’t be protected.
• There is debate over whether ring circuits plus fused plugs are clever engineering or a legacy hack with tricky failure modes.

International comparisons

• Schuko/Europlug: often praised as a good balance of size and robustness; detractors say tolerances vary, some sockets feel loose, and recessed design needs deeper boxes.
• US NEMA: widely criticised for plugs falling out, exposed live prongs when partly withdrawn, and awkward tamper‑resistant designs; defenders point to simpler, smaller hardware and lower‑current circuits.
• ANZ Type I: seen by some as best overall (compact, angled pins give good grip, optional grounding, can route cords parallel or perpendicular); others complain pins bend easily.
• Swiss, Singapore/China, and “universal” multi‑standard sockets are mentioned; universal types are considered convenient but mechanically weak and often lacking shutters.

Voltage, fires, and shock risk

• One subthread questions whether 230 V causes more electrocutions than 120 V; claims about fibrillation vs heart‑stopping effects conflict and remain unresolved.
• Several commenters clarify that fire risk is more about current and bad connections than nominal voltage, with examples showing how resistance heating scales.
• There is disagreement over whether UK’s high‑capacity rings plus plug fuses are safer than lower‑capacity, breaker‑only US branch circuits.

Meta, China, and Advertising

• Several commenters see Meta’s statement (“we don’t operate services in China”) as technically true but evasive, since Chinese firms can still buy ads targeting users abroad.
• Others argue that Chinese companies advertising on a U.S. platform is not itself evidence of leverage or “treason,” and that more concrete links are needed.
• One analogy raised: large advertisers (e.g., pharma) often buy influence over media through ad spend even when audiences can’t directly buy the products.

Whistleblower Claims, Evidence, and NDAs

• Some are frustrated that the whistleblower refers to “documents” without publicly releasing them, questioning credibility.
• Others note she’s likely constrained by NDAs and is effectively inviting Congress to subpoena the material to gain legal cover.
• There’s debate over how much NDAs can really block testimony: some insist NDAs cannot override cooperation with congressional or court proceedings; others point out that the threat of costly litigation and blacklisting is still very real despite legal protections.
• Meta’s alleged demand for punitive damages for each public mention of the company is viewed as a classic intimidation tactic, bordering on SLAPP-like behavior.

Undersea Cable and Factual Disputes

• A key point of skepticism: the allegation of a “pipeline” to China enabling CCP access to U.S. data, when the cited Pacific Light Cable route to Hong Kong was never completed in that form.
• Some see this mismatch as severely undermining her claims; others say her account focused on Meta’s intent and willingness to comply with Chinese demands, even if the project was later altered or blocked.
• Commenters ask for “extraordinary evidence” for any claim that the cable was planned as a deliberate backdoor, and note that many companies and governments were involved for more mundane strategic reasons.

Politics, China, and Double Standards

• Commenters question why certain politicians focus on Meta’s China ties while showing less interest in other major U.S. firms’ deep dependence on China.
• Broader debate emerges over U.S. vs. Chinese corruption, lobbying, and “legalized” influence-buying; some argue the U.S. is highly corrupt in normalized ways.

Adtech, Social Media, and National Security

• Multiple participants stress that, irrespective of China, large adtech platforms already pose serious risks: microtargeting, manipulation, opaque algorithms, and documented real-world harms.
• Some argue these platforms are inherently national-security issues and should be heavily regulated, similar to how TikTok has been treated.
• Others warn that regulating algorithmic amplification and political content risks sliding into state propaganda or heavy-handed control, yet acknowledge the current corporate control and lack of transparency are also dangerous.

Motives and Timing of the Whistleblower

• Skeptics highlight that the whistleblower left Meta years ago and is now releasing a book, reading the hearing as part of a PR tour.
• Defenders counter that fear of lawsuits, financial ruin, and career blacklisting can easily delay whistleblowing, and that lateness doesn’t inherently invalidate the substance—though contemporaneous documentation will be crucial.
• Some view her as partly complicit, given her senior role, but still see the disclosures as socially valuable regardless of personal motives.

Enduring cultural impact & quotability

• Many commenters treat Holy Grail (and broader Python) as near‑required viewing for adulthood, citing life‑changing first encounters and being “unable to take anything too seriously” afterward.
• Lines like “Help, help, I’m being repressed”, “strange women lying in ponds…”, “anarcho‑syndicalist commune”, Castle Anthrax, and interview‑question jokes (“unladen swallow”) are still widely referenced.
• Some note intergenerational transmission (teens quoting it, teachers laughing while classmates look blank), and see that as a marker of a shared but shrinking cultural canon.

Use in education and pop culture references

• One thread debates whether pop‑culture in‑jokes between teacher and student are inclusive; some schools explicitly forbid such references to keep focus on curriculum and “safe space” concerns.
• Others see pop culture (including Python) as a valuable shared language that can engage students.

Has it aged well? Mixed views

• Several argue the films and sketches remain “incredibly funny” because absurdity is timeless.
• Others feel parts of Python and Flying Circus have aged poorly: casual slurs, homophobic or misogynistic bits, and racist stereotypes (e.g., an Australian wine sketch) that still echo unpleasantly in real life.
• There’s disagreement over whether specific material (e.g., blackface, certain slurs) is “casual racism” or being misremembered; some recall being jarred rewatching old sketches.

Transgression, drag, and British comedy traditions

• Cross‑dressing is debated: some find “men in dresses” overused and unfunny; others stress it’s a long‑standing British stage/comedy tradition, often played straight rather than as a joke in itself.
• Commenters note Python’s broader transgression: nudity, attacks on class, religion, and authority; support for gay rights; and Life of Brian’s “blasphemy.”
• One view: they pushed boundaries “of their time” so sensibilities could move forward, even if some material now feels offensive.

Humor norms and online culture

• Multiple comments contrast Python’s celebrated irreverence with Hacker News’s strong discouragement of humor in threads.
• Some see this as necessary to keep discussion from degenerating into one‑liners; others miss spaces where “serious” talk and smart jokes can coexist (citing older Slashdot voting models).

What’s funny, and modern comparisons

• Some people simply don’t find Python funny, describing it as childish or obscure; others explain the appeal as the collision of high intelligence, absurdism, and furious social satire.
• Suggestions for “Python‑like” contemporary or later work include specific sketch shows, radio comedies, and live festival performances, with the caveat that what becomes legendary is often only recognized in hindsight.

Access, metadata, and side notes

• Complaints that much official Python content has disappeared from YouTube, reducing discovery for new audiences.
• A tangent on Hacker News auto‑editing titles (“How X became Y” → “X became Y”) and whether that’s helpful or distortive.
• One anecdote about Life of Brian being mislabeled “PG” on iTunes, leading to unintended early exposure for children.

Scope and Limits of the German Pilot

• Study is seen as good micro‑evidence that extra, no‑strings money improves well‑being and job satisfaction, and doesn’t make this specific group “lazy.”
• Many point out it’s not real UBI: small sample, three‑year horizon, not universal, and only middle‑income 21–40‑year‑olds living alone.
• Critics say behavior under a temporary experiment differs radically from behavior under a lifelong, guaranteed income; you don’t scrap a career for a 3‑year payment.
• Methodological questions raised: selection bias (only people with prior income and limited unemployment), how dropouts were handled, and whether attitudes toward UBI skewed the sample.

Work, Motivation, and Behavior

• Some cite experience of COVID benefits or early “retirement” as a personal golden age of creativity and productivity.
• Others counter with pensions and disability programs: once a guaranteed income appears, many do stop working; raising retirement ages is cited as evidence UBI wouldn’t be self‑financing.
• Debate over human nature: one side expects large numbers to coast if they can “just get by”; the other notes most people seek purpose, volunteer, or change jobs rather than do nothing.

Inflation, Housing, and Macroeconomics

• Biggest concern: UBI simply bids up prices, especially rent, canceling its benefit. Landlords capturing UBI via higher rents is a recurring fear.
• Proposed fixes include land value taxes, aggressive upzoning, and price/rent controls on essentials; skeptics say that drifts toward central planning.
• Others argue if UBI is tax‑funded, total money supply needn’t grow; redistribution toward poorer households can raise their real consumption even with some inflation.

Freedom, State Power, and Ideology

• Strong strand of distrust: UBI is framed by some as “signing your soul away” to a domineering state, with historical references to Soviet social control and self‑censorship.
• Opponents prefer reliance on markets and employment; supporters reply that corporations already control survival for many and the state already conscripts and regulates without UBI.
• Several note that generous redistribution schemes seem to require high social trust; policies can also erode that trust over time.

Alternatives and Complements

• Alternatives floated: universal basic services/provisioning (housing, food, health, education) instead of cash; job guarantees and large public works instead of income guarantees.
• Some see UBI as one piece of a larger transformation (wealth taxes, de‑financialized housing, stronger public services); others regard that as a stealth push toward socialism.

Positioning and Rebranding

• Widely recognized as a rebrand of Project IDX: a cloud/web IDE backed by Nix-based containers and a VS Code–like editor.
• Framed as Google’s entry in the “AI IDE / vibe coding” race, competing with Cursor, Replit, v0.dev, Lovable, Windsurf, Devin, etc.
• Multiple comments highlight confusion and overlap with other Google offerings (Google AI Studio, Vertex AI Studio, Firebase, Android Studio + Gemini), and recurring mistrust due to Google’s product graveyard.

AI Coding Experience and Reliability

• Many users report poor outcomes: failing to build simple apps (todo lists, chess clock, Angular SSR + i18n), generating broken configs, syntax errors, and getting stuck in “infinite loops” of failing fixes.
• For Flutter specifically, commenters from Google note that app-prototyping isn’t wired up yet, which explains some failures; workarounds via templates + chat exist but still fragile.
• Some say the AI feels like early GPT‑4: good-looking UIs that don’t work, missing core logic (e.g., I Ching coin-flip logic), and frequent need to manually debug and guide it.
• Others report strong successes for small, greenfield tasks: a Chrome extension in ~30 minutes, classroom utility apps, and praise for the underlying Nix environment, previews, and GitHub integration.
• There’s broad agreement (from experience with other tools too) that these systems work best when treated like a junior developer: step-by-step prompts, constant review, and willingness to revert or intervene.

Comparisons to Alternatives

• Several see it as significantly behind v0.dev, Replit, and specialized tools like Cursor/Zed for getting “real work” done.
• v0.dev is praised for one-shot pages plus one-click Vercel deploys; Lovable/Windsurf get positive mentions when carefully driven stepwise.
• Some note Gemini 2.5 performs very well in Google AI Studio, but Firebase Studio doesn’t seem to expose that “thinking” configuration clearly.

Lock‑in, Privacy, and Billing

• Concerns about “next level vendor lock-in”: AI + Firebase + Google Cloud as a tightly coupled stack.
• Worries about surprise billing and opaque Firebase/GCP pricing; others clarify Studio itself is free but prompts/data can be used for model training.
• Several commenters say they avoid environments tied to a single LLM provider given how quickly “best model” changes.

Firebase Ecosystem Feedback

• Multiple users say they’d rather have basic Firestore admin improvements (multi-select edit, simple text search) than AI layers.
• Firebase authentication is still seen as exceptionally easy and compelling, but lack of native search forces extra infra (Algolia/TypeSense, etc.).

AI‑Built SaaS and Perception

• A detailed anecdote about a non-technical founder building a full CRM with AI tools spurs debate:
  • Admiration for how far an individual can get in months using Replit/Cursor + LLMs.
  • Sharp criticism around security and PII handling by minimally technical builders.
• Some founders deliberately hide “made with AI” branding, fearing reputational harm and attackers probing for LLM-induced vulnerabilities.

UX, Design, and Misc

• Mobile Safari experience is poor (broken preview), disappointing to people who “vibe code” on phones.
• Some dislike the new Firebase Studio visual design, calling it chaotic and aesthetically worse.
• General sentiment: promising concept and infrastructure, but current AI behavior is inconsistent and often not production-ready compared to leading competitors.

Plans, Pricing, and Perceived Value

• Max is seen as mostly “more of the same”: higher rate limits and output sizes at $100–$200/month, but no clearly new capabilities or models.
• Several commenters say $200/month is reasonable for heavy professional use or vs API spend; others say that’s more than their monthly groceries and not justifiable for “no new features.”
• Many want capability-based tiers (e.g., Deep Research‑like tools, better coding, large-context or voice) rather than pure usage tiers.

Usage Limits, Reliability, and Transparency

• Strong frustration that “more usage” and free/pro limits are vague and seemingly dynamic; users feel limits have quietly tightened on Pro, possibly to make Max more attractive.
• Reports of frequent downtime, dropped responses, and “overloaded” API errors; some canceled Pro specifically over reliability.
• “Priority access during high traffic” is interpreted as an admission that lower tiers will be throttled more as higher tiers arrive.

Claude vs Competitors

• For coding:
  • Some say Claude 3.7 with extended thinking is roughly on par with o1‑pro for many tasks.
  • Others find Gemini 2.5 Pro significantly better for large codebases, long context, and test-passing reliability; still others say o1‑pro clearly wins.
  • A subset feel Claude 3.7 has regressed (“dumber,” over‑refactors, misunderstandings).
• Gemini 2.5 Pro + Deep Research is repeatedly cited as strong value, often making OpenAI/Anthropic subscriptions feel unnecessary.
• Claude is praised for design help and UI features but criticized as over‑censored by some.

Clients, UI, and Developer Workflows

• Some like Claude’s web UI (speed, diagrams, search in chats) and project organization; others think recent redesign broke basic UX (scrolling, selection, copy/paste).
• Heavy coders gravitate to editors and wrappers (Cursor, Roo Code, Cline, etc.), where API usage can easily reach tens or hundreds of dollars per week.
• Several argue that for $200/month, people are better off with pay‑as‑you‑go API plus their own UI, unless Max meaningfully improves reliability and exclusives.

Open Source and Self‑Hosting

• Rising subscription prices push interest toward open models (DeepSeek, Qwen, QwQ) and self-hosting with UIs like OpenWebUI, LibreChat, LobeChat.
• Commenters note that even open models require substantial hardware, so “free” isn’t actually free, but the control and predictability are appealing.

Market Reaction & Suspected Manipulation

• Commenters note huge intraday moves (many stocks +10–15%, options multiples) and argue anyone with advance knowledge could have made “zero-risk” profits.
• Widespread suspicion that insiders, political allies, or hedge funds were tipped off; several point to the “great time to buy” social post hours before the pause.
• Some think regulators could, in theory, unwind suspicious trades, but many doubt the SEC/DOJ will act in the current political environment.

Legality, Immunity & Conflicts of Interest

• Debate over whether public posts encouraging people to “buy” ahead of policy moves constitute market manipulation.
• Discussion of presidential immunity, “official acts,” and how recent court rulings and gutted enforcement (SEC, DOJ Public Integrity) make accountability unlikely.
• Historical contrast with earlier presidents placing assets in blind trusts and avoiding even small conflicts, versus today’s open entanglement with markets, crypto, and a public company.

Policy Whiplash & Business Planning

• Many emphasize that a 90‑day pause doesn’t fix the core problem: extreme volatility and unpredictability in trade policy.
• Businesses are seen as unable to plan factories, supply chains, or hiring when tariff levels can change within days; several expect postponed investment and layoffs.
• Personal investors oscillate between “buy the dip” opportunism and moving to cash or foreign assets to ride out the chaos.

Economic & Geopolitical Risks

• Commenters worry about lasting damage to U.S. credibility: allies questioning America as a dependable partner and reserve-currency issuer.
• Concerns that massive China tariffs (125%) remain, pushing supply chains into opaque transshipping and possibly tightening China–EU ties.
• Bond market turmoil and elevated yields are viewed as the real constraint that forced the partial retreat.

Trump’s Strategy: 4D Chess or Chaos

• One camp frames this as “Art of the Deal”: set extreme tariffs, then retreat to 10% to force negotiations and isolate China.
• Others see pure incompetence, ego, or a wealth-transfer grift, arguing the backtrack shows no coherent plan and deeply undermines trust in U.S. policy.

Foundational Skills vs. “Cheating with AI”

• Many comments argue the report understates how often Claude/LLMs are used to bypass learning rather than “collaborate.”
• Numerous anecdotes from CS and math courses: students paste assignments and error messages into ChatGPT, turn in code they can’t explain, and hit a “wall” in advanced courses because they never learned basics (structs, pointers, data types, etc.).
• Others see value in AI as a tutor/“spotter” for debugging or hints, but stress it must come after genuine struggle; otherwise students lose the “productive struggle” that builds deep understanding.
• There is concern about a generation that can only operate via prompt engineering, with little internalized knowledge or problem‑solving ability.

Universities, Assessment, and Adaptation

• Several argue universities should embrace AI as inevitable but radically change assessment: more in‑person, proctored, or handwritten exams; oral exams and tutorials; open‑book/AI‑allowed exams designed so only those who truly understand can pass.
• Others propose:
  • Homework largely ungraded or purely formative.
  • Grades based mainly on supervised exams and live performance.
  • Dual tracks: some tasks explicitly no‑AI, some explicitly AI‑enhanced and graded differently.
• There is pushback against “watering down” standards: degrees only remain meaningful if institutions actually ensure graduates can do the work without crutches.

Productivity, Cognition, and Long‑Term Effects

• Strong disagreement on whether LLMs increase productivity: some programmers report dramatic speedups; others say review, debugging, and lost understanding outweigh any gains.
• Broader worry that offloading thinking (not just lookup or arithmetic) will erode critical reasoning, memory, and metacognition—analogies to calculators, GPS, spell check, and search, but with higher stakes because LLMs sit closer to “thinking” itself.
• Counter‑argument: tools have always shifted which skills matter; perhaps humans should focus on higher‑level reasoning while AI handles routine parts—though it’s unclear whether that actually happens in practice.

Anthropic’s Report and Incentives

• Multiple commenters see the report as “AI‑washing”: category labels like “create and improve educational content” or “provide solutions for assignments” could easily hide large‑scale cheating.
• Noted conflict of interest: Anthropic both sells into universities and frames usage as largely benign; it has incentives to under‑report direct essay‑writing, test‑answering, and plagiarism.
• Methodological concern: using an LLM to classify millions of student chats imports LLM unreliability into the analysis, making fine‑grained distinctions (e.g., practice vs. cheating) dubious.

LHC Track Record and Expectations

• One camp sees the LHC as underwhelming post‑Higgs: no SUSY, no clear “new physics,” and questions whether a bigger machine is just “hoping something shakes out.”
• Others argue the LHC produced many important results: precise measurements, many hadrons, and statistically significant hints of Standard Model deviations.
• The absence of expected SUSY or other particles is itself viewed as a meaningful (if negative) discovery that killed dominant BSM scenarios, though some call this “weak” progress given the cost.

Purpose and Design of the FCC

• Clarification that the flagship near‑term project is an electron–positron collider (FCC‑ee) for precision Higgs and electroweak measurements; only later (around 2070) would the tunnel host a 100 TeV proton collider (FCC‑hh).
• Supporters emphasize:
  • e⁺e⁻ collisions at high energy give much cleaner measurements than proton–proton.
  • The main aim is to nail down anomalies and parameters in the Standard Model, not blindly hunt for unknown particles.
• Critics point out that official FCC documents do still talk about possible discoveries of dark matter candidates, axions, SUSY partners, etc., and see that as speculative padding.

Theory vs Experiment

• Debate over whether “theory should precede experiment.”
  • Some argue building a huge machine without a tight set of target theories is bad science and bad value.
  • Others note historically experiment has often led theory (e.g., anomalies prompting new particles), and that precision tests of an incomplete theory are valid science.

Cost, Funding Priorities, and Opportunity Cost

• Concern that a multi‑tens‑of‑billion project will “strangle” other areas of physics and neighboring fields for decades; suggestions to prioritize theory, novel accelerator R&D, or other sciences (materials, cancer research).
• Counter‑argument: spread over ~12 years, ~1–1.3B/year is modest versus national budgets and various political boondoggles; if we can waste that on less productive things, why not on core physics?
• Some former insiders now question whether such distant-from-application science is the best use of constrained university and research budgets.

Media, Public Perception, and HN Culture

• A popular YouTube critic of particle physics is heavily cited; several comments accuse her of algorithm‑driven contrarianism and oversimplification, while others defend her long‑standing skepticism.
• A linked Nature piece is used to show that doubts about the FCC are mainstream, not just YouTube drama.
• Meta‑thread laments perceived “anti‑intellectualism” and clickbait‑driven understanding of complex scientific policy.

Linux installers and everyday use of chroot

• Many distros rely on chroot internally or expose helpers:

  • Arch: arch-chroot (plus pacstrap) used for installs and recovery; Manjaro inherits similar behavior.
  • Gentoo: classic stage3 and handbook use chroot; people compile on a powerful machine in a chroot, then copy the result to weaker/remote machines.
  • Debian: debootstrap installs a minimal system into a directory (setup-time), then chroot is used for configuration; similar approach used for cross-arch image building.
  • NixOS: nixos-enter provides a chroot-like entry into NixOS roots.
  • Void, Linux From Scratch, and others also teach or rely on this approach.

• Common real-world uses:

  • Unbricking laptops, fixing failed boots, installing missing drivers, and general “rescue partition”/live-USB workflows.
  • Remote or mass provisioning by preparing disks in another machine via chroot.

---

Cross-architecture and advanced tools

• Using qemu-user + binfmt, users routinely chroot into foreign architectures (ARM phones, Raspberry Pi, Nvidia Orin BSPs, router/IoT firmware).
• This avoids cross-compiler complexity at the cost of runtime overhead; often “fast enough” on modern hardware.
• systemd-nspawn is frequently suggested as a “nicer chroot”:
  • Handles many mounts automatically and can --boot into a rescue target.
• Other tools mentioned: LXC, Bubblewrap, machroot, proot-docker, junest, container-shell.

---

Desire for a “better chroot” vs containers

• Some wish Linux had evolved chroot into a simple, first-class sandbox primitive instead of the Docker/cgroups/OCI stack.
• Others argue that Docker/Podman add important extras: images, overlayfs, networking, user namespaces, resource control.
• Comparisons to:
  • FreeBSD jails, DragonFly vkernels, Solaris/illumos Zones as “chroot++”.
  • Plan 9’s capability-style “mount what you need” model.
• Discussion of Linux namespaces:
  • Complaints that sandboxing is subtractive (start with everything, then drop) rather than additive.
  • unshare, mount namespaces, pivot_root suggested as safer/more controlled than raw chroot.

---

Limitations, gotchas, and best practices

• Correctly binding /proc, /sys, /dev, /dev/pts, /run is a common stumbling block; wrappers like arch-chroot exist mainly to automate this.
• Tools that accept a root-path (-R in Solaris-style tools) are seen as more robust than requiring a chroot when the target filesystem is broken or missing libraries.
• Chroot is not a strong security boundary; modern containment typically layers namespaces, cgroups, or full VMs.
• For remote reimaging of a root disk, people point to VM-based approaches or specialized tooling like nixos-anywhere.

Security, Model Theft, and Black-Box Concerns

• Many wonder if deployments will rely on confidential VMs and encrypted GPUs (SEV-SNP, TDX) to protect model weights; some speculate these could be eventually broken.
• Opinions split on leak risk: some say only large enterprises can run models this big and won’t risk lawsuits; others note it only takes one hacked org or state actor for weights to escape.
• Ideas floated: watermarking weights and heavy contractual liability to deter leaks.
• For air-gapped government-style deployments, insiders with SSH access are seen as a key exfiltration vector.

Privacy, Trust, and Government Access

• Strong skepticism toward a “black box in your DC” from a US advertising company, with fears of phoning home or covert access by intelligence agencies.
• Others respond that big customers can and will strictly monitor or block traffic, and that Google has incentives and contracts not to sabotage or spy on Fortune 50s.
• Debate over US government surveillance (FISA, PRISM, NSA tapping) leads some to argue that if it’s network-connected and run by Google, you shouldn’t treat it as fully private.
• Air-gapped Google Distributed Cloud offerings for Secret/Top Secret missions are cited as evidence this can be made offline.

Hardware Choices and TPUs vs Nvidia

• Notable that on-prem boxes use Nvidia Blackwell GPUs, not TPUs.
• Explanations offered: CUDA familiarity/portability for customers, limited TPU supply, and desire to keep TPU advantages (cost, efficiency, long context) inside Google’s own cloud.
• Gemini is said to be implemented in JAX/XLA, so it can target both TPUs and GPUs, though performance and cost differ.

Target Customers and Regulatory Drivers

• Seen as aimed at governments, defense, intelligence, banking, healthcare, and large financial firms with strict data rules or entrenched on-prem estates.
• Some argue true “must-be-on-prem” requirements mostly exist in government/adjacent sectors; others highlight broad corporate fear of data leaving the network, especially in the EU.
• A bank employee notes they’re currently banned from AI over privacy, suggesting strong demand.

Data, Training, and “LLM Slop”

• Discussion of Google’s proprietary data (Search, YouTube, Books) vs Common Crawl and LibGen: some see a moat, others emphasize that data quality, not human vs synthetic origin, matters.
• Concerns about Common Crawl being increasingly contaminated by LLM output; counter-argument is that filtering LLM output is just another quality-filtering problem, though some point to model collapse when training on LLM-generated data.

Business Strategy and Comparisons

• Some call this “government contract baiting” and a way to push Google Distributed Cloud, not just GCP.
• Debate over whether using Google Cloud is effectively supporting an “ad company,” versus a now-profitable, separate cloud org.
• Parallels drawn to the old Google Search Appliance: a mysterious but often better-than-alternatives yellow box, raising questions about opacity and logging.

Alternatives and Competition

• DeepSeek is mentioned as a contrasting model-you-can-self-host; however, it’s seen as not in the same capability class as top Gemini models and lacks enterprise support contracts.
• Microsoft’s Copilot and Azure’s earlier government approvals are noted as key competitive pressure.

Licensing and Pricing Model

• Uses a BSL-style license with limits (e.g., effectively one instance per service; broader features in the hosted offering).
• Some are wary of the delayed-open license and would prefer more permissive alternatives.
• “Maincloud” / “Energy” credit billing feels crypto-ish or needlessly abstract to several; comparisons made to Snowflake/Vercel units.
• Bandwidth pricing (~$0.28/GB) is called out as high, especially at scale; request for flat-rate options and fear of accidental large bills.
• Team says prices are conservative and intended to be lowered later.

Core Architecture and Goals

• In-memory relational DB with WASM “reducers” (stored procedures) meant to host all game/server logic.
• Promises: automatic client–server sync, elimination of server–DB sync layer, and simple deployment (upload WASM to the DB).
• Compared to “server in the DB” systems like Convex, and to in-memory SQLite plus a networking layer.
• Seen as a “universe brain reorg”: powerful if you fully buy into the new mental model, but requires rethinking architecture.

Scalability and Clustering Concerns

• Marketing claims very low latency and high throughput; commenters ask for concrete benchmarks and cluster design details.
• Internally, they say each DB is an actor in an Erlang/Elixir-style distributed system with inter-module communication; that piece isn’t yet public.
• Some argue that real MMO scaling is more about sharding, geography, and what you trust the client with than about raw DB speed.

Fit for Game Development and Netcode

• Many multiplayer devs say main challenges are lag masking, prediction, rollback, and cheat resistance, not CRUD or persistence.
• Current product doesn’t natively solve those; team plans future automatic client-side prediction by re-running server WASM on clients and reconciling.
• Skeptics note this is still a “TODO” without shipped examples; optimists see value even if lag masking remains app-specific.
• Several worry you’d need to significantly restructure servers, and possibly reimplement physics/pathfinding/animation outside Unity/Unreal, to fit the model.

Physics and Simulation in the DB

• Docs show very naive collision checks and assert “DB is fast” without numbers; physics/netcode specialists are unconvinced.
• Community experiments embed a Rust physics engine (Rapier) into WASM, then write results back into tables; interesting but early.
• Open question whether serious networked physics and engine integration are practical within this paradigm.

Developer Experience and Adoption

• Mixed feelings on tying schema directly to language structs/annotations (ORM-style) vs explicit SQL.
• Some report building small MMOs quickly and like the live query / push-based model.
• Others see significant risk in adopting a young, unconventional stack from a team whose flagship MMO isn’t yet released, citing past MMO backend hype that fizzled.

Formats: roff/mdoc vs HTML/Markdown

• Several comments argue roff/mdoc is already a “modern” semantic format compared to HTML, which is easy to parse but unpleasant to write, and Markdown, which is familiar but largely semantics‑free.
• Advocates of mdoc emphasize its explicit markup for flags, arguments, env vars, etc., which powers apropos/whatis queries, smart search, and hyperlinking. Converting to Markdown or plain HTML often discards this information.
• Others prefer authoring in Markdown/RST/AsciiDoc and generating man pages via tools (pandoc, Sphinx, Asciidoc, scdoc), but there is concern about:
  • Many incompatible Markdown dialects.
  • Loss of semantics and links.
  • Needing increasingly complex “markdown schema” and tooling that just re‑creates roff’s complexity.
• GNU info/texinfo is mentioned as an earlier attempt at richer, linked docs; technically good but hampered by poor standalone viewers and unfamiliar UI.

Readers and UX: terminals, editors, and GUIs

• Many suggestions to fix the reader, not the format:
  • Use man --html or web manpage mirrors; OpenBSD’s mandoc HTML is praised.
  • Emacs (man/info/woman), Vim/Neovim as MANPAGER, KDE Help Center, Dash/Zeal, pinfo, xman, and bat‑based pagers add links, search, highlighting, or nicer layout.
  • mandoc on BSD preserves semantics into less (ctags‑style jumps) and into HTML, including links directly to option definitions.
• Pain points with traditional man | less:
  • Line wrapping destroys indentation; no semantic “jump to flag” (people rely on regex searches like ^\s+-p).
  • Long pages (e.g., bash, tar) are hard to navigate; some print to PostScript/PDF (man -t) to see intended structure.
• Some highlight a skills gap: job control, less navigation, and info usage are not widely understood by newer users.

Content quality and examples

• Strong split:
  • Some praise manpages (especially BSD) as cohesive, dense, and sufficient when well‑written.
  • Others say manpages “suck”: too much dry option listing, too few examples and workflows, leading to tldr/cheat/bro pages, Stack Overflow, and now LLM wrappers.
• There is broad agreement that:
  • The framework is fine; the art of writing good documentation is the real bottleneck.
  • More examples, single‑page versions, and good printability would make man‑style docs much more useful.

Conventions and ecosystem

• Debate over relying on --help vs man:
  • --help is convenient but inconsistent, may require running the program, and often misuses stdout/stderr.
  • Manpages are more standardized across Unix systems, especially under distro policies.
• Underlying theme: the Unix/GNU/Linux documentation ecosystem is powerful but fragmented; attempts to “modernize” risk throwing away existing semantics unless they carefully preserve them.

Manufacturing Cost, Markups, and Nike Example

• The thread accepts that factory cost of a $100 retail Nike is relatively low (≈$20–25 landed), but stresses most of the price is downstream: brand, design, marketing, retail overhead, and risk.
• Retailers often work on percentage markups (commonly near 100% on wholesale), not fixed dollar margins, because many costs (inventory financing, insurance, shrink, returns, customer service) scale with ticket price and time-in-inventory.
• Some argue this “percentage logic” is not inevitable and partly reflects investor expectations and capital costs, not physical necessity.

Tariffs and Consumer Impact

• A key point: a 100% tariff at the border is magnified by percentage markups, so a $23 duty can push a $100 shoe toward $150, not $123.
• Several commenters liken tariffs to a hidden federal sales tax that consumers ultimately pay, with little net benefit beyond higher prices and reduced demand.
• Others say if tariffs were applied as a retail sales tax instead of an import duty, the final price increase could be smaller for the same tax revenue.

Onshoring vs. Self‑Sufficiency

• One side: bringing shoe manufacturing back to the US is economically “a fool’s errand” in a high-wage, low‑interest, automation-driven environment; tariffs won’t create many good jobs and will likely be rolled back under voter pressure.
• The other side: some degree of domestic production in critical sectors (chips, energy, defense, medical supplies, sometimes food) is strategically valuable, even if it’s costlier. Sneakers themselves are viewed by many as non‑strategic.

Automation, Jobs, and “Bullshit Work”

• Manufacturing that stayed in rich countries is already highly automated or very high value-add; new factories will be robot-heavy and employ far fewer people.
• Some argue “bringing back manufacturing jobs” is nostalgia; the real future work is designing, maintaining, and reconfiguring automated systems.
• A long subthread debates whether society should keep organizing around “jobs” at all versus some form of basic support or publicly funded creative work.

China, Global Trade, and Geopolitics

• Multiple comments stress that “Asia = cheap shoe labor” is outdated: China in particular is now competitive or leading in EVs, batteries, robotics, and other advanced tech, and rapidly automating factories.
• There’s disagreement over how much China’s demographic decline and political risks offset its strengths.
• Some see deep trade ties as stabilizing (“mutually assured economic destruction”), others as dangerous dependence that can be weaponized in a crisis (e.g., Taiwan).

Retail Concentration and Margins

• Several note that modern retail and cinema chains replaced small shops, increasing pricing power once local monopolies were established.
• Others counter that net profit margins for big retailers remain single digits once rent, wages, and logistics are included; high per‑item markups shouldn’t be confused with windfall profits.