Hacker News, Distilled

Why parsing ls is discouraged

• Core objection: ls is a human-oriented formatter; turning its output back into machine data is fragile and needless extra work.
• Typical anti-pattern: for f in $(ls) converts a list of filenames to a single string, then re-splits it, breaking on spaces, newlines, and control characters.
• Shell already exposes directory data structurally via globs (for f in *) and tools like find, so parsing ls is considered a “wrong obvious solution” beginners gravitate to.

---

Preferred Unix-side alternatives

• Use shell globs with options like nullglob / failglob to handle missing matches robustly.
• Use find for recursion, filtering, and safe execution: -exec, -print0 | xargs -0, -printf for custom formats, and sometimes -regex.
• For tricky pipelines, while read -r -d '\0' loops plus find -print0 are recommended, though commenters note these quickly become complex and easy to get subtly wrong.
• Some argue “always use find rather than for + glob” for serious scripts.

---

Structured/modern shells and higher-level languages

• Several posters advocate PowerShell, Nushell, or object-stream shells (e.g. Python-based) where ls returns structured objects instead of text, eliminating many parsing issues.
• Enthusiasts highlight easier filtering, sorting, and type-aware pipelines; critics say these shells create separate ecosystems, lack traditional features, and aren’t installed everywhere.
• Strong pro-Python contingent: shell is seen as a brittle glue language; Python (or similar) is preferred once scripts stop being trivial. Others counter that shell is faster to start, ubiquitous, and often “good enough”.

---

Filename edge cases and robustness

• Debate centers on how much to care about filenames with spaces, newlines, control chars, or odd Unicode.
• Some insist robust tools must handle all legal bytes except / and NUL; others argue such names are rare in their domains and not worth the complexity.
• There is interest in restricting filename character sets where possible (e.g., internal systems) but recognition that networked filesystems and other platforms prevent relying on that globally.

---

Standardized machine-readable output and teaching

• Multiple people want a uniform --json (or CSV-like) output mode for core utilities (ls, find, df, mount, stat, etc.) to simplify safe parsing; others respond that real programs should call APIs like readdir instead.
• Skepticism that a coordinated JSON standard across all Unix tools is realistically achievable.
• For teaching, ls | ... is acknowledged as a natural early pattern; later, instructors must “unteach” parsing ls and introduce globs, find, and safer patterns.

Dynamic / Surge Pricing Debate

• Many worry ESLs will enable “surge pricing” (e.g., higher prices for water or ice cream on hot days), seen as exploitative, especially for essentials.
• Some defend demand-based pricing as standard economics that helps allocate scarce goods and prevent shortages.
• Others counter that it doesn’t increase supply, just shifts goods to those with more money and raises profits on the same inventory.

Fairness, Trust, and Morality

• Several comments distinguish small vendors (e.g., umbrella sellers or flea-market stalls) from large chains, arguing the latter’s use of dynamic pricing feels more immoral due to scale and market power.
• There’s concern that opportunistic price hikes erode trust and foster antagonistic behavior (returns abuse, chargebacks, even theft).
• Some suggest competitive markets would push back: a retailer could undercut surge prices to win goodwill and loyalty.

Personalized / Discriminatory Pricing

• Strong fear that ESLs plus tracking (phones, Bluetooth, gait analysis) will lead to individualized pricing based on perceived wealth.
• Existing loyalty programs and geo-based pricing are cited as early forms of price discrimination; some note the possibility of “capturing consumer surplus.”
• Others argue competition and razor-thin grocery margins limit how far this can go.

Legal and Consumer Protection Issues

• Big concern: prices changing between shelf and checkout. Many note laws requiring the shelf price to be honored, or at least constraining how often prices can change.
• Proposals include: limiting automatic price changes per day, grace periods where the lowest recent price must be charged, or mandating that in-day changes can only go down.
• Some highlight weak or uneven enforcement of pricing laws (missing tags, mismatched sale prices).

Operational Reality and International Experience

• Multiple commenters from Europe and New Zealand say ESLs are already common and generally used for overnight updates and markdowns near expiry, not high-frequency dynamic pricing.
• Some think US lag is mostly inertia, not technical barriers.

Potential Benefits and Other Uses

• Benefits cited: reduced labor and errors in changing paper tags; easier markdowns; better handling of expiring goods.
• ESLs could support richer product information via scanning, though skeptics say UPCs already allow this without new hardware.
• Hobbyists mention reusing second-hand labels with open-source projects.

Role of Ruby in Shell Scripting

• Many commenters like Ruby as a “next step” when Bash becomes unwieldy: better data structures, cleaner syntax, easier to grow from one-liners into structured scripts.
• Ruby’s backticks and stdlib (e.g., Open3, Pathname, threads) are seen as strong primitives for system scripting.
• Some prefer mixing shell with embedded Ruby for the “hard bits,” though this can reduce readability.

Availability & Deployment Concerns

• A major objection: Ruby usually isn’t installed by default on common Linux distros, BSDs, Docker base images, or locked-down corporate environments.
• For many, “already on the system” is decisive; this favors sh/Bash, Python, or Perl.
• Others argue that installing Ruby via the package manager is trivial and shouldn’t block its use, but several note bureaucratic or air‑gapped environments where adding interpreters is non‑trivial.

Comparisons with Other Languages

• Perl is frequently cited as Ruby’s spiritual predecessor and still excellent for scripting, with the advantage of near‑universal availability and CPAN.
• Python “won the war” for non‑Bash scripting on Linux; even people who prefer Ruby often default to Python because it’s standard.
• Go and Rust are praised for single static binaries and predictable deployment; many modern CLI tools use Go for this reason despite less “scripty” ergonomics.
• Some mention D, Julia, Scala (Ammonite), Elixir, Nushell, and Java as alternative “scripting” environments.

Performance & Startup Time

• Older Ruby versions were widely perceived as slow; newer Ruby with YARV/YJIT is said to be much faster, comparable to or better than CPython in some benchmarks.
• There is disagreement about Ruby’s startup time: some see ~100ms as a serious drawback for small scripts; others report much faster startup or consider it irrelevant when shelling out to other programs.

Large Projects, Typing, and Tooling

• Several criticize Ruby (especially with Rails) for dynamic typing, heavy metaprogramming, global namespace issues, and implicit loading, which they say make large codebases hard to reason about.
• Others counter with successful long‑lived Ruby systems and argue that discipline, not language, is the main factor.
• Editor/tooling: Python is seen as having stronger LSP and IDE support (especially in VS Code); Ruby tooling is improving but is reported as less smooth and sometimes fragile.

Dependencies & Inline Deps

• Inline dependency mechanisms in Ruby (Bundler inline) are praised as powerful for single‑file scripts, but some worry about version pinning and reproducibility.
• Parallel discussion highlights similar emerging features in Python (PEP 723), Rust, and other ecosystems.

Scope of the Bias Finding

• Commenters see ChatGPT’s ableist behavior as unsurprising, paralleling known racial and gender biases in AI.
• Some test gender bias ad hoc (changing names/pronouns on identical CVs) and report no difference, but others point out this doesn’t test for cultural or name-based bias.
• Many emphasize that “don’t be biased” prompts may simply flip the bias or overweight disability signals rather than truly neutralize them.

AI as a Mirror of Human Bias

• Strong theme: LLMs reflect patterns in human data; if society and hiring are biased, the model will be too.
• Some argue this is “correct” in a statistical sense (it models reality), while others stress that “correct” for prediction is not “acceptable” for hiring or ethics.
• Several note this makes AI a powerful mirror of unconscious prejudice, but also a tempting way for organizations to outsource and deniably automate discrimination.

Technical and Methodological Limits

• Discussion of how instructions like “focus on disability justice” can swing rankings too far the other way, creating a new distortion.
• Multiple comments highlight that LLMs are black boxes, not transparent decision systems, and that deep learning makes detailed explanations of decisions essentially impossible.
• Others suggest limiting AI to narrow tasks (e.g., extracting structured data) and feeding that into explicit, auditable rules instead of free-form ranking.

Legal and Policy Concerns

• Debate over whether disparate impact (biased outcomes without explicit intent) is enough for legal liability; some say it’s hard to win in court, others note AI makes bias easier to empirically demonstrate.
• Concern that agencies and companies will adopt these tools quickly (e.g., child protection, housing, hiring) without understanding or controlling bias.

Disability, Productivity, and Hiring Practices

• Some question whether anti-discrimination laws “fight reality” if some disabilities affect productivity; others respond that reasonable accommodations often neutralize impact.
• Anecdotes: people omitting disability-related achievements or even “American Sign Language” from CVs due to suspected screening penalties.
• Several disabled or blind developers describe the dilemma of when to disclose and how little resumes convey about actual capability.

LLM Vocabulary “Tells” in Scientific Writing

• Commenters latch onto “delving” as a strong LLM shibboleth, noting a ~25x rise in abstracts and obvious echoes of marketing/corporate prose.
• Other frequently spiking words (“crucial,” “potential,” “significant,” “important,” “these”) are seen as common but overused in LLM output.
• Some think the paper’s use of “delving” in the title is a deliberate joke/meta-commentary.

Explanations and Confounders

• Several note possible confounders:
  • Autocorrect/grammar tools in Gmail and editing software.
  • Editors or journal workflows increasingly using LLMs, especially for non‑native authors.
  • Social contagion: people copy colleagues’ new phrasing even without direct LLM use.
• Others counter that year‑pair analyses in the paper and global usage spikes suggest a strong LLM effect, not just organic drift.

Use of LLMs in Academic Writing

• Many describe LLMs as ubiquitous in academia, especially among non‑native English speakers and in “publish or perish” environments.
• Reported uses range from light editing and clarity suggestions to first‑draft generation, translation, and heavy rewriting.
• Some editors and associate editors explicitly encourage LLM use to fix poor grammar before review.

Ethics, Plagiarism, and Norms

• Strong disagreement over whether using LLMs for prose assistance constitutes plagiarism or academic misconduct.
• One side: generating text for a paper (especially without acknowledgment) is tantamount to fraud, similar to AI art in graded/commissioned contexts.
• Other side: treating LLMs like advanced Grammarly or a helpful colleague is acceptable if findings aren’t fabricated; norms are still evolving.
• Dispute extends into analogies with calculators, Photoshop, and code assistants.

Style, Quality, and Human Voice

• Some lament increasingly generic, buzzword‑heavy prose and filler reviews that appear LLM‑generated.
• Others argue academic writing was already verbose and jargon‑laden; LLMs may not be uniquely to blame.
• Several advocate for clear, succinct writing and cultivating a distinctive personal style to stand out from LLM‑ish language.
• There is concern that people will learn to “sound like LLMs” because they use them as writing tutors.

Scope of the Idea

• Discussion centers on using a local SSH daemon (often on a Unix socket) as a replacement for sudo/setuid, not about exposing root SSH over the internet.
• Several note this is conceptually similar to systemd’s run0 and earlier work (e.g., LISA papers).

Security Model: SSH vs sudo/su

• One camp argues logging in directly as root via SSH (with keys) can be strictly more secure than user→sudo:
  • Fewer accounts that can escalate.
  • Avoids setuid binaries exposed to all users.
  • Reduces attack surface from user processes, daemons, and cron jobs that might have sudo access.
• Others push back:
  • sshd is larger and more complex than sudo; more code can mean more bugs.
  • su and sudo are also setuid but simpler than sshd; if the goal is less code, su/doas may be better.
  • SSH still has to handle environment, files, and sockets, so the “smaller surface” claim is disputed.

Setuid and Privilege Escalation

• Supporters emphasize the benefit of removing setuid/setgid entirely; elevating privileges is hard to secure in Unix.
• Critics respond that something must elevate privileges anyway; setuid isn’t uniquely bad, just another mechanism.
• Some highlight environment-based attacks on setuid/sudo (PATH, LD_PRELOAD) vs more “obvious” network data handling in daemons.

Auditing, Policy, and Usability

• Many value sudo for:
  • Fine-grained command control and argument restrictions.
  • Per-user auditing and accountability.
• SSH-based root sessions lose this granularity unless combined with ForceCommand or related tricks, which are less flexible.
• Several admit to using sudo bash and regretting poor audit trails, but still see sudo as better for policy enforcement.

Complexity and Operational Concerns

• Critics view “SSH as sudo” as added complexity and technical debt: extra daemons, sockets, crypto, and edge cases.
• Concerns raised about:
  • Single-user/recovery mode when root password is disabled.
  • Boot-time failures of sshd leaving the system inaccessible.
• Some argue that in modern immutable/golden-image setups, any interactive root shell (via SSH or sudo) is itself a failure mode.

Key Management and Access Control

• Central question: how to handle multiple admins?
  • Sharing root keys is discouraged; per-user keys plus SSH certificates are suggested but add management overhead.
• Traditional wheel/sudo groups and sudoers rules are seen as simpler for many environments.

Purpose & Effect of the Script

• Script automates visiting Facebook’s “advertisers who used a list to reach you” and sets each one to “Don’t allow.”
• Several commenters explain this disrupts the “refinement” step: third parties upload coarse data, Facebook enriches it with its psychographic data, and this linkage is what the script tries to cut.
• Others note this does not stop underlying data collection by trackers and data brokers; it mainly reduces how well that data can be exploited on Facebook.

Technical Behavior & Robustness

• Multiple users report errors (click on undefined, await syntax issues, timing problems) across Brave, Safari, Chrome.
• People share improved versions: adding async/await, explicit wait() functions, random or longer delays, and a clickAndWait helper.
• Some get temporarily rate-limited (“going too fast”), prompting suggestions to increase delays.
• Script is brittle: depends on specific DOM structure (width of list items, labels, text), which can change or vary by A/B tests, territory, and UI version.
• It may fail or misbehave when some advertisers are already set to “Don’t allow.”

Privacy, Tracking & Shadow Profiles

• Commenters stress that even logged-out or non-users are tracked via embedded Facebook code on third-party sites.
• Some argue shadow profiling is likely less accurate; others claim Facebook still uses such data and can share or sell insights.
• One view: even if collection continues, not feeding Facebook clean signals (by opting out / not using the platform) still matters.

Targeted vs Untargeted Ads

• Motivations against targeting: privacy, ethical objections to data sharing (e.g., banks or health providers uploading lists), and discomfort with “hyper-targeted” or “creepy” ads.
• Some see targeted ads as more effective at manipulating people into buying unneeded things.
• There’s debate on whether making ads less targeted leads to fewer or more ads:
  • One side: less targeting → less profitable inventory → fewer ads.
  • Other side: less targeting → lower revenue per impression → pressure to show more ads, though saturation and user annoyance already constrain this.

Legal / Policy & Platform Risks

• Concern that automated interaction might violate Facebook’s terms or, in extreme interpretations, computer misuse laws.
• Some fear permanent bans; others say a ban would be welcome or note earlier cases where similar tools led to lifetime bans.
• Suggestions to use advertiser lists to identify unlawful data sharing (e.g., by banks, health entities) and report to national Data Protection Authorities.

Alternative Strategies & Broader Concerns

• Alternatives mentioned: not using Facebook, using EU opt-outs, Firefox’s Facebook container, living in a country where the dominant language differs from the user’s, or using novelty language settings to break targeting.
• Some recommend tools like AdNauseam to both hide and mass-click ads to pollute targeting data.
• Broader worries about the “Internet of Behavior/Bodies” and pervasive tracking that may become nearly impossible to avoid.

Tesla vs Waymo and Why They Don’t Partner

• Many argue Tesla has strong incentives not to partner with Waymo: it would signal Tesla’s vision-only FSD is insufficient and undercut its high valuation premised on owning the autonomy stack.
• Waymo also gains little: it could source EVs from many automakers and avoid the reputational and bargaining risks of tying itself to a single, volatile partner.
• Some frame it like Google vs Apple in phones: competing stacks and business models where competition, not collaboration, is expected.

Technology Approaches (Sensors vs Vision)

• Waymo uses an expensive, lidar‑heavy sensor suite; Tesla pursues camera-only with heavy AI.
• Commenters note both rely on AI/computer vision; the real divide is “vision-only vs multi‑sensor.”
• Waymo’s hardware is seen as too costly and maintenance-heavy for mass consumer vehicles today.

User Experience and Safety Perception

• Multiple riders report Waymo in SF and LA as smooth, predictable, clean, and often “mundane in a good way”; some feel it now drives better than they do within its geofenced areas.
• Tesla FSD is described as improving but still anxiety‑inducing, like supervising a teenage driver.
• There are anecdotes of Waymos blocking traffic or entering a wrong-way alley; remote staff can give “advice” but not direct control, with roadside assistance as a fallback.
• One commenter dismisses Tesla FSD as essentially a live prototype with customers as test subjects.

Public Transit, Congestion, and “Leapfrog” Claims

• Strong skepticism that robotaxis can replace well‑run mass transit, especially in dense Asian/European cities.
• Critics stress geometry: low-occupancy cars, even autonomous, can’t match buses/trains for corridor capacity; self-driving fleets risk induced demand and more congestion.
• Others imagine shared, cooperative autonomous fleets (small shuttles, dynamic “flocks,” point‑to‑point minibuses) that could outperform today’s often-empty, infrequent U.S. buses.
• Debate over buses: some claim underused buses cause net congestion; others counter that at peak times they clearly remove car trips and that service must run off‑peak to be viable.

Economics, Scale, and Market Structure

• Disagreement over unit economics: some think cheaper hardware and high utilization will make robotaxis very cheap; others note current billion‑dollar losses and heavy ongoing engineering.
• One side argues autonomy is not a natural monopoly: multiple firms could reach similar safety levels.
• Another side claims scale-driven utilization advantages create winner‑take‑most dynamics similar to ride‑hailing.

Scope, Availability, and Endgame

• Waymo is praised but seen as geographically narrow; rollout to new cities and to harsher conditions (e.g., snow) is expected to be slow.
• Some see potential to reshape suburban commuting; others say that’s already possible with Uber and would mainly worsen sprawl.
• Speculation on Waymo’s endgame: Uber‑like service, tech provider to automakers or transit agencies, or a platform that Google could eventually monetize with advertising—prompting worries about future ad‑ridden rides.

Context & Original Case

• Original tweet: long-time Google Workspace/GSuite user with ~400 TB stored is being forced down to the 5 TB-per-user limit.
• They were on an older “unlimited” plan, later transitioned to a 5 TB plan, with enforcement only now arriving.
• User received months of warnings to reduce usage; data will become read-only and then be deleted after a grace period.
• They treat it more as an amusing “great deal while it lasted” than an injustice, and only plan to rescue ~20 TB.

Bandwidth and Data Export Practicalities

• Multiple commenters calculate that downloading 400 TB over a 1 Gbps link takes ~37–38 days, essentially the full 30‑day window, even before protocol overhead.
• There is debate over TCP/IP and request overhead; some note overhead is nontrivial but much less than 10%.
• Google offers export to a temporary Google Cloud Storage bucket, but it’s zipped and not selective, so tools like rclone are needed for partial retrieval.
• Some note that providers and consumer ISPs often cannot sustain full line rate at scale, and “unlimited” ISP plans can throttle heavy users.

“Unlimited” Storage and Contract Changes

• Strong disagreement over whether “unlimited” offers implicitly include reasonable limits:
  • One side: every “unlimited” offer has implicit limits; using hundreds of TB or PB is “abuse” of a flat-rate service.
  • Other side: if a company says “unlimited” without clearly stated caps, reducing or capping later is straightforward lying or bait-and-switch.
• Some argue that monthly/recurring contracts always allow term changes with notice; others counter that “unlimited” should not be used if not literally true or time-bounded.

Impact on Universities and Enterprises

• Universities previously got “unlimited” Google Drive; now many face strict institutional caps (e.g., 100 TB per university), hitting labs with multi‑TB or PB datasets.
• Similar retrenchments reported at Microsoft OneDrive; users see “unlimited” in UI but hit opaque, varying caps.
• Several note this exposes the risk of outsourcing core storage to cloud vendors instead of maintaining on‑prem systems.

Legal, Regulatory, and Ethical Views

• Some advocate regulatory penalties (e.g., FTC/ACCC actions) for misleading “unlimited” marketing, including escalating fines and even criminal liability for willful deception.
• Others suggest small claims or class actions, though there’s skepticism these would materially change big‑tech behavior.
• Broader sentiment: businesses often treat fines as a cost of doing business; more structural consequences or clearer advertising standards may be needed.

Rooftop Solar Experiences & Economics

• Multiple users report successful residential systems (3–20 kW), often covering or exceeding household demand, with strong production even at high latitudes and in India.
• Reported payback times vary widely (roughly 5–25 years) depending on latitude, shading, local electricity prices, installation costs, financing, and subsidies.
• Some argue rooftop solar is “obviously” economical because it avoids transmission infrastructure and billing overhead; others note insurance, roof age, and shade can make it marginal.

Residential vs Utility-Scale Solar

• Several argue utility-scale solar has better capacity factors and economics, and that some regions are at “diminishing returns” for more rooftop capacity.
• Others counter that once transmission and distribution upgrades are included, rooftop and distributed solar can be highly competitive, and offer resilience.
• Consensus: both scales are needed; non-rooftop sites (warehouses, parking lots, industrial land) are promising.

Grid Integration, Net Metering & Storage

• Concerns about midday export “gluts,” residential neighborhoods needing upgraded distribution, and the operational complexity of high rooftop penetration.
• Net metering is controversial: popular with homeowners but seen as effectively turning retailers into free batteries and shifting costs; Netherlands and UK cited as cautionary examples.
• Batteries are seen as key to reducing grid burden and providing backup, but are still expensive for multi-day outages; some propose generator-plus-battery hybrids.
• Ideas to shift flexible loads (e.g., industrial processes, pumping, data centers) to solar hours are discussed.

Climate Targets and Global Mix

• One side stresses we are far behind needed emission cuts; even rapid solar growth hasn’t bent the atmospheric CO₂ curve yet.
• Others claim the power sector is nearing a turning point, with ~40% low‑carbon electricity and coal likely peaking soon, though this is disputed.
• Debate over whether focusing on “100% clean” misses the bigger win of rapidly moving from 50% to 90%+ low‑carbon.

Nuclear, Fossil Fuels & Technology Mix

• Some frame “solar vs nuclear” as a false dichotomy, advocating a mix of solar, wind, nuclear, hydro, geothermal, and storage.
• Others criticize “rabid” pro‑nuclear factions for dismissing solar and downplaying nuclear costs and risks (e.g., Fukushima cleanup).
• Skeptics of high-renewables grids ask how multi-day low‑sun/wind events will be handled without massive backup plants or storage.

Policy, Lobbying & Public Perception

• Multiple comments allege fossil fuel–funded campaigns and think‑tank work undermined renewables, especially in coal‑exporting countries.
• Examples given of farmers posting anti‑solar/wind signs and of industry-funded misinformation on bird kills, toxicity, and lifecycle energy.
• Australian politics discussed: nuclear proposals portrayed by some as a stalling tactic to prolong coal; official cost studies there reportedly still favor renewables.

Future Outlook

• Many are highly optimistic: forecasts that solar will dominate electricity in the 2030s and total energy in the 2040s, driven by exponential cost declines and rapid build‑out.
• Others view some of these projections as over‑optimistic but still acknowledge dramatic, ongoing growth in solar, wind, and storage.

What HH70 Achieved

• First plasma in a high‑temperature superconducting (HTS) tokamak is seen as a real engineering milestone, but “first plasma” itself is a routine step many devices reach.
• Several commenters stress the main achievement is building a working HTS tokamak from scratch in ~3 years, not the plasma performance numbers (which are sparse).

Comparison to Other Fusion Efforts

• Compared to SPARC and other Western projects, HH70 looks “underpowered” (smaller radius, much weaker field), but HH70 actually exists and runs.
• Some argue this is analogous to an early SpaceX milestone: not unprecedented physics, but fast, competent execution that enables scaling.
• Others note private fusion startups (e.g., Commonwealth Fusion) and earlier HTS tokamak work in the West, so this is not uniquely first in all senses.

High-Temperature Superconductors & “Localization Rate”

• “High‑temperature” here is understood as relative to traditional superconductors; exact operating temperatures are not given.
• “Localization rate >96%” is widely interpreted as: almost all components and IP are domestically sourced, reducing sanction exposure and foreign dependence.

Plasma vs. Fusion, Q, and Power Extraction

• Multiple replies emphasize:
  • Plasma ≠ fusion; it’s just an ionized gas.
  • First plasma usually uses ordinary hydrogen, with little or no fusion.
  • Net gain is judged via Q (fusion power out vs. power into the plasma; and separately engineering Q vs. plant parasitic loads).
• Heat extraction in DT designs is expected to rely mainly on energetic neutrons heating a surrounding blanket and then steam turbines, similar in concept to fission plants.
• Several commenters note that no one has yet demonstrated a full, practical heat‑to‑electricity fusion cycle.

Geopolitics and Industrial Strategy

• Some see this as a potential “Sputnik moment” highlighting that China can rapidly execute complex, largely domestic high‑tech projects.
• Others counter that similar reactors and timelines exist in the West and that this press release alone shouldn’t be over‑interpreted.

Fusion vs. Renewables and Fission

• Long subthreads debate whether fusion will ever be economical versus rapidly falling costs for solar, wind, and storage.
• Pro‑fusion voices stress dispatchability, high energy density, and applications like shipping or heavy industry.
• Skeptics argue:
  • Reactors are extremely complex and likely very expensive.
  • Grid‑scale renewables with storage and advanced grid control may ultimately be cheaper and sufficient.

Safety, Proliferation, and Regulation

• Fusion devices are seen as much safer than fission in accident scenarios (no runaway chain reaction; plasma extinguishes if disrupted).
• One thread notes that intense neutron flux could still be used to breed weapons materials, so proliferation concerns are not automatically eliminated, though others argue similar neutron sources already exist.
• In the US, fusion is being regulated differently from fission, closer to large accelerator facilities.

Funding, Private Capital, and the MiHoYo Angle

• Historical magnetic‑fusion funding is characterized as “fusion never” levels relative to 1970s plans.
• Commenters discuss that many startups’ likely business model is selling IP/know‑how to eventual state‑backed projects rather than independently building commercial fleets.
• A widely noted anecdote: the HH70 company was early‑funded by a major Chinese game developer for tens of millions of dollars—“pennies” relative to its game revenue—illustrating how nontraditional private capital is entering fusion.

Automation, Productivity, and Working Time

• Many note that a 6‑day week was historically normal and still is in parts of the world; others see this as a regression given modern automation.
• Commenters recall past predictions of very short workweeks due to productivity gains; instead, most people still work full-time or more.
• Explanations offered:
  • Productivity gains have largely gone to executives and capital owners, not workers.
  • Social expectations and consumerism continually create new “needs,” so total work expands (Jevons‑paradox style).
  • Some argue people “like” working or need it for structure; others strongly reject this and prioritize less work.

Housing, Inequality, and Basic Needs

• Several focus on housing scarcity: rising productivity doesn’t matter if constrained housing drives up rents and mortgages.
• Debate over whether housing scarcity is “artificial” (through zoning, permitting, NIMBYism) or simply market forces.
• Broader critique: capitalism is good at producing wealth but poor at distributing it or guaranteeing shelter and food.

Greek Economic Context

• Greece is described as highly reliant on low‑wage, seasonal tourism and construction, with ~11–12% unemployment and significant emigration of young workers.
• Disagreement on tourism margins: some say hotels and beach concessions earn very high profits and underpay staff; others claim small hotels are barely viable, especially without “shady” practices.
• Widespread mentions of corruption, weak labor inspections, off‑the‑books work, and political clientelism; some liken Greece to a “cartel” economy.
• Comparison to other EU periphery countries (e.g., Croatia) facing similar brain drain and poor working conditions.

Details and Ambiguities of the New Law

• The article’s description of:
  • Up to two “unpaid” extra hours per day “in return for more free time.”
  • A 40% supplement for a sixth workday.
• Commenters find this wording confusing and possibly mistranslated:
  • Some interpret it as time‑in‑lieu and 140% pay for the 6th day.
  • Others argue in practice unpaid overtime and 6‑day weeks already exist and the law mainly legalizes/extracts more from workers.

Labor Shortages, Wages, and Training

• Thread questions how labor shortages coexist with high unemployment.
• One side: firms could raise wages or improve conditions; refusal to do so reflects collusion, culture, or spite.
• Others argue many businesses can’t raise pay without going under, especially in price‑sensitive tourism.
• Additional factors cited: emigration to better‑paying EU states, weak education for parts of the workforce, and employer reluctance to invest in training if workers can easily leave.

EU, Politics, and Systemic Critiques

• Some blame EU policies and bailouts for a “lost generation,” austerity, and weakened worker protections; others note EU partners also “bailed out” Greece.
• Broader trend noted: rising nationalism, strong corporate lobbies, and use of subsidies/tax breaks that enrich firms while cutting social benefits.

Robots, AI, and the Future of Work

• Mixed views on automation:
  • Some see robots/AI as still too immature to matter much; others think “serious automation” is being politically/economically suppressed.
  • General point: efficiency tends to unlock more work rather than reduce total work, unless society consciously chooses to trade gains for leisure.

Cost model and performance pitfalls

• Lambda bills for duration, so blocking on slow or hung downstream APIs can explode costs compared to always‑on servers.
• Timeouts are essential to cap cost, but aggressive timeouts plus retries can backfire and increase total work and spend.
• Cloudflare Workers are contrasted as billing on CPU time, not wall time, making network waits cheaper there.
• Thread notes Lambda’s default (3s) and practical limits on timeouts; some argue you want short timeouts to avoid “blank checks” to AWS.

When Lambda fits vs. when it doesn’t

• Good fit: low‑TPS internal apps, spiky or infrequent workloads, job/queue processing, webhooks, “fire sale” / telemetry‑like bursts, and pet projects where scaling to zero matters.
• Poor fit: consistently high load, long‑running or IO‑heavy calls, high DB connection counts, or where development iteration speed and observability are critical.
• Several commenters say ECS/Fargate or traditional containers solve many problems more cleanly once you have steady traffic.

Lambda Web Adapter and reuse of containers

• Adapter enables running the same HTTP container on Lambda, ECS/Fargate, or EC2, reducing duplicate codebases and easing lift‑and‑shift from one compute option to another.
• Some see this as improving security/compliance by reusing a single well‑known container; others worry it encourages keeping legacy systems alive instead of fixing underlying issues.
• One pattern: deploy the same image as both Lambda and Fargate service behind an ALB, then direct traffic to whichever is more cost‑effective for that workload.

Architecture patterns: async, APIs, and pooling

• Async/claim‑check patterns, webhooks, Step Functions, and queues (e.g., SQS) are suggested to avoid Lambdas waiting on long operations.
• Direct Lambda‑to‑Lambda blocking calls are viewed as an antipattern; event/queue‑based designs are preferred for cost and resilience.
• Connection pooling works per‑Lambda instance lifetime; for databases like Postgres, unbounded concurrency leads to too many connections without a proxy (e.g., RDS Proxy).

Developer experience and “programming the machine”

• Opinions diverge on iteration speed: some find Lambda cycles acceptable, others find build/deploy/debug painfully slow compared to local dev + containers/Kubernetes.
• Many note significant time spent on IaC and wiring AWS services (“programming the machine”) that can rival or exceed application coding.
• Scaling to zero is valued for per‑branch ephemeral environments and previews, but some argue the complexity often outweighs small cost savings.

Zen 4c / Bergamo Characteristics

• Zen 4c cores have similar per-cycle performance to Zen 4 but:
  • Less L3 cache per core.
  • Lower maximum clock, especially vs desktop Ryzen; not much lower than dense server Zen 4.
  • Slightly better performance per watt is suggested, but magnitude is unclear.
• Bergamo uses 8 compute chiplets; speculation/roadmap mention a follow-on part with 12 chiplets and 192 Zen 5c cores.

Caches, SRAM, and Coherency

• SRAM (especially cache) is seen as the primary scaling bottleneck: more cores imply more cache area.
• 3D V-Cache extends shared L3, but L1/L2 remain per core.
• Chiplet-based designs can scale cores and cache capacity, but:
  • Cache coherency across many dies is complex.
  • NUMA “cliffs” and memory latency can dominate as core counts rise.

Memory Models: x86 vs ARM

• x86 has a strong consistency model, making multithreaded programming easier but imposing higher synchronization and coherency costs.
• ARM offers weaker guarantees; can scale better but demands more careful software.
• Some argue these semantics may eventually limit how far x86 can scale in-package.

ISA and Decoder Complexity Debate

• One view: ISA choice barely matters; x86 decode occupies a small die/power budget (single-digit percent).
• Counterview: variable-length x86 instructions make wide, multi-instruction decode and branch prediction significantly more complex and power-hungry than fixed-width ARM/RISC-V.
• There is discussion of uop caches, instruction density, and complex multiplexing logic; no consensus on how decisive this is for future many-core designs.

Many-Core Futures and Software Limits

• Some see Bergamo-style “core spam” as the natural future: many simpler cores to hide memory latency, echoing earlier ideas like Sun Niagara and research projects.
• Others stress single-thread performance remains crucial; in many real workloads, most cores sit idle.
• Amdahl’s law and the difficulty of writing correct, scalable multithreaded code are repeatedly cited.
• Modern languages (Go, Rust, functional / reactive approaches) help but do not solve parallelism for typical applications.
• Removing bottlenecks like Python’s GIL is seen as necessary but far from sufficient for effective use of 100+ cores.

Networking, Caches, and Near-Memory Compute

• Some speculate about “RAM-less” or RAM-minimized servers that keep hot code and state entirely in cache, streaming data from NIC to CPU and back.
• Existing technologies (e.g., NIC-to-cache DMA, DPDK) already approximate “process in L3” models.
• There is interest in pushing more compute to or near the NIC, but current solutions (e.g., FPGA-based smart NICs) are expensive and specialized.

Value of Art and Creative Work

• Strong disagreement over whether art is socially “valuable.”
• One camp calls art a low-value luxury, primarily for status and entertainment, easily sacrificed when resources are scarce, and historically overvalued.
• Others argue art and entertainment (music, books, games, films, software) are central to quality of life, identity, and culture, pointing to massive global entertainment revenues as evidence of real demand.
• Some note that in rich societies, people want far more than subsistence; art is part of what makes a society “rich.”

Markets, Monopolies, and Pricing

• Debate over whether current creator pay reflects “the market” or distorted markets.
• Critics point to oligopolies/monopsonies (e.g., Spotify + major labels, big platforms) and regulatory capture; they argue prices and revenue shares are not set by real competition.
• Others insist that, aside from clear monopolies, market prices are the only coherent definition of economic value; oversupply of creativity makes marginal value low.

Middlemen, Platforms, and Distribution

• Broad agreement that intermediaries (labels, platforms, distributors) capture disproportionate value relative to individual creators.
• Some say customers pay for the packaged service (distribution, recommendation, UX) rather than raw content; packaging can be worth more than any single work.
• Others counter that without creators, platforms are worthless, so current revenue splits are unjust.
• Allegations that music companies sidestep royalties via non-cash compensation (e.g., equity), with questions about legality.

Technology, AI, and Barriers to Entry

• Cheap tools and internet distribution lowered entry barriers, but made it harder to earn a living due to oversupply and algorithmic dynamics (e.g., favoring frequent, lowest-common-denominator content).
• AI and digital tech are seen by some as exposing art’s “low real value” by flooding the market; others worry about homogenized, “elevator music” culture.

Proposed Fixes and Structural Ideas

• Suggestions: creators owning their own distribution, direct sales, small open-source platforms, collective bargaining, and stronger antitrust/regulated infrastructure.
• Concerns: collective bargaining faces transaction and agency problems; building non-gatekept internet structures remains an unsolved coordination challenge.

Impact of the textbooks

• Many commenters learned operating systems, networking, and computer architecture from these books in the 80s–2000s.
• Texts are described as unusually clear, approachable, and even “thriller-like,” yet dense and rigorous.
• Several anecdotes: failing students turning grades around by switching to these books; people still keeping them on shelves decades later; some used them later as instructors.
• “Structured Computer Organization,” “Operating Systems: Design and Implementation,” “Modern Operating Systems,” and “Computer Networks” are cited most.

MINIX: role, licensing, and Intel ME

• MINIX is praised as an elegant microkernel that influenced Linux and educated many, but never became the dominant free Unix-like system.
• Multiple comments dwell on the “what if” of MINIX having a FLOSS license earlier; some think it could have occupied Linux’s historical role.
• Discussion notes MINIX’s use inside Intel Management Engine, making it one of the most widely deployed OSes—though some are uneasy about this “hidden” role.
• Intel allegedly chose it partly for its BSD-style license and built-in services; alternative microkernels (L4, seL4) are mentioned but seen as less suitable or license-constrained.

MINIX3’s decline

• Consensus that MINIX3 development has effectively stalled; no commits in years and an unreleased 3.3.0 despite fixes.
• Contributors describe technical debt: 32‑bit, uniprocessor kernel, aging hardware support, heavy dependency on NetBSD source tree.
• Capabilities like fault-tolerant drivers and live-updatable services are praised as impressive research outcomes.

Microkernels vs monolithic kernels

• The old Tanenbaum–Torvalds debate is referenced, especially over microkernel vs monolithic designs.
• Some note that Linux was initially developed on MINIX and that criticism may have helped shape Linux; others point to modern performant microkernels as partial vindication.

Distributed OSes and systems research

• Several commenters lament that mainstream systems don’t provide a unified distributed OS where processes migrate seamlessly across personal devices.
• Others argue this is niche or “cool but not essential,” preferring higher-level, multi-system tools or cloud abstractions.
• Counterpoint: distributed OS research is still active; ideas appear in clouds, VM/container live migration, and academic microkernel projects.
• Debate becomes heated over whether an OS “should” provide transparent distributed access to resources; no consensus, but strong opinions on both sides.

Electoral-vote.com and broader influence

• Multiple comments highlight the long-running election analysis site run by Tanenbaum, now co-written, as a clear, data-driven daily political summary.
• Some credit his networking book and even a famous “sneakernet” quote with influencing thinking in industry (e.g., about Netflix’s early strategy).

Recognition and comparisons

• Many express that this award feels overdue given his influence on generations of students and practitioners.
• A few wonder why Linux’s creator hasn’t received the same ACM award; possible reasons (including personality) are speculated but unresolved.

Succession planning and family businesses

• Many note the absence of a formal succession plan, seeing it as common in small, founder‑led firms, especially where the business is tightly tied to the founder’s identity.
• Commenters discuss how children often don’t want or aren’t equipped to continue family businesses, in India and the US (e.g., HVAC, construction, manufacturing).
• Some share parallel stories of aging parents with no plan and anxiety about what happens to employees.

Sale dynamics, valuation, and “bottom feeders”

• Debate over small‑business sale multiples: some say 2–3x earnings can be fair for hands‑on, key‑person‑dependent firms; others think there’s systemic underpricing.
• Discussion of an “industry” of buyers who wait for owners to be ill or dying, then lowball.
• One side calls this exploitative, bordering on contracts under duress; another argues it is just supply/demand and sellers are not “entitled” to preferred prices.
• Some suggest markets for these niche businesses are inefficient and possibly tacitly collusive; others counter that if something consistently sells for 33% of “theoretical value,” that’s probably its real value given risk and information issues.

Why not internal CEO or co‑op?

• Multiple questions about promoting an internal manager or converting to a worker co‑op.
• Responses stress lack of broad business/generalist skills internally, limited access to capital/credit in that context, time constraints, and cultural/mental gap between “employee” and “owner.”
• Consensus: feasible in theory, but not practical here given location, scale, and urgency.

Employee welfare, culture, and PE vs strategic buyers

• Readers praise the priority given to employees and legacy when choosing a buyer.
• Some contrast this with a more purely financial “sell to highest bidder / private equity” approach, though others point out many founders do just that without guilt.
• Discussion around Indian tier‑2 cities: fewer alternative jobs, making employer responsibility feel heavier; “lifers” are common in manufacturing/industrial roles.

Co‑ops and ownership models

• Thread branches into debate over worker co‑ops vs “capitalist” ownership.
• Pro‑co‑op voices highlight large European co‑ops where workers share profits and governance.
• Skeptics list prerequisites: aligned ideology, willingness to bear risk, extra management overhead, and not all workers wanting ownership.

Personal reflections and reaction to the article

• Numerous commenters share emotionally similar experiences (death of a founder, chaotic handovers, businesses decaying under surviving spouses).
• The writing is widely described as unusually sincere and non‑self‑promotional; several suggest it has film or case‑study potential.

Inline assembly and “own constant folder” idea

• Core trick: wrap SIMD sqrt in inline assembly but special‑case compile‑time constants using __builtin_constant_p, so constants are folded while non‑constants emit the desired instruction.
• Several commenters see this as a narrow workaround for a specific Clang codegen quirk rather than a general “constant folder”.
• It’s noted that similar patterns are common in performance‑critical domains (e.g., HFT, codecs) where single‑instruction savings matter.

Clang vs GCC and sqrt optimization

• Clang under -ffast-math turns sqrtps(x) into x * rsqrtps(x) with Newton–Raphson refinement.
• Some call this a miscompile or performance bug; others argue it’s “working as intended” on older CPUs where this is faster.
• GCC is reported not to do this transformation in the shown example.
• Performance advantage is architecture‑dependent; newer Intel CPUs have much faster sqrt instructions, and using -mtune/-march can change Clang’s choice.

Fast-math flags: scope, hazards, and naming

• Strong pushback on global -ffast-math: it relaxes IEEE guarantees across the whole program and can even affect other translation units or shared libraries via global FP state.
• Some argue it’s fine when used only for specific numerical TUs; others cite bugs and surprising cross‑module effects.
• Preference expressed for more granular controls (per function, per expression, or pragmas), as in Rust’s unchecked_add or Julia’s @fastmath.
• Several commenters criticize the name “fast‑math” since it can be slower and less accurate; suggestions range from “unsafe math” to tongue‑in‑cheek renamings.

Determinism, correctness, and other languages

• Debate over whether floating‑point determinism matters, especially for games and multiplayer simulations.
• Examples from Rust, Julia, and Zig show similar low‑level tuning issues and different mechanisms for localizing unsafe or fast math.
• General consensus: all optimizing compilers can produce surprising code; you must check assembly for hot paths.

C/C++ complexity and when to dive this deep

• Many note this level of trickery is unnecessary for most projects.
• Guidance: focus first on architecture and hotspots; only then tweak compiler flags or write intrinsics/asm.
• Some see this as emblematic of C/C++’s power and complexity; others find it off‑putting and a motivation for safer languages.

Reality and Spectrum of Mental Imagery

• Many participants strongly affirm aphantasia is real; others express doubt or suspect miscommunication about what “seeing in the mind” means.
• Experiences are clearly on a spectrum: from complete absence of imagery to extremely vivid “hyperphantasia” that can feel like AR or movie-quality scenes.
• Several note different “planes” of imagery: some can superimpose on vision, others only in a separate inner space.

How People Experience Thinking

• Aphantasic participants describe:
  • No pictures when prompted (e.g., Mona Lisa, STOP sign, pink elephant).
  • Access only to concepts, words, spatial or logical relations.
  • “Knowing” shapes, routes, or schemas without any visual feel.
• Hyperphantasic participants describe:
  • Detailed, manipulable 3D scenes, sometimes with sound and mild tactile sense.
  • Strong daydreaming, intrusive imagery, difficulty reading because visuals dominate.
• Some report strong audio imagery but no visuals; others have neither images nor internal monologue, only “concepts.”

Impact on Work, Reasoning, and Daily Life

• Many with aphantasia report:
  • Little impairment in technical fields; strong abstract or spatial reasoning (math, physics, programming, navigation).
  • Reliance on diagrams, CAD, notes, and conceptual/graph-like models instead of mental pictures.
  • Difficulty with drawing from imagination, recognizing or describing faces, and following image-based directions.
• Some hyperphantasic people find their vivid imagery distracting or emotionally intense.

Dreams and Other Senses

• Mixed reports:
  • Some aphantasics dream vividly and only see images in sleep.
  • Others have non-visual or “book-like” dreams, or few/no recalled dreams.
• Several can vividly imagine sounds, music, and sometimes taste/smell, even when visual imagery is absent.

Memory and SDAM

• Multiple comments mention severely deficient autobiographical memory (SDAM): remembering life as facts, not replayable episodes.
• People describe weak recall of mundane details (e.g., breakfast, movie scenes) and poor episodic “re-living,” sometimes linked to aphantasia and possibly to depression.

Testing, Diagnosis, and Debate

• VVIQ (imagery vividness questionnaire) and simple prompts (imagined scenes, triangles with color, “ball on a table” questions) are discussed, but many find them too subjective.
• Objective signals mentioned: pupil response to imagined brightness, fMRI differences, visual memory tasks.
• Disagreement over whether many self‑diagnoses reflect true aphantasia versus high standards or semantics.
• Some discuss training methods (e.g., “image streaming”), meditation, or post‑Covid changes, but effectiveness and mechanisms remain unclear.

Cooking oil basics (smoke point, flavour, use-cases)

• Common rule:
  • For frying/sautéing: prioritize smoke point.
  • For dressings/toppings/sauces: prioritize flavour.
• Extra-virgin olive oil (EVOO) is seen as poor value for high‑heat frying: lowish smoke point, expensive, and heating destroys much of its distinctive flavour.
• Some use two oils: EVOO or light olive oil for salads/low–medium heat, and a neutral high‑smoke‑point oil for frying.
• Smoke vs steam: normal cooking gives steam; visible smoke with a sharp smell indicates oil has reached its smoke point and may degrade and taste bad.

Olive oil quality, fraud, and contamination

• Several posters distrust “extra virgin” labels, especially blended or white‑label imports. California single‑origin oils are viewed as more regulated in the US.
• Others say Spain has strict EVOO standards and passionate small producers, but note that generic brands and exports can be dubious.
• German product tests (ÖkoTest) reportedly found mineral-oil contamination in many European olive and rapeseed oils, including some organic ones, likely from processing/harvesting machinery.
• Commenters note Italian “EVOO” often includes bulk oil from Spain, Greece, or North Africa.
• One poster calls vegetable oils unsafe and most olive oil fake; others demand evidence and treat that as unsubstantiated.

Prices and supply

• Multiple reports of sharp price rises for olive oil (often ~50% in months).
• Cited reasons: drought, harvest failures, and global crop issues.
• Some people are turning to local mills, pooling small harvests for personal-use oil at lower effective prices.

Seed/vegetable oils and “seed oil” panic

• “Vegetable oil” labels can mean palm, soy, cottonseed, or other cheap, highly processed oils, depending on region and price.
• Several commenters think the current anti–seed‑oil/omega‑6 panic is overblown; one links to a long critique of anti–seed‑oil theories.
• Others speculate about linoleic acid’s possible role in obesity but admit the evidence is not decisive.

Health, breakfast, and obesity (tangential but extensive)

• Large subthread debates breakfast and meal timing:
  • One side cites observational and meta-analytic data associating habitual breakfast skipping with higher GI cancer, cardiovascular disease, and mortality; and argues that late eating worsens metabolic markers.
  • The other side cites reviews of randomized trials suggesting skipping breakfast does not harm weight or resting metabolism and notes industry influence in promoting breakfast.
  • Some nuance emerges: very early or very late eating may both be suboptimal; effects may depend on overall lifestyle, metabolic health, and total caloric intake.
• Another long subthread debates obesity causes:
  • Some insist the core driver is individual lifestyle (overeating, inactivity); “eat less, move more” framed as thermodynamics.
  • Others emphasize environment and systems: food quality, pricing, built environment (walkability vs car dependence), psychological factors, medications, and gut microbiota, arguing these strongly shape “choices.”
  • There is partial agreement that energy balance still matters, but implementation is hard and simplistic slogans are insufficient.

US vs Europe food culture and cooking

• One claim that Americans “can’t cook” is challenged by others who note large produce sections in US supermarkets and competent home cooking.
• European posters describe cheaper, higher‑quality “everyday” restaurant meals (e.g., in Portugal) compared to fast food, and more walkable cities, as likely contributors to lower obesity despite similarly sedentary jobs.

Information, marketing, and trust

• Multiple comments highlight how industries (tobacco, food, dairy) have historically shaped nutrition narratives (e.g., vegetable oil rebranding, “breakfast is the most important meal”).
• Some advocate skepticism toward marketing-driven health claims and encourage looking at primary evidence rather than alarmist or commercial messaging.