Hacker News
1S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic
- S&P Dow Jones Indices denied accelerated entry to the S&P 500 for SpaceX, OpenAI, and Anthropic, maintaining standard eligibility criteria including profitability requirements. This blocks potential billions in passive fund investments, though other indexes like Nasdaq-100 granted faster entry.
S&P 500 maintaining inclusion rules
Commenters praise S&P's decision to stick to established profitability requirements rather than making exceptions for SpaceX, OpenAI, and Anthropic. Many view this as protecting passive investors from forced exposure to unproven companies and preventing wealth transfers to billionaires through rule changes.Index fund integrity and passive investing
Discussion focuses on the importance of seasoning periods and GAAP compliance before index inclusion. Users emphasize that passive funds should remain truly passive, with concerns about AI companies' accounting practices and the need for proven profitability over multiple quarters.Valuation concerns and market speculation
Participants debate whether these AI companies' massive valuations are justified, with skepticism about their long-term viability and profitability. Many express relief that index funds won't be forced to participate in what some view as speculative bubbles built on hype rather than fundamentals.
2Gov.uk has replaced Stripe with Dutch provider Adyen
- The content contains only two URLs - one pointing to a UK government blog post about building for the future and making changes simple on GOV.UK Pay, and another linking to Adyen's press release about payments for gov.uk. Without access to the actual content of these pages, there is no meaningful information to summarize.
Payment processing costs and alternatives
Discussion reveals payment processing fees are viewed as excessive (up to 4% in US vs 0.3% EU cap), with examples of efficient government systems like Brazil's Pix ($10M/year) and India's UPI demonstrating lower-cost alternatives to traditional card networks.Stripe vs Adyen business models
Users compare Stripe's developer-friendly self-service approach targeting small businesses against Adyen's enterprise focus (€5M minimum), noting Stripe's superior marketing while Adyen serves larger clients like Microsoft and Meta.European payment sovereignty
Comments discuss concerns about relying on US companies for critical infrastructure, with mentions of Stripe's Irish founders, dual headquarters, and the broader trend of European governments seeking alternatives to US payment processors.
3New method turns ocean water into drinking water, without waste
- University of Rochester researchers developed an energy-efficient solar desalination system that produces fresh water without chemicals and converts leftover salts into useful materials like lithium. The technology uses laser-etched black metal panels that prevent salt buildup and extract nearly 100% of salts in solid form, avoiding harmful brine disposal.
Energy efficiency and theoretical limits
Discussion centers on fundamental thermodynamic limits of desalination, comparing this thermal method to reverse osmosis efficiency. Comments debate whether solar thermal can compete with solar panels powering conventional RO systems, noting RO is already near theoretical optimum.Practical implementation challenges
Focus on scaling from lab prototypes to real systems, particularly salt clogging prevention and durability. Users express skepticism about long-term operation without maintenance, comparing to fragile nanotechnology surface treatments that wear down over time.Salt disposal and byproduct value
Debate over whether crystalline salt output is advantageous or problematic compared to liquid brine disposal. Discussion includes potential value recovery from magnesium and other minerals, with environmental concerns about brine discharge versus solid waste management.
4Did Claude increase bugs in rsync?
- A statistical analysis of rsync software releases examined whether Claude AI-assisted development introduced more bugs compared to historical releases. The study analyzed 36 releases and found that the two Claude-assisted releases fell within normal distribution ranges, with statistical tests showing no significant increase in bug severity. The analysis was prompted by online outrage claiming AI assistance made the software less stable, but the data showed Claude releases performed similarly to or better than the historical average.
Statistical methodology and sample size concerns
Discussion centers on whether the analysis has sufficient data (only 2 Claude-authored releases) to draw meaningful conclusions, with debates over statistical power, attribution methodology, and whether absence of evidence equals evidence of absence.AI attribution and responsibility in code commits
Debate over whether developers should disclose AI usage in commits, with some arguing it creates unnecessary drama that will drive attribution underground, while others view non-disclosure as fraudulent misrepresentation.Developer experiences and polarization around AI coding tools
Split between developers who find AI tools transformative for productivity and those concerned about "vibe coding" quality, code review burden, and the broader implications of AI-assisted development on the profession.
5pg_durable: Microsoft open sources in-database durable execution
- pg_durable is a PostgreSQL extension that enables long-running, fault-tolerant SQL workflows with automatic checkpointing and resume capabilities. It allows teams to define durable execution workflows directly in SQL, eliminating the need for complex combinations of cron jobs, workers, and queues while keeping all state within Postgres for better reliability and operational visibility.
Code vs database logic placement
Developers debate whether queue/workflow logic should live in application code with Git version control versus in the database. Some prefer keeping business logic in code for better testing, versioning, and observability, while others see value in database-centric approaches for data locality and simplified architecture.Comparison to existing orchestration tools
Discussion around why one would choose pg_durable over established solutions like Temporal, Apache Airflow, or other DAG schedulers. Benefits cited include simplified infrastructure, better backup consistency, and reduced operational complexity, though many question the trade-offs.Database scaling concerns
Users express concerns about adding long-running jobs to PostgreSQL, traditionally one of the hardest infrastructure pieces to scale. Counterarguments highlight that databases are designed for fault tolerance and that such workloads already run against databases through external triggers.
6GrapheneOS user reported to authorities for using GrapheneOS
- A user reported that age verification company Yoti automatically flags GrapheneOS devices and reports them to authorities, sparking discussion about privacy rights. GrapheneOS developers dismissed this as fearmongering from customer support trying to close a ticket, noting nothing illegal about using the OS. Users debated whether this represents growing hostility toward privacy tools, with some suggesting maintaining separate devices for verification purposes to protect privacy while others called for contacting lawmakers about increasing surveillance requirements.
Credibility of the story and user's history
Commenters question the authenticity of the Reddit post, noting the original poster had hidden their profile after being called out for multiple posts about bypassing age verification systems. Some suggest the user may have been testing verification companies and got banned as a result.UK surveillance and police powers
Discussion centers on the UK's increasing surveillance state, with debates about police powers to compel passwords, comparisons to other countries' privacy rights, and concerns about reporting privacy-conscious users to authorities as suspicious activity.GrapheneOS detection and privacy implications
Technical discussion about why GrapheneOS doesn't spoof its identity to appear as stock Android, focusing on the philosophical approach of normalizing alternative OS usage rather than hiding it, and the broader implications for privacy-focused technology users.
7Astronauts told to return to ISS after sheltering over air leak repairs
- Astronauts aboard the International Space Station were told to shelter in a SpaceX spacecraft while Russian cosmonauts attempted repairs on persistent air leaks in the Zvezda service module. The leaks, which have been an ongoing problem since 2019, were losing about one kilogram of air per day. After repair work was paused for assessment, the crew was told to return to normal duties on the station.
Technical leak detection and repair challenges
Discussion of NASA's RELL robotic leak detector using mass spectrometer and ion vacuum pressure gauge. Users debated repair methods like paint/sealants, with explanations of why simple solutions don't work in space's complex environment.Safety protocols and station evacuation procedures
Extensive discussion about ISS structure lacking internal airlocks, why astronauts shelter in return capsules during emergencies, and escape vehicle capacity requirements ensuring enough seats for all crew members.ISS aging and future space missions
Conversations about the 24-year-old Russian Zvezda module's deteriorating condition, comparisons to potential Mars mission risks, and concerns about spacecraft reliability over extended periods without Earth support.
8Tracing a powerful GNSS interference source over Europe
- This paper identifies a space-based GNSS interference source causing widespread disruption across Europe, Greenland, and Canada since 2019. Using data from terrestrial reference stations, researchers developed detection methods and identified the source as Russian early warning satellites in Molniya orbits.
Russian GPS jamming impact
Users report widespread GPS jamming affecting construction projects near Romania and Poland, with daily interference from Russian sources. Discussion reveals Russia doesn't care about collateral damage to their own civilians, with ride-sharing apps warning users about location errors.Satellite-based interference identification
Researchers identified Russian satellite Cosmos 2546 from the Edinaya Kosmicheskaya Sistema constellation as source of GPS interference across Europe since 2019. Novel finding shows space-based jamming escalation beyond traditional ground-based interference.Response options and countermeasures
Discussion of potential responses including treaty violations, electronic warfare, shooting down satellites, or electromagnetic pulse attacks. Concerns raised about Kessler syndrome and retaliation risks from destroying orbital assets.
9Google will pay SpaceX $920M per month for compute
- SpaceX signed a $920 million per month deal with Google for AI compute access from October 2026 through June 2029, involving approximately 110,000 NVIDIA GPUs and related components. This follows a similar $1.25 billion monthly deal with Anthropic and comes just before SpaceX's historic IPO expected to raise $75 billion at a $1.75 trillion valuation.
Duplicate post management
A user identified a duplicate discussion and provided a link to the earlier post. A moderator confirmed they would consolidate comments to the original thread to avoid fragmentation of the discussion.
10Gemma 4 QAT models: Optimizing compression for mobile and laptop efficiency
- Google has released new Quantization-Aware Training (QAT) optimized checkpoints for Gemma 4 that dramatically reduce memory requirements while preserving model quality. The updates include Q4_0 format optimization and a specialized mobile quantization schema that reduces Gemma 4 E2B's memory footprint to just 1GB, enabling efficient local deployment on consumer hardware and mobile devices.
Model performance and capabilities
Users tested Gemma 4 QAT models locally, finding impressive multimodal capabilities (text, image, audio) in compact sizes. The 12B model runs on 8GB VRAM, 3.2GB model handles SVG generation, and even 0.8GB text-only version enables real-time conversations.Technical implementation challenges
Rapid release cycle creates integration difficulties for developers, with multiple model variants (QAT, MTP, different sizes) causing compatibility issues with llama.cpp and consumer apps. KV cache problems persist due to extreme activation ranges.Model comparison and ecosystem
Discussion of Unsloth vs Google's QAT implementations, with Unsloth claiming better quantization results. Users compare performance across different model sizes and debate the benefits of quantization-aware training over standard quantization.