Hacker News
1Tell HN: Help restore the tax deduction for software dev in the US (Section 174)
- US tech companies are struggling with a tax code change that requires software development expenses to be amortized over years instead of being immediately deductible. YC and others are leading efforts to reverse this change, urging US taxpayers to sign a letter to lawmakers supporting the reversal.
Section 174 explanation
Tax code change requires software engineering salaries to be amortized over 5 years instead of being immediately deductible, making engineers more expensive and potentially forcing companies to pay taxes even when unprofitable.Definition challenges
Debate over what constitutes "software development" under the law, affecting roles from QA engineers to data scientists, with IRS guidance attempting to clarify but leaving many edge cases unclear.Impact on businesses
Small companies and startups particularly affected, facing tax bills on phantom profits while cash-poor. Forces shift toward consulting/work-for-hire vs R&D, advantages VC-funded companies over bootstrapped ones.
2Kagi Reaches 50k Users
- Kagi Search is a privacy-focused search engine with 50,074 members, serving 760,600 queries per day. It offers features like custom domain insights, keyboard shortcuts, search operators, and query shortcuts. The platform includes specialized search tabs for images, videos, maps, and news, along with translation and calculation widgets.
Small vs growth
Discussion of Kagi's profitability at small scale, with many users supporting their choice to stay small and focused rather than pursuing aggressive growth. Debate around avoiding VC funding and resisting pressure to expand beyond core search.Pricing and accessibility
Debate around Kagi's pricing model, particularly its impact on users in developing countries. Some find it too expensive while others defend the pricing as necessary for sustainability. Discussion of potential regional pricing and family plans.Search quality and features
Users comparing Kagi's search quality to competitors like Google and DuckDuckGo. Discussion of Kagi's ability to avoid SEO spam, its maps functionality limitations, and integration of AI features into search results.
3Apple announces Foundation Models and Containerization frameworks, etc
- Apple announced new developer tools and technologies featuring Liquid Glass design, Foundation Models framework for AI capabilities, and enhanced Xcode 26 with ChatGPT integration. These updates aim to help developers create more beautiful, intelligent, and engaging apps across Apple platforms while maintaining privacy and offline functionality.
Containerization framework
Discussion of Apple's new container solution, with debates on whether it's truly native or VM-based, comparisons to existing solutions like Docker Desktop and Colima, and potential impact on developer workflows.Liquid glass design
Comments on Apple's new UI design approach, comparing it to past designs like Aqua and Windows Aero, with mixed reactions about readability concerns and the cyclical nature of flat vs. skeuomorphic design trends.iPad-Mac convergence
Debate about iPad becoming more Mac-like, with discussion of remaining limitations, Apple's strategic reasons for keeping platforms separate, and whether iPad can replace laptops for most users.
4Bruteforcing the phone number of any Google user
- A security researcher discovered a vulnerability in Google's account recovery system that allowed bruteforcing phone numbers through a combination of IPv6 address rotation and exploiting a no-JavaScript form submission method. The researcher developed a proof-of-concept tool that could identify account phone numbers using display names leaked through Google Looker Studio.
IPv6 blocking challenges
Discussion of how IPv6 complicates IP-based rate limiting and blocking, since users can get large blocks of addresses. Consensus that blocking by /64 prefix is common but may not be sufficient given availability of /48 blocks.Legacy system maintenance
Debate about challenges of maintaining old code/pages at large companies like Google, with examples of ancient UIs still in use. Some argue it's necessary work that's hard to staff, others say big companies should invest more.Phone number privacy
Concerns about how masked phone numbers across multiple services can be combined to reveal full numbers. Discussion of whether phone numbers should be considered private given their widespread collection and data breaches.
5Building supercomputers for autocrats probably isn't good for democracy
- OpenAI announced a partnership to build AI data centers in the UAE, claiming it promotes democratic values. However, critics point out the UAE's poor democratic record, with Freedom House giving it 18/100 for civil liberties. The deal raises concerns about empowering an autocratic government with significant AI capabilities.
Special treatment of OpenAI
Discussion about unfair focus on OpenAI vs other tech companies like Cisco/Oracle that have long histories of working with autocratic regimes, with some arguing OpenAI deserves scrutiny due to its non-profit origins and democratic claimsSurveillance capabilities
Concerns that LLMs enable Orwellian surveillance at scale for the first time in history, with AI-powered home devices potentially allowing governments to monitor citizens' thoughts and behaviors continuouslyBusiness with autocrats
Debate over whether tech companies should do business with autocratic regimes at all, weighing economic benefits against risks to democracy and human rights, with focus on Gulf states' AI initiatives
6Apple introduces a universal design across platforms
- Apple announces a new universal software design featuring Liquid Glass, a translucent material that adapts to content and context across iOS 26, iPadOS 26, macOS Tahoe 26, watchOS 26, and tvOS 26. The design brings greater focus to content while maintaining platform-specific features and familiar user experiences.
Accessibility and readability concerns
Many users worry the translucent glass design will harm readability and accessibility, especially for those with visual impairments. The low contrast and background bleed-through could make text hard to read and UI elements difficult to distinguish.Nostalgia for skeuomorphic design
Discussion of how previous skeuomorphic designs prioritized usability and learnability, with real-world metaphors helping users understand interfaces. Some feel the move away from this has made interfaces less intuitive and harder to teach.Performance and battery impact
Concerns about how the glass effects will impact device performance and battery life, especially on older devices. Some argue it's intentional to drive hardware upgrades, while others note Apple's track record of optimizing similar effects.
7FSE meets the FBI
- This post details the author's experience running a Fediverse server called FSE and their encounter with the FBI's data collection methods. The author discovered that the FBI contracts companies to scrape data, which is then scanned for keywords and analyzed through Facebook, revealing insights about law enforcement's digital surveillance practices.
Fediblock and instance blocking
Discussion about how instances block each other, with some defending blocking toxic users while others claim blocks are based on inaccurate information. Focus on whether blocks are justified or driven by ideology.FBI investigation handling
Debate about how the server admin handled FBI inquiries regarding violent threats, with some praising the transparency while others criticizing the dismissive attitude toward potential real threats.Server administration challenges
Exchange about technical and ethical challenges of running a community server, including handling spam/abuse, implementing protections, and balancing user experience with security measures.
8Finding Shawn Mendes (2019)
- The article analyzes Shawn Mendes' song "Lost in Japan" to humorously deduce his position on the Kuril Islands territorial dispute between Japan and Russia. Through detailed examination of lyrics, flight paths, time zones, and geographical data, the author playfully investigates Mendes' hypothetical location while making the song, ultimately connecting it to geopolitical implications.
Purpose and format
Discussion of the article's purpose, with readers agreeing it's an entertaining exercise in taking trivial topics seriously, serving as satire of conspiracy theories and fandom analysis while being engaging despite lacking traditional purposeSimilar content
Readers sharing other examples of similar long-form investigations of trivial topics, including a 14-hour TGI Friday's endless appetizer investigation and Nathan Fielder's comedy styleTechnical accuracy
Debate about the completeness of the analysis, with readers fact-checking hotel existence on Iturup Island and discussing alternative interpretations of the song's geographical references
9Why Android can't use CDC Ethernet (2023)
- The article explains why CDC Ethernet adapters often don't work on Android devices, focusing on the technical debugging process. It discusses how Android's compatibility with USB ethernet adapters depends on the device's kernel configuration and provides guidance on determining which adapters are supported on specific Android phones.
Android's interface naming restrictions
Discussion of Android's limitation to only recognize "ethX" named network interfaces, a MAC address bit workaround, and Google's attempts to fix this in recent Android versions through regex changes.Root access debate
Debate about whether root access is necessary for device ownership vs security concerns, with some arguing for user freedom while others warn about expanded attack surface and network traffic risks.Multiple network connections
Users frustrated with Android's inability to connect to multiple networks simultaneously, especially when one network lacks internet access, compared to Linux and Windows capabilities.
10LLMs are cheap
- The article explains that generative AI and Large Language Models (LLMs) are actually quite inexpensive to operate, contrary to common belief. Through price comparisons with web search services and detailed cost analysis, it demonstrates that LLM operations can be 25 times cheaper than comparable search services.
Cost vs profitability
Debate over whether LLM API pricing reflects true costs or is artificially low to gain market share. Some argue current prices are unsustainable loss-leaders, while others contend inference costs are genuinely low enough for profitability.Monetization challenges
Discussion of OpenAI's $5B loss and the difficulty of monetizing 500M users. Converting free users to paid ones is seen as extremely challenging, though advertising revenue could be an alternative path to profitability.Usage patterns and comparison
Debate over comparing LLM vs search engine costs, noting LLMs are used for complex tasks requiring more tokens while search engines primarily index websites. Critics argue the comparison overlooks fundamental differences in usage.