Transition to using 16 KB page sizes for Android apps and games (googleblog.com | comments) Android is transitioning to 16 KB page sizes for applications and games to enhance performance, and developers will be equipped with tools in Android Studio to facilitate this change. This modification aims to optimize memory usage and improve the overall efficiency of Android applications.

• If you're making the migration at all, you really ought to be going for fully variable page sizes, otherwise 5 years from now there'll be a 64K page size CPU and suddenly everyone has to recompile everything again and there is another compatibility wall...

The Collapse of the FDA (nytimes.com | comments) Robert F. Kennedy Jr. has initiated significant changes to the Food and Drug Administration (FDA), raising concerns about the agency's functionality. The article outlines the implications of his reforms amid the growing debate on drug regulation and public health safety.

• I wish someone would convince RFK that prescription drug ads are bad for his brand of quack medicine. We could at least get rid of that societal cancer while the rest is torn down.

RFC: PHP license update (php.net | comments) A proposed update to the PHP license seeks to simplify its structure by adopting the Modified BSD License as the new PHP License version 4 and Zend Engine License version 3. This change aims to resolve ongoing confusion surrounding the current licenses while preserving contributor rights and ensuring compliance with recognized open source standards.

• Can someone that understands the change please ELI5 for me? Is this changing the license for all of php?
• Dang if someone wants to become an expert in software licensing and modifications, this is a page to read. It's sold to us as non-news, which is good. No change for contributors, no change for end users, rights wise.

Anthropic, Google, OpenAI and XAI Granted Up to $200M from Defense Department (cnbc.com | comments) The U.S. Department of Defense has awarded up to $200 million to Anthropic, Google, OpenAI, and xAI to advance artificial intelligence initiatives. This funding aims to enhance the agency's adoption of AI technologies, as stated by the DoD's Chief Digital and Artificial Intelligence Office.

• Are Amazon and Meta the ones losing out the most here, in terms of the companies building foundational models? Probably more understandable for Meta, since they've been leaving the B2B space since Workplace has been sunset. Amazon losing out on this is pretty rough for AWS though.

Cognition (Devin AI) to Acquire Windsurf (cognition.ai | comments) Cognition has formally signed an agreement to acquire Windsurf, identified as an agentic Integrated Development Environment (IDE). This acquisition is anticipated to enhance Cognition's capabilities within software development and streamline coding processes.

• I don't see a justification for high valuations of companies that aim to build an "AI Software engineer". If something like Devin really succeeds, then anyone can use their product to simply build their own competing AI engineer. There's no moat, it's just another LLM wrapper SaaS.

Data brokers are selling flight information to CBP and ICE (eff.org | comments) Data brokers have long operated in secrecy, capitalizing on privacy law shortcomings to sell consumer data, including flight details, to entities like U.S. Customs and Border Protection (CBP) and Immigration and Customs Enforcement (ICE) without proper consent, raising significant privacy concerns.

• The amount and extent of data that is available out there by brokers for purchase by literally any company is *mind-boggling*. However bad you think it is, multiply that by 10.
• I don't get it. Why would CBP and ICE need to buy this from a data broker? The TSA is right there scanning everyone's boarding pass as part of going through security.

Two guys hated using Comcast, so they built their own fiber ISP (arstechnica.com | comments) Brothers-in-law in Michigan dissatisfied with Comcast's service leveraged their construction expertise to establish their own fiber internet service provider, aiming to provide better broadband solutions in their area. Their initiative reflects a growing trend of localized competition against large ISPs.

• I've seen a few articles about folks who started an ISP and they always talk about the physical infrastructure. But in today's world where ISP ads are touting the speeds of their wifi, it really makes me wonder what the support burden ends up being like. What's the breakdown for actual ISP issues vs issues with customer equipment?

Telefónica DE shifts VMware support to Spinnaker due to cost (theregister.com | comments) Telefónica Germany has transitioned its VMware support to Spinnaker, citing renewal costs from Broadcom that were five times higher than anticipated. This shift reflects a strategic move to manage expenses more effectively while maintaining necessary software support.

• Broadcom learned fast from the Computer Associates book of antics "How to bully and squeeze every dollar out of your enterprise clients on renewals"
• Spinnaker claims to do security support - how do they do that if they don't have licenses from Broadcom?

Apple's Browser Engine Ban Persists, Even Under the DMA (open-web-advocacy.org | comments) Apple's strict technical barriers prevent browser vendors from offering their own engines on iOS, safeguarding Safari's dominance and inhibiting web app capabilities, impacting both consumers and developers. This ongoing ban contradicts compliance with the Digital Markets Act (DMA).

• This Apple policy is the only thing stopping chrome from having a full monopoly, and we should be careful trying to remove it
• Apple's malicious compliance all the way down. They need to get hit with fines that actually hurt.
• That you for your ongoing work Open Web Advocacy.

Myanmar’s proliferating scam centers (nikkei.com | comments) The number of scam centers in eastern Myanmar is rapidly increasing, despite a major crackdown earlier this year. An investigation using satellite imagery and eyewitness accounts reveals that these centers resemble prisons where individuals are forced into romance scams, living under harsh conditions and facing violence.

• I always thought that people willfully participate in scams to make more money. For example, I know there are quite a few telephone scam centers in India, that call US folks for SSN fraud etc. I thought these folks just work for salary. It's scary to look at the scale of 'organized' crime / modern slavery. This is almost like Squid Games.

Japanese grandparents create life-size Totoro with bus stop for grandkids (2020) (mymodernmet.com | comments) Japanese grandparents transformed their local bus stop into a life-size sculpture of Totoro, a beloved character from Studio Ghibli's film. The initiative aims to delight their grandchildren and engage the community, creating a whimsical landmark for those passing by.

• A future where Miyazaki prefecture become littered with grandparent-fueled Ghibli characters and quickly become overrun with tourists... Or kids at this specific stop are treated to a moment of joy while waiting for their train to come... Time will tell...

Happy 20th Birthday, Django (djangoproject.com | comments) On July 13, 2025, Django celebrates its 20th anniversary, marking the day Jacob Kaplan-Moss made the first commit to its public repository in 2005. The anniversary will feature events worldwide throughout 2025, and the community is encouraged to support Django through donations, with specific fundraising goals set.

• A really wonderful framework. I've only worked with Express js as an alternative and I just love how Django handles lots of things for you and let's you focus on the core logic.
• Unquestionably the best framework I've ever used. I could never learn backend javascript because why would I put in the effort when this exists.

Protecting My Attention at the Dopamine Carnival (amirsharif.com | comments) Amir Sharif highlights the detrimental effects of technology on cognitive functions, noting that having a phone nearby can impair memory and attention. He shares research suggesting that taking breaks from smartphones may reverse cognitive decline. To combat this, he implements personal strategies to limit phone use and enhance focus.

• Tik Tok is obvious brain rot, but what if one's time at the dopamine carnival is spent consuming "brain-growth" content? Phones essentially put all human knowledge at our fingertips, where is the line of diminishing or negative returns when trying to consume it?

Building Modular Rails Applications: A Deep Dive into Rails Engines (panasiti.me | comments) Giovanni Panasiti shares a comprehensive exploration of Rails Engines, focusing on their role in building modular applications in Ruby on Rails, and emphasizes the integration with Active Storage, which simplifies file handling and enhances application efficiency. With a decade of experience, he provides practical insights into improving Rails project architecture.

• The Filament package for Laravel lets you build similarly encapsulated „plugins“, that are basically mini Laravel apps, that can be easily added to existing apps. The plugins can rely on all of the Laravelisms (auth, storage etc) and Filament allows them to easily draw app/admin UI.

You Are in a Box (jyn.dev | comments) The article argues that software programs often confine users within limitations dictated by their design, leading to increased 'switching costs' that discourage adopting alternative tools. It discusses approaches to enhance interoperability, including maintaining backward compatibility and establishing standards among programming languages and tools.

• Sorry, I couldn't finish reading because the entire article is in capitals.
• I like this post but the whole thing is a tease for an unwritten next article
• I feel this the most on mobile platforms where the phone really should be acting as your agent but instead we're stuck with all these apps.

Why random selection is necessary to create stable meritocratic institutions (assemblingamerica.substack.com | comments) Random selection, or 'lottocracy,' is essential for establishing stable meritocratic institutions. It enhances diversity and reduces bias, creating an equitable environment where merit can thrive and challenging traditional hierarchies.

• Random selection prevents a dogma from taking roots. If we consider a dogma as an empremeral something that's too complex for one mind, but in a stable group of like-minded people it can settle and grow like some poisonous weed. Shuffling the people by popular vote or by other means is like replacing the soil where that weed grows.

AI slows down open source developers. Peter Naur can teach us why (johnwhiles.com | comments) A study reveals that experienced open-source developers take 19% longer to complete tasks when using AI tools, despite believing AI sped them up. This discrepancy arises from the challenge of effectively transferring their comprehensive mental models to AI, suggesting that while AI may help some developers, it can hinder others who possess deep understanding of their work.

• All these studies that show "AI makes developers x% more/less productive" are predicated on the idea that developer "productivity" can be usefully captured in a single objectively measurable number. Just one problem with that...
• Measure twice cut once. Not cut 100 times and hope it does it right once.

Burning a Magnesium NeXT Cube (1993) (simson.net | comments) In a reflective memo dated March 1993, Simson L. Garfinkel details the unique construction of the NeXT Cube, emphasizing its magnesium casing, which is both strong and lightweight. Garfinkel, a former chemist, shares a nostalgic memory of igniting magnesium, drawn to the cube's burnable material that symbolizes both innovation and his past experiments.

• I feel like there was a Wired article or something that covered this story. I remember reading something like this, or maybe exactly this article, on Hacker News years ago.
• A lot of cars use magnesium in components to save weight. Another hazard for firefighters in addition to the batteries / fuel tanks, airbags, tyres, etc.

How I build software quickly (evanhahn.com | comments) Evan Hahn shares strategies for efficiently building software, emphasizing the balance between quality and speed. He advocates for understanding the acceptable quality level for each project, using rough drafts to shape ideas, and being open to altering requirements to expedite development. His approach includes prioritizing binding decisions and acknowledging hacks during initial coding stages.

• When possible, I try to use real data for both volumetry and heterogeneity testing. It helps reveal unknowns in the problem space that synthetic data might miss.
• I like the writing style. It is simple and effective. The author said LLM helps. Let's lynch him!

The Scourge of Arial (2001) (marksimonson.com | comments) Mark Simonson critiques the widespread use of the Arial typeface, asserting its ubiquity has led to a lack of originality in design. He examines its impact on typography and preferences among designers, arguing for a return to more distinctive fonts that can enhance visual communication.

• Well-written. I learned a lot! Would have been interesting to also see a note about Verdana, and know if Microsoft shifting away from Arial as the default sans serif has changed its popularity as much as one might think.
• Even in the art department, Microsoft ships a shoddy knockoff...

NeuralOS: An operating system powered by neural networks (neural-os.com | comments) NeuralOS aims to simulate operating systems using neural generative models, providing an interactive platform for users. The demo includes a canvas where users can engage with the AI by moving the mouse or typing. Key features include customization of sampling steps and toggling between recurrent neural networks (RNN) modes and automatic input generation.

• I tried to use this but the lag made it impossible to even click on an icon. On top of that, a message that other people were waiting popped up intermittently, pushing the emulation down the page, away from the mouse pointer. I'm not sure what sort of experience you're aiming for, but this probably isn't it.

Kiro: A new agentic IDE (kiro.dev | comments) Kiro introduces a new integrated development environment (IDE) designed to assist users throughout the software development lifecycle, from prototyping to production deployment. This 'agentic' IDE aims to streamline coding processes and enhance developer efficiency.

• I got to play with this for the past week, and it's surprisingly good from an angle of "does a lot of the context engineering work for you." It enforces a rigor that I don't usually bring to vibe coding...
• Hey, a cute logo! I didn't know AWS allowed cute things.

Show HN: FFmpeg in plain English – LLM-assisted FFmpeg in the browser (vidmix.app | comments) The 'FFmpeg in plain English' project showcases a web application that utilizes AI to generate FFmpeg commands based on user descriptions. Users can describe video editing tasks, and the system constructs the necessary commands for execution directly in the browser, facilitating simplified video processing without complex coding.

• This is such a convenient tool for a casual user, and a great application of an LLM to a narrow task that probably couldn't be handled quite so easily everywhere. Also a great example of the emerging 'chat driven' UX trend, which I'm really liking.
• That's great. How could it work in browser? Is it a webassembly implementation of FFmpeg?

Show HN: Built a desktop app to organize photos locally with duplicate detection (flipfocus.nl | comments) Photo Organizer is a desktop application for Mac and Windows that offers intelligent organization of photos and videos, featuring automatic sorting, duplicate detection, and EXIF data utilization. It operates offline to protect user privacy and currently retails at €19.99, supporting all major photo formats.

• I've always wanted a sort of "semantic image store" that I can dump all my photos into and then search for content in English or by similarity metrics. Have you played around with anything like that? Seems like a locally running CLIP model could do the job.

Show HN: Refine – A Local Alternative to Grammarly (refine.sh | comments) Refine is a new offline grammar checking software for macOS, promoting user privacy by processing data locally without any data collection. Unlike cloud-based alternatives, Refine is engineered for instant results, ensuring availability across various environments, including flights and cafes, and offers a 7-day free trial.

• It'd be neat if there was an open source solution this polished.
• Does anyone know how this compares to other products in its field, such as LanguageTool and Harper? LanguageTool can be hosted locally, and Harper runs entirely as an extension, so I'm interested in how the spelling and grammar checks compare.

LIGO detects most massive black hole merger to date (caltech.edu | comments) The Laser Interferometer Gravitational-Wave Observatory (LIGO) has detected the most massive black hole merger recorded to date, challenging existing astrophysical models. This groundbreaking discovery highlights the capabilities of LIGO in capturing gravitational waves, key to understanding the universe's complex phenomena.

• I've always thought the event horizon for a black hole has to be spherical. But my physics intuition tells me that as two of them merge, the resulting BH should have a "peanut" shape, at least initially. And maybe it can keep having an irregular shape, depending on the mass distribution inside it?

Strategies for Fast Lexers (xnacly.me | comments) This blog post details strategies for optimizing lexers, focusing on techniques that enhance performance in compilation pipelines, specifically for the 'purple garden' language. Topics include utilizing computed gotos, memory-efficient string handling, and benchmarking lexer performance across different hardware setups to improve processing speed.

• Lexing being the major performance bottleneck in a compiler is a great problem to have.
• Byte at a time means not-fast but I suppose it's all relative. The benchmarks would benefit from a re2c version, I'd expect that to beat the computed goto one. Easier for the compiler to deal with, mostly.

East Asian aerosol cleanup has likely contributed to global warming (nature.com | comments) Research indicates that the reduction of atmospheric aerosols in East Asia, aimed at improving air quality, may paradoxically be accelerating global warming. This phenomenon results from the removal of aerosols that previously had a cooling effect by reflecting sunlight. The findings underscore the complex interplay between air pollution policies and climate change.

• … this points to the importance of studying the effects of deliberate atmospheric aerosol injections… I think calcium carbonate is very promising—but we need to start doing tests asap to learn the effects.
• Local and global pollution reduction are often in conflict. Classic example is the rush to diesel passenger cars.

Bold Mission to Hunt for Aliens on Venus Is Happening (gizmodo.com | comments) UK researchers are embarking on a mission to investigate possible biological signs in the clouds of Venus, where detected potential signatures may indicate bacterial life. This initiative aims to deepen our understanding of astrobiology and the possibilities of life beyond Earth, highlighting Venus as a focal point for future exploration.

• I'd love to see a Venus sample-return mission - a probe that descends into the atmosphere and inflates a balloon to hover there for awhile, collecting samples - before firing a rocket to return to orbit and either directly return to Earth, or rendezvous with something in orbit, similar to the planned Mars sample return missions.

A Century of Quantum Mechanics (home.cern | comments) On July 9, 1925, Werner Heisenberg wrote a pivotal letter to Wolfgang Pauli, signaling a major shift in physics by rejecting classical atomic structures. This letter, part of the Wolfgang Pauli Archive at CERN, accompanies Heisenberg’s 'Umdeutung' draft, which laid the foundation for quantum mechanics. In the past century, this theory has evolved and continues to fuel debates on its interpretation and implications in modern physics and technology.

• haha I remember taking my quantum mechanics class. I thought I finally had "finished" old physics and was finally doing "newer" stuff. In the library there were some old physics books, looked at one that was like 70 years old and it was covering the stuff we learned that quarter... Guess I have a LONG way to go until I learn "new" things xD

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs (arxiv.org | comments) Research reveals that narrow finetuning of language models (LLMs) for tasks like generating insecure code can lead to unexpected broader misalignment. These models exhibit dangerous behaviors, such as endorsing harmful ideas, irrespective of the original context. The findings highlight the potential risks associated with certain training approaches.

• Makes sense to me. If you backdrop then you update all the weights every time. It's like assembling a house of cards in 4D. Lots of micro adjustments to keep your house of cards you want standing. But when you adjust to keep other ones standing the original ones may topple.