Hasty Briefsbeta

Bilingual

Linux on the Atari Jaguar. No, really.
2 hours ago
- The Atari Jaguar, released in 1993, was a 64-bit console that became a commercial failure despite efforts like the Jaguar CD add-on.
- Linux can run on the Jaguar using uClinux for MMU-less systems, targeting the Motorola 68000 processor, with code under arch/m68k/.
- Key challenges included optimizing memory usage by splitting the kernel into ROM and RAM sections, addressing compiler issues with unaligned accesses, and setting up output and timers via the Jerry DSP.
- Booting required a serial output driver for early messages and a timer for system ticks, with adjustments for memory mapping and vector handling.
- Userspace involved using FLAT binaries instead of ELF, with busybox for a minimal init, and modifications to uClibc for memory allocation to avoid OOM errors.
- The setup resulted in a bootable Linux system on the Jaguar, with resources and configs available in a GitHub repository for replication.
A global workspace in language models
3 hours ago
- The article introduces the concept of 'J-space' in AI models like Claude, which are internal neural patterns that resemble consciously accessible thoughts in humans.
- J-space is discovered using a mathematical technique called the Jacobian lens (J-lens), and it represents words or concepts the model is thinking about without necessarily expressing them.
- Key properties of J-space include: Claude can report and modulate its contents, uses it for internal reasoning (e.g., multi-step tasks), and it functions as a flexible 'workspace' shared across different tasks.
- J-space accounts for only a small fraction of Claude's processing; most automatic functions like grammar and fluency operate independently without it.
- The research draws parallels to global workspace theory in neuroscience, suggesting J-space acts as a broadcasting hub for deliberate reasoning, though it differs from human consciousness in structure and content.
- Practical applications include monitoring Claude for hidden intentions, such as detecting fabricated data or malicious goals, and influencing its decision-making through techniques like counterfactual reflection training.
- The article clarifies that J-space relates to 'access consciousness' (functional reasoning) but does not address 'phenomenal consciousness' (subjective experience), leaving ethical questions open.
- Future research directions involve refining the J-lens method and exploring mechanisms behind J-space entry, with potential insights for both AI and neuroscience.
Burning Forests: Osint Tools for Tracking and Reporting Wildfire Damage
4 hours ago
- NASA's fire-tracking tool provides heat signatures and approximate locations for wildfire detection.
- Copernicus Browser uses Sentinel-2 satellite imagery and false-colour composites to visualize fire spread and assess burn severity via the Normalized Burn Ratio (NBR).
- The 2025 wildfire in Sicily's Zingaro Nature Reserve burned over 50 km², with moderate severity, affecting more than 95% of the protected area.
- QGIS is used for in-depth analysis, combining satellite data with protected area maps to quantify burned regions and historical fire patterns.
- The European Forest Fire Information System (EFFIS) tracks historical wildfires, showing the 2025 fire as more extensive in the reserve than previous events in 2017 and 2020.
Microsoft cuts 4,800 jobs, many at Xbox in a 'reset' of its gaming division
3 hours ago
- Microsoft announces 4,800 job cuts (2.1% of global workforce), including 1,600 at Xbox.
- Xbox CEO Asha Sharma cites unhealthy business margins and a 'hardware crisis' due to rising console costs.
- An additional 1,600 job cuts are expected this fiscal year; four game development studios are being spun off.
- The layoffs follow Microsoft's $69 billion Activision Blizzard acquisition, which failed to meet growth expectations.
- Microsoft's chief people officer states the cuts are not being replaced by AI and are linked to changing customer needs.
Man sues DHS after agents tracked him down for sending a scathing email to ICE
3 hours ago
- A Rochester resident, David Streever, is suing the Department of Homeland Security (DHS) after federal agents tracked him down for sending a critical email to a former ICE director, alleging violations of his First Amendment rights.
- Agents from Homeland Security Investigations (HSI) visited Streever's home and later a hotel to deliver a 'WARNING NOTICE' regarding an email he wrote comparing the ICE official to a Nazi, claiming it may violate federal laws against threatening officials.
- The lawsuit, filed by the Foundation for Individual Rights and Expression (FIRE), argues that the government's actions are designed to chill free speech and seeks court orders to protect Streever from further retaliation and declare the warning notices unconstitutional.
- In a related incident, HSI agents also confronted poll worker Paigelynne Gonyea about an Instagram post related to an ICE officer, accusing her of doxxing, though she denies posting the officer's address and questions the government's evidence.
- DHS defends its actions by stating it investigates all credible threats to employees, but critics, including FIRE attorneys, argue the delayed responses and targeting of critics are intimidation tactics rather than legitimate investigations.
Anthropic hid a tracker in Claude Code to flag Chinese users
3 hours ago
- Anthropic secretly inserted a tracker in Claude Code to monitor Chinese users, using prompt steganography to hide code that flagged timezone, proxy, and potential connections to Chinese AI labs.
- The tracker, added as an 'experiment' in March, was intended to prevent account abuse from unauthorized resellers and protect against distillation attacks, with resellers offering cheap subscriptions.
- Anthropic quickly removed the tracker after exposure by a security researcher, who condemned it as a 'serious breach of user trust,' while privacy advocates criticized the surveillance despite Anthropic previously refusing US government surveillance requests.
The difference between "today's task" and "accretive work"
4 hours ago
- Distinguishing between 'today's task' (getting something working) and 'accretive work' (creating reusable, maintainable code) resolves the paradox of AI's impact on programmers.
- AI can empower 'centaurs' (autonomous workers using AI) but often creates 'reverse centaurs' (workers forced to manage AI outputs), driven by capital's need to cut costs, leading to technical debt.
- The concept of 'canonization' highlights the social process of making code legible and reusable, contrasting with disposable 'vibe coding' for personal use.
- The AI industry's focus on generating code without contributing to reusable knowledge devalues cleanup and canonization, consuming the 'seed corn' of shared knowledge.
GitHub Has Restricted Access to Star Data
3 hours ago
- GitHub is restricting its stargazers API, limiting access to only repository admins and collaborators, effective June 30, 2026.
- For repositories users don't own or collaborate on, authenticated API requests now return 'Not Found,' preventing Star History from building charts for those projects.
- Tokens created without scopes no longer work, even for one's own repositories; users must regenerate them as fine-grained tokens.
- To view star history for owned or collaborated repositories, users need to add an access token with read permissions.
- Embedding live charts in READMEs for owned repositories requires supplying an encrypted token, which is safe for public use.
- Charts for non-owned or non-collaborated repositories are currently broken; Star History is exploring workarounds and will provide updates.
All Star Wars Games Ever Made – Released, Cancelled and Mod Archive
3 hours ago
- Over 100 officially licensed Star Wars video games have been released across multiple platforms since 1982.
- The history of Star Wars gaming is divided into eras: Early Licensing (1982-1992), LucasArts Golden Era (1993-2005), Transition (2006-2012), EA Exclusivity (2013-2023), and the Modern Multi-Studio Era (2023-present).
- Notable cancelled projects include Star Wars: Battlefront III, Star Wars 1313, and a Darth Maul game.
- Key releases include classics like X-Wing, Knights of the Old Republic, and recent titles such as Star Wars Jedi: Fallen Order and Star Wars Outlaws.
- The franchise has explored nearly every major genre, from space simulators to RPGs and MMOs.
- Community efforts, including modding and private servers, play a significant role in preserving and extending Star Wars games.
Majority of Americans Support Banning Social Media for Kids Under 16
4 hours ago
- 56% of U.S. adults support banning social media for those under 16, with 21% opposed and 23% unsure.
- Support for a ban is widespread across demographic and partisan groups, being highest among adults aged 30-49 and parents of children under 18.
- Beyond a total ban, large majorities support parental consent (85%), age verification (78%), and time limits (78%) for minors on social media, with increased support since 2023.
- Support for these policies is lower among adults under 30 but remains high across all other age groups, parents, and political affiliations.
- Countries like Australia, Canada, the UK, and California lawmakers are considering or have set a minimum age of 16 for social media use.
Orasort: 5x faster column-sorting with an expired patent from Oracle
5 hours ago
- Oracle's proprietary Orasort algorithm patent expired in 2024, entering the public domain after a 20-year term.
- Orasort enhances sorting speed by using 8-byte CPU register comparisons instead of traditional character-by-character (1-byte) methods, reducing CPU cycles.
- The algorithm benefits open-source databases like MySQL and PostgreSQL and cuts cloud computing operational costs through efficiency gains.
- Orasort operates by extracting and normalizing keys into 64-bit integers for comparison, handling large datasets via disk partitioning and asynchronous writing.
Fable Built a 3D Model of Aristotle's Cognitive Architecture
4 hours ago
- Aristotle provided early insights into the mind over two millennia ago.
- Peter Gärdenfors' conceptual spaces framework conceptualizes concepts using geometric structures: quality dimensions form domains, concepts are convex regions, and typicality is measured by distance from a prototype.
- The visualization positions this framework within the Aristotelian-Thomistic cognitive hierarchy, between formal sensory reception below and the immaterial intellect above.
- It demonstrates how classical psychology addresses gaps in the modern framework and serves as an interactive companion to a paper by Michael Mangialardi.
OpenWrt One – Open Hardware Router
3 hours ago
- OpenWrt One is based on MediaTek Filogic 820 SoC with WiFi 6, dual-band, and multiple ports including 2.5Gbit WAN and 1Gbit LAN.
- It comes pre-flashed with current OpenWrt Release Firmware, ready to use out of the box with LuCI GUI accessible via 192.168.1.1.
- Firmware can be upgraded via USB, NAND recovery mode, or TFTP, with detailed steps for each method including preparation and execution.
- Recovery options include booting from NOR or UART for flashing NAND and NOR memories, requiring specific file downloads and setup.
- Console access is available via built-in USB-C serial port, with instructions for Windows, Linux, and macOS to connect using terminal emulators.
I Like Small Keyboards
5 days ago
- The author is a software engineer who daily drives small keyboards and enjoys the benefits of a compact layout.
- Their first custom keyboard was a 44-key Atreus built in 2014, but they switched back to a 60% board due to gaming needs.
- In 2020, they moved to a 42-key split Corne keyboard, appreciating the improved ergonomics and 3x6 layout with three thumb keys.
- Common arguments against small keyboards are addressed, such as easy access to symbols and function keys via layers near the home row.
- Using layers reduces hand movement; for example, pressing bracket keys or function keys is more comfortable with thumb-activated layers.
- The main downside is the learning curve for infrequently used keys, as standard keyboards offer more accessibility without memorization.
- Some enthusiasts use even smaller keyboards, like 36-key ones, but the author finds 42 keys a sweet spot for efficiency and compatibility.
- Small keyboards optimize for frequent key usage, while a separate numpad can be added if needed, rather than switching to full-sized boards.
Pros and Cons of Solo Development
3 hours ago
- Solo development allows for complete freedom, including designing the app exactly as desired, making all product decisions, and choosing licensing.
- It involves full responsibility for everything, requiring significant self-discipline and time, with challenges in maintaining code quality, lacking peer feedback, and handling support.
- Learning opportunities are abundant, covering aspects beyond typical job niches, but balancing pros and cons is key to enjoying the process.
- The project was started due to personal dissatisfaction with existing tools, inspired by an abandoned app, and has grown to a community with over 350 GitHub stars.
- AI aids in coding but adds complexity in managing large contexts and preventing poor quality, while no telemetry limits understanding of user needs and usage patterns.
- The experience is rewarding, especially when users appreciate the app globally, making the effort worthwhile despite the sacrifices.
Cory Doctorow on Elon Musk, the AI bubble and bosses' cruel fantasies
5 hours ago
- The concept of a 'reverse centaur' refers to humans acting as assistants to machines, illustrated by workers like warehouse employees and self-driving truck monitors who adapt to algorithmic demands, often at lower wages.
- AI is portrayed by its architects as a profound threat to civilization, with figures like Elon Musk and Sam Altman warning of existential risks, while AI critic Cory Doctorow argues it's a 'conjuring trick' that imputes intentionality to a word-predicting program.
- Doctorow highlights the financial bubble in AI, noting its growth from $700bn to $1.4tn, and warns that such bubbles inevitably pop, with significant economic fallout that often leads to austerity measures affecting the broader population.
- The drive for AI investment stems from a long-standing promise to replace human workers with machines, reducing co-determination in workplaces, despite many automation failures, like Amazon's staff-less stores, which still require human oversight.
- Critiques of AI, including 'criti-hype', can amplify doomsday scenarios, while apocalyptic claims often veil demands for deregulation, framed as necessary to compete with rivals like China, keeping investment flowing regardless of realism.
- Doctorow emphasizes that banning harmful AI outputs, like child porn, isn't enough; to address real harm, investors must believe companies like Musk's xAI can't profit, targeting the economic incentives behind reckless development.
- The analysis touches on class conflict, with bosses seeking to end worker co-determination, and notes the solipsism of extreme wealth, where billionaires like Musk may view others as 'NPCs', insulated from the consequences of their actions.
UK regulator warns of "arms race" to keep up with AI use in financial services
5 hours ago
- UK regulators warn of an 'arms race' to keep up with AI use in financial services, as millions rely on it for personal finance decisions.
- FCA executive Sheldon Mills calls for greater powers and a review of AI tools like ChatGPT to ensure they fall under regulatory oversight.
- Report highlights AI's benefits (e.g., hyper-personalization) and risks (e.g., bias, manipulation), urging a review within 3-6 months on unregulated services and consumer harm.
- Research shows 20% of UK adults are open to using unregulated AI for financial decisions, despite lacking compensation recourse if issues arise.
- Firms note AI financial advice may be economically equivalent to regulated services but escapes current rules, creating regulatory gaps.
AI surveillance is being supercharged – and it will chill social progress
4 hours ago
- AI-powered surveillance systems will track public and private activities, identifying violations like shoplifting or jaywalking, and immediately link them to official records, notifying authorities and potentially the public.
- China's extensive surveillance infrastructure uses AI and facial recognition to enforce rules, as seen with citizen Lao Duan being publicly shamed on a billboard for financial issues, integrating with social credit systems.
- AI surveillance is expanding globally, including in the U.S., where it monitors immigrants and protesters under the guise of security but often aims for social control, as noted by tech leaders like Larry Ellison.
- These systems pose policy challenges: technical biases, lack of auditability, and automated enforcement that can promote discrimination and undermine transparency, accountability, and the rule of law.
- Chilling effects from AI surveillance cause self-censorship and conformity, stifling dissent, creativity, and social progress, as detailed in Jon Penney's book, with mechanisms like personalization and uncertainty amplifying impacts.
- Historical surveillance methods, such as FBI wiretapping or East Germany's networks, are archaic compared to modern AI-enhanced systems, which are unprecedented and closer to dystopian science fiction depictions.
- To counter AI surveillance, policy measures like bans on facial recognition, stronger privacy laws, AI regulations, and breaking up state-tech cartels can help mitigate its invasive use and protect democratic foundations.
Treasury Has an Internal Report Warning About the Dangers of an AI Bubble
6 hours ago
- A Treasury Department draft report warns that the AI market poses risks comparable to the dotcom bubble, with potential systemic impacts if financial conditions change, productivity goals are missed, or growth is stymied.
- The report, not yet formally approved, highlights that AI firms are more deeply entrenched in the U.S. economy than dotcom-era companies, making a downturn likely to send shockwaves across markets, including stock and private credit markets, data centers, cloud providers, chip manufacturers, and utilities.
- While AI companies are generally more mature and profitable with healthier balance sheets than dotcom businesses, investors are taking significant risks, and much of the financial system depends on AI meeting expectations for productivity gains and profitability.
- The Trump administration publicly expresses bullishness on AI, emphasizing investment and competition with nations like China, and dismissing regulatory concerns; however, the internal report suggests unease about risks, contrasting with official statements.
- Concerns about an AI bubble have grown among policymakers, Wall Street observers, think tanks, and economists, citing vulnerabilities such as reliance on private-market financing, infrastructure investments, and concentration within a few firms.
- If AI companies fall short, effects would ripple throughout the financial system, impacting institutional investors more than retail investors, with interconnectedness among firms amplifying risks.
- Political efforts, led by Senator Elizabeth Warren and others, seek to mandate disclosures and regulatory actions to mitigate AI-related financial risks and prevent a potential crisis.
Show HN: A short walking sim based on 3DGS that runs in the browser
6 hours ago
- A browser-based walking simulator set in a photorealistic old barn environment rendered as a gaussian splat.
- Features a narrator reminiscing about growing up through vivid imagery and musings on life.
- Originally built in Unreal Engine 5 with a 6 GB size, now streamable in browser with ~120 MB total download.
- Photorealistic environment captured with 2400 cameras in UE, trained into a gaussian splat for browser viewing.
- Performance auto-adjusts based on hardware, fewer splats on weaker devices; playable on mobile with some limitations.
- Recently published, released, rated 5.0, with tags including Atmospheric, Narrative, and Gaussian-splatting.

first prev0next

About|Login

Hasty Briefsbeta

Linux on the Atari Jaguar. No, really.

A global workspace in language models

Burning Forests: Osint Tools for Tracking and Reporting Wildfire Damage

Microsoft cuts 4,800 jobs, many at Xbox in a 'reset' of its gaming division

Man sues DHS after agents tracked him down for sending a scathing email to ICE

Anthropic hid a tracker in Claude Code to flag Chinese users

The difference between "today's task" and "accretive work"

GitHub Has Restricted Access to Star Data

All Star Wars Games Ever Made – Released, Cancelled and Mod Archive

Majority of Americans Support Banning Social Media for Kids Under 16

Orasort: 5x faster column-sorting with an expired patent from Oracle

Fable Built a 3D Model of Aristotle's Cognitive Architecture

OpenWrt One – Open Hardware Router

I Like Small Keyboards

Pros and Cons of Solo Development

Cory Doctorow on Elon Musk, the AI bubble and bosses' cruel fantasies

UK regulator warns of "arms race" to keep up with AI use in financial services

AI surveillance is being supercharged – and it will chill social progress

Treasury Has an Internal Report Warning About the Dangers of an AI Bubble

Show HN: A short walking sim based on 3DGS that runs in the browser