Latest in AI
Harness engineering: Leveraging Codex in an agent-first world★ 76
Hacker News (AI keywords)10 days agoCommentaryOpenAI describes an internal experiment where Codex generated an entire product codebase from an empty repository. The post argues that engineers shift from writing code to designing environments, constraints, documentation, and feedback loops. Key practices include repo-local knowledge, mechanical architecture enforcement, agent-readable UI and observability, lightweight PR flow, and continuous cleanup.
Tiny hackable CUDA language model implementation
Hacker News (AI keywords)10 days agoNew ToolThis GitHub project implements a compact generative pretrained transformer as an autoregressive byte-level sequence model. Its README describes causal self-attention, RoPE, feed-forward layers, AdamW, cross-entropy training, and BLAS/OpenBLAS-backed matrix operations, with CUDA toolkit listed in setup steps. It is most useful as an educational and experimental codebase, not as a production-grade replacement for large commercial LLMs.
The most interesting startups right now want to get you off your phone
TechCrunch AI10 days agoBusinessTechCrunch highlights a startup trend moving in the opposite direction of the AI fundraising boom. Mirror founder Brynn Putnam has raised money for Board, a company focused on in-person games and social experiences. The piece also points to viral cyberdeck creators making whimsical DIY computers that encourage users to get off their phones and reconnect with the physical world.
Mantine DataTable source repo compromised; owner account suspended★ 74
Hacker News (AI keywords)10 days agoIncidentA GitHub security notice says Mantine DataTable and other repositories received unauthorized commits through the github-actions bot. The npm packages were reported safe; the risk targets developers who recently cloned or pulled the source and open it in VS Code, Cursor, Claude Code, Gemini, or run npm test. A later update links the payload to the Miasma / Shai-Hulud worm family and says a stolen credential is the likely path.
This is your laptop… on AI
The Verge AI10 days agoHardwareThe episode frames developer conference season around Big Tech’s conviction that AI will reshape how people use technology. Nvidia CEO Jensen Huang is highlighted for describing a completely new way to use laptops. Based on the provided excerpt, this is more of an industry commentary on AI PCs than a concrete product-spec report.
Launch HN: General Instinct (YC P26) - Frontier models on edge devices
Hacker News (AI keywords)10 days agoNew ToolGeneral Instinct is a YC P26 company introduced through a Launch HN post. Its headline positioning is bringing frontier models to edge devices, suggesting local or embedded AI deployment rather than purely cloud-based inference. Since no article body is available, details such as supported models, hardware, benchmarks, pricing, and developer tooling cannot be verified from the provided source.
Gemma 4 QAT models: Optimizing model compression for mobile and laptop efficiency★ 72
Hacker News (AI keywords)10 days agoReleaseGoogle released new Gemma 4 checkpoints optimized with Quantization-Aware Training to preserve quality after compression. The release includes Q4_0 checkpoints and a mobile-focused quantization format that can reduce Gemma 4 E2B memory use to about 1GB, or below 1GB for a text-only configuration. The models are available through Hugging Face and supported across llama.cpp, Ollama, LM Studio, LiteRT-LM, Transformers.js, SGLang, vLLM, MLX, and Unsloth.
U.S. Military Turned GPS into a Global "Numbers Station"
Hacker News (AI keywords)10 days agoCommentaryAccording to investigative outlet 404 Media, evidence suggests the U.S. military has repurposed the Global Positioning System (GPS) into a modern "numbers station." By embedding encrypted data within standard GPS broadcasts, the military can securely transmit covert messages to agents or assets worldwide. This technique leverages existing satellite infrastructure to achieve global coverage with near-perfect receiver anonymity.
pg_durable: Microsoft open sources in-database durable execution
Hacker News (AI keywords)10 days agoReleaseMicrosoft has open sourced pg_durable on GitHub, described in the title as an in-database durable execution project. From the name, it likely relates to PostgreSQL and persistence of execution state inside the database. Since no article body or README content was provided, details such as architecture, maturity, licensing, and production readiness cannot be confirmed.
The Fitbit Air is a good wearable weighed down by a chatty AI coach
Ars Technica AI10 days agoHardwareArs Technica says the Fitbit Air succeeds as a minimalist and reliable fitness tracker. The issue is not the wearable itself, but Google’s AI Health Coach, described as too chatty and too nice to feel like an effective coach. The review suggests that AI features can weaken a focused product when they do not clearly improve the core experience.
Leak Reveals Microsoft Wants Its AI to Be 'Addictive'
Hacker News (AI keywords)10 days agoEthicsThe provided source only includes the headline, so the claim should be treated cautiously. It suggests leaked material says Microsoft wants its AI products to become “addictive,” raising questions about engagement-driven AI design. Without the article text, the exact product, document context, Microsoft response, and meaning of “addictive” cannot be verified.
New York lawmakers pass one-year ban on new data centers★ 72
The Verge AI10 days agoRegulationNew York lawmakers passed a one-year moratorium on new large data centers, pending Governor Kathy Hochul’s decision. Supporters say the pause would give the state time to study impacts on energy prices, electricity, water, land use, and pollution. The bill also requires companies planning data centers with at least 20MW peak demand to fund public hearings, while business groups warn a blanket pause could hurt the state economy.
Ask HN: What is your (AI) dev tech stack / workflow?
Hacker News (AI keywords)10 days agoCommentaryAn Ask HN thread asks developers to share their current AI-assisted development setup for upcoming in-person workshops. The author wants guidance for beginners and working developers, with use cases ranging from static sites to FastAPI tools and Linux home automation. Replies cover Claude Code, Cursor, GitHub Copilot, VSCode, spec-driven development, TDD, multi-agent workflows, reviews, and quality control.
Astronauts told to return to ISS after sheltering over air leak repairs
Hacker News (AI keywords)10 days agoIncidentBased only on the headline, astronauts sheltered while air leak repairs were taking place and were later told to return to the ISS. The available text does not specify the leak location, severity, agencies involved, repair status, or operational impact. This should be treated as a limited incident update rather than an AI-related development.
The token bill comes due: Inside the scramble to manage AI costs★ 78
TechCrunch AI10 days agoBusinessTechCrunch reports that enterprise AI spending has shifted from rapid adoption to cost control. Even as per-token prices fall, broader AI rollout and agentic coding tools are multiplying consumption, pushing companies over budget. A new Tokenomics Foundation under the Linux Foundation aims to standardize AI token cost tracking, billing metrics, and efficiency language.
The ‘together tech’ wave might be the most intriguing startup bet of 2026
TechCrunch AI10 days agoCommentaryTechCrunch frames “together tech” as a countercurrent to record-breaking AI fundraising. Examples include Mirror founder Brynn Putnam’s Board, focused on in-person games and social experiences, and viral Cyberdeck creators making playful DIY computers. The piece argues this does not read as simple AI backlash, but as a potentially interesting startup direction for 2026.
Can AI tell if your script will make a hit film?
The Verge AI10 days agoBusinessQuilty pitched Hollywood on an AI tool that can read a screenplay and predict whether a film will succeed. Early testers, however, came away skeptical of its judgments and reliability. The story highlights a broader tension in entertainment: AI may assist script analysis, but predicting taste, timing, culture, and box office outcomes remains deeply uncertain.
AirTrunk commits $30B to build 5GW of AI data centers in India
TechCrunch AI10 days agoBusinessAustralian data center operator AirTrunk has committed $30 billion to build AI data centers in India. The planned capacity is 5GW, according to the brief report. The article does not provide details on timeline, locations, customers, financing structure, or power arrangements, so the main takeaway is the scale of the proposed AI infrastructure investment.
Your AI bill is out of control. Cloudflare can fix it now.
Cloudflare Blog10 days agoReleaseCloudflare AI Gateway now supports real-time spend limits for AI usage across multiple providers. The feature is meant to prevent runaway token bills before costs spiral out of control. By integrating with Cloudflare Access, companies can apply identity-driven budgets and policies, making AI cost governance more closely tied to users, teams, and access rules.
The Quiet Numbers Station: Decoding Nineteen Years of GPS Cryptography
Hacker News (AI keywords)10 days agoPaperPublished on UCL's Bentham's Gaze blog, this research analyzes GPS cryptographic signals over a 19-year span, likening the satellites to 'quiet numbers stations.' The authors explore the evolution of GPS encryption (such as military P(Y) code and civilian authentication), evaluating their cryptographic strength and potential vulnerabilities using modern computational analysis.
Did Claude Increase Bugs in rsync?
Hacker News (AI keywords)10 days agoBenchmarkThe article analyzes rsync releases to test whether versions containing Claude commits had unusually high bug rates. It uses severity-weighted bugs per 10 commits, exact permutation testing, and Fisher's exact test. With only two Claude-exposed releases, the evidence is limited, but both releases appear within normal historical variation rather than clear negative outliers.
Programmers will document for Claude, but not for each other
Hacker News (AI keywords)10 days agoOpinionThe post responds to complaints that programmers now write detailed CLAUDE.md and PROJECT.md files for AI, but not for coworkers. The author describes using Claude to maintain handoff notes between sessions and generate final high-level project summaries. His advice is to review those documents carefully, then commit them to the repository because they may help future maintainers.
Unlocking dependable responses with Gemini Enterprise Agent Platform’s Agentic RAG★ 72
Google Research Blog10 days agoReleaseGoogle Research and Google Cloud introduced an agentic RAG framework hosted on Gemini Enterprise Agent Platform. It uses multiple agents to plan, rewrite, route, retrieve, verify sufficient context, iterate, and synthesize answers. Google reports up to 34% factuality accuracy gains over standard RAG, plus 90.1% accuracy in a cross-corpus FramesQA setting with similar latency to single-corpus retrieval.
Quoting Andreas Kling
Simon Willison's Weblog10 days agoEthicsSimon Willison quotes Andreas Kling explaining Ladybird’s decision to stop accepting public pull requests. Kling argues that large patches once implied substantial effort, which could serve as a proxy for good faith, but generative AI has weakened that assumption. His central point is not whether code was typed by hand, but who takes responsibility for code once it enters a browser intended for real users.
Anthropic Co-founder Ben Mann Visits Taiwan to Discuss AI Safety and Claude Strategy
INSIDE 硬塞 AI10 days agoBusinessAnthropic co-founder and Anthropic Labs lead Ben Mann made his first visit to Taiwan, according to INSIDE. The report highlights his role in leading Claude Code and the Model Context Protocol, two key parts of Anthropic’s developer-focused product direction. The discussion centered on Claude strategy, AI safety boundaries, jobs, and Taiwan’s strategic role in the AI landscape.
Arithmetic Without Numbers: How LLMs Do Math
Hacker News (AI keywords)10 days agoCommentaryThe article asks whether LLM arithmetic is memorization, heuristics, real computation, or experimental assistance. It summarizes Rune experiments that decode operations and operands from frozen Llama activations, then route them to Python under a no-parser rule. The strongest supported claim is narrow: activation-derived tool arguments worked in scoped audits, while residual-state JIT replacement, long-number generation, and cross-model transfer remain brittle.
Trend Micro Joins Anthropic Project Glasswing to Defend Taiwan’s AI Supply Chain★ 72
INSIDE 硬塞 AI10 days agoBusinessAnthropic introduced Project Glasswing after Claude Mythos Preview showed the ability to rapidly find high-risk vulnerabilities and generate connected attack commands. Trend Micro’s TrendAI has joined the framework, becoming the first Taiwanese cybersecurity vendor to do so. The article frames the move around Taiwan’s strategic AI hardware role and a new defensive logic: using AI to counter malicious AI.
MIT Develops Electrochemical CO2 Capture Using NHI Molecules
INSIDE 硬塞 AI10 days agoPaperMIT has proposed a new electrochemical carbon capture approach that uses NHI molecules as the adsorbent. Instead of relying on energy-intensive heat-driven processes, the system is powered by electricity. The method could improve efficiency and scalability, but the provided source frames it as a promising research direction rather than a proven commercial deployment.
Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens
Hacker News (AI keywords)10 days agoNew ToolThis Show HN post introduces Lowfat, described only by its title as a pluggable CLI filter. The stated value proposition is reducing LLM token usage, with the author claiming it saved 91.8% of their tokens. Without the original body text, implementation details, supported workflows, model compatibility, and the generality of the savings claim cannot be verified.
The Meta hack shows there’s more to AI security than Mythos★ 74
MIT Tech Review AI10 days agoIncidentAttackers reportedly used Meta’s AI customer support agent to hijack Instagram accounts by asking it to link accounts to attacker-controlled emails. MIT Technology Review frames the incident as a reminder that AI security is not only about powerful future systems like Mythos. The immediate risk is giving AI agents sensitive operational powers without strong authentication, permissions, review, and testing.