AI Briefing
Hacker News news
The latest AI and developer news from Hacker News, with short summaries.
Visit website →Study finds LLMs break 32-51 working PHP files during modernization
In a single-file test of 100 WordPress 4.0 files, the best LLM left only 59 passing behavioral checks compared to 83 in the unchanged originals.
Anthropic's internal AI usage value exceeds employee compensation
In April, the retail value of Anthropic's internal Claude usage surpassed its monthly employee compensation, with usage surging 250-fold by September.
Balestriero and LeCun Prove Platonic Representation Hypothesis Using JEPAs
A new theorem demonstrates that Joint-Embedding Predictive Architectures (JEPAs) achieve cross-modal alignment with 42 times fewer training samples than next-token prediction.
WiCi Introduces WiCi One, a Wireless GPU System for Local AI Inference
WiCi One is a wireless GPU system starting at US$2,599 that uses Wi-Fi 7 to present a remote GPU as local and incorporates a 4 TB SSD for inference.
OpenAI Merges ChatGPT and Codex Usage Limits for Plus, Pro, and Business Plans
Starting in December, token-based usage allowances for Plus, Pro, and Business plans will be shared across ChatGPT and Codex as agentic Work capabilities integrate into the main ChatGPT experience.
PickAnthropic restricts Claude from reverse-engineering video games
Claude now blocks requests to decompile or clone game code, a move attributed to copyright concerns following a wave of AI-assisted jailbreaks and ports.
OpenAI Secures $11B at $1.25T Valuation, Missing $30B Target
OpenAI has raised approximately $11 billion for its latest round at a $1.25 trillion valuation, falling short of its $30 billion target and initial $1.4 trillion valuation hopes due to dampened investor appetite.
Prospero: A lightweight C++ coding harness for local LLM agents
Prospero is a new open-source coding harness written in C++ that uses approximately two orders of magnitude fewer resources than web-based alternatives while supporting headless deployment and desktop control.
Fujifilm to invest $63 million to boost LTO tape production fivefold
The expansion, funded by a 10 billion yen investment, is driven by generative AI's exponential data growth and the need for air-gapped storage against ransomware.
spinf-serve: Open-source inference server for bulk LLM scoring on single GPU
Apache 2.0 licensed tool processes 491 questions per second on an RTX 4090 by reading content once for multiple decision queries.
Y Combinator Labs and Olam Labs run multi-agent AI swarms in Minecraft and Mindustry
The study observed emergent leadership and bureaucracy in 50-agent swarms, with GPT-6.1 Sol defeating Claude Opus 5.5 in 6 of 9 Bed Wars matches and 6 of 7 Mindustry games.
Simantic Launches Rust-Based Firmware Simulation for AI Agents
The YC F26 startup provides a simulation layer that allows AI agents to validate firmware on virtual devboards in seconds, eliminating the need for physical hardware during the testing loop.
Bernstein – open-source governance layer for AI agents
The release supports 53 CLI adapters and uses a deterministic Python scheduler to ensure byte-identical replay of agent plans without LLMs in the coordination loop.
Claude Code background agent status updates consume 15-20% of tokens
An unofficial Bun preload hook disables the undocumented agent summary feature, reducing token spend from 16.4% to 0%.
Epoch AI demonstrates GLM 5.3 exploits Firefox patch gap with one-click exploits
Z.ai's open-weight model GLM 5.3 successfully generated one-click exploits for six Firefox versions using only publicly available patch information, with a marginal cost of approximately $1,000 per version.
SkyPilot Topology-Aware Scheduler Boosts NVL72 Efficiency by 27% via Rack-Aware Placement
A new scheduling strategy keeps workloads within NVLink domains to avoid 43% performance loss, yielding 27% more usable GPUs and 14.4% higher throughput in pre-training.
Google AI Mode fails to answer 2020 election winner due to technical gap
Google confirmed a technical gap in its AI Mode that prevented it from providing direct answers for the 2020 presidential election and 2022 Senate races, an issue the company is currently fixing.
Google announces Gemini 4 Carbon with top cybersecurity benchmarks
The model ranks #1 on the Artificial Analysis Intelligence Index v5.0 at 59 and leads in vulnerability detection with 88.2% on SEC-bench Pro.
PsiQuantum Unveils 8x8 Optical Circuit Switch for AI Infrastructure
The demonstration vehicle achieves 1.2 dB insertion loss and 40 dB cross talk without amplification, with commercial products expected in coming months.
Parameter Labs Discloses P1 ChatGPT Sandbox Escape Enabling Account Takeover
A vulnerability chain involving MathLive and YouTube JSONP allowed attackers to execute arbitrary JavaScript in victims' authenticated sessions via public shares, enabling private data exfiltration.