tl;dr: OpenAI has released `@openai/codex-security`, a CLI and TypeScript SDK for scanning repositories to find, validate, and fix security vulnerabilities, with support for tracking findings over time and integrating into CI pipelines. It requires Node.js 22+ and Python 3.10+, and authenticates via either ChatGPT sign-in or an `OPENAI_API_KEY` (the latter taking precedence in non-interactive contexts).

tl;dr: HNewhere is a userscript that displays Hacker News discussions in a resizable sidebar alongside any article, eliminating the need to switch between tabs. It auto-detects matching HN stories via the Algolia search API, tracks HN-originated links, and supports collapsible comments with direct reply links back to HN. Requires a userscript manager like Tampermonkey or Violentmonkey.

tl;dr: Writers who make Substack their primary home are digital sharecroppers—vulnerable to policy changes, algorithm shifts, and corporate decisions, just as previous generations learned with Medium, Tumblr, and Twitter. The author advocates for the POSSE model (Publish on Own Site, Syndicate Elsewhere): own a domain, treat your website as the source of truth, and use platforms like Substack purely as distribution channels. John Scalzi's 28-year run on whatever.scalzi.com is cited as the ideal approach.

tl;dr: Half-Life has been ported to PowerPC Macs running Mac OS 9, 28 years after release and reviving a version Valve originally cancelled in 1999. GitHub user doctashay built the port using a fork of Xash3D FWGS (a GoldSrc engine reimplementation), with full single-player, multiplayer, and both Blue Shift and Opposing Force expansions playable on G3/G4 machines.

tl;dr: Kimi K3 is a scaled-up (2.8T parameter) production version of Kimi Linear, now the largest open-weight model, incorporating efficiency-focused components like LatentMoE (down-projected MoE, similar to Nemotron 3 Ultra), multi-head latent attention, and Kimi Delta Attention. Notably, it drops RoPE entirely in favor of NoPE across all layers—a first for a frontier-level model—and adds attention residuals that weight cross-layer residual connections via attention scores for ~4% training cost. It also introduces native multimodal support.

tl;dr: SBCL 2.6.7 adds a new SB-MANUAL contrib module that embeds the manual in docstrings for interactive exploration, extends SB-SIMD to ARM64, and adds AVX512 support on x86-64. The release includes miscellaneous bug fixes (SAP-REF-N on ARM64, MULTIPLE-VALUE-CALL miscompilation, CONCATENATE type errors) and optimizations including SIMD-accelerated UTF-8 conversions, cons-free handling of constant complex numbers in local calls, and a tuned sparse-set implementation in the compiler.

tl;dr: Summary not available.

tl;dr: UNA is a Scotland-built GPS smartwatch designed for repairability and longevity, featuring a user-replaceable 280 mAh battery, USB-C charging, and screwdriver-accessible internals for swapping components like the screen. It includes dual-frequency GPS, continuous heart rate monitoring, a barometric altimeter, 10-day battery life, and Strava sync via a companion iOS/Android app. An open SDK lets developers build custom apps, activities, and watch faces with direct access to sensors and data.

tl;dr: Zig's incremental compilation, now usable on x86_64-linux via `zig build --watch -fincremental`, achieves 50-70ms rebuilds by tracking a dependency graph of "analysis units" (declaration types, values, function bodies, type layouts) that get selectively re-analyzed when source hashes change. The tightly-integrated linker uses a `MappedFile` abstraction to patch machine code directly into the output binary, resizing regions with exponential growth to amortize costs. Most remaining overhead comes from a reference-graph traversal during flush that could be optimized further; the feature requires Zig master (or 0.17.0) and still has rough edges.

tl;dr: Anthropic researchers used Claude Mythos Preview to discover new cryptographic attacks: one that halves the effective key strength of HAWK (a NIST post-quantum signature candidate) and another that speeds up attacks on 7-round AES by 200-800×. Neither affects production systems—HAWK isn't deployed and the AES attack only works on a reduced-round variant—but the results demonstrate that frontier LLMs can now perform novel cryptanalysis at expert-researcher level, with each discovery costing roughly $100,000 in API usage.

tl;dr: Summary not available.

tl;dr: Summary not available

tl;dr: Summary not available

tl;dr: Apple's macOS Tahoe 26.6 patches dozens of security vulnerabilities, including multiple root privilege escalations, sandbox escapes, kernel memory corruption bugs, Gatekeeper bypasses, and issues allowing apps to access sensitive user data or fingerprint users. Notable fixes include kernel-level arbitrary code execution (CVE-2026-64747), code signing bypass (CVE-2026-43813), and a file quarantine/Gatekeeper bypass (CVE-2026-64708). Credited researchers span independent hunters and firms like Blackwing Intelligence, SpecterOps, NVIDIA AI Red Team, and Nosebeard Labs.

tl;dr: Anthropic CEO Dario Amodei clarifies that the company does not support banning open-weights models, contrary to recent accusations, arguing such bans wouldn't address his real concerns about authoritarian AI dominance and misuse. Instead, he advocates for three measures: restricting chip exports to China, cracking down on industrial-scale distillation of frontier models, and mandatory safety testing for all sufficiently capable models regardless of whether they're open or closed.

tl;dr: A 38-year-old disabled writer reflects on the shame of no longer being able to work consistently and having to ask family for financial help, tying her experience to the Protestant work ethic and capitalism's demand for constant productivity. She highlights structural traps facing disabled Americans—SSI payments below the poverty line, benefits cliffs, and the hidden costs of chronic illness—and argues that disability exposes the myth of the self-sufficient worker. The essay questions how to decouple self-worth from economic output in a culture that equates the two.

tl;dr: A microbiologist grew bacteria from Bigelow's probiotic tea and confirmed via PCR that BC30 (Bacillus coagulans) survives boiling water because it's a heat-resistant spore-forming organism. However, a properly steeped cup yields only ~30,000 CFUs—0.003% of the 1 billion CFU dose used in clinical trials—raising doubts about whether such probiotic products deliver meaningful benefits, especially given conflicting evidence about GI tract survival and undefined targets for a "healthy" microbiome.

tl;dr: Kimi Linear is a hybrid attention architecture combining Kimi Delta Attention (KDA)—a gated linear attention variant with hardware-efficient Diagonal-Plus-Low-Rank transitions—with Multi-Head Latent Attention (MLA) in a layerwise mix. A 48B-parameter (3B active) model reportedly beats full MLA attention across short-context, long-context, and RL tasks while cutting KV cache by up to 75% and delivering 6x faster decoding at 1M context. The team has open-sourced the KDA kernel, vLLM integration, and model checkpoints.

tl;dr: Researchers at La Jolla Institute, Scripps Research, and IAVI developed an HIV vaccine using a "germline targeting" approach that trains naive B cells through a priming shot and sequential boosters to produce rare broadly neutralizing antibodies. In rhesus macaque trials, roughly 44% of animals generated these antibodies at high levels — the strongest such response recorded in primates. The priming immunogen is already in Phase 1 human trials (IAVI G004), with plans to test the full regimen in humans next.

tl;dr: Summary not available.