Skip to content
AI NEWS JOURNAL
AI, IN PERSPECTIVE

Search AI news

Search headlines, summaries and articles to connect the dots.

Newest first

Show HN: SrcWranglr

I spend a lot of time diving deep into C/C++ code bases for various reasons. Normally for Vulnerability Research. I like to get a good feel for the software, it's data structures, control flow, etc. Are all usually important.In the past we have SourceTrail, this is now defunct and its abilities were somewhat limited. If you have thousands to spend you can grab a Understand license... Wait maybe thats SourceNavigator. Either way I wrote SrcWranglr. Its my very first Rust project (congrats to the insufferable Rusteans, I finally cracked and learned Rust).Anyhow SrcWranglr is wicked free, wicked fast, has a modern UI, and features that assist in finding bugs/vulns. You can run semgrep rules over the codebase, or have a fuzz harness generated for you by the program with ease.. Its still super early in development but it already helped me discover a 0day :-).. Give it a shot if you want. I am open to any suggestions and if it is in your heart to contribute let me know. I only get to work on personal projects here and there on account of work... But yea enough babbling.https://github.com/jdefrancesco/SrcWranglr Comments URL: https://news.ycombinator.com/item?id=50380536 Points: 1 # Comments: 0

Same Rule, Different Verdicts: Rubric Language and Clause Structure in LLM Judges for Korean Support Dialogues

arXiv:2611.04091v1 Announce Type: new Abstract: Services that operate in Korean often write their evaluation rubrics in Korean and hand them to an LLM judge. We test whether a judge's verdict depends on the language of the rubric and on how a rubric item is phrased. We write 12 criteria for support dialogues in four clause structures (affirmative, negated, conditional, exception) such that two pairs of structures have the same meaning, and we give each item in Korean and in English. Six low-cost models from three providers, plus one reference model, judge 60 translated ESConv dialogues and 144 constructed boundary cases under every combination (51,744 requests across the seven models). The analysis plan was fixed and committed before data collection. Cross-provider agreement is lower with Korean rubrics than with English rubrics (Cohen's kappa 0.687 vs. 0.745; difference -0.058, 95% CI [-0.072,-0.045]). Disagreement concentrates in negated, conditional, and exception items (odds ratio 5.6 relative to affirmative items). A single model changes its verdict between two items with the same meaning in 11-15% of cases. Two hypotheses are not supported: models from the same provider do not agree more than models from different providers, and adding paired examples does not reduce the language gap. We measure consistency only; no human labels are used, and all materials were drafted by an LLM.

v0.44.0

ComfyUI v0.44.0 introduces multiple new partner nodes for tools like Patina, VEED, BFL, Ideogram, Pruna, MiniMax, Topaz, Sonilo, and OpenAI, adding features such as higher resolution video upscaling, context window editing, and new models. Enhancements include support for ProRes and custom chroma subsampling in video saving, improved frontend handling with pinned versions, optimized tensor comparisons, and better VAE compute dtype control. Performance optimizations address redundant concatenations and memory management, while bug fixes resolve crashes related to VAE decoding and prefetch queue cleanup. The release also updates workflow templates, embedded documentation, and default settings, with contributions from new developers.

Google’s AI-powered search is challenging its advertising business, with third-quarter search ad revenue falling short of expectations.

During the Q3 2026 earnings call, Google disclosed that its AI Mode generative search is disrupting traditional search advertising, as AI-generated summaries and answers reduce the commercial monetization potential of search results. Chief Business Officer Philipp Schindler stated that AI responses are reshaping how search results are presented, making it more difficult to monetize ads. In the third quarter, search ad revenue reached $63.4 billion, up 12% year over year, but fell about $1.6 billion short of Wall Street estimates, with growth also lagging behind expectations.

谷歌AI搜索冲击广告业务,第三季度搜索广告收入不及预期
Timeline and related coverage

Lipi-Ghor: A Large-Scale, Diarized, Multi-Domain Bengali Speech Corpus and a Scalable Pipeline for Building It

arXiv:2610.36184v1 Announce Type: new Abstract: Bengali, spoken by over 230 million people, remains severely under-resourced in spoken language research. Existing open Bengali ASR corpora are confined to short, read or crowd-sourced speech and lack speaker diarization and long-form audio. We present Lipi-Ghor, a large-scale weakly-supervised Bengali speech corpus comprising 885 hours of audio sourced from 998 YouTube videos across 596 unique channels, spanning 17 diverse domains including audiobooks, drama, religious sermons, news, and regional dialects. Each recording is annotated in SSTT format (Speaker, Speech, Transcription, Timestamp), which combines automated diarization with Bengali YouTube caption tracks. To our knowledge, Lipi-Ghor is the first Bengali speech corpus to pair large-

Timeline and related coverage

Calibrated for Which Policy? Conformal Uncertainty for Fixed World Models Under Intervention

arXiv:2610.35877v1 Announce Type: new Abstract: Changing a deployment policy can invalidate a world model's uncertainty while leaving the physical transition mechanism and predictor unchanged. For one-step prediction, if source and target laws share the conditional mechanism $P(Y|X)=Q(Y|X)$, then the intervention is identified by the pre-outcome change of measure $w(X)=dQ_X/dP_X$. For policy-controlled dynamics this ratio factors into state-occupancy and action-propensity terms. Weighted conformal calibration transports marginal coverage to the target law; with an estimated ratio $\hat w$, $P_Q\{Y \in C_{\hat w}(X)\} \ge 1-\alpha-TV(Q,\hat Q) \ge 1-\alpha-E_P|w-\hat w|$. Population log-loss regret for a source-versus-target domain classifier bounds the same ratio error. If $h$ is its clipp

Timeline and related coverage