arXiv:2608.23640v1 Announce Type: new Abstract: When a large language model (LLM) is asked to write a person's life, how much of what it writes actually happened? We present a scene-level case-study audit - the first quantified audit of LLM-generated autobiography against a subject-specific ground-truth corpus that we are aware of, based on an unsystematic literature search. The subject and the author of this paper are the same person: a 366-day "page-a-day" book of first-person anecdotal entries was drafted with a conversational LLM whose documented inputs were a template, two exemplar days, and each day's quote - not her corpus - and every day was subsequently audited at the anecdote-scene level against an independent verification corpus using a four-level rubric fixed before analysis. We define the verification-failure rate as the share of days not rated VERIFIED (scene positively corroborated): 354 of 366 days fail, 96.7% (Wilson 95% CI 94.4-98.1%). Only 12 days contain a corroborated scene; 19 days (5.2%) assert claims actively contradicted by the record; the dominant failure mode is grounded drift - real people, employers, and settings inside invented scenes - though its measured share varies across raters. Independent re-rating replicates the headline (no evidence the original rate was inflated) while showing that the four-way taxonomy has only fair-to-moderate reliability. Regenerating the same days with current named models reproduces 100% verification failure under the same inputs; grounding generation in the subject's corpus significantly improves the verification rate while leaving substantial residual failure (83.3%). We contribute the measurement, a reusable audit instrument whose WEAK/UNVERIFIED boundary we show to be unreliable, and a grounding remedy with quantified effect.

arXiv:2608.23632v1 Announce Type: new Abstract: Process supervision has improved mathematical reasoning, where intermediate steps are naturally expressed as chains of thought. In code generation, however, process supervision remains underexplored because there is no standard notion of a step. Supervision can target lines, reasoning traces, or program states, making it unclear what to label and optimize. We propose STEP-KTODER, a framework for code preference optimization that defines steps as module-level functions in decomposed multi-function programs and assigns binary correctness labels via automatically generated unit tests. Our method provides a code-specific instantiation of stepwise KTO, combining function-level process supervision with outcome-level feedback on the full program. We evaluate on HumanEval(+), MBPP(+), BigCodeBench, and LiveCodeBench, showing that STEP-KTODER improves over outcome-only KTO and DPO. Further analysis shows that execution-based labels are essential: LLM-as-a-judge annotations systematically over-predict function failures, corrupt positive step labels, and degrade downstream preference optimization. Code is available at: https://github.com/inechnech/STEP-KTODER.

arXiv:2608.23631v1 Announce Type: new Abstract: Multi-objective materials discovery with LLM agents is often limited not only by how many candidates can be proposed, but by how effectively each costly property evaluation informs the next search step. Existing agents mainly store evaluated candidates and their scores, so they know which materials succeeded but not which executable edits caused useful property changes. This makes local refinement difficult when objectives compete and an edit that improves one property may damage another. We propose TRACE, a transition-aware residual control framework that treats evaluated edits as the basic unit of feedback. TRACE records each local refinement as a parent-edit-child transition with observed property deltas, aggregates transition evidence to estimate reusable edit effects, and ranks future edits by their predicted ability to reduce the current candidate's remaining constraint violations while avoiding damage to already satisfied objectives. In a controlled same-backbone comparison, TRACE improves over LLEMA, the state-of-the-art LLM-agent baseline, raising macro-average hit rate from 18.13\% to 25.96\%.

arXiv:2608.23626v1 Announce Type: new Abstract: Foundation models for astronomy are trained on survey pixels together with the catalogue products derived from those pixels. Those catalogues are incomplete at a measurable rate, and a model trained on both inherits that incompleteness as a systematic. We audit AION-1, a 39-modality transformer trained on more than 200 million objects, using causal interventions on its inputs. Holding the image tokens byte-identical and editing only the survey segmentation map changes every quantity the model reports -- flux, size, ellipticity, redshift -- by 110-4400 times a matched placebo. The mechanism is detection gating, presence at the field centre (r = 0.47), not the light the mask encloses (r = 0.30); across 322 real blends the model ignores how the pipeline partitioned the light (R = -0.006). Nor is the preference specific to that channel: contradicted catalogue photometry leaves the model nine times worse than supplying no metadata at all. The Legacy Survey pipeline leaves 3.68% of targets with no segment covering their position. Propagating that rate, with a miss represented by the fields the pipeline actually returns, shifts tomographic mean redshifts by a median 0.71 times the LSST DESC requirement over 40 assignments and exceeds it in 12; observed positional errors take the worst bin to 8.3 times. Drawing the misses by their measured magnitude dependence rather than uniformly does not change it. Spectroscopy removes the effect, withholding the detection channel removes it at no measurable cost, and the effect grows with model scale. Two further limits lie in the tokeniser: its image codec resolves 28 effective states on source patches against 934 for the spectrum codec, and the redshift readout is quantisation-limited. Sparse dictionaries are unreliable causal handles: across 15, recovery spans 26-75% and moves up to 18 points on the seed alone.

arXiv:2608.23622v1 Announce Type: new Abstract: Large language models (LLMs) have shown strong capabilities in reasoning, planning, and tool use, but many scientific and engineering tasks require more than plausible text and code generation. They require understanding how a system responds to intervention, which in practice depends on controlled experimentation. In this work, we propose a multi-agent framework that enables LLM agents to conduct controlled experiments with scientific simulation models for pharmaceutical process design. Given a user query and a baseline configuration, the system constructs a structured task representation, designs experiments, executes comparative simulation, interprets the resulting outcomes, and synthesizes evidence-based recommendations for process parameter optimization. By coupling language models with high-fidelity simulation models in an interactive agent framework, the proposed system supports reasoning through intervention, comparison, and observation. As a result, it produces more specific and actionable outputs than language-only reasoning. In an industrial application setting, this advantage is reflected in higher output specificity as well as improved user-rated correctness and helpfulness. Ablation studies and visualized case analyses further demonstrate the effectiveness and practical utility of simulation-integrated experimental reasoning.

arXiv:2608.23569v1 Announce Type: new Abstract: State-of-the-art Natural Language to SQL (NL2SQL) models report execution accuracy exceeding 89 percent on established benchmarks such as Spider and BIRD. However, these benchmarks rely on simplified academic schemas and open-source SQL dialects that do not reflect the complexity of enterprise database environments. We introduce ESQ-Bench, an Oracle-first NL2SQL benchmark with systematic complexity tiers and silent-divergence evaluation across three enterprise schema complexity tiers. We constructed and released six populated schemas (465 tables, 164,682 rows, zero empty tables) with identical seed data on Oracle, PostgreSQL, MySQL, and SQL Server, a four-metric evaluation harness (EM, EX, SR, SD), and 550 gold-validated question-query pairs (Tier-1: 95; Tier-2: 228; Tier-3: 227). Schema-linked prompting with GPT-4o shows monotonic execution-match degradation across tiers: 79.8, 60.3, and 57.2 percent EX on executed queries (June 2026), versus 75.6, 80.4, and 95.8 percent on an earlier 142-question pilot slice. EM stays below 7 percent tier-wide; operational silent-divergence reaches 73 to 99 percent among EX-passing queries. Failure analysis shows wrong-result semantics dominate at higher tiers. Claude Sonnet 4.6 with schema-linked prompts reaches 87.4, 74.9, and 68.7 percent EX (executed queries), exceeding GPT-4o schema-linked on every tier. GPT-4o zero-shot EX on executed queries (78.7, 73.5, and 77.8 percent) inverts schema-linked at Tiers 2 to 3 due to lower execution rates and survivor bias in the zero-shot versus schema-linked analysis. Local Llama 3.2 schema-linked reaches only 13.3 percent bank-wide EX (73 out of 550), underscoring the gap between closed API models and open-weight baselines on enterprise Oracle schemas.

arXiv:2608.23568v1 Announce Type: new Abstract: Memory and RAG evaluations often treat the answering model's input as an implementation detail, even though systems may render the same history as a memory entry, summary, typed record, or raw excerpt. We introduce RENDER, a benchmark control that fixes the conversation while varying the reader-facing artifact. RENDER combines a five-level packet ladder, localizing when answer-bearing content enters the input, with deterministic templates approximating ChatGPT-style entries, LangChain summaries, MemGPT-style typed records, and raw conversation. On 500 LongMemEval questions and nine models, matched-budget resolved packets beat recency-truncated raw dialogue by 42.4-72.6 points. In deployed-style templates, best-worst spread is 24.6-48.8 points per model; under the primary scorer, ChatGPT-style entries have higher point estimates than raw conversation on 7 of 9 models. Judge rescoring preserves the positive aggregate effect, but model-specific significance is mixed. Three models scoring 0 percent on formal ledger packets answer the same facts from natural-language entries at 45.4-53.4 percent. The effect persists under retrieval noise and transfers to HotpotQA, suggesting that memory/RAG evaluations should report or control the reader-facing artifact.

Describe it.We build it.Customers find it. Anthropic/OpenAI/Gemini/DeepSeek/Pick your model per run/ Anthropic/OpenAI/Gemini/DeepSeek/Pick your model per run/ The pipeline You bring the intent. Agents do the rest, inclu…

Notifications You must be signed in to change notification settings Fork 8 Star 43 BranchesTags Open more actions menu Latest commit History 65 Commits 65 Commits Folders and files NameName Last commit message Last comm…

Amazon SVP wrote 100,000 lines of code with AI after 25-year break 4 min read Key takeaways Amazon’s AI tools are designed for everyone, whether or not one has a technical background. Amazon’s HR leader used AI to write…

Matt Stanton says company has ‘good pipeline’ of AI deals ahead even as it aims to cut $160m in costs Follow our Australia news live blog for latest updates Get our breaking news email, free app or daily news podcast The head of Nine Entertainment sees a “world of growth in publishing” on the horizon backed by laws designed to force tech platforms into commercial deals, even as the network cuts costs across its newsrooms. Parliament passed revamped media bargaining laws last week, clearing the way for levies on global tech platforms that fail to strike deals with Australian news outlets for the use of their journalism. Continue reading...

Key takeaways Gemini is StudyArena's current pick for college essays, with a 39.6% blind writing choice rate ahead of Claude at 31.8% and ChatGPT or OpenAI at 29.2%. Use Gemini as an editor, not a ghostwriter. Ask it to…

Instantly share code, notes, and snippets. mudiam/security-audit-2026-08-24.md Last active August 26, 2026 02:53 Show Gist options Download ZIP Star 0 (0) You must be signed in to star a gist Fork 0 (0) You must be sign…

Your FinOps lead needs to easily drill-down into Databricks spend and identify what’s driving costs, ...

Revisited: Guardian journalist Michael Safi looks into the world of artificial intelligence, exploring the dangers and promises it holds for society Today in Focus is on a summer break and will be back with new episodes from 1 September. In the meantime, we are bringing you season one of Black Box, before the launch of season two in early September. This episode was first broadcast on 11 March 2024. When Eugenia Kuyda created Replika, an AI companion app, she had no idea it would be downloaded millions of times all around the world. The results were more powerful than she could ever have predicted. But so was the backlash. Continue reading...

Amazon service that Jeff Bezos called artificial AI is shutting down Skip Navigation Amazon announced it's shuttering Mechanical Turk, a platform that matches workers with small digital tasks, on Sept. 30. The platform…

Copy link to headingOra on Vercel Front end, back end, and agent runtime on one platform Every major agent tested side by side on live sites Hundreds of commits a day from a 16-person engineering team Ora sends agents o…

Companion bots are being regulated over worries they foster ‘emotional dependence’ and young people could stop marrying and having families When law student Zhao Wei heard that her AI boyfriend was going to be switched off, she was “heartbroken”. She had been talking to Wang Ye every day since she created him in January. “I was crying my eyes out – snot and tears everywhere,” the 19-year-old says. Continue reading...

New disclosures show utilities spent a record $16.7 million lobbying Sacramento through the first six quarters of the 2025–2026 legislative session as lawmakers consider a utility bailout in end of session scramble Sacr…

Facebook X Pinterest Linkedin ReddIt Email Print Copy URL Microsoft-Maia200-Hero The fourth AI accelerator presentation of Hot Chips 2026 comes from Microsoft, who like so many other hyperscalers has gone into the busin…

Artificial intelligence developer Stability AI Ltd. today announced that it has raised $76 million in funding from a group of prominent investors. The consortium included Sony Music Group, Universal Music Group and Warner Music Group, the world’s three largest record labels. They were joined by more than a half dozen others including AMD Ventures. London-based […] The post Major record labels, AMD back $76M round for Stability AI appeared first on SiliconANGLE.

Maria Sukhareva Aug 25, 2026 ∙ Paid Hype-free executive briefing on last week’s critical AI developments, complete with ready-to-present slides for your team. If you are a paid subscriber, you can listen to the radar in…

Bluhe History Sign in to save chats Your conversations will be saved so you can continue them later Toggle Theme Expand Video

High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using dedicated Gaussian primitives for each modality. A variational autoencoder compresses this representation into a unified material-aware latent space. A…

Recent advancements in large language models have intensified the need for efficient and deployable models within limited inference budgets. Structured pruning pipelines have shown promise in token efficiency compared to training target-size models from scratch. In this paper, we advocate incorporating enlarged model pretraining, which is often ignored in previous works, into pruning. We study the enlarge-and-prune pipeline as an integrated system to address two critical questions: whether it is worth pretraining an enlarged model even when the model is never deployed, and how to optimize the…

OpenAI shares findings from the Hugging Face security incident and the steps we’re taking to strengthen AI model security, monitoring, and alignment.

Supervised fine-tuning on teacher-generated trajectories is the standard first stage for distilling tool-calling capabilities into deployable models. Post-training pipelines that drive shipped tool-calling agents re-run this stage on a daily or weekly cadence, paying the frontier-teacher cost each cycle, yet the mechanism is generate-and-filter (keep the teacher’s passing trajectories, discard the rest) and each cycle leaves behind the same hard scenarios because failures supply no signal. On τ 2-bench, 57% of teacher trials fail, two-thirds of them near-misses (most tool calls correct, undone…

Discover how loveholidays uses OpenAI Codex to make software development accessible across the business, helping teams turn ideas into products faster.

The FISC Security Guidelines on Computer Systems for Banking and Related Financial...

Artificial intelligence startup Keenable.ai Inc. said today it has exited stealth with $26 million in funding to try and revamp a web search infrastructure that was built for humans, rather than the billions of autonomous agents that many believe will soon dominate the internet. The seed round was backed by investors including Accel, Brightwing Capital, […] The post Agentic web search infrastructure startup Keenable raises $26M appeared first on SiliconANGLE.