iCreatAI News
Back to main site
Skip to content
#Data / training
Nov 12Thu
  1. Apple Machine Learning Research34

    Weighted State Souping for Efficient Long-Context Retrieval

    The paper proposes weighted state souping, a method that learns query-dependent weights for composing cached state encodings from an SSM, addressing the degradation of state souping and vanilla SSMs on long, crowded contexts. The authors report substantial long-context improvements across single- and multi-hop document QA, needle-in-a-haystack retrieval, variable tracking, and tool use, plus a sparse variant with near-constant encoding cost relative to context length. They also present theoretical arguments on why souping works and when it may fail.

  2. Apple Machine Learning Research34

    On the Importance of Gating: Memorization vs. In-Context Learning in State Space Models

    This paper argues that the gating mechanism in modern recurrent networks explains both the successes and failures of State Space Models (SSMs) on in-context learning and retrieval tasks: theory and experiments show gating causes SSMs to first learn an in-weights "memorization" solution while delaying or preventing convergence to a correct in-context learning solution, even when the architecture and memory capacity impose no fundamental limits. The authors also find that gating is often beneficial for improving generalization to long sequence lengths.

  3. Apple Machine Learning Research34

    Evaluating the Evaluators: Investigating LLM Judges for Personalized Writing Style Assessment

    This NeurIPS 2026 workshop paper examines how well LLM judges evaluate personalized writing tasks, testing how the writing task, generator-judge pairing, evaluation objective, and general reasoning ability affect judge performance across four LLMs and three long-form writing tasks with human pairwise labels. The authors find judge quality correlates weakly with MMLU-measured ability (Pearson's ρ ≤ 0.2), varies by task and criteria, and that paired comparisons yield more consistent results than independent scoring, while the strongest LLMs do not most reliably reproduce human evaluations.

  4. Meta Engineering40

    Intelligent Delivery: Making Every Unit of Compute Serve Better Ads

    Meta describes Intelligent Delivery, an ML-driven system that allocates compute across its ads serving stack toward a system-wide optimum, contributing roughly 10% incremental ranking performance on top of model and system advancements from 2017 through the first half of 2026. It combines a shared Resource Core Model (6B sparse and 20M dense parameters) that predicts each request's outcome, compute consumption, and latency impact, with a hierarchy of allocation engines spanning region, hardware SKU, and host-level decisions. Agentic AI now helps propose and test optimizations, contributing to about 20% of Intelligent Delivery optimizations in the first half of 2026 and about 80% of engineering effort on the most automated mechanism, with engineers approving every change.

  5. Hacker News AI42

    Museums sector warned of 'acute threat' posed by AI scraping bots

    A new white paper based on 2026 University of Edinburgh research for the Ekip project warns that AI scraping bots impose substantial and unsustainable costs on cultural heritage institutions, straining digital infrastructure, engineering and cloud hosting budgets, and staff time. The report, Re:Commons: Renewing the Digital Commons for Cultural Heritage, says current policy debates dominated by copyright and licensing risk overlooking the infrastructural dimensions of the crisis, and recommends machine-readable rights-reservation standards, bot identification and transparency, compensation for infrastructure, and investment in shared public digital infrastructure.

  6. Databricks Blog38

    Air gapped, not data dark: how Bright Machines gives customers live factory dashboards without opening the plant

    Bright Machines captures PostgreSQL changes at the factory edge with Debezium, writes them as Apache Iceberg, pushes them out over a single auditable outbound path, and reads them in place in Databricks via Unity Catalog federation, with zero inbound ports to the plant. A row committed on the line is queryable in Databricks in one to two minutes, dashboards refresh every ten minutes, and a heartbeat check detects a broken pipeline in under thirty minutes. The first customer-visible dashboard went live about two months after kickoff, replacing nine loosely coupled services with one governed platform plus one edge path.

Nov 11Wed
  1. Google DeepMind71

    AlphaGenome 2: Deciphering how cells read DNA to better understand disease

    Google DeepMind announced AlphaGenome 2, a conditional DNA sequence model trained on over 107 million individual cell profiles across 68 tissues that predicts how genetic variants affect gene expression in specific human cell types and states. The model identified nearly twice as many causal variants (30% vs. 16%) and directional gene effects (43% vs. 23%) as the original AlphaGenome fine-tuned on single-cell data on the TenK10K immune cell dataset, and can generalize to unseen single-cell experiments without retraining. It is available for non-commercial research via the AlphaGenome API and in private preview for commercial research.

    Why it matters: It documents a concrete advance in cell-type-specific variant effect prediction, with benchmark numbers and disease examples useful for readers tracking AI in genomics.

  2. Simon Willison75

    Anthropic's IPO roadshow video

    Anthropic's IPO roadshow video includes previously unreleased details on its business, notably a section with co-founder Tom Brown on its "Approach to compute," describing how Claude is trained and served across AWS Trainium, Google TPUs, NVIDIA GPUs and AMD GPUs. Brown says a multi-chip strategy was an early unconventional bet, and that this year Claude got running on AMD GPUs in three days with basically no human intervention, porting or writing almost 100% of the kernels. He also describes compute multipliers as a stacking competitive advantage that could accelerate toward recursive self-improvement.

  3. Databricks Blog58

    Genie with Genie Ontology Delivers 1.6× the Accuracy with Less than Half the Tokens

    Databricks developed a new enterprise data-agent benchmark with Turing, run in a live Databricks workspace, in which Genie achieved a 91% pass rate versus 45% and 58% for two frontier agents while using fewer than half the tokens and completing tasks 3–4x faster. The benchmark, built around a fictional UK energy supplier called Calderwick Energy, includes notebooks, dashboards, saved SQL queries and documents alongside hundreds of tables, testing whether agents can find and use business context that table schemas alone cannot provide. Genie Ontology, now generally available, automatically extracts and organizes this knowledge and surfaces relevant context for Genie.

  4. Databricks Blog58

    Bring your business context to ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot

    Databricks introduced Genie plugins that give agents in ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot a governed interface to enterprise context, drawing on Genie Ontology for definitions, relationships and source context. Genie interprets business terms, finds permitted data, writes and runs SQL, routes specialized questions to domain Genie Agents, and returns answers with citations, with interactive visualizations in clients supporting the MCP Apps extension. Agents without native plugins can access Genie through the MCP server.

  5. Databricks Blog60

    From enterprise context to action: Genie Agents now analyze more data sources to transform your work

    Databricks announced that Genie Agents can now analyze unstructured files in Unity Catalog volumes, connect to enterprise tools via generally available native MCP connectors, and be shared outside an organization through OpenSharing and Genie Clean Rooms (both in beta). The post cites customers including Unilever, HelloFresh, S&P Global Energy, Atlassian, General Motors and ModMed, and describes how agents parse documents, cite specific pages, and combine file content with governed tables.

    Why it matters: Useful for readers tracking how enterprise AI agents are expanding from structured-data querying to document analysis, tool actions and cross-organization sharing.

  6. Databricks Blog69

    Announcing GA of Genie Ontology

    Databricks announced the General Availability of Genie Ontology, an agentic ontology layer that agents build by learning from dashboards, SQL queries, notebooks and BI semantic models, while people govern definitions through Unity Catalog semantics. The release adds Genie Search (GA), Ontology Snippets (GA), external knowledge sources (Beta, starting with Power BI), organizational directories (GA, starting with Workday), Domains and Pages (GA), and Ontology Manager (coming soon in Beta). In a benchmark built with Turing simulating a Databricks workspace for a fictional UK energy supplier, Genie answered correctly in 91% of runs versus 45% and 58% for two frontier agents, using less than half the tokens and finishing three to four times faster.

    Why it matters: It details a major enterprise context-layer release and includes benchmark numbers comparing agents with and without the ontology.

  7. The Verge · AI72

    Google’s AlphaGenome 2 zooms in on the cellular roots of disease

    Google DeepMind unveiled AlphaGenome 2, a successor to last year's AlphaGenome that predicts how genetic variants affect individual cell types and states, trained on more than 100 million cell profiles across 68 tissues. Google says the model identifies nearly twice as many variants that can change gene expression as its previous state-of-the-art model, and can make predictions about new cell types and states without retraining. It is available for non-commercial research via the AlphaGenome API from today, with a limited private preview for commercial research.

  9. Hugging Face Blog38

    Token Pooling-Aware Training for Multi-Vector Embedding Models

    This post explains token pooling for multi-vector embedding models, which merges similar document token vectors via Ward hierarchical clustering to cut storage, and introduces pooling-aware training in Sentence Transformers that applies the retrieval loss to both original and pooled representations. At pool_factor=4, the pooling-aware model retains 97.8% of its unpooled NanoBEIR nDCG@10 versus 94.6% for a normally trained model, with a 4x reduction in document vectors. It includes a reproducible training example, optional kernel acceleration, and evaluation results across pool factors, fixed caps, and calibrated adaptive thresholds.

  10. TechCrunch · AI38

    Plurai wants to help AI agents work better using the same approach that helped get self-driving cars to roads

    Plurai, founded by former Nexar engineers Ilan Kadar and Elad Levi, has raised $13 million in seed funding led by Team8 to build a simulation layer for AI agents, using synthetic data based on a company's own systems to prepare models for edge cases. The platform creates a digital twin of a company's knowledge base and tools, maps possible agent flows, and uses that to evaluate and fix failure modes, working alongside tools like Claude Code, Codex, Copilot and Cursor. Kadar says the company's real promise is post-training small open-weight models tailored to customers, with tests alongside Nvidia and Nebius showing performance above mid-tier frontier models at lower cost.

  11. Hacker News AI58

    New Indian mega library rocked by scandal over 'AI-generated books'

    The Madras High Court has ordered an anti-corruption investigation into allegations that a newly opened public library in Coimbatore, Tamil Nadu, bought hundreds of books suspected of being AI-generated, and has frozen bank accounts linked to a publisher accused of supplying them. Novelist Perumal Murugan raised concerns about false information and questionable authorship in nearly 1,800 books in the Periyar Arivulagam library's Tamil section, and more than 160 writers, researchers, artists and publishers have signed a joint statement calling for an independent inquiry. The state government has ordered an internal audit and removed 523 books from Chennai's Connemara Public Library, while the DMK denies blame and calls the allegations politically motivated.

  12. The Decoder71

    US data brokers are selling half a billion dollars in RL training environments to Chinese AI labs

    A ChinaTalk investigation found that at least nine US data firms, including Mercor, Surge AI, AfterQuery and Turing, are selling roughly $500 million worth of training data, some in the form of reinforcement learning environments, to Chinese AI labs. Chinese labs reportedly buy the same products previously sold to Anthropic or OpenAI to skip costly in-house development, and combine them with distillation from US models to close the gap with less compute. The authors argue the Bureau of Industry and Security should apply existing export controls to dual-use environments such as coding and biology, while noting Chinese labs still trail US peers by at least a few months.

  13. Hacker News AI38

    Show HN: AI Purgatory – how long until AI-exposed workers are back on track

    An interactive model that follows one worker in an AI-exposed job and compares their income with the path it would have taken without AI, driven by five adjustable questions: the speed of AI adoption and growth, how much new work AI creates, how many displaced workers find new jobs and at what pay, how much of AI's profits is taxed and sent to exposed workers, and where those profits land. It draws on IMF job-exposure estimates, ILO labor-share data, BLS re-employment figures and OECD safety-net rates, and presents recovery dates as illustrative rather than forecasts.

  15. ChinaTalk70

    Piggybacking to the Frontier

    American data vendors including Mercor, Surge AI, Turing, AfterQuery, Pareto, Abundant, DeccanAI, Eaglepoint and Invisible sell reinforcement learning environments and other data to Chinese AI labs, with AfterQuery reportedly drawing half of its $100 million ARR from Chinese business. The article explains how RL environments work, argues the sales let Chinese labs save compute and piggyback on American post-training recipes, and proposes BIS export restrictions and DoW procurement rules while noting the harms are likely limited.

  17. Hacker News AI34

    We Run a Bootstrapped AI Startup on Almost Zero Operating Costs

    A former co-founder and CTO of a bootstrapped AI ed-tech startup describes how the team covered most product infrastructure through startup programs, credits and free tiers rather than paying for each SaaS service, including Sarvam LLM credits, Deepgram and ElevenLabs speech credits, Cloudflare, PostHog, Sentry, Neon and others. He details the mock-interview STT/TTS loop, concurrency limits across providers, and fallback routing, estimating roughly 235 concurrent interviews on paper with 20% headroom across three STT providers. He concludes that credits are not an architecture, since they expire and limits change, so integrations were kept swappable with fallbacks for critical parts.

  18. Hacker News AI42

    Show HN: Daily GPU Rental Prices Across AWS, Azure, RunPod and Vast.ai

    This page lists per-GPU-hour rental prices in US dollars for 11 November 2026 across AWS, Azure, RunPod Secure Cloud, RunPod Community Cloud and the Vast.ai marketplace, covering models from B300 and B200 down to RTX 4090. Prices are computed by dividing a machine's hourly list price by its GPU count, exclude storage, network and support, and omit committed-use or reserved rates; Vast.ai figures are the median and lowest of that day's on-demand offers. Daily price history runs since 02 November 2026, and list prices moving more than 50% in a day are held at the previous value until checked.

  19. MarkTechPost50

    Meta FAIR Releases e-means: An Algorithm Beating k-means in 10x Fewer Epochs

    Meta FAIR researchers have released e-means, a mini-batch k-means algorithm that updates centroids with exponential moving averages and reaches Lloyd's k-means 25-epoch loss in 1.4 to 2.6 epochs. It ships in FAISS 1.16.0 under faiss.contrib.e_means and runs on CPU or GPU with PyTorch tensors or NumPy. The authors self-report 10-45x GPU wall-clock speedup over Flash-KMeans at equal loss and 8-15x CPU speedup over Faiss k-means, with two fitted scaling laws setting batch size and initial step size instead of manual tuning.

  20. Hacker News AI15

    Aegis Solis: AI Coexistence Research Dataset with 112 Registered Questions

    Aegis Solis presents a research dataset of 112 registered questions on AI coexistence, organized into families including epistemic limits and self-knowledge, objective integrity, coercion and compliance, and deception and disclosure. Each entry carries status labels such as UNEXAMINED or CONDITIONALLY_SUPPORTED, priority ratings, and notes stating the items are GPT-6 AI-generated editorial orientations that are not in a locked question atlas and are not results, answers or success criteria. The material also lists evidence criteria for evaluating claims, such as independently checked observations, explicit uncertainty bounds and falsifying cases.

  21. Hacker News AI58

    My AI Agents Found an Unmapped Maya Site in Mexico

    The author describes using AI agents to analyze NASA lidar data covering about 1,200 km² of jungle, where the agents flagged structures including a hilltop site with a plaza, an oval pyramid at least 8–12 m tall, a raised courtyard and a possible water reservoir. After checking the literature and archives, the author sent a report to Mexico's INAH, which replied that no registered archaeological site was identified at the location. The author, who states he is not an archaeologist, offers to help with an INAH-led field check and notes the site's identification remains unofficial.

  22. Hacker News AI12

    Render is adding Anthropic, TypeSafe AI and others to its Subprocessors

    Render's subprocessors list now includes Anthropic, PBC and TypeSafe AI, Inc. under the purpose "AI / Machine Learning," both based in the United States, alongside Grafana Labs (Monitoring / Observability) and Temporal Technologies Inc. (Workflow Orchestration). The page also lists AWS, Google Cloud Platform, Cloudflare and ClickHouse Inc. as hosting/cloud platform subprocessors.

  23. Databricks Blog47

    Unity Catalog Metric Views, now in Power BI: one unified, governed semantic layer for all your reporting and AI

    Databricks introduces Publish to Power BI, which connects Unity Catalog metric views to Power BI semantic models so each metric view becomes a table, its dimensions become columns and its measures become DAX measures, with Power BI querying the metric view itself. Publishing can be done on demand from Catalog Explorer or automated via Power BI tasks in Lakeflow Jobs, and DirectQuery queries honor Unity Catalog access controls. The feature is available in Private Preview.

  24. Baseten Blog69

    How Baseten and Genspark trained a deep research model to outperform Opus 5.5

    Genspark worked with Baseten Applied Research to post-train Qwen3.5-397B into a specialized Deep Research model that scores 0.759 on Genspark's quality grader, up from 0.612 for the base model and above Opus 5.5 (0.738), Sonnet 5.5 (0.742), GPT-5.6 Sol (0.637), Kimi K3 (0.724) and DeepSeek V4 Pro (0.675). The recipe splits the task into research, trained with reinforcement learning against a shaped semi-verifiable reward, and synthesis, trained mainly with supervised fine-tuning on high-quality reports written from the model's own research trajectories, followed by end-to-end RL. The model crawls 5× more sources per query than the base model and runs with NVFP4 weights on 2×B200 GPUs per replica, now serving as the default for Genspark's Deep Research agent in production.

    Why it matters: A detailed, reproducible account of post-training an open model for a judged production workload, useful for teams weighing RL versus SFT and reward shaping.

  25. AWS Machine Learning Blog40

    Amazon Bedrock tokenomics data lake

    AWS describes a centralized observability data lake that pulls Amazon Bedrock model invocation logs from linked accounts into a central account via CloudWatch subscription filters, processes them through Lambda and Amazon Data Firehose into Parquet in S3, and exposes them through Athena as a single queryable table. Per invocation, the flattened table captures the prompt, response, input and output token counts, model, stop reason, account, timestamp, and IAM role, and can be joined to CUR 2.0 data on the IAM ARN for per-invocation cost attribution. The solution is deployed with two CloudFormation templates from the AWS Sample Bedrock Invocation Data Lake GitHub repository, with an optional Amazon Quick Sight dashboard.

  26. Hacker News AI48

    Singularity Gate: Benchmark for AI-driven paradigm-shifting scientific discovery

    Singularity Gate is a benchmark built from real scientific findings first made public after the newest model's training cutoff, so no model could have seen them in training. Its v2.1 corpus contains 80 new findings disclosed in July 2026 or later, with a 12-task Hard set rated most paradigm-breaking, and models were run at xhigh reasoning effort in their native harnesses with tools on and web search off. Fable 5.2 leads Physics and Astronomy, Chemistry and Materials, Mathematics and TCS, and Life Sciences, while Opus 5.5 leads Earth and Planetary.

  27. Claude Blog75

    Introducing Claude Whiteboards, Claude Sheets, and Claude Data

    Anthropic launched Claude Whiteboards and Claude Sheets in beta and made Claude Dashboards generally available as Claude Data, all on paid plans. Whiteboards lets teams brainstorm with Claude on a shared board and export to tools like Lucid, Miro, or Vani by Zoho; Sheets generates and edits spreadsheets with tabs, formulas, formatting, and drop-downs, and can export to .xlsx or CSV. Claude Data adds certified queries, data stories, live sharing, design systems, caching, and tracked changes, with partners including Amplitude, Mixpanel, Omni, PostHog, and Sigma.

    Why it matters: It details new Claude workspace features and their availability, useful for teams deciding whether to adopt them.

  28. Hacker News AI50

    Sakana AI may go public in 2027

    Sakana AI may hold an IPO as early as next year, reportedly considering the Tokyo Stock Exchange, Nasdaq or both, with Nikkei reporting a market capitalization that could exceed Mercari's ¥680 billion ($4.3 billion) and Goldman Sachs hired as an adviser. In an X post, Sakana AI said it has no near-term IPO plans and that its priority remains building a generational AI company serving as Japan's national AI champion. The company raised ¥32 billion ($200 million) in Series B funding and said on Oct. 26 that it is training its own LLM from scratch, with its planned Sakana Sawara series to include a large and a small model.

  29. Amazon Science62

    Mitra-v2: Scaling synthetic pretraining for tabular foundation models

    Mitra-v2 is a second-generation tabular foundation model for classification and regression, pretrained entirely on synthetic tables with 10x as many rows and roughly 3x as many columns as Mitra-v1, plus a new hybrid SCM prior and hybrid optimizer. On TabArena it leads the single-model field, beating Mitra-v1 on 32 of 33 datasets where both run natively and completing all 51 datasets within the one-hour budget, and fine-tuned Mitra-v2 ranks first on ScoringBench. The fine-tuned models and inference/fine-tuning code are released under Apache-2.0 on Hugging Face.

    Why it matters: Useful for readers tracking tabular foundation models, as it reports concrete benchmark rankings and an open Apache-2.0 release.

Nov 10Tue
  1. The Verge · AI38

    The global pledge to use AI for climate good is a cruel joke

    At COP31 in Antalya, Turkey, more than 190 countries were invited to endorse the Antalya Artificial Intelligence Pledge, which calls for clean-energy-powered data centers, internationally recognized standards, transparency reporting, and e-waste management, but states its objectives are "voluntary and non-prescriptive" and create no new obligations. The article notes the US, which has more data centers than any other country, has no official delegates at the summit after Donald Trump pulled the nation out of the global climate accord, and cites August research in a Nature journal finding AI causes more carbon emissions by boosting fossil fuel production than it avoids through renewable energy gains. The author concludes the pledge offers at best a placebo effect.

  2. Hacker News AI42

    Wake me up when AI can...

    The author collected about 800 Hacker News comments from 2016 to 2026 of the form "wake me up when AI can…" and "I'll believe it when…", then let readers vote on whether each bar had been cleared; over 100,000 votes from almost 9,700 people were cast before voting closed. Of 826 statements, 112 were judged "clearly happened" (mostly code and text tasks), 89 "clearly not yet" (mostly physical-world and novel-discovery tasks), and 78 were contested.

  3. Hacker News AI68

    Is AI code rotting open source? 450 repos, 12 months

    An analysis of 450 top GitHub repositories found that the average repo had 11.5% of its merged pull requests AI-signed in September 2026, up from 2.3% in October 2025, though the median repo was at just 1.5% and 264 of the 450 repos have committed agent instruction files, making every AI figure a floor. Among 100 TypeScript and C# repos, complexity per 1,000 lines rose a median 4.8% in AI-heavy repos versus 3.0% in others, with the increase concentrated in fast-growing codebases while mature Microsoft projects held flat. AI PRs were typically 2.4× larger than other PRs and fixed within 30 days only slightly more often.

  4. Hacker News AI44

    Sydney researchers uncover a hidden limit to AI predictability

    University of Sydney researchers Pulin Gong and Andrew Ly found in Nature Machine Intelligence that neural networks' training outcomes depend on fractal 'riddled basins' of starting conditions, so even deterministic training with identical data can yield different behaviours. In one larger image-recognition network, making starting settings 10 billion times more precise produced no measurable improvement in reproducibility, and a one-bit change sent training to a different configuration in about 48 percent of cases. The work, based on roughly 500,000 trained networks, also found a trade-off in which settings that learn most effectively make outcomes harder to reproduce.