iCreatAI News
Back to main site
Skip to content
#Data / training
Nov 6Fri
  1. Hacker News AI30

    The Future of Knowledge Work with AI

    The article describes knowledge workers who now spend their days managing and reviewing the output of AI agent "swarms" that draft copy, write code, and produce strategy, and argues that when every company in an industry runs its thinking through the same small set of LLMs, competitive differentiation erodes and skills atrophy. Drawing on the Kissimmee River straightening, Kyle Chayka's "Airspace," and research on LLM homogenization, it predicts a barbell-shaped future in which cheap high-volume AI output and premium verifiably human craft both thrive while the middle hollows out.

  2. Hacker News AI58

    Physical AI can't learn from the internet. We build the worlds it learns from

    Physicl launched Studio, a beta app that lets researchers describe a failure case in plain words and then proposes a matching scene from a library of 10,000 environments and 100,000 assets, suggesting variations to cover it. It renders images and video with ground truth computed from the 3D scene on every frame, including depth, surface normals, segmentation, boxes and camera pose. Studio is free during the beta and opening to research teams in waves.

  3. Hugging Face Blog58

    ToyBlocks: A controlled world for robot learning and evaluation

    ToyBlocks is an open simulation framework built around a robot arm and blocks on a tabletop that automatically generates tasks and demonstrations, using reinforcement learning and motion planning to produce skills such as pushing, flipping, and stacking. It supports studies of low-level action model generalization, high-level planning with vision-language models, agentic systems using tool calls, action-conditioned world models, and zero-shot sim-to-real transfer, with an initial release of 85,336 long-horizon demonstrations covering nearly 1,600 hours and 144 million frames. The authors present ToyBlocks as the first of their FoundationWorlds and plan a second release with a larger dataset, more skills, and integrated RL training.

  4. Hacker News AI38

    The Experts Being Recruited to Teach AI Their Professions

    AI training providers are recruiting experienced professionals—biostatisticians, air traffic controllers, engineers, lawyers and others—to design difficult tasks, create verified reference solutions, annotate specialist data and write detailed grading rubrics for evaluating AI systems. The article notes that many of these are project-based freelance engagements rather than salaried jobs, with no guarantee of placement or steady work, and that contributors are valued for professional judgment rather than AI-building skills.

  5. TechCrunch · AI38

    Physicl raises $10M to make the digital world real enough for robots

    Physicl, founded by Alex de Vigan, has raised $10 million in seed funding from Inovia Capital to build more realistic simulations for training physical AI and robot models. The startup is releasing a product called Studio with 10,000 3D environments and 100,000 assets that lets users create scenes to generate synthetic training data for difficult edge cases. World model builder SpAItial collaborated with Physicl as an early customer to improve the fidelity of its products.

  6. Artificial Analysis Articles40

    VIALS-AA: different paths to scientific image reasoning

    VIALS-AA analyzes how 19 models reason over scientific images, classifying 9,177 tool calls from its original test set: 38.3% were image viewing and 47.3% prepared or measured images. Image measurement made up just 5–6% of tool calls for GPT-6 Astra (max) and GPT-6.1 Sol (max), versus 25–27% for Claude Opus 5.5 (max) and Claude Sonnet 5.5 (max). The report also notes that Gemini 3.8 Flash (high) accounts for 86.7% of commentary/no-op calls, that more intelligent models tend to call fewer tools, and that cost per task varies by nearly 100× across models.

  7. Hacker News AI58

    Closed AI models dominate the price-performance frontier

    Village's LLM Atlas finds closed models hold almost every budget on the price-performance frontier, with open-weight models the best buy at only 3% of budgets, the lowest share measured, down from 41% in June. DeepSeek V4 Flash and Zhipu's GLM-5.2 once led large budget bands, but OpenAI's GPT-6 Luna and GPT-6.1 Terra and Anthropic's Claude Haiku 5.5 reclaimed them. The piece attributes the shift to closed labs distilling their own frontier models, releasing full model families, and serving inference at greater scale.

  8. Hacker News AI55

    Paper: Nubank cuts AI agent update cycles from 3 weeks to under 5 days

    The paper presents a hypothesis-driven simulation workflow using the Snowglobe simulator to screen candidate customer-experience agents before deployment, tested on Nubank's Card Delivery agent and its successor Card Management. Across 4 deployed versions, simulated and production version-level binary evaluator scores showed high correlation, and simulation-guided iteration raised transactional net promoter score by 36.69 points in a live A/B test. Screening open-weight configurations in over 16,000 simulated conversations led to a model that increased self-service rate by 8.82 percentage points in a subsequent live A/B test, with no statistically significant change in tNPS.

  9. vLLM Blog42

    VeRL-Omni v0.3.0: Async Training and Broader Multimodal Post-Training

    VeRL-Omni v0.3.0 adds asynchronous and decoupled RL execution paths across multimodal workloads, makes the TransferQueue/ReplayBuffer-based V1 trainer the default, and moves the rollout stack to vLLM 0.30.0 and vLLM-Omni 0.30.0. It includes end-to-end recipes for MiniMax H3 audio-video post-training with DiffusionNFT and MiniCPM-o 4.5 AVQA GSPO, where a 4×H800 experiment found separate-async about 9% slower per prompt at natural response lengths but roughly 6.7% faster at 1024 tokens. The release also expands model and algorithm coverage, adds Ascend and AMD ROCm Dockerfiles, and notes upgrade changes including V1 as default, named diffusion media artifacts, and fail-closed attention backends.

  10. MarkTechPost67

    Baidu Releases ERNIE 5.5 Flash Base: A 280B MoE Base Model With 14B Active Parameters

    Baidu has released ERNIE 5.5 Flash Base, a 280B-parameter text-only sparse MoE pre-training checkpoint that activates 14B parameters per token, with 128K training context and up to 1M for fine-tuning and inference. It uses hybrid attention (37 HCA layers caching 1 entry per 128 tokens plus 6 MLA+DSA layers), 512 routed experts with top-10 plus one shared expert, and ECHO in place of MTP, and was pre-trained from scratch on 33.5T tokens with the Muon optimizer. Baidu reports best scores on 24 of 30 base-model benchmarks in its unified pipeline, including MMLU-Pro 78.4 versus 69.2 for DeepSeek-V4-Flash-Base, with weights on Hugging Face and ModelScope in FP8 and BF16 and serving documented for SGLang and vLLM.

  11. Hacker News AI38

    Joules-to-GDP: How Much Economic Value Do AI Systems Produce per Unit of Energy?

    The paper introduces Joules-to-GDP, a framework measuring Net Value per Joule (NVJ) and Time Saving per Joule (TSJ), and applies it to 220 GDPval tasks across 44 occupations. Among five tested models, GLM 5.2 delivers the most net value at $72.27 per task on average, while Gemma 4 31B reaches 56% of that with 23% of the energy, for 2.5x the NVJ and 2.8x the TSJ; eight H100s use 3.2x the energy of one to cut completion time by only 22%. Measured against the work of experts who use AI, GLM 5.2's NVJ falls by 70% and its TSJ by 65%.

  12. TechCrunch · AI47

    Meta warns that election backlash against data centers could make AI expansion harder

    Meta President Dina Powell McCormick said at the WSJ Tech conference that Democratic midterm wins will make it harder for companies to build data centers, and that the industry has not done enough to explain the benefits of these investments. She said backlash against data centers is tied to how people feel about AI overall, and that Meta aims to demonstrate value to local communities through measures like paying for its own electricity and directing tax surpluses to teachers and first responders. Meta has data centers in 23 states, five planned at one gigawatt or more, and expects more project review going forward, while saying it receives daily calls to lease its compute.

  13. Hacker News AI22

    Benefits of Performance Testing with AI-Generated Synthetic Data

    This SDV tutorial compares real, fake (DayZSynthesizer), and synthetic (GaussianCopulaSynthesizer) credit card transaction data, showing synthetic data scores 70.98% versus 29.41% on SDV quality evaluation. In a simulated fraud-detection runtime scenario, synthetic data reproduces the real data's multi-modal runtime distribution (KSComplement ~0.86) while fake data does not (~0.30). The authors conclude synthetic data is a better substitute for production data in testing, development, and simulation.

  14. The Decoder66

    Baidu releases ERNIE 5.5 Flash Base with 280 billion parameters as an open-source foundation model

    Baidu has released ERNIE-5.5-Flash-Base, an open foundation model shipped without fine-tuning or post-training, with a post-trained version said to follow. It uses a mixture-of-experts architecture with 280 billion parameters, 14 billion active per input, and was pre-trained on 33.5 trillion tokens with a training context of up to 128,000 tokens, designed to support up to one million tokens during fine-tuning and inference. Baidu says it competes with similarly sized open foundation models like DeepSeek V4 Flash and MiMo V2.5 on reasoning, code, and long-context benchmarks, and the weights are available on Hugging Face in FP8 and BF16.

  15. AWS Machine Learning Blog29

    How Amazon Marketplace evaluates content quality with Amazon SageMaker AI

    Amazon describes a system that renders a full A+ Content page as a single snapshot and scores it with a fixed Vision-Language Model across four rubric dimensions, then fuses those scores with four OCR readability signals in a CatBoost classifier trained on expert annotations. On a 20 percent holdout set of 5,478 product detail pages from 10 categories, the fused model reached 0.94 accuracy and 0.95 F1, versus 0.80 and 0.84 for an averaged-score baseline, and SHAP values plus an LLM turn the results into prioritized recommendations. The pipeline is implemented on Amazon SageMaker AI using Processing Jobs, Inference endpoints, Feature Store, Training Jobs, Model Registry and monitoring, with a diffusion model generating improved image assets.

  16. METR Blog78

    Independent Assessment of AI R&D Acceleration at Anthropic in October 2026

    METR conducted a high-access evaluation at Anthropic from August 12 to October 29, 2026, concluding that Anthropic is likely improving model capabilities on well-specified tasks at around twice its March 2024–October 2025 rate, with AECI benchmark scores growing at an estimated 1.9x the pre-October 2025 rate (95% CI: 1.5–2.8x). METR estimates that a majority (perhaps ~80%) of the current acceleration is due to AI, found no clear bottlenecks to continued progress over the next 12 months, and cautioned that Anthropic redacted most of the evidence it found most informative.

    Why it matters: It offers a rare public accounting of how frontier AI R&D acceleration is measured, including explicit uncertainty ranges and disclosure of what redactions prevented the public from judging.

  17. Apple Machine Learning Research38

    Faster Retrieval for Better, More Efficient RAG

    Apple researchers present amortized maximum inner product search (MIPS), training small neural networks offline on a fixed key database to learn the support function of the key set, which is convex and whose gradient returns the best-matching key. SupportNet learns per-cluster support functions to route queries more accurately than centroid-based routing, while KeyNet rewrites queries before a standard IVF index; on Natural Questions, SupportNet reached 95% routing recall about 1.9× faster than centroid routing, and KeyNet cut the time to 99% recall by up to ~8×. The work was presented at ICML 2026.

  18. Hacker News AI40

    Which AI labs pay the most for engineers? Disclosed base pay, October 2026

    Anthropic had the highest median disclosed base pay for engineer roles among 49 AI labs tracked in October 2026, at $402,500 across 154 roles, followed by MatX at $380,000 and OpenAI at $339,000. The figures are midpoints of posted base salary ranges in US dollars, excluding equity, bonuses and offers, and labs are ranked only where at least ten open engineer roles disclosed a usable range. Frontier labs had the highest median by kind at $330,000, while defense companies were lowest at $190,000, a figure driven almost entirely by Anduril and Shield AI.

  19. Hacker News AI60

    Vetto ScienceBench: Expanding Scientific Capabilities of Frontier AI Models

    Vetto ScienceBench introduces 50 new data-analysis tasks in the Terminal-Bench Science format across 14 scientific disciplines, each reviewed for quality and audited per trial, where the best of 13 frontier models passes 49% of trials and most fail over 80%. Tasks require agents to explore data, choose a method, check its assumptions and deliver an answer or pipeline graded against held-out ground truth, with data features like instrument artifacts or selection effects that defeat textbook approaches. In about two thirds of failures the model had already seen the problem in its own output and then dismissed it as noise or corrected it with the wrong model, and a one-line warning raised GPT-6.1 Sol's pass rate from 29% to 39% without improving outcomes once the problem was noticed.

  20. Anthropic Research58

    AI for scientific discovery: How researchers use Claude

    Anthropic researchers used privacy-preserving classifiers to label 207,049 Claude and Claude Code research sessions (August 2–September 30, 2026) by field and task, finding heavier use in formal sciences like computer science, mathematics, and statistics than in health and lab-based physical sciences, with data analysis and writing dominating. Claude use rises with a constructed "theoretical AI capability" score, which correlates with reinforcement-learning suitability and cognitive work and negatively with physical, interpersonal, or regulated work. Scenario modeling suggests current capabilities could double research output in some fields, and full automation of non-physical work could raise output up to 15-fold in fields like computer science and statistics, versus three- or fourfold in bench sciences.

  21. Databricks Blog58

    From Delta Sharing to OpenSharing: what changed, what's next

    In June 2026 Databricks and the Linux Foundation announced that Delta Sharing has evolved into OpenSharing, an open, vendor-neutral protocol under the Linux Foundation AI & Data Foundation that extends sharing beyond tables to volumes, agent skills, and AI models, and adds Apache Iceberg REST Catalog support and on-premises or private cloud reach. Existing Delta Sharing clients and integrations remain compatible, and enabling OpenSharing on a Unity Catalog metastore is a configuration step rather than a migration. Databricks also introduced capabilities on top of the protocol, including SecureConnect (generally available), Global Distribution (private preview), Genie Agent Sharing (beta), and Lakebase table and change data feed sharing (public preview).

  22. Databricks Blog22

    Which Data Collaboration Tool: OpenSharing, Clean Rooms, Marketplace

    This Databricks blog compares three data-sharing methods: OpenSharing, which gives recipients direct read-only access to live assets such as tables, views, notebooks and AI models; Clean Rooms, where up to ten parties combine data and run only mutually approved code without exposing full underlying datasets; and Marketplace, a public storefront for discovering and publishing data and AI products. It advises choosing based on the relationship with the recipient and notes the tools are not mutually exclusive.

  23. TechCrunch · AI58

    Axiom Math wants to keep AI slop out of mathematics

    Axiom Math, a $1.6B startup founded in March 2025 by Stanford math Ph.D. dropout Carina Hong, announced it has formally verified the Stacks Project, translating its roughly 7,654 pages into about 16 million lines of computer-checked Lean code using a fleet of AI agents. The company argues that verified results become building blocks for future proofs and that the approach can extend to software written by AI agents.

  24. Hacker News AI34

    Five Grand Challenges in Material Science for AI to Solve

    The announcement lists five long-standing, unsolved materials science problems—predictive theory of cuprate superconductivity, interfacial electrical properties from processing history, breaking the speed–accuracy barrier in quantum chemistry, tuning materials with light, and predicting heat transport across solid interfaces—with a $1 million prize pool for solutions. Submissions open in November 2026 and are sent by email, with each problem page specifying required contents and submissions bound by the Official Rules.

  25. The Decoder58

    No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark

    SemiAnalysis released GDPx, a benchmark of 206 tasks across eleven domains that places AI agents in simulated work environments operating professional software for up to eight hours. Claude Opus 5.5 leads at 12 percent, followed by Gemini 4 Carbon at 10.9 percent and GPT-6 Astra at 9.6 percent, and no model fully solves the finance and legal tasks. Critics including developer Josh Purtell call GDPx "more noise than signal," citing provider-specific agent software and small sub-domain task counts.

  26. Waymo Blog36

    Read more: How the Waymo Driver Manages Severe Winter Weather

    Waymo describes how its Driver handles winter conditions, using cameras, radar and lidar with heaters and micro-wipers to keep sensors clear, and measuring real-time traction to adjust steering, braking and acceleration. The company says it may slow, limit or pause service when weather exceeds the boundaries defined by its Safety Framework, drawing on fleet-shared road data, forecasts and local officials. Waymo says it has accumulated tens of thousands of real-world miles in snowy regions including Upstate New York, Michigan's Upper Peninsula and the Sierra Nevada, and will conduct closed-course evaluations and heavy snowfall data collection in Alaska this season.

  27. Pragmatic Engineer22

    Scaling Slack and GitHub with Maude Lemaire

    In this episode of The Pragmatic Engineer, Maude Lemaire discusses scaling large products at Slack, GitHub and Cursor, including Slack's load testing with the Go-based Koi Pond tool, thundering herds and event floods, and GitHub's Rails monolith and database sharding limits. She also covers how she uses AI in her work while keeping judgment calls in her own hands, and suggests rotation between platform and product engineering roles.

  28. NVIDIA Technical Blog22

    Five Ways NVIDIA Co-Design Maximizes the Value of Every Byte in the AI Factory

    NVIDIA outlines five co-design approaches for maximizing the value of every byte in an AI factory, where high bandwidth memory (HBM) is described as the fastest, most expensive and most contested memory tier. The piece argues that bytes spent on redundant weight copies, idle sessions' KV cache or recomputed prefixes are bytes not producing tokens, shifting the question from how much memory a GPU has to how much useful work each byte delivers.

Nov 5Thu
  1. The Verge · AI22

    The hidden supply chain costs of the AI boom

    In this Decoder interview, Harvard Business School economist Willy Shih discusses his book Where Does My Stuff Come From? and argues that global supply chains are shaped by economics, so tariffs tend to push manufacturers to reroute production through places like Vietnam and Mexico rather than reshore it. He explains why China remains a manufacturing powerhouse through industrial policy, regional competition and heavy automation investment, and why the iPhone illustrates the difficulty of moving complex production.

  2. Amazon Science18

    Hamed Karimi wins ECML PKDD Test-of-Time Award

    Amazon senior applied scientist Hamed Karimi and coauthors Julie Nutini and Mark Schmidt won the 2026 ECML PKDD Test-of-Time Award for their 2016 paper on linear convergence of gradient and proximal-gradient methods under the Polyak–Łojasiewicz condition. The paper showed that the largely overlooked 1963 inequality can prove fast linear convergence for standard training methods even when many solutions work equally well, mapped how previously scattered weaker-condition results relate, and extended the analysis to nonsmooth landscapes and algorithms including coordinate, sign gradient, and stochastic gradient descent. It has been cited over 2,000 times and is used in graduate textbooks.

  3. TechCrunch · AI58

    Dropbox acquires Tigris Data for $350M to bolster its AI infrastructure

    Dropbox announced a $350 million acquisition of cloud technology company Tigris Data, which it says is expected to strengthen its AI infrastructure capabilities and support DBX Agent Cloud, a new storage platform for businesses and developers building AI models and agents. The deal is expected to close in the fourth quarter of 2026, and Dropbox said it has no current plans to reduce headcount as part of the acquisition. Tigris Data, founded by former Uber engineers, provides cloud storage technology that positions data closer to the users and computing resources that need it.

  4. The Decoder58

    Apple slashes internal AI budgets, with deeper cuts still planned

    According to Bloomberg, Apple cut some engineers' daily token budget from $300 to $200 last month, with plans to potentially lower it to $100 a day and as low as $25 for some non-technical teams, and employees needing more can request exceptions from management. Employees have access to Anthropic's Claude, OpenAI's ChatGPT, and Google's Gemini, and some teams will be shifted to cheaper models or AI that runs locally on devices. The cuts fit the direction of new CEO John Ternus, who has pushed to make the company leaner, even as Apple's R&D spending jumped 32 percent in fiscal year 2026, driven by higher infrastructure and AI costs.

  5. Baseten Blog22

    Baseten joins the NVIDIA Nemotron Coalition to advance custom AI in production

    Baseten has joined the NVIDIA Nemotron Coalition, expanding its collaboration with NVIDIA on open frontier models, post-training, RL environments, and dedicated inference. The post cites CodeRabbit's Nemotron 3.5 Lightning customization, which raised routing-policy agreement from 75.8% to 80.7% on 1,000 tasks and cut inference cost 50.4% at saturated throughput, and notes Harvey post-trains Nemotron on Baseten for legal workflows. Baseten Loops now supports Nemotron 3.5 Super VL and Nemotron 3 Ultra, with customers retaining ownership of trained weights.

  6. Cerebras Blog50

    What It Took to Build Wafer-Scale, and Why It Lasts

    Cerebras CEO Andrew Feldman rebuts a Culper Research short-seller report claiming its wafers wear out after about two years, saying the company does not use the PariPoser material and that its patented assembly keeps temperatures steady and even while flexible connections absorb thermal expansion. He cites TSMC qualification for 10 years of continuous operation at 105°C and an average junction temperature below 65°C across deployed CS-3 systems, arguing Arrhenius modeling with those inputs implies 33 years at maximum temperature and over a century at average. Feldman also says fewer than 0.5% of deployed CS-3 wafers have been decommissioned since 2024 and defends the company's 3-5 year depreciation schedule as conservative.

  7. TechCrunch · AI40

    A YC startup is putting $1M behind 5 unsolved materials science problems for AI

    Discovered Materials, a San Francisco startup in Y Combinator's spring 2026 batch, announced a $1 million prize pool for AI models that solve five open materials science problems, including a theory of cuprate superconductivity and numerical methods for silicon-silicon dioxide interfaces and Density Functional Theory. The company said frontier lab models given internet access and other resources worked autonomously for days on each problem without making much headway. The startup, which develops new materials for semiconductors and partners with chip giants, said its timeline for generating a useful material has accelerated beyond its earlier one-year hope.

  8. Databricks Blog36

    Push-based CDC into Unity Catalog with Zerobus Ingest and Debezium

    Debezium Server 3.7 ships an open-source Zerobus Ingest sink, built by Databricks with NVIDIA's contribution of envelope mode, that pushes change events directly into Unity Catalog Delta tables without a Kafka cluster between capture and the lakehouse. Envelope mode lands each full Debezium change event in a fixed nine-column table with source position and an idempotency key, and the sink waits for Zerobus to confirm durability before Debezium advances its offset, giving at-least-once delivery. Landed events can then be shaped into Type-1 current-state tables via AUTO CDC sequenced by source LSN, or Type-2 history tables.

  9. Hacker News AI65

    We measured the impact of AI on 46 OSS repos (200k commits, 16M lines)

    Kerno's first-edition report analyzes 201,576 commits across 46 open-source projects from October 2024 to October 2026 and finds new code carries 8.1 decision points per 100 lines after AI adoption, up from 5.2, with complexity growth shifting from −0.4% to +3.1% a month. Each engineer ships 71% more production code per week, while AI-marked commits are revisited 43% more often than other commits and are 17% more likely to add a hard-to-test function. Claude signs 69% of AI-marked commits, which are 54% bigger than unmarked commits, and on the current trend complexity per 100 new lines would rise another 42% by late 2028.

  10. Anthropic Research66

    Investigating vault particles with Claude

    Anthropic's biology research lab reports the first 3D structure of a vault particle from a prokaryotic cell—a bacteriophage-derived vault built from 96 copies of a single protein, L05, solved by cryo-EM. Under researchers' direction, Claude searched protein databases, designed DNA constructs, optimized a size-exclusion chromatography isolation method that cut purification from about three days to roughly 90 minutes, and processed microscopy data, yielding a first structural model in 34 hours. The team also engineered altered and enlarged vault variants, and notes the work demonstrates using Claude across every step of fundamental biology research.

    Why it matters: Describes a concrete AI-assisted structural biology result, including the methods, speedups, and limitations the team encountered.

  11. Hacker News AI48

    Novel AI framework runs on optical hardware

    Researchers including a University of Tokyo team combined an SS-DFA training method with a simpler "single-map framework" architecture and demonstrated it in an optical computing system using a 200-meter optical fiber and few electronic components. The simplified design matched or in some tests outperformed more complex neural networks while requiring far less computation, measurement and data storage. The team says future work must test scaling to the largest AI models and address miniaturization, stability and large-scale manufacturing.

  12. Hacker News AI60

    Agent Delegability Index: How much of the work can AI agents take over?

    Balam Labs' Agent Delegability Index v2026.1 rates 18,796 occupational task statements, 433 occupations and 146 countries, finding that 35.2% of task time in world employment is at least partly delegable to AI agents in 2026, while only 2.5% can be handed over in full. Office work is the most delegable major group (managers 81.8%, clerical support 78.3%) and elementary occupations the least (12.3%), with each log point of GDP per capita adding 7.23 percentage points of delegable work. Full delegation peaks for tasks of about 11.8 minutes, and delegability correlates with observed AI use across 101 countries (r 0.86).