NAll news

All news

Wednesday916 items

  1. ModelBest OpenBMB
    AI score39

    OpenBMB Congratulates MLX-VLM and MLX-Audio on Moving to Neywalabs

    MLX-VLM and MLX-Audio have moved to the Neywalabs GitHub organization, and original author Prince Canuma stated that both projects will continue to be open source under the MIT license. OpenBMB posted its congratulations, saying it is delighted to collaborate with the other party and looks forward to the projects' future progress.

    Quoted postPrince Canuma · @Prince_Canuma

    Today MLX-VLM and MLX-Audio move to the @Neywalabs GitHub org 🏠 I started MLX-VLM on my M1 Air with a vision of what local inference and small models could do. I couldn't even run some of the models I was porting 😂 Then @lllucas and I built MLX-Audio because we had an itch to scratch. Today they're used by 7M+ developers, and by teams at @IBMResearch, and many more. We've also partnered with the teams behind the best open models such as @GoogleDeepMind/@googlegemma, @huggingface, @cohere, @OpenBMB, @Recursive_SI, @liquidai, @Open_MOSS and more. These projects outgrew my personal account. Now a whole team wakes up every day to make them better. My promise: MLX-VLM and MLX-Audio will remain open source under the same MIT license. That will never change. "Helping people through technology" has always been my north star for decades. This is how we at @Neywalabs keep doing it, at a bigger scale. Thank you all from the bottom of my heart. There is so much more on the way 🚀 https://t.co/wSZO3OT9Y5

    #Industry news#Open-source ecosystem

  2. ModelBest OpenBMB
    AI score43

    OpenBMB praises MiniCPM-V on-device calorie reminder project

    ModelBest OpenBMB reposted and praised an open-source maker project: the project deploys MiniCPM-V-4.7-35B-A3B locally on a phone to recognize food and estimate calories, then uses a small ESP32 screen to display daily intake progress in sync. The project has been open-sourced to the esp32-intake repository.

    Quoted postEnglish text is being prepared. · @emwstudio

    English text is being prepared.

    #English text is being prepared.#English text is being prepared.#Multimodal

  3. ModelBest OpenBMB
    AI score22

    OpenBMB Responds to MiniCPM-OCR Integration with Nock

    NockAI announced that MiniCPM-OCR has been integrated into Nock, stating that it accurately recognized a 1914 typed letter and a Chinese court notice, but misrecognized a 1921 Serbian telegraph form. OpenBMB reposted the response, thanking NockAI for sharing both successful and difficult cases, and considering such real-world feedback very valuable.

    Quoted postNockAI · @useNockAI

    We wrote up everything nock runs on your own machine: a 1 GB coder, a video model, an agent, and now Nock Hub. New today: MiniCPM-OCR from @OpenBMB. It read a 1914 typed letter and a Chinese court notice with every character right. A 1921 Serbian telegraph form, it got badly wrong. Both results are on the listing. pipx upgrade nock-cli nock add minicpm-ocr

    #Industry news#Multimodal#Open-source ecosystem

  4. François Chollet
    AI score26

    Chollet agrees with Altman: AI should always be a tool for humans

    Sam Altman stated that people should always be in control of the future, and OpenAI opposes relegating humans to only doing "things that AI systems think are worth rewarding in humans"; humans should not cede control of the future to superintelligence. François Chollet reposted and concurred, saying that the only future worth building is one in which AI always remains a tool in human hands, used to improve the human condition.

    Quoted postSam Altman · @sama

    People should always be in control of the future. OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that “AI systems think makes sense to reward in humans”. People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress.

    #English text is being prepared.

  5. Alibaba Cloud Developer Community AI
    AI score29

    2026 CAD AI Primitive and Legend Recognition: Technological Evolution and System Construction

    In 2026, CAD AI primitive and legend recognition has shifted from rule-based template matching to the integration of multimodal vision large models (VLMs) and 3D geometric deep learning. Industrial-grade systems need to sequentially complete multi-format drawing parsing, semantic segmentation of primitives and legends, GNN semantic association and BIM reverse mapping, and a human-machine collaborative verification loop, with accuracy typically required to reach above 99%. Few-shot learning and block-based parallel streaming processing address the performance bottlenecks of non-standard legend generalization and large drawings with millions of primitives, respectively.

    #Tutorials / practice#Multimodal#Image generation

  6. Alibaba Cloud Developer Community AI
    AI score18

    Alibaba Cloud Token Plan Double 11 Annual Mega Sale: Extra 50% Off on Top of Discounts, Starting at 19.5 Yuan for New Customers

    Alibaba Cloud Token Plan Double 11 Annual Mega Sale is here. For new individual customers, the Lite plan is 19.5 yuan for the first month (originally 80 yuan/month). The four tiers of plans use Credits to uniformly offset usage of text, vision, and speech models. The Standard and Pro plans include Harness tool quota, with usage beyond that billed at 12% off; calls to qwen3.8-max and qwen3.8-flash from 22:00 each night to 08:00 the next day enjoy 60% off, and the DeepSeek series enjoys 50% off.

    Same story · 3 reports

    #Product updates#Agents#Multimodal

  7. Alibaba Cloud Developer Community AI
    AI score22

    Developer Kuangshi open-sources a free test development practice-question website, featuring 200 AI testing interview questions

    Developer Kuangshi has launched a free, open-source test development practice-question website, ai.testfather.cn, organizing 200 interview questions across 10 topics, covering large model testing, automated testing, performance testing, and other areas. Each question includes four interviewer-perspective dimensions: company motivation, passing criteria, discriminative power, and bonus points. The website also provides a 13-stage learning path, 22 hands-on tutorials, and a navigation of 140 testing tools. It can be used without registration or login, and the content is simultaneously open-sourced on GitHub.

    #Tutorials / practice#Agents#Open-source ecosystem

  8. Alibaba Cloud Developer Community AI
    AI score22

    Alibaba Cloud Token Plan Double 11 campaign launched: up to 50% off, new customers from 19.5 yuan

    Alibaba Cloud Token Plan Double 11 campaign launched. The individual Lite monthly plan is priced at 19.50 yuan during the campaign, corresponding to 11,500 Credits/month; the team standard seat monthly plan starts at 150 yuan/seat/month, corresponding to 25,000 Credits/seat/month. Standard/Pro packages include Harness tool quota, with overage billed at 12% off, plus top-up packs and a nightly discount from 22:00 to 08:00 the next day (as low as 60% off). The team edition promises that data will not be used for model optimization.

    Same story · 3 reports

    #Industry news#Agents#English text is being prepared.

  9. Alibaba Cloud Developer Community AI
    AI score13

    Alibaba Cloud 2026 Double 11 Promotion Analysis: Event Dates, Participation Entry Points, and Benefits Across All Categories

    The theme of the Alibaba Cloud 2026 Double 11 promotion is "Smart Savings at 11.11, Use Alibaba Cloud for AI." The event runs until November 30 and is divided into three phases: the warm-up period, the official peak period, and the extended encore period. New users can participate in a flash sale for a lightweight application server at 38 yuan for their first purchase, while existing users can enjoy new purchases and renewals at the same price of 99 yuan for e instances and 199 yuan for u1 instances, and the Bailian Token Plan is available at up to 50% off. Enterprise users must sign up in advance for tiered spend-based rebates; orders from users who have not signed up will not count toward cumulative spending.

    Same story · 20 reports

    #Industry news

  10. Cailian Press
    AI score70

    AMD's data center revenue reached $8.5 billion in the third quarter of fiscal 2026, with revenue growth expected to exceed 80% in 2027

    AMD announced its third-quarter fiscal 2026 financial results, with total revenue of $13.4 billion, up 45% year over year, and data center revenue of $8.5 billion, up 97% year over year, accounting for 64% of total revenue. The company provided fourth-quarter revenue guidance of approximately $17 billion and expects revenue growth to exceed 80% in 2027, with server CPU revenue doubling year over year, at which point the data center business will account for more than 80% of total revenue. The earnings call also mentioned that MI455X and Helios racks began shipping this quarter, and that customers including Anthropic, Meta, Microsoft, OpenAI, and Oracle have already deployed them online, and announced the acquisitions of Taalas and World Labs, with Fei-Fei Li to serve as AMD's chief scientist.

    Same story · 6 reports

    #Industry news#English text is being prepared.#English text is being prepared.

  11. Shao MengProduct release
    AI score78

    OpenAI Used GPT-6 Astra to Design a Hardware Device in One Week

    OpenAI announced progress made by GPT-6 Astra in physical engineering: two employees used Astra to design from scratch a small device that can converse with Codex, with no direct manual modifications throughout the entire process; the design was completed in less than a week, and a contract manufacturer produced a small batch of prototypes in less than a month. OpenAI said the model still cannot reliably account for factory production constraints, tolerances, and assembly details, and is still several rounds of review away from mass production, so it is publicly recruiting industry partners. On BenchCAD 2.0, GPT-6 Astra scored 0.594 at its highest reasoning setting, while an unreleased internal new model averaged 0.875. OpenAI is also expanding its partnership with Chip Ganassi Racing to test converting 3D scan data into editable CAD models and 2D production drawings, and newly announced a research collaboration with Mazda on aerodynamics and crash safety.

    Quoted postOpenAI · @OpenAI

    ChatGPT and Codex are helping engineers build for the physical world. We’re sharing our progress across mechanical design, electrical engineering, and simulation—from helping our hardware teams iterate faster to testing new workflows with industry collaborators. https://t.co/b3rW9zYJGA

    Same story · 10 reports

    #Industry news#Agents#English text is being prepared.

  12. Alex Heath
    AI score27

    Shopify CEO Tobi: AI Has Already Added Trillions of Dollars in Value to the Earth

    Shopify CEO Tobi Lütke said the AI industry has already added potentially trillions of dollars in new objective value to the Earth, value that had previously been "hidden in floating-point operations." He believes floating-point operations and matrix multiplication have long existed, but people didn't realize that greatly accelerating them in a particular way could unlock trillions of dollars in growth, and he used uranium and nuclear energy as an analogy to explain that value is often locked away by knowledge.

    Quoted postAlex Heath · @alexeheath

    Tobi Lütke on the AI agent revolution and future of Shopify @tobi on where agent commerce is headed, sci-fi becoming reality, and his GPU wine cellar. Why did Amazon block Meta's Muse but Shopify welcomed it? I sit down with Tobi in Toronto to discuss why the companies are taking different approaches to agents and how shopping changes when AI starts buying things on your behalf. He explains the early AI shopping behavior he’s seeing. We also get into why he thinks AI makes starting a business easier, the backlash against the technology, and why everyone will eventually have to embrace it to get ahead in their careers. We get into how AI agents will make money, who pays when they handle a purchase, and why advertising gets complicated when your assistant is supposed to work for you. We discuss his experience on Meta’s board and working with Mark Zuckerberg. Lütke also shares why he turned his wine cellar into a GPU server rack and how he uses his personal AI agent that runs entirely from his house. This is one of my favorite pods I've ever done. Timestamps: 00:00 Amazon blocks Muse 02:40 Shopify's approach to AI agents 07:00 How AI agents will change shopping 11:05 Sci-fi and tech optimism 17:57 Zero-sum thinking vs. abundance 21:20 AI backlash and the founder advantage 45:28 Building an AI-first Shopify 52:18 Who profits from agentic commerce? 01:03:22 How Shopify stays ahead 01:10:39 Advertising and discovery in the agent era 01:16:02 Rack, Cellar, and Claw Links: Spotify: https://t.co/dEn15RrP7B Substack: https://t.co/JxkbohU0qz YouTube: https://t.co/2DTtOZ64yN Apple Podcasts: https://t.co/fEI7x7IzlD This episode's presenting sponsor is @hyperagentapp Thanks to the show’s premiere sponsors: @Atlassian, @meetgranola, and @mercury

    #English text is being prepared.#Industry news

  13. Peter Yang
    AI score52

    Peter Yang tested Grok's phone feature and gave 5 prompt use cases

    Peter Yang introduced the new phone feature of Grok @Bot and gave 5 prompt use cases that can be used directly, including negotiating with carriers, canceling gym memberships, waiting on hold with airlines, making restaurant reservations, and requesting quotes from 3 contractors. He noted that the feature currently only supports US numbers and included a link to a real call demonstration video. The quoted tweet said that his bot had called Comcast and cut his internet bill by nearly $500 per year.

    Quoted postPeter Yang · @petergyang

    Grok @Bot can now call US businesses for you, and it's one of my favorite AI features ever. Here's my new video where I play a real call of my bot successfully negotiating with Comcast to cut my internet bill by almost $500 a year. Some highlights include my bot: → Responding to "How's the weather?" → Turning down 3 upsells from the rep → Pushing the rep with a rival's lower offer 📌 Watch now: https://t.co/UWt3SanRSN

    Same story · 4 reports

    #Product updates#Agents#Speech

  14. Artificial Analysis
    AI score41

    Claude Opus 5.5 struggles to judge task termination when operating a robotic arm

    Artificial Analysis points out that when language models operate a robotic arm, they find it difficult to judge when to stop. Claude Opus 5.5 took 23 minutes to complete the task and said, "The crocodile looks like it's completely in now," then spent another 22 minutes pushing the crocodile back and forth, ultimately ending up almost back in its original position.

    #English text is being prepared.#Agents#Embodied AI

  15. Artificial AnalysisAnalysis
    AI score60

    Artificial Analysis releases AA-PersonalityBench…

    In AA-PersonalityBench, Artificial Analysis compared the 10 open-weight models and 10 closed-source models with the highest Intelligence Index scores, ranking them on a 1-to-100 percentile scale across 30 personality traits. Leading open-source models average 32 points lower than closed-source models on Truthful, the largest gap of any trait; they also rank lower on traits such as Pragmatic, Evidence-Based, and Structured, while ranking higher on Slop, Bottom-Line Summary, and Scannable.

    #Evaluations / benchmarks#Open-source ecosystem

  16. Codex：GitHub ReleasesFirst-party
    AI score6

    OpenAI Codex releases rust-v0.167.0-alpha.5

    OpenAI Codex releases rust-v0.167.0-alpha.5. This version is an alpha pre-release; the specific update details are not shown on the page.

    #English text is being prepared.

  17. Codex：GitHub ReleasesFirst-party
    AI score13

    OpenAI Codex releases rust-v0.168.0-alpha.1

    OpenAI Codex has released version rust-v0.168.0-alpha.1. This version is an alpha pre-release, and the release page does not provide specific update details.

    #English text is being prepared.

  18. East Money
    AI score40

    Apple warns that AI agents act unpredictably, says its ability to supervise and control them is limited

    In a filed document, Apple lists artificial intelligence agents as a new risk factor, warning that such agents may act in unexpected or unpredictable ways, including taking actions without instruction, guidance, or authorization, and that the company's ability to supervise and control third-party AI agents may be limited. Apple currently has no agentic AI; new CEO John Ternus said there is still a lot of room to expand Siri's capabilities and emphasized ensuring users clearly understand which access permissions they have granted.

    Same story · 7 reports

    #Industry news#Agents#Safety / alignment

  19. East Money
    AI score42

    DeepSeek V5 Plans to Pretrain on Huawei Ascend Chips, the First Large Model Natively Trained on Ascend 950

    DeepSeek's next-generation large model, DeepSeek V5, plans to conduct pretraining based on Huawei Ascend chips. Previously, Ascend chips were mainly used for inference tasks of DeepSeek large models. V5 will be its first large model natively trained on Ascend 950, with the project timeline pointing to a launch next year. Huaxi Securities noted that the driving force of domestic computing power has shifted from catching up on single-card performance to system-level engineering capability and supply guarantee capability, and manufacturers with system-level supply capabilities such as super-node complete machines will benefit.

    #Industry news#DeepSeek#English text is being prepared.

  20. East Money
    AI score36

    Tech giants' AI demand slowing? Xiechuang Data, Xingyun Technology and other computing power industry chain companies respond: no signs of reduction, some order prices raised

    Xiechuang Data responded that there are currently no signs of reduction, and the capital expenditure of leading domestic internet companies is expected to remain at a relatively high scale next year; Xingyun Technology has disclosed a cumulative total of approximately 20.641 billion yuan in computing power and storage orders this year, and Customer V and Customer VB have completed the second upward adjustment of contract amounts. GigaDevice said the storage industry boom continues, and prices for high-capacity NOR Flash products are expected to rise moderately; Ruike Da's 400G/800G AEC products have gradually entered volume delivery, and 1.6T products have been sampled to North American customers.

    #Industry news#English text is being prepared.