<?xml version='1.0' encoding='utf-8'?>
<rss version="2.0"><channel><title>Gnosly Feed - deepseek</title><link>https://gnosly.com/feed/deepseek/</link><description>Business Intelligence Console and Prediction Algorithms</description><item><id>3179</id><handler>https://www.linkedin.com/in/vincentius-liong/</handler><handler_name>Vincentius Liong/Leong 梁国豪</handler_name><external_id /><title>Huawei engineers have reportedly been working at DeepSeek’s offices to improve chip performance using the startup’s feedback. The goal is to train advanced AI models entirely on Chinese-made hardware.</title><text>&lt;p&gt;DeepSeek is reportedly testing Huawei’s Ascend 950DT chips and could begin using them for training by the end of 2026. &lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://lnkd.in/p/dudfA9te</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-10-21T05:03:26.643694+00:00</created_at><source><id>1</id><name>linkedin.com</name></source><tags><tag>ai</tag><tag>huawei</tag><tag>deepseek</tag></tags><media><media_item><id>3036</id><link>https://media.licdn.com/dms/image/v2/D5622AQHeaRiGjiNsYA/feedshare-shrink_800/B56aFaAg0qIUAc-/0/1792511446573?e=1794441600&amp;v=beta&amp;t=Aupilv-ITN6QnEbi65Pex7KBRJyFRiHlneLv41IeOF4</link><alt>View image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-21T05:03:28.617790+00:00</created_at></media_item></media></item><item><id>3164</id><handler>@deepseek_ai</handler><handler_name>DeepSeek</handler_name><external_id /><title>DEEPSEEK: Introducing DeepInfer, our in-house inference framework, now fully open source, powering DeepSeek-V4.1-Flash at scale with high throughput and low latency</title><text>&lt;p&gt;Source: DeepSeek (X)&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://x.com/deepseek_ai/status/2112476496283775135</url><published_at /><credits>0.0040</credits><parent_id /><referenced_id /><created_at>2026-10-20T16:56:39.888225+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>ai</tag><tag>open-source</tag><tag>deepseek</tag></tags><media><media_item><id>3023</id><link /><alt /><media_type>IMAGE</media_type><file>/media/content/3164/image.png</file><file_name>image.png</file_name><file_ext>png</file_ext><created_at>2026-10-20T17:02:06.088152+00:00</created_at></media_item></media></item><item><id>3139</id><handler>@ValsAI</handler><handler_name>Vals AI</handler_name><external_id /><title>Vals AI: "We found that DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program. We have been building a cybersecurity benchmark based on simulated corporate networks."</title><text>&lt;p&gt; Source: ValsAI (X)&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://x.com/ValsAI/status/2112471082397778083</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-10-20T12:54:53.311994+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>cybersecurity</tag><tag>ai</tag><tag>deepseek</tag></tags><media><media_item><id>2995</id><link>https://pbs.twimg.com/media/HVD946nbcAAam5i?format=jpg&amp;name=small</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-20T12:54:54.278492+00:00</created_at></media_item></media></item><item><id>3097</id><handler>https://www.linkedin.com/in/msukhareva/</handler><handler_name>Maria Sukhareva</handler_name><external_id /><title>DeepSeek is banned from US federal civilian government systems. The ban also covers third-party products and services built using DeepSeek implementations - Probably because DeepSeek is fast and cheap</title><text>&lt;p&gt;Source: Maria Sukhareva (X)&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://www.linkedin.com/feed/</url><published_at /><credits>0.0090</credits><parent_id /><referenced_id /><created_at>2026-10-20T07:18:34.556089+00:00</created_at><source><id>1</id><name>linkedin.com</name></source><tags><tag>deepseek</tag><tag>ai</tag></tags><media><media_item><id>2950</id><link /><alt /><media_type>IMAGE</media_type><file>/media/content/3097/20261020-deepseek-us-ban.jpg</file><file_name>20261020-deepseek-us-ban.jpg</file_name><file_ext>jpg</file_ext><created_at>2026-10-20T07:21:38.815578+00:00</created_at></media_item></media></item><item><id>2960</id><handler>@deepseek_ai</handler><handler_name>DeepSeek</handler_name><external_id /><title>DeepSeek-V4.1-Flash-1017 release</title><text>&lt;p&gt;Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API.&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://x.com/deepseek_ai/status/2111375640717033777</url><published_at /><credits>0.0070</credits><parent_id /><referenced_id /><created_at>2026-10-17T12:03:46.929556+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>deepseek</tag><tag>ai</tag><tag>china</tag></tags><media><media_item><id>2791</id><link>https://pbs.twimg.com/media/HU0b65KbsAAypqr?format=jpg&amp;name=360x360</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-17T12:03:47.939619+00:00</created_at></media_item><media_item><id>2792</id><link>https://pbs.twimg.com/media/HU0b7UfbwAAGXNH?format=jpg&amp;name=360x360</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-17T12:03:47.965516+00:00</created_at></media_item></media></item><item><id>2446</id><handler>@socialcapital</handler><handler_name>Social Capital</handler_name><external_id /><title>In the last week of September, DeepSeek models processed more tokens on OpenRouter than OpenAI, Google, Anthropic and xAI models combined.</title><text>&lt;p&gt;Source: socialcapital (X)&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://x.com/socialcapital/status/2107933993173954859</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-10-08T07:10:14.466586+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>deepseek</tag><tag>ai</tag><tag>openrouter</tag></tags><media><media_item><id>2258</id><link>https://pbs.twimg.com/media/HUDiy4jagAATvFv?format=jpg&amp;name=small</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-08T07:10:14.932694+00:00</created_at></media_item></media></item><item><id>2380</id><handler /><handler_name>Anniek Bao</handler_name><external_id /><title>DeepSeek considers doubling latest funding round to up to $15 billion, sources say</title><text>&lt;p&gt;Source: Anniek Bao (CNBC)&lt;/p&gt;</text><analysis /><region /><search_query /><result_type /><url>https://www.cnbc.com/2026/10/06/deepseek-funding-round.html</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-10-07T06:09:56.822761+00:00</created_at><source><id>26</id><name>cnbc.com</name></source><tags><tag>ai</tag><tag>deepseek</tag></tags><media><media_item><id>2194</id><link>https://image.cnbcfm.com/api/v1/image/108373192-1791288371612-gettyimages-2196098758-arriens-aiillust250130_nprzf.jpeg?v=1791288400&amp;w=240&amp;h=180&amp;vtcrop=y</link><alt>DeepSeek considers doubling latest funding round to up to $15 billion, sources say</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-10-07T06:09:57.104552+00:00</created_at></media_item></media></item><item><id>867</id><handler /><handler_name /><external_id /><title>Gnosly aggregates real-time data from news feeds, ...</title><text>Gnosly aggregates real-time data from news feeds, email accounts, X (formerly Twitter), weather, satellite imagery, and uploaded files or images into thematic streams. Pricing • Docs • Discord. The information is analyzed using the latest AI models from OpenAI, Anthropic, Google, and DeepSeek.</text><analysis /><region /><search_query /><result_type /><url>https://gnosly.com/</url><published_at /><credits>0.0040</credits><parent_id /><referenced_id /><created_at>2026-09-17T09:01:44.403035+00:00</created_at><source><id>28</id><name>gnosly.com</name></source><tags><tag>gnosly</tag><tag>deepseek</tag><tag>x</tag><tag>google</tag><tag>openai</tag><tag>anthropic</tag></tags><media /></item><item><id>603</id><handler>@victormustar</handler><handler_name>Victor M</handler_name><external_id /><title>DeepSeek V4.1 Flash + Blender = ❤️ No need to inst...</title><text>DeepSeek V4.1 Flash + Blender = ❤️ No need to install MCP (or anything other than Blender): just ask and it will write code to build the scene in headless Blender. Then repeatedly render to visually compare and refine until satisfied (and it's good at it).</text><analysis /><region /><search_query /><result_type /><url>https://x.com/victormustar/status/2098439548275409385</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-09-12T08:35:43.604548+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>deepseek</tag><tag>blender</tag><tag>render</tag></tags><media><media_item><id>533</id><link>https://pbs.twimg.com/media/HR8npOLbYAAyfeI?format=jpg&amp;name=360x360</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-09-12T08:35:43.791736+00:00</created_at></media_item><media_item><id>534</id><link>https://pbs.twimg.com/media/HR8npOkbEAEnz7I?format=jpg&amp;name=small</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-09-12T08:35:43.809551+00:00</created_at></media_item><media_item><id>535</id><link>https://pbs.twimg.com/media/HR8npNhWIAscpyw?format=jpg&amp;name=small</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-09-12T08:35:43.828384+00:00</created_at></media_item></media></item><item><id>571</id><handler>@deepseek_ai</handler><handler_name>DeepSeek</handler_name><external_id /><title>Introducing DeepSeek-V4.1-Flash: smarter, faster, ...</title><text>Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. Introducing the smallest model in our new architecture family, with native visual understanding. Designed for greater capability, faster inference, higher throughput, and scaling to larger models. 1/6</text><analysis /><region /><search_query /><result_type /><url>https://x.com/deepseek_ai/status/2097930608790167907</url><published_at /><credits>0.0050</credits><parent_id /><referenced_id /><created_at>2026-09-11T05:20:20.852953+00:00</created_at><source><id>2</id><name>x.com</name></source><tags><tag>v4.1</tag><tag>architecture</tag><tag>deepseek</tag></tags><media><media_item><id>507</id><link>https://pbs.twimg.com/media/HR1UyHiaAAAtpqw?format=jpg&amp;name=small</link><alt>Image</alt><media_type /><file /><file_name /><file_ext /><created_at>2026-09-11T05:20:21.124638+00:00</created_at></media_item></media></item><item><id>557</id><handler /><handler_name /><external_id /><title>Chinese AI Giants Accused of Sending Millions of User Queries to U.S. Models</title><text>Anthropic says DeepSeek and Moonshot used thousands of fake accounts and millions of real user queries in “distillation” effort to clone its AI capabilities.</text><analysis /><region /><search_query /><result_type /><url>https://www.wsj.com/tech/ai/chinese-ai-giants-accused-of-sending-millions-of-user-queries-to-u-s-models-768c9d26</url><published_at /><credits>0.0040</credits><parent_id /><referenced_id /><created_at>2026-09-10T17:38:02.146237+00:00</created_at><source><id>23</id><name>wsj.com</name></source><tags><tag>anthropic</tag><tag>deepseek</tag><tag>moonshot</tag></tags><media /></item></channel></rss>