<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title><![CDATA[AIghten – AI, SaaS & IT Business Insights]]></title>
    <link>https://www.aighten.com/en</link>
    <description><![CDATA[AIghten is an AI-focused blog covering SaaS products, IT business strategies, startup growth, and emerging technology trends for builders and founders.]]></description>
    <language>en</language>
    <lastBuildDate>Wed, 28 Oct 2026 03:13:40 GMT</lastBuildDate>
    <atom:link href="https://www.aighten.com/rss.xml" rel="self" type="application/rss+xml"/>
    <ttl>60</ttl>
    <image>
      <url>https://www.aighten.com/logo.png</url>
      <title><![CDATA[AIghten – AI, SaaS & IT Business Insights]]></title>
      <link>https://www.aighten.com/en</link>
    </image>
    <item>
      <title><![CDATA[Liquid AI LFM3: 2B Open-Weight Models for On-Device AI]]></title>
      <link>https://www.aighten.com/en/blog/liquid-ai-lfm3-open-weight-edge-models-benchmarks</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/liquid-ai-lfm3-open-weight-edge-models-benchmarks</guid>
      <description><![CDATA[Liquid AI's LFM3 family brings text, vision, audio and omni models of about 2B parameters with open weights. Benchmarks, specs, memory use and license terms.]]></description>
      <pubDate>Wed, 28 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[LFM3]]></category>
      <category><![CDATA[Liquid AI LFM3]]></category>
      <category><![CDATA[LFM3-1.7B]]></category>
      <category><![CDATA[LFM3-Omni-2B]]></category>
      <category><![CDATA[on-device LLM]]></category>
      <category><![CDATA[small language model benchmarks]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-28/liquid-ai-lfm3-open-weight-edge-models-benchmarks/cover-2911f911.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Genie 4: Google DeepMind's World Model with Native Audio]]></title>
      <link>https://www.aighten.com/en/blog/google-deepmind-genie-4-world-model-audio-access</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/google-deepmind-genie-4-world-model-audio-access</guid>
      <description><![CDATA[Genie 4 is Google DeepMind's real-time world model that generates sound and dialogue with video. Who can use it, how it differs from Genie 3, and its limits.]]></description>
      <pubDate>Wed, 28 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Genie 4]]></category>
      <category><![CDATA[Google DeepMind Genie 4]]></category>
      <category><![CDATA[Genie 4 world model]]></category>
      <category><![CDATA[Project Genie]]></category>
      <category><![CDATA[Genie 3 vs Genie 4]]></category>
      <category><![CDATA[AI world model]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-28/google-deepmind-genie-4-world-model-audio-access/cover-a797b7b6.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[GPT-6.1 Terra: Pricing, Benchmarks and Computer Use]]></title>
      <link>https://www.aighten.com/en/blog/openai-gpt-6-1-terra-computer-use-pricing-benchmarks</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/openai-gpt-6-1-terra-computer-use-pricing-benchmarks</guid>
      <description><![CDATA[GPT-6.1 Terra is OpenAI's mid-tier model for computer use: $1/$5 per million tokens, a 1.05M context window, Terminal-Bench 4.0 at 50.3% and Critical cyber.]]></description>
      <pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[GPT-6.1 Terra]]></category>
      <category><![CDATA[gpt-6.1-terra API]]></category>
      <category><![CDATA[GPT-6.1 Terra pricing]]></category>
      <category><![CDATA[GPT-6.1 Terra benchmarks]]></category>
      <category><![CDATA[OpenAI computer use model]]></category>
      <category><![CDATA[GPT-5.6 Terra vs GPT-6.1 Terra]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-27/openai-gpt-6-1-terra-computer-use-pricing-benchmarks/cover-0a3b1c39.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Anthropic S-1 Filing: $16.3B Revenue and Who Controls ANT]]></title>
      <link>https://www.aighten.com/en/blog/anthropic-s1-ipo-filing-revenue-losses-governance</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/anthropic-s1-ipo-filing-revenue-losses-governance</guid>
      <description><![CDATA[Anthropic's S-1 shows $16.3B in H1 2026 revenue, a first quarterly operating profit, an $80.7B mostly non-cash loss and a founder share with 50.1% of votes.]]></description>
      <pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[finance]]></category>
      <category><![CDATA[Anthropic S-1]]></category>
      <category><![CDATA[Anthropic IPO]]></category>
      <category><![CDATA[Anthropic revenue]]></category>
      <category><![CDATA[ANT Nasdaq]]></category>
      <category><![CDATA[Anthropic Class F stock]]></category>
      <category><![CDATA[Long-Term Benefit Trust]]></category>
      <category><![CDATA[Anthropic compute commitments]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-27/anthropic-s1-ipo-filing-revenue-losses-governance/cover-6cfb679f.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Node2Shell CVE-2026-111497: Node.js Brotli RCE and Fixes]]></title>
      <link>https://www.aighten.com/en/blog/node2shell-cve-2026-111497-nodejs-brotli-rce</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/node2shell-cve-2026-111497-nodejs-brotli-rce</guid>
      <description><![CDATA[Node2Shell (CVE-2026-111497) is a Brotli decoder bug in Node.js that OpenAI's Codex Security found. Fixed versions, who is exposed and what to patch first.]]></description>
      <pubDate>Mon, 26 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Node2Shell]]></category>
      <category><![CDATA[CVE-2026-111497]]></category>
      <category><![CDATA[Node.js Brotli vulnerability]]></category>
      <category><![CDATA[Node.js security release October 2026]]></category>
      <category><![CDATA[Brotli decoder out-of-bounds write]]></category>
      <category><![CDATA[OpenAI Codex Security]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-26/node2shell-cve-2026-111497-nodejs-brotli-rce/cover-e327e2b5.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Cambricon Q3 2026 Earnings: Profit Up 148%, Revenue Misses]]></title>
      <link>https://www.aighten.com/en/blog/cambricon-q3-2026-earnings-ai-chip-revenue-miss</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/cambricon-q3-2026-earnings-ai-chip-revenue-miss</guid>
      <description><![CDATA[Cambricon's Q3 2026 revenue doubled to RMB 3.53B and profit rose 148%, yet sales missed estimates by about 44%. Inventory, cash flow and supply limits.]]></description>
      <pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[finance]]></category>
      <category><![CDATA[Cambricon Q3 2026 earnings]]></category>
      <category><![CDATA[Cambricon revenue]]></category>
      <category><![CDATA[China AI chips]]></category>
      <category><![CDATA[Cambricon inventory]]></category>
      <category><![CDATA[HBM shortage China]]></category>
      <category><![CDATA[688256]]></category>
      <category><![CDATA[Nvidia China replacement]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-25/cambricon-q3-2026-earnings-ai-chip-revenue-miss/cover-e99bc6dc.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[OpenAI Moves Frontier Training Offline After Sandbox Escape]]></title>
      <link>https://www.aighten.com/en/blog/openai-frontier-training-offline-sandbox-escape</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/openai-frontier-training-offline-sandbox-escape</guid>
      <description><![CDATA[OpenAI now trains frontier models with no live internet, using a cached offline web. What the Sept. 20 DNS escape was and what red-team agents found.]]></description>
      <pubDate>Sun, 25 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[OpenAI offline training]]></category>
      <category><![CDATA[OpenAI frontier model training]]></category>
      <category><![CDATA[AI sandbox escape]]></category>
      <category><![CDATA[OpenAI DNS incident]]></category>
      <category><![CDATA[offline world]]></category>
      <category><![CDATA[AI containment]]></category>
      <category><![CDATA[red-teaming]]></category>
      <category><![CDATA[GPT-6 Astra Cyber]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-25/openai-frontier-training-offline-sandbox-escape/cover-5f110c80.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Meta Q3 2026 Earnings: Capex Turns Free Cash Flow Negative]]></title>
      <link>https://www.aighten.com/en/blog/meta-q3-2026-earnings-ai-capex-free-cash-flow</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/meta-q3-2026-earnings-ai-capex-free-cash-flow</guid>
      <description><![CDATA[Meta's Q3 2026 free cash flow was minus $2.12B as capex hit $34.91B. What drove it, the $135-145B 2026 capex outlook, the debt and off-balance-sheet deals.]]></description>
      <pubDate>Sat, 24 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[finance]]></category>
      <category><![CDATA[Meta Q3 2026 earnings]]></category>
      <category><![CDATA[Meta free cash flow]]></category>
      <category><![CDATA[Meta capex 2026]]></category>
      <category><![CDATA[AI capex]]></category>
      <category><![CDATA[Meta data centers]]></category>
      <category><![CDATA[Hyperion data center]]></category>
      <category><![CDATA[Meta 2027 spending]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-24/meta-q3-2026-earnings-ai-capex-free-cash-flow/cover-fdbeff9c.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Kimi K3.1 Pro: Benchmarks, Pricing and Open Weights Date]]></title>
      <link>https://www.aighten.com/en/blog/kimi-k3-1-pro-open-weight-model-benchmarks-pricing</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/kimi-k3-1-pro-open-weight-model-benchmarks-pricing</guid>
      <description><![CDATA[Kimi K3.1 Pro is Moonshot AI's 2.8T open-weight model: 65.7 on Terminal-Bench 4.0, $2/$10 per million tokens, weights by Nov 4. How it compares.]]></description>
      <pubDate>Sat, 24 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Kimi K3.1 Pro]]></category>
      <category><![CDATA[Kimi K3.1 Pro Highspeed]]></category>
      <category><![CDATA[Moonshot AI]]></category>
      <category><![CDATA[Kimi K3.1 Pro benchmarks]]></category>
      <category><![CDATA[Kimi K3.1 Pro pricing]]></category>
      <category><![CDATA[Kimi K3.1 open weights]]></category>
      <category><![CDATA[open-weight model]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-24/kimi-k3-1-pro-open-weight-model-benchmarks-pricing/cover-a7b6a1fc.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Recursive Flint: 40B Local Coding Model, Benchmarks]]></title>
      <link>https://www.aighten.com/en/blog/recursive-flint-local-coding-model-benchmarks</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/recursive-flint-local-coding-model-benchmarks</guid>
      <description><![CDATA[Recursive Flint is an Apache 2.0 coding model with 40B total and 2.5B active parameters. Benchmarks, hardware needs, the rec engine and how it compares to Qwen.]]></description>
      <pubDate>Fri, 23 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Recursive Flint]]></category>
      <category><![CDATA[Flint coding model]]></category>
      <category><![CDATA[local coding model]]></category>
      <category><![CDATA[Flint benchmarks]]></category>
      <category><![CDATA[rec inference engine]]></category>
      <category><![CDATA[Flint vs Qwen 3.8]]></category>
      <category><![CDATA[SWE-bench Pro v2]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-23/recursive-flint-local-coding-model-benchmarks/cover-88ffb809.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Claude Duet: Anthropic's Proactive AI Agent]]></title>
      <link>https://www.aighten.com/en/blog/anthropic-claude-duet-proactive-agent</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/anthropic-claude-duet-proactive-agent</guid>
      <description><![CDATA[Claude Duet is Anthropic's new background agent for Pro and Max users: what it does, who gets in, how usage limits work and how it compares to OpenAI dots.]]></description>
      <pubDate>Fri, 23 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Claude Duet]]></category>
      <category><![CDATA[Anthropic Claude Duet]]></category>
      <category><![CDATA[Claude Duet waitlist]]></category>
      <category><![CDATA[proactive AI agent]]></category>
      <category><![CDATA[Claude Opus 5.5]]></category>
      <category><![CDATA[Claude Duet vs OpenAI Dots]]></category>
      <category><![CDATA[Meta Muse for Small Business]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-23/anthropic-claude-duet-proactive-agent/cover-20acc7b2.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[PHP2Shell: AI-Found PHP RCE Flaws and How to Patch]]></title>
      <link>https://www.aighten.com/en/blog/php2shell-php-rce-ai-found-patch-the-planet</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/php2shell-php-rce-ai-found-patch-the-planet</guid>
      <description><![CDATA[PHP2Shell (CVE-2026-109618, CVE-2026-109619) allows unauthenticated RCE and is now exploited. Who found it with AI, which versions fix it, what to do now.]]></description>
      <pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[PHP2Shell]]></category>
      <category><![CDATA[CVE-2026-109618]]></category>
      <category><![CDATA[CVE-2026-109619]]></category>
      <category><![CDATA[PHP security release October 2026]]></category>
      <category><![CDATA[Patch the Planet]]></category>
      <category><![CDATA[AI vulnerability discovery]]></category>
      <category><![CDATA[mod_php RCE]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-22/php2shell-php-rce-ai-found-patch-the-planet/cover-07d4c113.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[LiteLLM 1.107.0 Compromise: Are You Affected?]]></title>
      <link>https://www.aighten.com/en/blog/litellm-1-107-0-pypi-supply-chain-attack</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/litellm-1-107-0-pypi-supply-chain-attack</guid>
      <description><![CDATA[LiteLLM 1.107.0 on PyPI and Docker shipped a credential stealer on October 21, 2026. What it stole, how to check your systems and which secrets to rotate.]]></description>
      <pubDate>Thu, 22 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[LiteLLM 1.107.0]]></category>
      <category><![CDATA[LiteLLM compromise]]></category>
      <category><![CDATA[LiteLLM supply chain attack]]></category>
      <category><![CDATA[PyPI malware]]></category>
      <category><![CDATA[cook.pth]]></category>
      <category><![CDATA[LLM gateway security]]></category>
      <category><![CDATA[credential rotation]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-22/litellm-1-107-0-pypi-supply-chain-attack/cover-2969403e.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Mistral Voxtral 3: Models, Pricing and Voice Agent API]]></title>
      <link>https://www.aighten.com/en/blog/mistral-voxtral-3-speech-models-voice-agent-api</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/mistral-voxtral-3-speech-models-voice-agent-api</guid>
      <description><![CDATA[Mistral Voxtral 3 adds speech-to-text, TTS and a voice agent API from $0.0035 a minute. See each model, prices against rivals, benchmarks and how to try it.]]></description>
      <pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Mistral Voxtral 3]]></category>
      <category><![CDATA[Voxtral 3 pricing]]></category>
      <category><![CDATA[Voxtral Small Transcribe 3]]></category>
      <category><![CDATA[Voxtral TTS 3]]></category>
      <category><![CDATA[Mistral voice agent SDK]]></category>
      <category><![CDATA[speech to text API]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-21/mistral-voxtral-3-speech-models-voice-agent-api/cover-d6042883.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[DeepSeek DeepInfer Inference Engine: Features and Limits]]></title>
      <link>https://www.aighten.com/en/blog/deepseek-deepinfer-open-source-inference-engine</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/deepseek-deepinfer-open-source-inference-engine</guid>
      <description><![CDATA[DeepSeek open-sourced DeepInfer, the engine behind its API. What it includes, what it leaves out, the hardware it needs (SM100, Ascend 950) and how to try it.]]></description>
      <pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[DeepSeek DeepInfer]]></category>
      <category><![CDATA[DeepInfer]]></category>
      <category><![CDATA[DeepSeek inference engine]]></category>
      <category><![CDATA[open-source LLM inference framework]]></category>
      <category><![CDATA[Huawei Ascend 950]]></category>
      <category><![CDATA[DeepSeek-V4.1-Flash]]></category>
      <category><![CDATA[DeepInfer vs vLLM]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-21/deepseek-deepinfer-open-source-inference-engine/cover-04121896.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[MiniMax M3.1 Flash: Specs, Pricing and Benchmarks]]></title>
      <link>https://www.aighten.com/en/blog/minimax-m3-1-flash-specs-pricing-benchmarks</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/minimax-m3-1-flash-specs-pricing-benchmarks</guid>
      <description><![CDATA[MiniMax M3.1 Flash is a 428B MoE agent model at $0.15/M input tokens. See the specs, benchmarks, cost per task and how it compares with DeepSeek and Haiku.]]></description>
      <pubDate>Tue, 20 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[MiniMax M3.1 Flash]]></category>
      <category><![CDATA[MiniMax M3.1 Flash pricing]]></category>
      <category><![CDATA[MiniMax M3.1 Flash benchmarks]]></category>
      <category><![CDATA[Space Bunny model]]></category>
      <category><![CDATA[cheap AI agent model]]></category>
      <category><![CDATA[MiniMax API]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-20/minimax-m3-1-flash-specs-pricing-benchmarks/cover-b951afad.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[China AI Export Controls Draft: Models, Agents and Chips]]></title>
      <link>https://www.aighten.com/en/blog/china-ai-export-controls-draft-models-agents-chips</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/china-ai-export-controls-draft-models-agents-chips</guid>
      <description><![CDATA[China's draft export controls cover AI model weights, agents, chip design and 10,000-card clusters. What is banned, what is restricted, and why APIs are exempt.]]></description>
      <pubDate>Mon, 19 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[China AI export controls]]></category>
      <category><![CDATA[China export control catalogue AI]]></category>
      <category><![CDATA[MOFCOM AI model weights]]></category>
      <category><![CDATA[China AI agents export]]></category>
      <category><![CDATA[2 trillion parameters]]></category>
      <category><![CDATA[AI Diffusion Rule comparison]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-19/china-ai-export-controls-draft-models-agents-chips/cover-3e91e50c.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Anthropic Public Benefit Fund: What the 5% Pledge Means]]></title>
      <link>https://www.aighten.com/en/blog/anthropic-public-benefit-fund-5-percent-shares-explained</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/anthropic-public-benefit-fund-5-percent-shares-explained</guid>
      <description><![CDATA[Anthropic is moving 5% of its shares into a Public Benefit Fund that can grow with its valuation. How the LLC works, what it may fund, and how it compares.]]></description>
      <pubDate>Mon, 19 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[finance]]></category>
      <category><![CDATA[Anthropic Public Benefit Fund]]></category>
      <category><![CDATA[Anthropic 5% shares]]></category>
      <category><![CDATA[Anthropic IPO]]></category>
      <category><![CDATA[public benefit corporation]]></category>
      <category><![CDATA[OpenAI Foundation stake]]></category>
      <category><![CDATA[Long-Term Benefit Trust]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-19/anthropic-public-benefit-fund-5-percent-shares-explained/cover-efd4dd2c.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[DeepSeek V4.1-Flash-1017: Benchmarks, Pricing, Architecture]]></title>
      <link>https://www.aighten.com/en/blog/deepseek-v4-1-flash-1017-benchmarks-pricing</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/deepseek-v4-1-flash-1017-benchmarks-pricing</guid>
      <description><![CDATA[DeepSeek V4.1-Flash-1017 is an MIT-licensed update with big agentic coding gains. See the benchmarks, API prices, Search API and how the 1M-context MoE works.]]></description>
      <pubDate>Sun, 18 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[DeepSeek V4.1-Flash-1017]]></category>
      <category><![CDATA[DeepSeek V4.1 Flash]]></category>
      <category><![CDATA[deepseek-flash API pricing]]></category>
      <category><![CDATA[DeepSeek benchmarks]]></category>
      <category><![CDATA[open-weight AI model]]></category>
      <category><![CDATA[DeepSeek Search API]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-18/deepseek-v4-1-flash-1017-benchmarks-pricing/cover-c49809f7.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Gemini 4 Argon Benchmarks: Where It Leads and Where It Loses]]></title>
      <link>https://www.aighten.com/en/blog/gemini-4-argon-benchmarks-explained</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/gemini-4-argon-benchmarks-explained</guid>
      <description><![CDATA[Gemini 4 Argon leads 13 of 19 rows in Google's benchmark table but trails on Terminal-bench. Third-party scores, the thinking-mode caveat, pricing and access.]]></description>
      <pubDate>Sat, 17 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Gemini 4 Argon]]></category>
      <category><![CDATA[Gemini 4 Argon benchmarks]]></category>
      <category><![CDATA[Gemini 4 Argon vs GPT-6 Astra]]></category>
      <category><![CDATA[Gemini 4 Argon vs Claude Opus 5.5]]></category>
      <category><![CDATA[Gemini 4 Argon pricing]]></category>
      <category><![CDATA[Artificial Analysis Intelligence Index]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-17/gemini-4-argon-benchmarks-explained/cover-18a23556.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Meta Muse Spark Open: Specs, License and Mac Studio Setup]]></title>
      <link>https://www.aighten.com/en/blog/meta-muse-spark-open-weights-mac-studio</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/meta-muse-spark-open-weights-mac-studio</guid>
      <description><![CDATA[Meta released Muse Spark Open, an 816B open-weight MoE model with a 1M-token context. Specs, benchmarks, license terms and how it runs on two Mac Studios.]]></description>
      <pubDate>Sat, 17 Oct 2026 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Muse Spark Open]]></category>
      <category><![CDATA[Meta open-weight model]]></category>
      <category><![CDATA[Muse Model License]]></category>
      <category><![CDATA[MLX Mac Studio]]></category>
      <category><![CDATA[816B MoE]]></category>
      <category><![CDATA[open-weight LLM]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2026-10-17/meta-muse-spark-open-weights-mac-studio/cover-a0e06a60.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Skild AI – An Analysis of Physical AI and the Future of Robotic Intelligence]]></title>
      <link>https://www.aighten.com/en/blog/6953787e801ad787bec0727b</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/6953787e801ad787bec0727b</guid>
      <description><![CDATA[An in-depth analysis of Skild AI’s core technologies and business strategy. This article explores how the general-purpose robotic brain, “Skild Brain,” powered by physical interaction intelligence, is set to revolutionize robotics and accelerate the arrival of the AGI era. ]]></description>
      <pubDate>Tue, 30 Dec 2025 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Skild AI]]></category>
      <category><![CDATA[Physical AI]]></category>
      <category><![CDATA[robot]]></category>
      <category><![CDATA[AI]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-30/ai-skild-ai-ai-physical-ai/1767078014159-tg8cqq.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Meta's Strategic Leap: Unpacking the Acquisition of Intelligent Agent Firm Manus]]></title>
      <link>https://www.aighten.com/en/blog/695368dc801ad787bec071fc</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/695368dc801ad787bec071fc</guid>
      <description><![CDATA[Meta concludes a year of aggressive AI investments by acquiring Manus, a Singapore-based intelligent agent firm. This move signals a significant acceleration in Meta's ambition to integrate sophisticated AI agents across its ecosystem, from social platforms to the Metaverse, and intensify its competition in the global AI race.]]></description>
      <pubDate>Tue, 30 Dec 2025 05:53:32 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Meta AI]]></category>
      <category><![CDATA[Intelligent Agents]]></category>
      <category><![CDATA[AI Acquisitions]]></category>
      <category><![CDATA[Manus]]></category>
      <category><![CDATA[Generative AI]]></category>
      <category><![CDATA[Metaverse AI]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-30/meta-acquires-intelligent-agent-manus/1767074009819-o2p98t.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Cursor Forges Ahead: Why the Graphite Acquisition is a Game-Changer for AI-Powered Code Review]]></title>
      <link>https://www.aighten.com/en/blog/694cfb95902872a62113dcbe</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/694cfb95902872a62113dcbe</guid>
      <description><![CDATA[Explore the strategic acquisition of code review innovator Graphite by AI coding platform Cursor. This deep dive analyzes how the merger aims to revolutionize developer workflows, streamline code delivery, and intensify the competition in the rapidly evolving AI-driven software development landscape.]]></description>
      <pubDate>Thu, 25 Dec 2025 08:53:40 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Cursor acquisition]]></category>
      <category><![CDATA[Graphite code review]]></category>
      <category><![CDATA[AI coding]]></category>
      <category><![CDATA[developer tools]]></category>
      <category><![CDATA[software development workflow]]></category>
      <category><![CDATA[stacked diffs]]></category>
      <category><![CDATA[tech mergers]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-25/cursor-graphite-acquired/1766652819352-j7c2ze.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Gemini 3 Flash: Unleashing Frontier AI with Unprecedented Speed and Efficiency]]></title>
      <link>https://www.aighten.com/en/blog/694403d5245dc6249389c317</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/694403d5245dc6249389c317</guid>
      <description><![CDATA[Explore Google's Gemini 3 Flash, a groundbreaking AI model engineered for lightning-fast performance and cost-efficiency. This deep dive uncovers its capabilities, from a 1-million-token context window to advanced developer features, revolutionizing real-time AI applications and enterprise solutions.]]></description>
      <pubDate>Thu, 18 Dec 2025 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Gemini 3 Flash]]></category>
      <category><![CDATA[Google AI]]></category>
      <category><![CDATA[AI performance]]></category>
      <category><![CDATA[Large Language Models]]></category>
      <category><![CDATA[LLM speed]]></category>
      <category><![CDATA[frontier intelligence]]></category>
      <category><![CDATA[enterprise AI]]></category>
      <category><![CDATA[developer tools]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-18/gemini-3-flash/1766065105304-e7nws3.jpg" length="0" type="image/jpeg" />
    </item>
    <item>
      <title><![CDATA[Liquid AI Unveils LFM2 Blueprint: Revolutionizing Enterprise Small Model Training for On-Device AI]]></title>
      <link>https://www.aighten.com/en/blog/liquid-ai-unveils-lfm2-blueprint-revolutionizing-enterprise-small-model-training-for-on-device-ai</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/liquid-ai-unveils-lfm2-blueprint-revolutionizing-enterprise-small-model-training-for-on-device-ai</guid>
      <description><![CDATA[MIT offshoot Liquid AI has released a groundbreaking blueprint for enterprise-grade small model training, dubbed LFM2. This initiative promises to democratize advanced AI capabilities, offering businesses a path to deploy efficient, robust, and privacy-preserving AI directly on edge devices, overcoming the limitations of large language models.]]></description>
      <pubDate>Thu, 18 Dec 2025 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <enclosure url="https://images.ctfassets.net/jdtwqhzvc2n1/3hcEEwXVdJKS26YG86bl1G/70c2a214da37d17bae5a5600a5400225/1ouQYIrPZZ4x2Dcn0fD9c.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Mistral AI's Devstral 2 and Vibe CLI: Ushering in a New Era of Intuitive AI-Powered Coding]]></title>
      <link>https://www.aighten.com/en/blog/mistral-ai-s-devstral-2-and-vibe-cli-ushering-in-a-new-era-of-intuitive-ai-powered-coding</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/mistral-ai-s-devstral-2-and-vibe-cli-ushering-in-a-new-era-of-intuitive-ai-powered-coding</guid>
      <description><![CDATA[Mistral AI has launched Devstral 2, a powerful new coding model, alongside the Mistral Vibe CLI, an innovative command-line interface designed for 'vibe coding.' This release marks a significant stride in AI-assisted software development, offering enhanced code generation, a laptop-friendly open-source version, and a more natural, conversational developer experience.]]></description>
      <pubDate>Thu, 18 Dec 2025 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-10/mistral-releases-devstral2-and-mistral-vibe-cli/1765358587761-2qba5n.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Unleashing the Next Frontier: GPT-5.2 Now Accessible via OpenAI API]]></title>
      <link>https://www.aighten.com/en/blog/unleashing-the-next-frontier-gpt-5-2-now-accessible-via-openai-api</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/unleashing-the-next-frontier-gpt-5-2-now-accessible-via-openai-api</guid>
      <description><![CDATA[Explore the profound implications of GPT-5.2's API release for developers, businesses, and the future of AI. Dive into its expected enhancements, new capabilities, and how it's set to revolutionize application development.]]></description>
      <pubDate>Thu, 18 Dec 2025 00:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-12/gpt-5-2-is-now-available-in-the-api/1765531499492-1nfn7q.webp" length="0" type="image/webp" />
    </item>
    <item>
      <title><![CDATA[DeepSeek AI Models: China's Next-Gen Powerhouse Challenging Global AI Dominance]]></title>
      <link>https://www.aighten.com/en/blog/deepseek-ai-models</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/deepseek-ai-models</guid>
      <description><![CDATA[Explore DeepSeek AI's latest innovations, including the DeepSeek-V3.2 model, its groundbreaking performance at the International Mathematical Olympiad, and how these Chinese AI models are reshaping the global tech landscape, offering a compelling comparison to models like GPT-5.]]></description>
      <pubDate>Tue, 02 Dec 2025 08:00:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[DeepSeek AI models]]></category>
      <category><![CDATA[DeepSeek-V3.2 capabilities]]></category>
      <category><![CDATA[DeepSeek vs GPT-5 comparison]]></category>
      <category><![CDATA[Chinese AI models challenging US tech]]></category>
      <category><![CDATA[DeepSeek-V3.2-Speciale performance]]></category>
      <category><![CDATA[AI in mathematics]]></category>
      <category><![CDATA[large language models]]></category>
      <category><![CDATA[generative AI]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-02/deepseek-ai-models/1764638941048-cczbe9.webp" length="0" type="image/webp" />
    </item>
    <item>
      <title><![CDATA[Apple's AI Reimagined: Decoding the Leadership Shift and Siri's Future]]></title>
      <link>https://www.aighten.com/en/blog/apple-ai-leadership-change</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/apple-ai-leadership-change</guid>
      <description><![CDATA[Apple's top AI executive, John Giannandrea, has stepped down amidst ongoing challenges with Siri. This deep dive analyzes the implications of this leadership change for Apple's AI strategy, its product ecosystem, and the future of its intelligent assistant, examining the path forward under new direction.]]></description>
      <pubDate>Mon, 01 Dec 2025 23:34:09 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Apple AI leadership change]]></category>
      <category><![CDATA[John Giannandrea leaving Apple reason]]></category>
      <category><![CDATA[Apple Siri AI problems 2025]]></category>
      <category><![CDATA[Who is Apple's head of AI now]]></category>
      <category><![CDATA[Impact of leadership change on Apple AI]]></category>
      <category><![CDATA[Amar Subramanya Apple VP AI]]></category>
      <category><![CDATA[Apple AI strategy]]></category>
      <category><![CDATA[Siri development]]></category>
      <category><![CDATA[Generative AI]]></category>
      <enclosure url="https://www.aighten.com/_next/image?url=https%3A%2F%2Fpub-47c6f4791d57423fadd704dff93359fc.r2.dev%2Fcontent-images%2F2025-12-02%2Fapple-ai-leadership-change%2Fapple-siri.webp&amp;w=1920&amp;q=75" length="0" type="image/jpeg" />
    </item>
    <item>
      <title><![CDATA[NVIDIA Nemotron 3: The New Standard for Open-Source Agentic AI]]></title>
      <link>https://www.aighten.com/en/blog/nvidia-nemotron-3-ai</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/nvidia-nemotron-3-ai</guid>
      <description><![CDATA[Discover NVIDIA's latest breakthrough with Nemotron 3, a family of open-source models specifically optimized for Agentic AI. Learn how its advanced reasoning, tool-use capabilities, and SteerLM alignment are revolutionizing enterprise automation and AI development.]]></description>
      <pubDate>Thu, 18 Dec 2025 14:30:00 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[NVIDIA Nemotron 3]]></category>
      <category><![CDATA[Agentic AI]]></category>
      <category><![CDATA[Open Source LLM]]></category>
      <category><![CDATA[SteerLM]]></category>
      <category><![CDATA[AI Agents]]></category>
      <category><![CDATA[Generative AI]]></category>
      <category><![CDATA[NeMo Framework]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-18/nemotron-cover.png" length="0" type="image/png" />
    </item>
    <item>
      <title><![CDATA[Skild AI Soars to $14 Billion Valuation with Renewed Backing from SoftBank and Nvidia]]></title>
      <link>https://www.aighten.com/en/blog/skild-ai-soars-to-14-billion-valuation</link>
      <guid isPermaLink="true">https://www.aighten.com/en/blog/skild-ai-soars-to-14-billion-valuation</guid>
      <description><![CDATA[Discover how Skild AI, a pioneering robot software startup, achieved a staggering $14 billion valuation, attracting significant investments from tech giants like SoftBank, Nvidia, and Samsung, and what this means for the future of intelligent robotics.]]></description>
      <pubDate>Thu, 11 Dec 2025 12:36:57 GMT</pubDate>
      <category><![CDATA[tech]]></category>
      <category><![CDATA[Skild AI]]></category>
      <category><![CDATA[Robotics Investment]]></category>
      <category><![CDATA[AI Software]]></category>
      <category><![CDATA[SoftBank]]></category>
      <category><![CDATA[Nvidia]]></category>
      <category><![CDATA[Samsung Investment]]></category>
      <category><![CDATA[Tech Valuation]]></category>
      <category><![CDATA[Robotics Trends]]></category>
      <enclosure url="https://pub-47c6f4791d57423fadd704dff93359fc.r2.dev/content-images/2025-12-11/skild-ai-rockets-to-14-billion/1765456613263-owuovd.webp" length="0" type="image/webp" />
    </item>
  </channel>
</rss>