<?xml version="1.0" encoding="UTF-8"?>
<!-- Neural Brew (it) : articles publiés depuis le 2026-11-01T08:16:19+01:00 ; build 2026-11-03T08:16:19+01:00 -->
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9" xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/hume-audio-decision-api-domande-audio-probabilita/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-03T08:11:00+01:00</news:publication_date>
    <news:title>Hume lancia la sua Audio Decision API, che risponde con probabilità a domande su una registrazione</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/itch-io-dichiarazione-ia-generativa-filtro-ricerca-rimozione-contenuti/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-03T08:11:00+01:00</news:publication_date>
    <news:title>Itch.io estende la dichiarazione sull'IA generativa ed elimina terabyte di contenuti «slop»</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/openai-pando-backend-git-monorepo-agenti/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-03T08:11:00+01:00</news:publication_date>
    <news:title>Pando: OpenAI ha ricostruito il server Git del suo monorepo per reggere il ritmo dei suoi agenti</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/openai-sprint-sicurezza-cathedral-bug-bounty-premi/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-03T08:11:00+01:00</news:publication_date>
    <news:title>OpenAI prepara un nuovo sprint di sicurezza con Cathedral e quadruplica i premi del suo bug bounty</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/anthropic-flashprove-prova-crittografica-inferenza-open-source/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-03T08:10:00+01:00</news:publication_date>
    <news:title>Anthropic pubblica FlashProve, un prototipo open source per dimostrare quale modello ha prodotto una risposta</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/anthropic-s1-emendato-soglia-accelerazione-ia-rsp-terzo-trimestre/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T20:08:00+01:00</news:publication_date>
    <news:title>Anthropic scrive nel prospetto per la quotazione che potrebbe raggiungere «in modo imminente» la soglia di accelerazione dell'IA tramite l'IA</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/openai-lifeval-test-attivita-quotidiane-assistente-personale/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T19:04:00+01:00</news:publication_date>
    <news:title>OpenAI presenta LifeVal, il suo test per gli agenti IA sulle attività quotidiane</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/deep-principle-hydrogen-1-0-modello-scientifico-pesi-aperti/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T17:10:00+01:00</news:publication_date>
    <news:title>Deep Principle pubblica Hydrogen-1.0, un modello scientifico a pesi aperti derivato da GLM-5.3</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/meituan-longcat-2-5-modello-agenti-multimodale/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T14:11:00+01:00</news:publication_date>
    <news:title>Meituan lancia LongCat-2.5, un modello multimodale da 1.600 miliardi di parametri per agenti di codice</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/tencent-weviscut-agente-montaggio-video-open-source/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T07:54:00+01:00</news:publication_date>
    <news:title>WeVisCut: Tencent pubblica in accesso libero un agente di montaggio video da 9 miliardi di parametri</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/semianalysis-gdpx-benchmark-agenti-claude-opus-dodici-percento/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T07:52:00+01:00</news:publication_date>
    <news:title>GDPx, il nuovo test di SemiAnalysis: persino Claude Opus 5.5, primo, si ferma al 12%</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/abu-dhabi-openai-quattro-accordi-adnoc-mubadala/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-02T07:50:00+01:00</news:publication_date>
    <news:title>Abu Dhabi sigla quattro accordi con OpenAI: ADNOC, Mubadala e un'amministrazione dalle regole «leggibili dalle macchine»</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/xai-grok-api-watermark-testo-novembre/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-01T16:13:00+01:00</news:publication_date>
    <news:title>L'API di Grok applicherà di default un watermark ai testi generati dal 7 novembre, con disattivazione possibile per modello</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/anthropic-watermark-testo-claude-vecchi-modelli/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-01T11:11:00+01:00</news:publication_date>
    <news:title>Anthropic estende il watermark invisibile dei testi di Claude a quattro vecchi modelli dal 3 novembre, secondo un'e-mail ai clienti</news:title>
   </news:news>
  </url>
  <url>
   <loc>https://neuralbrew.ai/it/articoli/2026/11/deepseek-dsh-superagent-agente-codice-proattivo-open-source/</loc>
   <news:news>
    <news:publication>
     <news:name>Neural Brew</news:name>
     <news:language>it</news:language>
    </news:publication>
    <news:publication_date>2026-11-01T10:10:00+01:00</news:publication_date>
    <news:title>DeepSeek pubblica dsh-superagent, un prototipo open source di agente di codice proattivo per GitHub</news:title>
   </news:news>
  </url>
</urlset>
