Најди сличен содржај

▎大模型入门指南 一篇关于大语言模型（Large Language Model）的入门指南。文章介绍了大模型的发展历程，包括 BERT、GPT 和 LLAMA 等模型。文章还讨论了大模型的部署和量化技术，以及如何在 macOS 上进行大模型的实验。最后，文章介绍了一些免费的大模型产品，供读者参考。 标签：#AI 链接：https://liujiacai.net/blog/2023/08/12/llm-quickstart/ 频道：@yi1de

Национальный Банк опубликовал доклад об искусственном интеллекте на финансовом рынке. https://nationalbank.kz/ru/news/informacionnye-soobshcheniya/16693?fbclid=IwZXh0bgNhZW0CMTEAAR1qummpUMPcqczW79loRr727fK_yzweY9dRhklP_gMm50dGue7iQ8w2pys_aem_AZPGTGv5kA5PiCM89XVeX8V4DzeDQHj-tEdB0LeOen6B2pjfQUX91j7W0SE605t8FeXQL5TAWSo7ffxbHrR4Ljzm #ai

AI Protections Secured for Animation Voice Actors Voice actors in the realm of animated shows have achieved a significant milestone in their ongoing quest for recognition and fair compensation, as a deal has been struck to safeguard their rights in the age of artificial intelligence. Members of the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA) voted in favor of a three-year contract with the Alliance of Motion Picture and Television Producers (AMPTP). Under the terms of the agreement, voice actors are assured compensation if their performances are utilized to create AI-generated clones for recognizable characters. This monumental provision ensures that the invaluable contributions of voice actors are duly acknowledged and remunerated. The agreement also mandates producers to engage in regular consultations with the union to explore mechanisms for monitoring and tracking the utilization of digital replicas. The agreement is effective retroactively from July 1, 2023, to June 30, 2026. #AI

Michael Cohen's Court Motion Denied After AI Citation Blunder A federal judge has decided not to sanction Michael Cohen and his lawyer, David M. Schwartz, for including three fake citations generated by Google Bard AI in a court filing. While Cohen avoids punishment, the judge rejects his motion to end supervised release. Cohen, Trump's former attorney, filed a motion citing fictional cases in an attempt to terminate his supervised release early. The citations, however, backfired as they involved non-existent defendants and undermined Cohen's credibility. Despite the citation blunder, the judge declines to sanction Cohen, noting his reliance on legal counsel. However, the judge deems Schwartz's actions as embarrassing and negligent but stops short of imposing sanctions. This incident mirrors a similar AI citation blunder last year, resulting in fines and reprimands for the lawyers involved. In Cohen's case, the judge emphasizes his contradictory testimony, which undermines his motion for early release. #AI

#ai DeepSeek 🐂 之前基于claude 3.5 调试了很久的超长文本处理 prompt 打算部署之后面向熊言熊语的会员读者提供。 考虑到claude 3.5 的价格，打算找一个可基本平替的国内模型。 试了一圈下来，不改 prompt 的前提下，昨天刚刚发布的 DeepSeek V2.5 给出了基本匹配 claude 3.5 的一致输出效果。以及，DeepSeek 本来亲民的价格，再加上不需要任何设置的cache hit，真的是太强了！

#AI 不对劲，关掉声音

#AI 现在的ai啊……唉不说了，提醒我了

🔵 Что-то вас генерация статичного сайта из мокапа не сильно впечатлила.)) А как вам такое молодые Илоны Маски - генерация скриптов/игр/форм/чтопридумаетееще. Не нужно писать тз и объяснять разработчикам, что вы от них хотите, просто пришлите им готовый код. И да - это тооолько самое начало и вы тому свидетели) Как мы знаем возможности #AI ограничены только вашим воображением 😘 💙 Перешли своему фронтенд разработчику, пусть задумается)

#Ai

#ai 气笑了

#ai

US & UK Drive Global AI Security Standards Hello, everyone! The UK's National Cyber Security Agency and the US's Cybersecurity and Infrastructure Security Agency have jointly unveiled groundbreaking AI security guidelines. This initiative, endorsed by 17 nations, focuses on a "secure-by-design" approach, ensuring cybersecurity is integrated into every stage of AI development. Australia, Germany, Japan, and others join forces to establish global AI security standards. Core Focus Areas: Secure Design, Development, Deployment, and Operation & Maintenance. #AI