OpenAI Upgrades AI Agent Development Tools to Support TypeScript and Improves Voice Interaction

Defeat GPT-4o! Phonely Voice AI Breaks 99.2% Accuracy, Layoffs of 350 Customer Service Workers

A key technical hurdle in the AI voice customer service field—delay and unnatural conversation experience—has finally been overcome. The phone AI system with human-like real-time response capabilities was jointly developed by Phonetic AI company Phonely, inference optimization platform Maitai, and chip manufacturer Groq. This advancement propels voice AI into a new era of human simulation. According to disclosures, Phonely’s response speed increased by over 70%, and the AI dialogue accuracy rate rose from 81.5% to

Stanford's Latest Evaluation: DeepSeek R1 Medical AI Model Outperforms Google and OpenAI with High Scores

Recently, Stanford University released a comprehensive evaluation of clinical medical AI models. DeepSeek R1 stood out as the champion among nine leading large models, achieving a 66% win rate and a macro average score of 0.75. The highlight of this evaluation is that it not only focuses on traditional medical license exam questions but also delves into the daily work scenarios of clinical doctors, providing more practical assessments. The evaluation team developed an integrated assessment framework called MedHELM, which includes 35 benchmarks covering 22 subtasks in medicine.

Yuánbǎo Integrates with Tencent Maps: Mobile and Web Versions Officially Launched

Recently, Yuánbǎo has officially integrated with Tencent Maps, bringing users more accurate and convenient location search and navigation services. Leveraging the powerful features of Tencent Maps, Yuánbǎo performs exceptionally well in terms of location searches, meeting diverse user needs. Now, when a user asks questions like "Where can I find shoe repair nearby?" or "What are some mid-range Cantonese restaurants nearby?", Yuánbǎo can provide quick and accurate recommendations by combining map information. Moreover, Yuánbǎo's responses will automatically display map cards, allowing users to initiate navigation with just one click, significantly enhancing the user experience.

Tencent Coins Officially Integrate with DeepSeek R1-0528 Latest Version

Tencent Coins officially announced the integration with the latest version of DeepSeek R1-0528, providing early access on both desktop and web versions. This model update brings three significant changes, offering users a more premium and efficient AI experience. In terms of reasoning ability, the new model demonstrates stronger capabilities. It can handle complex code issues and conduct in-depth analyses of various problems quickly and accurately, providing reliable solutions for users. This improvement makes Tencent Coins more practical in work and learning scenarios.

AI Daily: DeepSeek releases new version R1-0528; ByteDance releases image Agent Xiaoyunque AI; Keeling 2.1 is officially launched

DeepSeek R1 Model Release 0528 Version: Comprehensive Improvement in All Aspects

DeepSeek officially announced that its R1 model has completed minor version trial upgrades and is now open for user testing. This upgraded model is named "DeepSeek-R1-0528", showing amazing performance in various tests. The new DeepSeek-R1-0528 has significantly improved programming capabilities, aesthetic design, and code completion. During testing, with just simple instructions, the model can quickly generate high-quality code and design works. In addition, R1-0528 has excelled in generating complex pre-

OpenAI Sora's Second Anniversary: Classic Animated Short Film 'Critterz' Gets AI Remake

Recently, OpenAI celebrated the one-year anniversary of the release of its text-to-video model, Sora. To commemorate this milestone, OpenAI has launched a remastered version of the animated short film 'Critterz'. This short film uniquely showcases Sora's astonishing video generation capabilities. It presents a side-by-side comparison, with the original 'Critterz' on the left and a brand new version entirely generated by Sora on the right. Viewers can intuitively feel Sora's powerful strength and realistic effects in video generation.

AI Expert Tan Xu Joins Dark Side of the Moon, Responsible for Developing End-to-End Speech Models

According to media reports, Tan Xu, former Chief Research Manager at Microsoft Research Asia, joined the startup Dark Side of the Moon in August, primarily responsible for developing end-to-end speech models. One important goal after Tan's joining is to create a speech experience similar to GPT-4o for the company. Dark Side of the Moon has not yet responded, but insiders have confirmed Tan's employment, and specific position details have not been finalized. Public information indicates that Tan Xu's research areas include deep learning, language/speech/music processing, and AI content generation, with publications in AI academic conferences.

Gemini Fully Opens Imagen 3 Image Generation Feature, but Free Users Cannot Generate Portraits

Google Gemini recently announced on its official Twitter account that the highly anticipated Imagen 3 image generation feature is now available to all Gemini users worldwide. This major update marks an important step for Google in the field of AI image generation, providing users with more powerful and flexible creation tools. Imagen 3, as the latest version in the Imagen series, brings several significant improvements compared to its predecessor, Imagen 2. According to Google's official introduction, these improvements mainly focus on...

Small but Mighty! A 10-Person Team Develops the First Fine-Tuned Llama 3.1 405B

A start-up company with only 10 people, Nous Research, has challenged tech giant Meta with its AI model called Hermes3. Fine-tuned from the Llama3.1 model, Hermes3 boasts powerful performance, optimizing VRAM and disk requirements, capable of running on a single node. The model excels in dialogue capabilities, text understanding quality, and integrated agent capabilities, showcasing advanced abilities that transcend traditional language modeling. Nous Research focuses on open source.