Install
RedditVote46FlipShareTweet46 SharesRedditVote46FlipShareTweet46 Shares
- 1,174articles · 365d
- 6+ hour agolatest article
- Sep 14, 2025earliest in window
- 97%with images · 26 videos
- 339avg words
- Science & Technology 1,152
- Software Dev. 1,001
- Computers & Electronics 898
- Science & Nature 146
- Jobs & Education 100
- STEM 82
- News 34
- Business & Industrial 17
Please confirm you are human
This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.
A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.
News
IBM Releases Granite 4.2: Bringing Native Reasoning and Agentic RL to Open Enterprise Models
2+ week, 5+ day ago (303+ words) Yes, All three Granite 4.2 language models ship under Apache 2.0, so download, fine-tuning, and commercial production use carry no licensing gate. Granite 4.2 is a decoder-only dense transformer, not a hybrid or MoE design. Core components are Grouped Query Attention with 8 KV…...
Meet S1-mini: Superwhisper's 462 MB Open-Weights Text Normalizer That Turns Raw ASR Transcripts Into Clean Written Text
3+ week, 3+ day ago (323+ words) Yes, but only S1-mini. S1-mini is published on Hugging Face under Apache 2.0 plus a naming clause. S1-Voice and S1-Language are Superwhisper-hosted services, so they are consumable, not self-hostable. S1-mini is a text normalizer, not a transcriber and not a chat…...
Tencent Open-Sources AngelSpec: A Unified Training Framework for MTP and Block-Parallel Speculative Decoding on Hy3 Models
1+ mon, 2+ week ago (445+ words) Most speculative-decoding work searches for one drafter that scores well on an averaged benchmark mixture. Real serving traffic does not look like that mixture. AngelSpec treats workload heterogeneity as a first-class design constraint, and specializes structure, training data, and verification…...
Poolside Releases Laguna S 2.1, an Open-Weight Agentic Coding Model Punching Above Its Weight Class on SWE-Bench Multilingual
1+ mon, 3+ week ago (455+ words) Poolside has released Laguna S 2.1, a 118B-parameter open-weight model built for agentic coding. It is a Mixture-of-Experts (MoE) model with 8B activated parameters per token. It supports a context window of up to 1M tokens in both thinking and no-thinking modes. The…...
Alibaba's Tongyi Lab Releases Qwen-Audio-3.0-TTS, a Hosted Text-to-Speech Model in Flash and Plus Tiers Across 16 Languages
1+ mon, 3+ week ago (519+ words) Alibaba’s Tongyi Lab has released Qwen-Audio-3.0-TTS, a production-oriented text-to-speech (TTS) system. The model ships in two variants from the same lineage. Flash targets real-time interaction. Plus targets high-quality generation. Both are delivered as hosted models through Alibaba Cloud Model…...
Google Cloud's Always-On Memory Agent Replaces RAG and Embeddings With Continuous LLM Consolidation on Gemini 3.1 Flash-Lite
1+ mon, 3+ week ago (323+ words) Most AI agents forget. They process a request, answer it, then drop the context. Google Cloud’s generative-ai repository now ships a sample that tackles this directly. It is the Always-On Memory Agent, a reference implementation that treats memory as a…...
Moonshot AI Releases Kimi K3: A 2.8 Trillion Parameter Open MoE Model With Kimi Delta Attention and 1M Context
1+ mon, 4+ week ago (492+ words) Moonshot AI just released Kimi K3. It is a 2.8-trillion-parameter model with native vision and a 1-million-token context window. Moonshot calls it the world’s first open 3T-class model. Kimi K3 is a sparse Mixture-of-Experts (MoE) model built on two architectural updates. Those…...
OpenAI Details GPT-Red: An Internal Automated Red-Teaming Model That Beat Human Red-Teamers 84% To 13% On Prompt Injection
1+ mon, 4+ week ago (650+ words) This week, OpenAI published details of GPT-Red, an internal-only automated red-teaming model. Its job is to attack OpenAI’s own models and find prompt injection vulnerabilities. OpenAI gives two reasons. Human red-teaming is time-intensive and does not scale. Commonly used robustness…...
Soofi Consortium Releases Soofi S 30B-A3B: An Open Hybrid Mamba-Transformer MoE Foundation Model For German And English
1+ mon, 4+ week ago (342+ words) A German research consortium has published the pretraining report for Soofi S 30B-A3B. It is an open base model for German and English. Training ran end to end on Deutsche Telekom’s Industrial AI Cloud in Munich. Preview weights are on Hugging…...
Meituan Releases LongCat-2.0: A 1.6T-Parameter Open MoE Model with Native 1M Context and LongCat Sparse Attention
2+ mon, 1+ week ago (383+ words) Meituan has released LongCat-2.0, a large-scale Mixture-of-Experts (MoE) language model. It carries 1.6 trillion total parameters and activates about 48 billion per token. The model targets agentic coding: code understanding, generation, and execution inside agent workflows. Two facts stand out. First, LongCat…...