#speculative decoding
165 related articles
·3 min
Laguna S 2.1: The Open-Source Coding Model That's 14x Smaller Yet Stronger
Poolside releases Laguna S 2.1, an open-source agentic coding model: 118B total params with only 8B active, scoring 40.4% on DeepSWE — ~4.5x DeepSeek V4 Pro Max. Supports 1M-token context, deployable on a single workstation.
Read more →
·5 min
Why Are China's Open-Source AI Models Getting Stronger? A Deep Dive Into Three Comeback Strategies
Facing US chip bans and closed-source monopoly, how do China's open-source AI models keep fighting back? A deep dive into three core paths: open-source pricing games, optical interconnect positioning, and on-device scenarios.
Read more →
·5 min
Why Are China's Open-Source AI Models Getting Stronger? A Deep Dive into Three Comeback Paths
Facing US chip bans and closed-source monopolies, how do China's open-source AI models keep striking back? A deep dive into three core paths: open-source pricing-power games, optical interconnect positioning, and edge-side use cases.
Read more →
·3 min
How China's Open-Source LLMs Are Closing In on America's Closed-Source Giants
Kimi, Alibaba Qwen, and DeepSeek leverage MoE architecture, home-field data advantages, and the open-source flywheel to close in on OpenAI and Anthropic at ultra-low cost. A deep dive into China's open-source LLM catch-up.
Read more →
·3 min
The Rise of General AI: Breakthroughs in Mathematical Reasoning and Competitive Programming
Frontier AI is going general: costs are dropping, general models are beating specialized ones in math and competitive programming, and multi-agent workflows are maturing fast.
Read more →
·3 min
Is Gemini Getting Worse? A Deep Dive into the 'Model Degradation' Phenomenon
Reddit users report Gemini feeling "worse over time." This deep dive explains the real causes—silent updates, safety over-alignment, and compute throttling—plus how to rationally evaluate AI model performance.
Read more →
·7 min
Auditing Cryptographic Libraries with AI: What We Found in Cloudflare's CIRCL
An in-depth look at using AI LLMs to audit Cloudflare's open-source cryptographic library CIRCL, covering constant-time detection, side-channel vulnerabilities, and human-AI collaboration.
Read more →
·6 min
Hands-On with Panel AI On-Premises Deployment: One-Click Enterprise Adoption and Unified Compute Cluster Scheduling
In-depth review of Panel AI v1.1.1: second-level installation, no-public-IP networking, batch compute cluster management. Learn how enterprise AI on-premises deployment barriers are dramatically lowered.
Read more →
·7 min
AI Daily: T-Head Open-Sources SAIL Software Stack at WAIC, Flurry of Large Model Releases
T-Head open-sources AI software stack T-Head SAIL at WAIC to lower the barrier for domestic chip development; Kimi K3 tops the WebDev leaderboard; Qwen 3.8 Max Preview cuts prices aggressively; Moonshot prepares a Hong Kong IPO; and Oracle switches its data center to a fuel cell microgrid.
Read more →
·4 min
Claude Usage Anxiety: The Hidden Cost Trap of AI Coding Tool Subscriptions
Real Reddit user rants reveal AI subscription pain points: Claude, Sol, and other tools consume usage at wildly different rates—does faster mean pricier? A deep dive into AI billing logic, usage transparency, and platform trust.
Read more →
·2 min
GPT-Live Full-Duplex Voice Model: Listen While Speaking + Reasoning Delegation Architecture Explained
OpenAI's GPT-Live full-duplex voice model enables natural simultaneous conversation with a reasoning delegation architecture pairing real-time dialogue with GPT-5.5 deep reasoning. Now live for 150M users.
Read more →
·3 min
What Is a Large Language Model, Really? The Core Principles Explained Through the Lens of Functions
Demystify large language models using middle-school math: LLMs are complex functions, training solves for parameters, and inference predicts next-token probabilities.
Read more →
·2 min
GPT-Live Real-Time Voice Interaction: How "Listen and Speak Simultaneously" Enables Natural Conversation
GPT-Live breaks the "turn-based" limit with millisecond real-time voice interaction. Explore the tech behind simultaneous listen-and-speak, multilingual interpretation demos, and AI voice's shift from tool to conversational partner.
Read more →
·3 min
Running a 27B Model on iPhone: How 1-bit Quantization Achieves Extreme Compression
PrismML's Bonsai compresses a 27B model from 54GB to 3.9GB, running at ~11 tokens/sec on iPhone. A deep dive into QAT, knowledge distillation, and speculative decoding.
Read more →
·4 min
DSpark Open Source: DeepSeek's Inference Acceleration Tech That Doubles LLM Speed
DeepSeek and Peking University open-source DSpark, an inference acceleration technology using semi-autoregressive architecture and dynamic scheduling to boost LLM speed by 50%+ and double GPU concurrency without quality loss.
Read more →
·2 min
OpenAI Reportedly Planning Major Price Cuts: A Direct Strike at Anthropic's API Market
OpenAI is reportedly in internal discussions about major API token price cuts targeting Anthropic's user base. We analyze the market logic, developer impact, and where AI infrastructure competition is headed.
Read more →
·3 min
LongCat 2.0: Meituan Trains a 1.6 Trillion Parameter LLM on Domestic Chips
Meituan open-sources LongCat 2.0, a 1.6T-parameter MoE model trained on 50,000+ custom chips without NVIDIA GPUs or Google TPUs, rivaling OpenAI and Google.
Read more →
·3 min
DeepSeek Open-Sources DSpark: 50%–400% LLM Inference Speedup Without Retraining
DeepSeek open-sources DSpark, delivering 50%–400% LLM inference speedup with no retraining or quantization, via semi-autoregressive drafting and confidence-scheduled verification.
Read more →
·3 min
Perplexity AI Deep Dive: The Perfect Fusion of Search Engine and AI Q&A
Deep dive into Perplexity AI: real-time web search + multi-model AI, transparent citations, Focus Modes, PDF chat, and collaborative Collections. Can it replace Google and ChatGPT?
Read more →
·4 min
Running Qwen3.6 27B Locally on a Single RTX 3090: A Complete Guide to Cloud-Rivaling Performance
One used RTX 3090, one 16.8GB GGUF file, and Qwen3.6 27B runs locally offline. SWE-bench score of 77 rivals Claude Sonnet. MTP boosts speed to 59 tok/s. Full local AI coding assistant deployment guide.
Read more →