63 links · newest first
This video explains DeepSeek's Manifold-Constrained Hyper-Connections (mHC), a new technique that widens internal information flow in AI models instead of simply scaling layers, parameters, or data. It contrasts mHC with residual connections and earlier hyper-connection attempts that became unstabl…
Visit originalThis video explains Meta FAIR's VL-JEPA (Vision-Language Joint Embedding Predictive Architecture), led by Yann LeCun, which predicts meaning directly instead of generating text token by token. It contrasts VL-JEPA with standard vision-language models, arguing embedding-based prediction is more effi…
Visit originalThis Medium article examines Google's 'Titans' neural long-term memory architecture, which claims to enable 2M+ token context windows with O(n) complexity. It cites reported benchmark results, including 98.8% needle-in-haystack accuracy at 8K context versus Mamba-2's 31%, while raising skepticism a…
Visit originalThis Medium article by Aakash Gupta claims that after reviewing 1,500 academic papers on prompt engineering, much popular social-media advice is actually counterproductive. It argues that research-backed practices favor structure over length, direct instructions over unnecessary examples, and conti…
Visit originalThis is the same Dwarkesh Patel podcast episode with Ilya Sutskever (accessed via a different share link) covering SSI's research strategy, the limits of pre-training and scaling, and why AI models currently generalize worse than humans. Topics include reinforcement learning, self-play, continual l…
Visit originalThis YouTube video is Dwarkesh Patel's podcast interview with Ilya Sutskever, discussing SSI's strategy, the limitations of pre-training, and why AI models generalize worse than humans. The conversation covers reinforcement learning, value functions, self-play, continual learning, and Sutskever's v…
Visit originalIn this Machine Learning Street Talk episode, Transformer co-inventor Llion Jones and researcher Luke Darlow of Sakana AI argue the Transformer architecture may be trapping AI research in a local optimum, limiting genuine reasoning progress. They introduce Continuous Thought Machines (CTM), a biolo…
Visit originalZDNet reports that Beijing-based Moonshot AI released Kimi K2 Thinking, an open-source reasoning model the company claims outperforms OpenAI's GPT-5 and Anthropic's Claude Sonnet 4.5 on benchmarks like Humanity's Last Exam. The model uses a Mixture-of-Experts architecture, is available on Hugging F…
Visit originalThis Medium piece by Alberto Romero argues Silicon Valley is fixated on the wrong kind of AI, framing the current AI boom as bubble-like and comparing it to historical manias such as railroad speculation. It introduces Tiny Recursion Models (TRMs) as an interesting alternative approach worth more a…
Visit originalPhysicist and YouTuber Sabine Hossenfelder argues that current AI models suffer from three fundamentally unsolvable problems: being narrowly purpose-bound rather than generally intelligent, persistent hallucinations and prompt-injection vulnerabilities, and an inability to extrapolate beyond traini…
Visit originalAI Stupid Level is an independent, real-time benchmarking platform that scores large language models like GPT-5, Claude Opus 4, Grok 4, and Gemini 3 on coding, reasoning, tool-calling, and speed. It also functions as an intelligent API router that can direct requests to the best-performing model an…
Visit originalThis Substack post from Kilo Code analyzes GLM-4.6, a Zhipu AI model whose usage reportedly grew nearly 100-fold on Kilo's coding leaderboard within 12 days. It covers the model's architecture, benchmark performance, low pricing, and Zhipu AI's shift toward domestic Chinese hardware as evidence of …
Visit originalCNBC reports that Chinese AI startup DeepSeek released DeepSeek-V3.2-Exp, an experimental model built on V3.1-Terminus that improves efficiency and lowers costs using a new 'DeepSeek Sparse Attention' mechanism for handling long documents and conversations. The article discusses performance compara…
Visit originalWindows Central reports on SpikingBrain1.0, a brain-inspired large language model built by Chinese researchers that claims to run up to 100 times faster than typical LLMs by focusing on nearby words rather than comparing every word in a sequence. The model reportedly was tested on a domestically pr…
Visit originalChina's DeepSeek released its v3.1 'hybrid' model, which outperforms OpenAI's hyped GPT-OSS open-source release on fiction, logic, and coding tasks. The article frames this as a comeback for open-source AI, with DeepSeek regaining ground even as OpenAI tries to re-enter the open-weight space.
Visit originalThis Medium article discusses a paper from Google DeepMind, KAIST AI, and Mila researchers titled 'Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation,' framed as a potential 'Transformers killer.' It claims the new architecture achieves 2x inference speed …
Visit originalThis Medium article by Jordan Gibbs argues that the single most valuable ChatGPT prompting technique is instructing the model to ask clarifying questions before answering. The author says this approach prevents ChatGPT from rushing to a response and instead helps it gather more useful context from …
Visit originalThis Medium article claims GPT-5 would arrive in July 2025 and mark a major shift in AI capability, citing a 2024 comment from Sam Altman that GPT-5 would follow GPT-4.5 'in months, not years.' It argues GPT-5 would significantly advance reasoning, multimodal understanding, coding, and autonomous a…
Visit originalThis LessWrong post argues that large language models can generalize beyond their training data, produce novel outputs, and perform causal or counterfactual reasoning without human-like 'understanding.' It uses analogies like linear regression, calculators, and self-driving cars to argue that obser…
Visit originalThis DataCamp blog post explains Retrieval Augmented Generation (RAG), describing how it connects LLMs to external data sources to produce more accurate and current answers. It walks through the RAG pipeline — data collection, chunking, embeddings, retrieval, and generation — with examples like cus…
Visit original