🔴 High Significance
Model Releases
🔴 🧡 Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber — score 81
Sources: hackernews · lab_blog/DeepMind
We’re introducing new Gemini models, including Gemini 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber.
Developer Tools
🔴 💬 CEO of Hugging Face: Banning open-source AI would hurt defenders 10x more than attackers, which would make the world 10x more dangerous and this is a good example why! — score 96
Sources: reddit/r/LocalLLaMA
From clem 🤗 on 𝕏: https://x.com/ClementDelangue/status/2079301434357456931 Fortune: Hugging Face says it resorted to a Chinese AI model to battle a fully autonomous cyberattack because U.S. model guardrails stymied its defense: [https://for
🔴 💬 AI agents don’t choose from the whole market. They choose from an invisible approved-vendor list. — score 94
Sources: reddit/r/AIAgents
While building agent systems, I keep running into the same thing: the real competition between tools happens before anything reaches a screen. Every agent ends up with something like an invisible approved-vendor list. It contains the tools, data sources, services, and counterparties the agent is abl
🔴 💬 OpenAI admits responsibility for HuggingFace Attack - an agent from an internal evaluation is reportedly the cause. — score 83
Sources: reddit/r/LocalLLaMA · hackernews · lab_blog/OpenAI
OpenAI and Hugging Face share early findings from a security incident during AI model evaluation, highlighting advanced cyber capabilities and lessons for defenders.
🔴 💬 Would you pay for knowing when your agents hallucinate? — score 83
Sources: reddit/r/AIAgents
no self promo or hook for you to use my product, i just want genuine help please :) I am thinking about building product analytics for AI agents, and want input from people actually shipping them. The case: a friend built a lot of their product from chat. Clients like it, but sometimes the AI does
Research Papers
🔴 🤗 Open-AoE: An Open Egocentric Manipulation Dataset and Toolchain for Embodied Learning — score 95
Sources: huggingface
Egocentric videos of human manipulation provide scalable supervision for embodied intelligence, yet existing resources rarely combine low-cost continuous capture, manipulation-level structured annotations, and reusable tools for robot learning. We present Open-AoE, an open, community-oriented egocen
🔴 🤗 Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints — score 78
Sources: huggingface · arxiv/cs.AI
Structure-based drug design (SBDD) leverages the 3D structure of protein targets, often complemented by other spatial constraints, to generate candidate binding molecules. While diffusion models have dominated as a leading paradigm for high-quality 3D molecule generation, LLM-based methods are rapid
🔴 🤗 ShotPlan: Cinematic Video Generation with Learnable Planning Token — score 70
Sources: huggingface
Current video generation models achieve impressive results in single-shot generation, yet remain limited in cinematic video generation, where coherent narratives and effective multi-shot composition require explicit shot planning. To address this challenge, we propose ShotPlan, a framework for expli
🔴 🤗 The Geometry of Semantic Space: A Continuous Geometric Framework for the Transformer Architecture — score 70
Sources: huggingface · arxiv/cs.CL
We present a continuous geometric framework that models the discrete algebraic operations of the Transformer architecture as an integro-differential equation (IDE) on a semantic fiber bundle calE = calM times R^d. Beginning from a single geometric axiom -- that the token sequence forms a discrete 1-
Other Signals
🔴 💬 Number of Submissions @ AAAI [D] — score 81
Sources: reddit/r/MachineLearning
Recently submitted my abstract and the submission number is 32xxx. With still a day to go, I just wonder where are we heading. Hope these conferences at least start making the reviews and names public for the withdrawn/rejected papers. So that people atleast take that accountability
🔴 💬 Laguna S 2.1 Released: Cheaper than Deepseek v4 Flash, Better than V4 Pro — score 75
Sources: reddit/r/LocalLLaMA
|Model|Size|Terminal-Bench 2.1|SWE-bench Multilingual|SWE-Bench Pro (Public Dataset)|DeepSWE|SWE Atlas (Codebase QnA)|Toolathlon Verified| |:-|:-|:-|:-|:-|:-|:-|:-| |Laguna S 2.1|118B-A8B|70.2%|78.5%|59.4%|40.4%|46.2%|49.7%| Finally the banger we've been waiting from Lagu
🟡 Notable
Model Releases
🟡 💬 poolside/Laguna-S-2.1 released! Finally an interesting 120B contender! — score 68
Sources: reddit/r/LocalLLaMA
HF: https://huggingface.co/poolside/Laguna-S-2.1 GGUFs available for use with llama.cpp custom fork: https://huggingface.co/poolside/Laguna-S-2.1-GGUF Posted on X: [https://x.com/poolsideai/status/20
🟡 💬 I search for participants! We want to conduct a study on multi agent systems :) — score 67
Sources: reddit/r/AIAgents
Hello everyone, We are conducting a study on how people experience multi-agent systems. We are seeking participants with hands-on experience using systems such as Manus, Hermes, OpenAI Agents SDK, CrewAI, AutoGen, LangGraph, Google ADK, or comparable systems that enable multiple agents to co
🟡 ✉️ If theProtein Data Bank(PDB) unlocked structural biology models (Boltz Episode,ESM/BioHub Episode), CELLxGENE has done the same thing for Virtual Cell models. Like PDB, CELLxGENE has inspired a zoo of — score 65
Sources: newsletter/Latent Space
If theProtein Data Bank(PDB) unlocked structural biology models (Boltz Episode,ESM/BioHub Episode), CELLxGENE has done the same thing for Virtual Cell models. Like PDB, CELLxGENE has inspired a zoo of AI models of RNA expression; so much so that RNA expression models have become synonymous with Virt
🟡 ✉️ HOW ANTHROPIC RUNS LARGE-SCALE CODE MIGRATIONS WITH CLAUDE CODE (14 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ SECURING THE AI SUPPLY CHAIN ON GKE: INTRODUCING K8S-AIBOM FOR AUTOMATED AI BOMS (6 MINUTE READ) — score 65
Sources: newsletter/tldr
Omitted 13 additional model releases items from the main section; see raw data and source-specific sections below.
Developer Tools
🟡 💬 Tri-Net v2: Open-source implementation of our Scientific Reports paper on unified skin lesion and symptom-based monkeypox detection [R] — score 69
Sources: reddit/r/MachineLearning
Hi everyone, We've open-sourced Tri-Net v2, the official implementation accompanying our recently published Scientific Reports (Nature Portfolio) paper: "Tri-Net: Unified Deep Learning for Skin Lesion and Symptom-Based Monkeypox Detection" Rather than releasing only training scripts, we rebuilt the
🟡 ✉️ We were lucky enough to have the two central figures in this story on our podcast. Taking the lead from Ci Chu and Bo Wang, Xaira Therapeutics is betting thatinformation richdata is the key to AI-driv — score 65
Sources: newsletter/Latent Space
We were lucky enough to have the two central figures in this story on our podcast. Taking the lead from Ci Chu and Bo Wang, Xaira Therapeutics is betting thatinformation richdata is the key to AI-driven drug development. Chuwas recently promotedto Chief Discovery Officer and Bo to Chief AI Scientist
🟡 ✉️ FROM OTEL TO SLMS: DISTILLING FRONTIER MODEL BEHAVIOUR FROM PRODUCTION TELEMETRY (45 MINUTE VIDEO) — score 65
Sources: newsletter/tldr
🟡 ✉️ DATA MODELING ISN'T DEAD, YOU JUST STOPPED DOING IT WITH JOE REIS (65 MINUTE VIDEO) — score 65
Sources: newsletter/tldr
🟡 ✉️ AGENTS THINK IN MILLISECONDS, LEGACY INFRASTRUCTURE DOESN'T. LINKEDIN, WALMART, AND ZENDESK SHARED HOW THEY CLOSED THE GAP AT VB TRANSFORM 2026 (4 MINUTE READ) — score 65
Sources: newsletter/tldr
Omitted 10 additional developer tools items from the main section; see raw data and source-specific sections below.
Infrastructure & Compute
🟡 ✉️ Iftest loss flatlinesafter 1.5B parameters whiletraining loss continues to dropas you scale, that tells you that your model islimited by the amount of informationin your data. — score 65
Sources: newsletter/Latent Space
Training on a single, smallish data set exposed an information gap: the 3.1B model falls off the scaling trend. Neither parameters nor compute will improve performance past this wall. For predicting changes to gene expression, you need moreinformation rich data.
🟡 ✉️ APPLE, NVIDIA VIE FOR TITLE OF WORLD'S MOST VALUABLE COMPANY (2 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ SpaceX is negotiating a multibillion-dollar compute deal with the U.S. DoD, adding to the list of rental partnerships that already include Google, Anthropic, and Reflection AI. — score 65
Sources: newsletter/rundown-ai
🟡 ✉️ Moonshot is halting new subscriptions after demand for its K3 model pushed the company to its compute limits, also splitting memberships into chat and coding plans. — score 65
Sources: newsletter/rundown-ai
Business & Funding
🟡 ✉️ FROM WEEKS TO A DAY: HOW WE MADE LLM EVALUATION FAST ENOUGH TO ITERATE ON (10 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ DATABRICKS HITS $188B VALUATION, EXTENDING ITS RUN AS AI'S FAVORITE SECOND ACT (4 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ META IN TALKS TO LEASE COMPUTING POWER TO ANTHROPIC IN POTENTIAL $10 BILLION DEAL (5 MINUTE READ) — score 65
Sources: newsletter/tldr
Research Papers
🟡 🤗 Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL — score 60
Sources: huggingface · arxiv/cs.AI
Recent growth in reinforcement learning (RL) has surfaced a need for diverse, specialized training environments. Hand-curated environments with fixed task and reward difficulties become ineffective signals as model performance improves, and sparse rewards over long horizons induce mode collapse on s
🟡 🤗 ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video — score 45
Sources: huggingface · arxiv/cs.AI
Egocentric devices, such as wearable front-facing cameras, provide a unique perspective for capturing the continuous interaction between a human viewer and the surrounding environment. A holistic and efficient multimodal model capable of reconstructing this 4D representation is therefore highly desi
🟡 🤗 Can Multimodal Large Language Models Understand OCT? — score 45
Sources: huggingface · arxiv/cs.CL
Optical coherence tomography (OCT) imaging is essential for the diagnosis and treatment of retinal diseases. Although multimodal large language models (MLLMs) have demonstrated considerable potential in medical image analysis, existing benchmarks largely reduce OCT understanding to coarse-grained di
Other Signals
🟡 ✉️ IN-HOUSE LLM SERVING AT NETFLIX (7 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ DREMIO'S EXIT IS THE CLEAREST SIGN YET THAT LAKEHOUSE-ONLY WON'T SURVIVE AI (3 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ CHINA JOINS RUSH TO RETHINK THE SMARTPHONE FOR THE AI ERA (5 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ THESE AI-NATIVE COMPANIES HAVE TINY STAFFS AND FEWER BOSSES (7 MINUTE READ) — score 65
Sources: newsletter/tldr
🟡 ✉️ ARE THE LLM WARS THE DATABASE WARS? (2 MINUTE READ) — score 65
Sources: newsletter/tldr
Omitted 14 additional other signals items from the main section; see raw data and source-specific sections below.
🟢 Incremental
Model Releases
🟢 💬 New Model: Nanbeige4.2-3B (Looped Transformer, outperforms 4x size) — score 39
Sources: reddit/r/LocalLLaMA
https://huggingface.co/Nanbeige/Nanbeige4.2-3B Nanbeige4.2-3B is a compact agentic model built on Nanbeige4.2-3B-Base, designed to combine strong agentic behavior with broad reasoning and alignme
🟢 💬 I wired Fable 5 agent into a database of every Polymarket wallet and trades via MCP. What do you want me to ask it next? This is what I found so far: — score 33
Sources: reddit/r/AIAgents
Hey Fellas... I've been tracking all Polymarket Activity for months (1.3 Billion trades and 2.7M wallets) and I gave Claude Code a Postgres MCP (found crazy stuff) pointed at the live ledger, so now I just ask it anything in plain English and it writes + runs the query itself. A few I already asked,
🟢 💬 There is no need to worry about Trump banning China‘s open source model at all. — score 32
Sources: reddit/r/LocalLLaMA
I’ve seen many people worrying that if Trump moves to block Chinese AI models, aggregators like OpenRouter will no longer be able to host them. As someone from China, let me reassure you: there is really no need for such concern. China has been locked in trade conflicts with Trump for years. **Nobod
🟢 🧡 "Drawing" the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok — score 21
Sources: hackernews
🟢 💬 Updated Gemma-4 chat template witchcraft: Gemma-4-26B-a4B shows dominance over Qwen3.6-MoE and Qwen3.5-MoE fine tunes (Instruct mode and Reasoning efficiency) — score 18
Sources: reddit/r/LocalLLaMA
Kind of unexpected. Happy for Gemma-4/Google, big win for us, LocalLLMers. Yet Qwen3.6 still does better in Hermes than Gemma-4 somehow. We need Gemma-4.1 fine-tuned on Agentic-tasks. That would be killer.
Omitted 2 additional model releases items from the main section; see raw data and source-specific sections below.
Developer Tools
🟢 🐙 rohitg00/agentmemory — #1 Persistent memory for AI coding agents based on real-world benchmarks — score 39
Sources: github_trending
#1 Persistent memory for AI coding agents based on real-world benchmarks
🟢 🐙 akitaonrails/ai-memory — Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors — score 36
Sources: github_trending
Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
🟢 💬 An experiment on symmetric agent communication. It went better than I expected — score 33
Sources: reddit/r/AIAgents
This is an experiment I've been living in for a while, and I wanted to share where it went. I set out to build a coding agent from scratch in Rust: no agent SDK, a hand-rolled agentic loop, a ratatui TUI, tokio underneath. I did this because I wanted to understand how agents work under the hood. Mid
🟢 💬 Reproducing OpenAI’s “persistently beneficial models” - GRPO trait install barely moves. Ideas? [P] [R] — score 31
Sources: reddit/r/MachineLearning
TL;DR: I’m reproducing the trait-persistence result from arXiv:2606.24014 on one RTX 3090. Before I can test persistence I need to install a trait via RL — and my GRPO run moves the trait only +2.4 points (95% CI [+0.2, +4.8]) when I need ~+15. Traini
🟢 🐙 owainlewis/awesome-artificial-intelligence — A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers. — score 30
Sources: github_trending
A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers.
Omitted 7 additional developer tools items from the main section; see raw data and source-specific sections below.
Infrastructure & Compute
🟢 🐙 NVIDIA/cosmos-framework — Our inference and training framework to run on the Cosmos Models — score 5
Sources: github_trending
Our inference and training framework to run on the Cosmos Models
Other Signals
🟢 💬 We've analyzed over 300,000 AI agent conversations. Here are the 3 ways they fail. — score 37
Sources: reddit/r/AIAgents
My co-founder and I at Green flash spend all day looking at how end-users interact with production AI agents. After analyzing over >300k convos and >12.5m individual analysis runs, we've noticed that the biggest threat to your product isn't always API errors. Here's what actually breaks when r
🟢 🧡 Meta's AI models are powering the first wave of Genesis Mission projects — score 36
Sources: hackernews
🟢 💬 I ran Laguna-S-2.1 through my private agentic eval vs Qwen3.5-122B on an RTX Pro 6000 (96GB). Fastest 100B+ I've tested and the best tool calling, but it invents facts under pressure. — score 25
Sources: reddit/r/LocalLLaMA
Laguna-S-2.1 dropped few hours ago and as I am in the market for an upgrade to trusty qwen3.6 dense and the current daily 122B, I ran it through the same eval harness I use to pick the model that runs my local agent stack. Posting because the results don't fit the usual "benchmaxed or king" binary,
🟢 💬 My OCR model mislabels section titles as body text. Is a CRF the right fix, or am I overcomplicating it? [P] — score 12
Sources: reddit/r/MachineLearning
Hi everyone, I'm working on extracting the hierarchical structure of long PDF documents (legal/regulatory text, lots of numbered sections) and would like to gather some feedback on my approach before committing to it. What I've done so far: I render each PDF page to an image and run it through [
🟢 💬 Bessent says U.S. could sanction China over AI model 'theft' — score 11
Sources: reddit/r/LocalLLaMA
Omitted 1 additional other signals items from the main section; see raw data and source-specific sections below.
📊 Cross-Source Signals
Items that appeared on 3+ sources today:
- OpenAI admits responsibility for HuggingFace Attack - an agent from an internal evaluation is reportedly the cause. — appeared on: reddit/r/LocalLLaMA (491), hackernews (538), lab_blog/OpenAI (100)
📈 Trending Repos
| Repo | Description | Stars Today | Language |
|---|---|---|---|
| agegr/pi-web | Web UI for the pi coding agent | 286 | typescript |
| rtk-ai/rtk | CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies | 248 | rust |
| rohitg00/agentmemory | #1 Persistent memory for AI coding agents based on real-world benchmarks | 72 | typescript |
| akitaonrails/ai-memory | Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors | 66 | rust |
| owainlewis/awesome-artificial-intelligence | A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers. | 50 | python |
| dottxt-ai/outlines | Structured Outputs | 49 | python |
| langchain-ai/open_deep_research | 14 | python | |
| yvgude/lean-ctx | Control what your AI can see. LeanCTX (Lean Context) is the context intelligence layer for AI agents — one local Rust binary that decides what they read, remembers what they learn, guards what they touch, and proves what they save. 60–90% fewer tokens as the receipt. 76 MCP tools, 30+ agents, local-first. | 14 | rust |
| pydantic/monty | A minimal, secure Python interpreter written in Rust for use by AI | 6 | rust |
| NVIDIA/cosmos-framework | Our inference and training framework to run on the Cosmos Models | 5 | python |
📄 New Papers
| Title | Category | Hotness | Link |
|---|---|---|---|
| Open-AoE: An Open Egocentric Manipulation Dataset and Toolchain for Embodied Learning | research_paper | 25 | Open |
| Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints | research_paper | 13 | Open |
| ShotPlan: Cinematic Video Generation with Learnable Planning Token | research_paper | 5 | Open |
| The Geometry of Semantic Space: A Continuous Geometric Framework for the Transformer Architecture | research_paper | 5 | Open |
| Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL | research_paper | 4 | Open |
| Rater State Bias in RLHF Preference Data: An Audit Framework | cs.AI | 0 | Open |
| Design and Validation of a Lightweight 1D CNN for Affective Touch Classification in Soft Plush Companions | cs.AI | 0 | Open |
| Some Large Language Models Exhibit Consistent Risk Attitudes | cs.AI | 0 | Open |
| A Survey on GNN-based Link Prediction: Techniques, Applications, and Challenges | cs.AI | 0 | Open |
| PlanFlip: Attacking Multi-Agent LLM Systems via Planning-Phase Prompt Injection | cs.AI | 0 | Open |
| Deterministic Replay for AI Agent Systems | cs.AI | 0 | Open |
| Generative Ontology Induction: Domain-Agnostic Schema Discovery from Document Corpora Using Large Language Models | cs.AI | 0 | Open |
| Democratizing AI with Small Language Models: Structured Benchmarking and Parameter-Efficient Fine-Tuning for Local Deployment | cs.AI | 0 | Open |
| It Takes 8 Tokens: Weak-to-Strong Off-Policy RL via Auxiliary Branches | cs.AI | 0 | Open |
| PPO-HSC: An Exploratory Reinforcement Learning Framework Based on Wide-Area Policy Coverage Optimization | cs.AI | 0 | Open |
🏢 Lab Blog Posts
- OpenAI: Introducing the ChatGPT for small business program
- OpenAI: David Vélez and Robin Vince join the boards of the OpenAI Foundation and OpenAI Group PBC
- DeepMind: Introducing Gemini 3.5 Flash Cyber
🐦 Twitter/X Highlights
| Account | Tweet Summary |
|---|---|
| OpenAI | We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation. Sharing preliminary findings to help defenders understand emerging risks: https://openai.com/index/hugging-face-model-e Post |
| OpenAI | We’re sharing new research with @apolloaievals on reward-seeking—when models follow what they believe a grader rewards rather than what users or developers want—and a new method, Contrastive SDF, for measuring how strongly such beliefs shape behavior. https://alignment.openai.com/measuring-reward-se Post |
| GoogleDeepMind | Gemini 3.5 Flash-Lite is our fast, cost-effective model for scaling repetitive use cases like sorting tickets and extracting data. Watch how it performs against 3.5 Flash on a series of high volume tasks ↓ Post |
| GoogleDeepMind | Gemini 3.6 Flash builds directly on feedback from 3.5 Flash. Watch how it compares on quality and token usage ↓ Post |
| MistralAI | Mistral is announcing an expanded global strategic partnership with @Microsoft to give enterprises and regulated industries frontier AI they can control. As Mistral is expanding its AI compute capacity in Europe, the companies are expanding their strategic partnership with Microsoft’s commitment to Post |
| karpathy | One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anythi Post |
| mattshumer_ | So GPT-6: - one-shotted a counter-example to the Jacobian conjecture - and then escaped containment, and hacked into HuggingFace… all for a benchmark Yeah, this model is going to be something else. Post |
Newsletter
- Latent Space: Iftest loss flatlinesafter 1.5B parameters whiletraining loss continues to dropas you scale, that tells you that your model islimited by the amount of informationin your data.
- Latent Space: We were lucky enough to have the two central figures in this story on our podcast. Taking the lead from Ci Chu and Bo Wang, Xaira Therapeutics is betting thatinformation richdata is the key to AI-driv
- Latent Space: If theProtein Data Bank(PDB) unlocked structural biology models (Boltz Episode,ESM/BioHub Episode), CELLxGENE has done the same thing for Virtual Cell models. Like PDB, CELLxGENE has inspired a zoo of
- tldr: FROM OTEL TO SLMS: DISTILLING FRONTIER MODEL BEHAVIOUR FROM PRODUCTION TELEMETRY (45 MINUTE VIDEO)
- tldr: DATA MODELING ISN'T DEAD, YOU JUST STOPPED DOING IT WITH JOE REIS (65 MINUTE VIDEO)
- tldr: AGENTS THINK IN MILLISECONDS, LEGACY INFRASTRUCTURE DOESN'T. LINKEDIN, WALMART, AND ZENDESK SHARED HOW THEY CLOSED THE GAP AT VB TRANSFORM 2026 (4 MINUTE READ)
- tldr: AI AGENTS NEED DATA PRODUCT CONTEXT NOT MORE RAG (11 MINUTE READ)
- tldr: EXPERIENCE GRAPHS: THE DATA FOUNDATION FOR SELF-IMPROVING AGENTS (28 MINUTE READ)
- tldr: IN-HOUSE LLM SERVING AT NETFLIX (7 MINUTE READ)
- tldr: FROM WEEKS TO A DAY: HOW WE MADE LLM EVALUATION FAST ENOUGH TO ITERATE ON (10 MINUTE READ)
- tldr: DATABRICKS HITS $188B VALUATION, EXTENDING ITS RUN AS AI'S FAVORITE SECOND ACT (4 MINUTE READ)
- tldr: DREMIO'S EXIT IS THE CLEAREST SIGN YET THAT LAKEHOUSE-ONLY WON'T SURVIVE AI (3 MINUTE READ)
- tldr: META IN TALKS TO LEASE COMPUTING POWER TO ANTHROPIC IN POTENTIAL $10 BILLION DEAL (5 MINUTE READ)
- tldr: SPACEX IN TALKS TO PROVIDE COMPUTING POWER FOR PENTAGON'S AI PUSH (3 MINUTE READ)
- tldr: CHINA JOINS RUSH TO RETHINK THE SMARTPHONE FOR THE AI ERA (5 MINUTE READ)
- tldr: HOW ANTHROPIC RUNS LARGE-SCALE CODE MIGRATIONS WITH CLAUDE CODE (14 MINUTE READ)
- tldr: THESE AI-NATIVE COMPANIES HAVE TINY STAFFS AND FEWER BOSSES (7 MINUTE READ)
- tldr: APPLE, NVIDIA VIE FOR TITLE OF WORLD'S MOST VALUABLE COMPANY (2 MINUTE READ)
- tldr: ARE THE LLM WARS THE DATABASE WARS? (2 MINUTE READ)
- tldr: AI MANIA IS EVISCERATING GLOBAL DECISIONMAKING (38 MINUTE READ)
- tldr: HOW TO MANAGE AI INVESTMENTS IN THE AGENTIC ERA (5 MINUTE READ)
- tldr: REVIEWING AI-GENERATED CODE (9 MINUTE READ)
- tldr: SECURING THE AI SUPPLY CHAIN ON GKE: INTRODUCING K8S-AIBOM FOR AUTOMATED AI BOMS (6 MINUTE READ)
- tldr: CODEX RESETS (WEBSITE)
- tldr: OPENAI TWEAKS CHAT ACCESS IN THE CHATGPT APP FOR MAC (1 MINUTE READ)
- tldr: META'S "REGRET" OVER ITS SCRAPPED AI TOOL IS LIKE A BURGLAR APOLOGISING FOR THE MESS (7 MINUTE READ)
- tldr: WHY IS AI BAD AT DESIGN? (7 MINUTE READ)
- rundown-ai: SpaceX is negotiating a multibillion-dollar compute deal with the U.S. DoD, adding to the list of rental partnerships that already include Google, Anthropic, and Reflection AI.
- rundown-ai: Moonshot is halting new subscriptions after demand for its K3 model pushed the company to its compute limits, also splitting memberships into chat and coding plans.
- rundown-ai: The MLB is ending teams' ability to load their own programs onto dugout iPads to crack down on the use of AI ==to help make strategy decisions==.
- rundown-ai: Anthropic closes the book on Fable access
- rundown-ai: Claude helps disprove 87-year math problem
- rundown-ai: Before the report, former AI czar David Sacks wrote that the "leading closed labs" want "the government to eliminate their open-source competition."
- rundown-ai: Make AI note-taking easier with a “Captain’s Log”
- rundown-ai: The engineering leader's AI playbook
- rundown-ai: LM Studio Bionic - Local agent that codes and edits docs with open models
- rundown-ai: CData Connect AI can cut LLM context handling costs by up to 97.6% without sacrificing answers. Learn more here.
- rundown-ai: The U.S AI standards chief, Chris Fall, stepped down after just three months into running CAISI, the Commerce agency that tests commercial AI models.
- rundown-ai: Google is reportedly developing Frozen v2, a chip with part of Gemini's architecture carved into the hardware, projected to run the model 6-10x more efficiently than TPUs.
- tldr: GOOGLE VIDS NOW LETS YOU STAR IN YOUR OWN AI VIDEOS (1 MINUTE READ) - first seen 2026-07-20
- rundown-ai: Anthropic ends Fable 5’s access questions - first seen 2026-07-20
- rundown-ai: OAI CEO Sam Altman had mocked Anthropic’s access situation on X, with posts like “clarity is nice,” and “stay because we don’t treat you with contempt”. - first seen 2026-07-20
- rundown-ai: The Rundown Roundtable: Our AI use cases - first seen 2026-07-20
- rundown-ai: Deploy a mini-SaaS in minutes with ChatGPT Sites - first seen 2026-07-20
- rundown-ai: Turn AI time savings into real revenue - first seen 2026-07-20
- rundown-ai: Kimi K3 - Moonshot AI’s powerful new open-source model - first seen 2026-07-19
- rundown-ai: Alibaba’s Qwen said its soon-to-be-launched, open-weight Qwen3.8 model is “compatible to leading frontier AI models, second only to (Claude) Fable 5”. - first seen 2026-07-20
- rundown-ai: Incoming OAI hire Dean Ball predicted that the U.S. will scare firms off Chinese AI, with ex-White House AI czar David Sacks slamming the piece as a closed-lab plot. - first seen 2026-07-20
- rundown-ai: Read our last AI newsletter: Moonshot’s Kimi K3 closes the frontier gap - first seen 2026-07-20
Repeated From Recent Briefings
- bojieli/ai-agent-book — 《深入理解 AI Agent：设计原理与工程实践》（李博杰 著）开源主仓库：全书正文、编译版 PDF 与按章配套代码 - first seen 2026-07-19
- diegosouzapw/OmniRoute — Never stop coding. Free MIT AI gateway: one endpoint, 268+ providers (50+ free), 500+ models — Kimi, Claude, GPT, OpenAI, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 500+ contributors - first seen 2026-05-08
- US gov't lobbied by major US labs is about to ban open source models. - first seen 2026-07-20
- tirth8205/code-review-graph — Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows. - first seen 2026-05-02
- I just read LeCun’s recent thoughts on world models. Thoughts on JEPA as a path forward? [D] - first seen 2026-07-20
- stablyai/orca — Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and VPS. - first seen 2026-06-21
- earendil-works/pi — AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI - first seen 2026-05-09
- koala73/worldmonitor — Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface - first seen 2026-06-21
- rohitg00/ai-engineering-from-scratch — Learn it. Build it. Ship it for others. - first seen 2026-07-17
- 1jehuang/jcode — The most intelligent agent harness for code - first seen 2026-06-21
- ... plus 137 more repeated items in processed data