Anthropic's S1 filing shows an $80.7B H1 net loss and a November IPO is targeted.
DeepSeek open-sourced DeepInfer, an offline inference engine supporting DeepSeek-V4.1-Flash and parallel decoding optimizations.
Anthropic and Citrini Research used an internal model to design antiscorpion-venom minibinders; 17 of 211 met criteria and all data were released.
The U.S. Army Cyber Command is developing self-spreading ECHOTRIBBLE AI cyberweapons per internal documents.
According to an S1 filing obtained by The Wall Street Journal, Anthropic lost $80.7 billion in the first half of this year and could spend $750 billion on computing power in coming years. The company is preparing to go public.
BloombergDetails →
French government · Holds emergency meetings with TikTok and other platforms over protest footage and AI-altered images, while resale sites including Vinted and eBay sign a deal to block looted goods.
US Army Cyber Command · Is developing AI cyber agents codenamed ECHOTRIBBLE that spread through target networks like worms and can reach air-gapped systems, with experts warning of collateral damage.
Chinese government · Establishes a state-owned data group directly administered by the central government to strengthen data infrastructure and bolster its AI ambitions.
Mass General Brigham · Researchers trained AI on over 40 million facial photos to estimate biological age and predict cancer survival from a single selfie, with tools named FAHR-FaceAge and FAHR-FaceSurvival.
Ethan Mollick · announces his new book, Co-Existence, out today, a guide to living and working in a world where AI is sometimes but not always better than you.
Aravind Srinivas · announces that Perplexity Computer can now run fully local on a Mac, with no cloud inference and even no Wi-Fi needed.
MiniMax (official) · announces that MiniMax M3.1 Flash is available free in Cline, aimed at long agentic coding runs where cost adds up fast.
Jev CardsA picture quiz game where you race the vision model Jev, with live matches and open-source code.
PenmentorA pen-side math mentor where Claude, Gemini or GPT marks handwritten work with hints, not answers.
Exa 3.0A major search API upgrade giving AI agents richer data at lower cost
Composio InstantLets AI agents pay for their own tools like web search and voiceovers, no signup.
Mintlify AutopilotAn always-on docs agent that watches code and Slack, drafts doc updates, and opens PRs for review.
Exa 3.0Smarter search API giving AI agents richer data at lower cost.
ToDo.isAn always-on AI agent that completes written to-dos, researches with sources, and reports back across web and messaging.
bondsMessaging app where every friends' chat gets an AI friend with superpowers, not an assistant
ESP-MosaicoA modular ESP32 dev board that lets coding agents create and test real hardware
Kling 4.0Fast, affordable AI video model for creators with keyframe control and built-in lip sync
Harness²Plug-and-play recursive framework letting LLM agents auto-evolve their execution harnesses.
Jotform App Builder for ChatGPT & ClaudeBuild and edit no-code apps by chatting inside ChatGPT and Claude
Asimov 1A highly customizable humanoid robot that founders and developers can train and build on
receptionNative in-app iOS support chat with AI answers to reduce churn
RunInfraFlat-rate monthly plans giving coding agents fast access to open models.
InflightCapture pinned design feedback on live prototypes and hand it to coding agents via MCP
snyviFree open-source desktop app that organizes multiple passion projects with coding agents side by side
Lights-on software factory boilerplateA boilerplate for running a Claude Code software factory that cuts token costs by ~80%
Own 1Fanless personal AI computer running two local LLMs and eight agents on your desk.
Jailbreak tool for DeepSeek Harness with swappable per-model prompts.
Makes your AI agent code like the laziest senior dev, cutting code and tokens.
Coding-agent skill for multi-phase security audits with independently verified findings.
Turns any codebase, docs, SQL schemas and PDFs into a queryable knowledge graph.
A complete AI agency with specialized expert agents each having personality and deliverables.
Open-source plugin repo for knowledge workers using Claude Cowork across many domains.
High-throughput, memory-efficient inference and serving engine for LLMs.
Local agent-first control plane for catalog discovery and planning over 2,600+ skills.
Claude Code-based job application framework to evaluate postings, tailor CVs and prep interviews
Run a fleet of coding agents with your own subscription on desktop, mobile and remote
Open source coding agent that helps developers write and edit code in the terminal
Modular diffusion model GUI, API and backend with fast local inference for creators
Long-term memory for coding agent CLIs plus handoff between different agent vendors
Framework making all software agent-native through CLI access
Agent harness to deploy multi-agent swarms and coordinate autonomous workflows
Open-source foundation model for the language of financial markets
Compress tool outputs and logs before the LLM, cutting token usage while keeping answers
Agentic coding tool in the terminal that understands codebases via natural language
Nvidia is in talks to invest hundreds of millions of dollars in Australian AI cloud compa…
South Korean data center firm Elice fell 32% in its market debut, reviving fears about th…
Eka Robotics, founded by MIT professor Pulkit Agrawal and former DeepMind researcher Tuom…
Salesforce has agreed to acquire Braze, an AI-powered marketing platform, for almost $5.7…
OpenAI is reportedly rushing to reach a $1.4 trillion valuation before Anthropic launches…
OpenAI is acquiring Orchestra Bio to bring AI-powered operations to life sciences R&D, ai…
The A.I. Daily
