AI — Page 40
Models, agents, infra, applied AI.
- Insilico's AI-discovered lung fibrosis drug enters Phase III
Rentosertib will test whether Insilico's end-to-end AI drug discovery thesis can survive late-stage clinical scrutiny.
- ZML's Steeve Morin releases free LLMD to loosen AI inference from Nvidia
The Paris startup's closed-source inference server runs Llama, Gemma, Qwen and Mistral models across Nvidia, AMD, TPU, Intel and Apple backends.
- OpenAI says GPT-5.6 Sol Ultra produced a proof of a 50-year graph theory problem
Ethan Knight said the run used 64 subagents in under an hour, pushing OpenAI's new ultra mode into peer-review territory.
- Melius wants agents, not chatbots, to run creative production
Joowon Kim's startup is betting that designers and marketers will manage agent work on a canvas instead of inside a chat window.
- Drafted V2 turns room placement into AI-generated home plans
Nick Donahue's AI home-design startup added partial regeneration, exterior options, saved projects, and faster floor-plan generation.
- GPT-5.6 Sol Turns Blender Into an AI Speedrun
A viral X clip showed OpenAI's flagship model driving Blender at a claimed 750 tokens per second in Very Fast mode.
- Mistral moves into robotics with a single-camera navigation model
Robostral Navigate is an 8B embodied AI model that Mistral says can follow language instructions without LiDAR or depth sensors.
- Stan's 14-day AI build turned customer knowledge into $3 million ARR
John Hu says Stanley began with manual interviews, public posts and AI-written cold emails before expanding from LinkedIn to Instagram.
- GPT-5.6, OpenAI's flagship model, helps build itself
Shruti Gandhi's July 9th post spotlights a deeper shift behind Sol, Terra and Luna: the frontier model is becoming production infrastructure.
- Anthropic Puts a Meter on Claude Fable 5
Starting July 12th, paid Claude subscribers will owe API-style usage fees for Anthropic's top consumer model, a test of whether flat-rate AI subscriptions can survive agentic compute costs.
- Cells2Pixels gives neural cellular automata an 8x resolution path
Ehsan Pajouheshgar's SIGGRAPH 2026 project keeps the automata grid coarse and uses a small local decoder to render detail.
- GPT-5.6 Built a Tiny iMessage Transformer From One Prompt
Pietro Schirano says Sol generated the code, pipeline, tokenizer, training loop and local reply model in about 30 minutes.
- Google trains SensorFM on a trillion minutes of wearable data
The July 9th release turns Fitbit and Pixel Watch signals into a reusable health model evaluated across 35 prediction tasks.
- GPT-5.6 Sol runs a week-long voxel Manhattan build in Matt Shumer demo
The HyperWrite co-founder says OpenAI's flagship preview model completed the job autonomously, with no prompt, repo, logs, or cost disclosed.
- OpenAI launches GPT-5.6 with Sol, Terra and Luna tiers
Sam Altman called GPT-5.6 OpenAI's best model as the company pushes developers toward agentic work priced by capability tier.
- OpenAI sets a 10 a.m. Pacific livestream and leaves the subject blank
The post landed 38 minutes before showtime, one day after OpenAI started rolling out GPT-Live for ChatGPT Voice.
- Meta opens Muse Spark to developers with a low-priced coding model and 1M context window
Mark Zuckerberg says Muse Spark 1.1 adds a 1M-token context window, parallel sub-agents and computer-use training across desktop, mobile and browser.
- RoboDojo puts generalist robot policies through a tougher sim-and-real test
Tianxing Chen's team evaluated 30 policies across 42 simulation tasks and 18 real-world tasks, with top success rates still in single digits.
- Grok 4.5 gives Musk an Opus-class price weapon
SpaceXAI launched the coding-focused model on July 8th with $2 input and $6 output pricing, while Musk cast speed as the real edge.
- Artificial Analysis brings final-state scoring to agent workflow benchmarks
Micah Hill-Smith and George Cameron are testing agents against SaaS workflows where the database state matters more than the answer text.
- SpaceXAI's Grok 4.5 ranks fourth on Artificial Analysis benchmark
The benchmarker put Grok 4.5 behind Fable 5, GPT-5.5 and Opus 4.8, with cost as the clearest part of SpaceXAI's pitch.
- MeetKai signs Albania MOU as James Kaplan pushes sovereign AI into government workflows
The July 6th deal is a planning agreement with AKSHI, with no disclosed value, model architecture, or timetable.
- SpaceXAI releases Grok 4.5 with Cursor-trained coding push
Michael Truell says the model has replaced Composer 2.5 for many on the Cursor team as SpaceXAI prices it for broad developer use.
- OfficeCLI Gives AI Agents a Local Handle on Word, Excel and PowerPoint
The open-source iOfficeAI project is building the document layer agents need after chat, with no verified company, founder, or funding disclosed.