[research] ·
ExtractBench: A New Benchmark for Schema-Guided Document Extraction
LlamaIndex releases a benchmark that tests AI systems on extracting structured data from complex enterprise documents, with a focus on completeness and grounding.
[launches] · · By ByteBulletin Editors
OpenAI Launches ChatGPT Desktop App for Linux, Covering Every Major Desktop OS
The preview release brings ChatGPT, ChatGPT Work, and Codex to Ubuntu, Debian, and Fedora, filling a long-standing gap in OpenAI's desktop coverage.
Read the story →TOP STORIES
editor’s picks
[launches] ·
The Billion-User AI Chatbot Race Is Over — Now the Real Work Begins
ChatGPT and Gemini have both crossed the billion-user mark, shifting the AI competition from adoption to monetization and retention.
[research] ·
Anthropic's unreleased model makes significant progress on the Riemann hypothesis
A new AI model from Anthropic, left to work autonomously, improved the known bounds on one of math's oldest unsolved problems, raising questions about the role of AI in mathematical discovery.
THE WIRE
Full archive →[funding]By ByteBulletin Editors
River AI raises $1.1B to rebuild AI from scratch with personally trainable agents
The xAI co-founder's 2-month-old startup lands a massive seed round to reinvent model training, promising agents that are truly yours.
[tooling]By ByteBulletin Editors
Ungate Lets You Use Claude, ChatGPT, and MiniMax Subscriptions Inside Cursor
A new open-source extension proxies Cursor to your existing AI subscriptions, avoiding per-token API costs.
[research]By ByteBulletin Editors
Why AI Agents Aren't as Skilled as They Seem: New Research Highlights Inconsistency Problem
A new arXiv study reveals that AI agents show troubling inconsistency in skill execution, raising questions about their reliability in real-world tasks.
[research]By ByteBulletin Editors
Survey Maps the Multimodal Frontier: A New Taxonomy of Safety and Vulnerability in Multimodal LLMs
A comprehensive survey categorizes emerging risks in multimodal LLMs, from cross-modal attacks to evaluation gaps, offering a framework for safer AI development.
[research]By ByteBulletin Editors
New DoTime Benchmark Measures How Well AI Agents Manage Real-World Tasks
Researchers introduce DoTime, a benchmark that evaluates AI agents on time-sensitive, real-world activities to push beyond static coding tests.
[launches]By ByteBulletin Editors
Lumabri Lets You Run Huge Mixture-of-Experts Models Across a Swarm of Peer Machines
A new pure-C, dependency-free system pools CPU and disk from any machines on a network to serve massive MoE models, with byte-identical output and no upfront downloads.
[models]By ByteBulletin Editors
Meta doubles down on open-weight AI with Muse Glimmer release and Zuckerberg manifesto
Meta releases a 30B-parameter open model for local use, promises Spark 1.2 weights, and makes a philosophical case for decentralized AI.
[launches]By ByteBulletin Editors
OpenAI and Jony Ive’s AI device is a doughnut-shaped smart speaker without a screen
The long-rumored AI hardware from OpenAI and Jony Ive is reportedly a battery-powered, doughnut-shaped smart speaker with no display, priced over $300 and slated for 2027.
[research]By ByteBulletin Editors
SFT Alignment Transfer: A New Frontier in Efficient Model Specialization
Researchers propose a method to transfer alignment from one fine-tuned model to another, cutting training costs while preserving safety and task performance.
[models]By ByteBulletin Editors
Meta's Glimmer model brings 'personal superintelligence' to your laptop
With the open-weight Muse Glimmer, Meta offers a local, privacy-conscious AI agent for consumer hardware—and a glimpse of where it draws the line between open and closed AI.
[research]By ByteBulletin Editors
New Framework Lets AI Coding Tools Explain Their Reasoning
A new arXiv tool helps developers see how AI models reach clinical-style decisions, promising greater transparency in AI-assisted workflows.
[research]By ByteBulletin Editors
RL or SFT? New Research Teases Apart What Actually Drives Reasoning in LLMs
A new arxiv study compares reinforcement learning against supervised fine-tuning to isolate which training method truly boosts reasoning performance in large language models.
[research]
43ExtractBench: A New Benchmark for Schema-Guided Document Extraction
Anthropic's unreleased model makes significant progress on the Riemann hypothesis
Why AI Agents Aren't as Skilled as They Seem: New Research Highlights Inconsistency Problem
Survey Maps the Multimodal Frontier: A New Taxonomy of Safety and Vulnerability in Multimodal LLMs
[launches]
35OpenAI Launches ChatGPT Desktop App for Linux, Covering Every Major Desktop OS
The Billion-User AI Chatbot Race Is Over — Now the Real Work Begins
Lumabri Lets You Run Huge Mixture-of-Experts Models Across a Swarm of Peer Machines
OpenAI and Jony Ive’s AI device is a doughnut-shaped smart speaker without a screen
[models]
16Meta doubles down on open-weight AI with Muse Glimmer release and Zuckerberg manifesto
Meta's Glimmer model brings 'personal superintelligence' to your laptop
Escha Labs ships 2-bit Qwen3.6-35B-A3B build that runs on a 24GB GPU
ByteDance Is Training a 10 Trillion-Parameter Model, Leaning Into Scale to Chase Anthropic
[funding]
14River AI raises $1.1B to rebuild AI from scratch with personally trainable agents
OpenAI Acquires NextSlide to Bolster ChatGPT’s Presentation Capabilities
SpaceX's AI pivot pays off: compute deals with Anthropic and Google fuel revenue surge, but losses persist
Prentis, a new AI lab co-founded by Reid Hoffman and Mark Pincus, seeks $100M for computer-use agents
Get the signal, not the noise.
One short email when it matters. No recaps of recaps.