TII’s Falcon-H1R 7B Outperforms 47B Models on Math Reasoning While Running on a 16GB Laptop
A 7-billion parameter model just scored 88.1% on AIME-24 math reasoning, beating models with 47 billion parameters. The parameter count arms…
UiPath Screen Agent Hits 53.6% on OSWorld Benchmark—First Enterprise RPA Tool to Claim #1 Ranking for Agentic Automation
UiPath just outperformed OpenAI’s own agents on the most rigorous test of autonomous computer operation—and they did it by swapping…
OpenAI Signs $10 Billion Cerebras Deal for 750 Megawatts of AI Inference Infrastructure Through 2028
OpenAI just mass-ordered 750 megawatts of specialized silicon that isn’t made by NVIDIA. The $10 billion Cerebras deal signals that the…
Google’s 12B TranslateGemma Outperforms Its Own 27B Model: Open Translation Hits 55 Languages with MetricX Score of 3.60
Google’s smaller translation model just beat its larger sibling on standardized benchmarks, forcing us to reconsider everything we…
Sequoia Backs Anthropic’s $25B Round at $350B Valuation—While Still Funding OpenAI and xAI
Sequoia just broke Silicon Valley’s oldest rule: never fund your portfolio’s competitors. Their bet on Anthropic—while backing…
xAI Raises $20 Billion at $230 Billion Valuation, Plans 50 Million H100-Equivalent GPU Deployment by 2030
xAI just raised more in one round than most AI companies are worth—and plans to deploy 50× more compute than it has today. The math is…
Google Launches Universal Commerce Protocol: Open Standard Backed by Shopify, Walmart, Visa, and Mastercard Enables AI Agents to Execute Purchases Across Any Retailer Without Custom Integrations
Six days ago, Google announced the infrastructure layer that makes AI-powered shopping actually work at scale—and they convinced Visa,…
Illumina Launches 1 Billion Cell Atlas: 20 Petabytes of CRISPR Data Built with AstraZeneca, Merck, and Lilly to Train Next-Gen Drug Discovery AI
Illumina just gave AI drug discovery something it’s never had: ground truth data for what happens when you toggle every human gene on…
TII’s Falcon-H1R 7B Outperforms 47B Models on Math Reasoning While Running on a 16GB Laptop
A 7-billion parameter model just scored 88.1% on AIME-24 math reasoning, crushing NVIDIA’s 47B Nemotron at 49.7%. The assumption that…
Meta Announces $600 Billion AI Infrastructure Spend Through 2028, Creating ‘Meta Compute’ Division to Build Tens of Gigawatts of Data Center Capacity
Meta just committed more capital to AI infrastructure than the entire GDP of Finland, Belgium, or Thailand combined—and they hired a former…
