#Large Language Models
AIGemini 3.5 Pro Targets July 17 - and Google Rebuilt It From Scratch to Get There
Gemini 3.5 Pro targets July 17 with a 2M token context window. Google rebuilt it from scratch. DeepSeek's deepseek-chat and deepseek-reasoner aliases stop responding July 24.
AIWhat Is Loop Engineering - and Why It Replaced Prompt Engineering in 2026
Loop engineering is how developers design AI agents that run without supervision. It's the skill that replaced prompt engineering as the real bottleneck in 2026.
AIUN Convenes 193 Countries in Geneva for First Global AI Governance Dialogue
UN AI governance talks opened July 6 in Geneva. 193 countries. Big Tech CEOs. A scientist warning of catastrophic harm. No binding rules from this week.
AIAmazon Mechanical Turk Closes to New Customers July 30 After 21 Years
Amazon is closing Mechanical Turk to new customers July 30. No new features planned. Existing users keep access to the 21-year-old crowdsourcing platform.
AILongCat-2.0: Meituan Open-Sources a 1.6T Coding Model Trained Entirely on Chinese Chips
LongCat-2.0 is a 1.6T open-source coding model from Meituan, trained on 50,000 domestic Chinese ASICs. MIT license, 1M context, SWE-bench Pro 59.5.
AIGLM-5.2: 1M Tokens, MIT License, and Better Coding Scores Than GPT-5.5 at One-Sixth the Cost
GLM-5.2 is the best open-weight AI model on earth right now. MIT license, 1M-token context, and coding benchmarks that beat GPT-5.5 at one-sixth the cost.
AIMicrosoft Frontier Is a $2.5 Billion Bet on Enterprise AI Deployment With 6,000 Engineers
Microsoft launched Frontier Company today - 6,000 engineers and $2.5 billion focused on getting enterprise AI past the proof-of-concept stage.
AIFable 5 Returns With 50% Plan Limits and a Safety Classifier That May Block Your Code
Fable 5 returned July 1 with 50% plan limits through July 7, then usage credits. A new safety classifier may silently switch your coding sessions to Opus 4.8.
StartupsReflection AI Is Paying SpaceX $150 Million a Month. No Model Has Shipped.
Reflection AI started paying SpaceX $150 million a month on July 1 for Nvidia GB300 access at Colossus 2. No model has shipped yet.
TechMeta Is Building a Cloud Business to Sell AI Compute, and CoreWeave Just Took the Hit
Bloomberg reported July 1 that Meta plans a cloud business to sell surplus AI compute. Meta stock jumped 10%. CoreWeave dropped 14%. Nebius 16%.
AIClaude Sonnet 5 vs Opus 4.8: Benchmarks, Pricing, and Which One to Run
Claude Sonnet 5 vs Opus 4.8 - benchmarks, real deployment data, and pricing through the August 31 deadline. Here is which model to run and why.
AIClaude Fable 5 Returns July 1 After US Lifts Export Controls on Anthropic Models
Anthropic restores Claude Fable 5 on July 1 after the US Department of Commerce lifted export controls on both Fable 5 and Mythos 5 on June 30, 2026.