AI NEWS: Meta just tried poaching from Mira Murati's new startup, with offers going up to $1B. Plus, more news from OpenAI, Microsoft, Google, Ideogram, Cursor, Writer, and Moonvalley. Here's everything you need to know:
LLMS
-
Qwen3-30B-A3B-Instruct MLX 8-bit Model Released
By
–
https://
huggingface.co/lmstudio-commu
nity/Qwen3-30B-A3B-Instruct-2507-MLX-8bit
… in LM Studio, its 32GB -
Smart Mode on Copilot rolling out with GPT-5
By
–
The time has come! Some users started receiving Smart Mode on Copilot. Earlier, GPT-5 was spotted as powering Smart Mode internally. Considering the earlier announcements, the router part of it could be rolling out already. * I don't have it yet, screenshot is for reference
-
Iterative LLM Reflections Outperform Heavy Reinforcement Learning
By
–
Iterative reflections for LLMs can outperform heavy RL? This paper shows that having the LLM reflects on its own trajectories, rewrite its own prompts, and evolve a diverse pool of candidates beats RL w/ GRPO so far on four reasoning tasks . 10% improv with 35x fewer rollouts!
-
LLM Stock Market Results: Statistical Bias in Spectacular Outcomes
By
–
If you let 1000 LLMs play the stock market, half of them will yield positive returns, and a few will yield spectacular ones. The same is at play in all the other amazing LLM results you see.
-
AI Copilots for Insurance Underwriting: Enterprise Readiness Benchmark
By
–
Can an AI copilot underwrite insurance? We built a benchmark with real underwriters to find out. Top models struggled with tool use, hallucinations, and basic reasoning. See what it takes to build enterprise-ready AI https://
snorkel.ai/blog/evaluatin
g-ai-agents-for-insurance-underwriting/
… #AI #InsuranceTech #SnorkelAI -
ReAcT Pattern Experience: Prompt Engineering Techniques
By
–
Yeah I had fun with the ReAcT pattern for that a couple of years ago
-
Training weights in machine learning models explained
By
–
sorry, by train i mean actually updating weights, but i appreciate the comment!
-
Training and Fine-tuning Models: Community Discussion Opportunity
By
–
If you train/fine-tune models, I'd love to chat with you. Please comment or DM and I'll reach out!