Terfi

@aadilthecoder.bsky.social

Software developer, medical doctor, pianist

Snapshot isolation gives each transaction a consistent view of the database, reducing blocking and non-repeatable reads—great for most OLTP workloads! 🚦 But watch for update conflicts: if two transactions change the same data, only one wins at commit time. #databases #isolation

HMD’s new Key phone brings AI to the budget market! At just £59, this Android 14 Go device offers AI-powered camera modes like Portrait and Night mode. Virtual memory boosts performance, while 2-day battery life keeps you going. Affordable tech has never been smarter! #HMDKey #AIonaBudget

No surprise that AI is better at vignettes with multiple choice answers. That is not clinical medicine. Real medicine is nuanced. What questions to ask, interpreting a response of “not really”, building trust so a patient shares “the rest of the story”. And that’s just part of step 1, the history

Scott McGrath@smcgrath.phd · 2y ago

🧪 While LLMs have done well in medical skill comparisons with multiple choice questions, a new study finds GPT-4 scored lower than human doctors in free-text assessments of primary care cases, highlighting its limitations in complex clinical reasoning. 🩺🖥️ #MLSky

🚀 Setting up a local large language model server: Choose a Model: Select a pre-trained LLM (like llama 3.2 on ollama). Install Dependencies: Use Python with libraries like Hugging Face's Transformers. Set Up Environment Deploy: Run the model using FastAPI Enjoy! #AI #MachineLearning #ollama