• Seattle Skeptics on AI
Seattle Skeptics on AI

Tag: AI capabilities

Parameter Counts in Large Language Models: Why Size and Scale Matter for Capability
Parameter Counts in Large Language Models: Why Size and Scale Matter for Capability

Tamara Weed, Dec, 20 2025

Parameter count in large language models determines their reasoning power, knowledge retention, and task performance. Bigger isn't always better-architecture, quantization, and efficiency matter just as much as raw size.

Categories:

Science & Research

Tags:

large language models parameter count LLM size model scaling AI capabilities

Recent post

  • Multi-GPU Inference Strategies for Large Language Models: Tensor Parallelism 101
  • Multi-GPU Inference Strategies for Large Language Models: Tensor Parallelism 101
  • Implementing Generative AI Responsibly: Governance, Oversight, and Compliance
  • Implementing Generative AI Responsibly: Governance, Oversight, and Compliance
  • Access Controls and Audit Trails for Sensitive LLM Interactions: How to Secure AI Systems
  • Access Controls and Audit Trails for Sensitive LLM Interactions: How to Secure AI Systems
  • Chain-of-Thought in Vibe Coding: Why Explanations Before Code Work Better
  • Chain-of-Thought in Vibe Coding: Why Explanations Before Code Work Better
  • Trustworthy AI for Code: How Verification, Provenance, and Watermarking Are Changing Software Development
  • Trustworthy AI for Code: How Verification, Provenance, and Watermarking Are Changing Software Development

Categories

  • Science & Research

Archives

  • March 2026
  • February 2026
  • January 2026
  • December 2025
  • November 2025
  • October 2025
  • September 2025
  • August 2025
  • July 2025

Tags

vibe coding large language models AI coding tools prompt engineering generative AI LLM security AI compliance AI governance AI coding transformer models AI code security GitHub Copilot AI development LLM deployment AI coding assistants prompt injection AI code vulnerabilities GPU utilization LLM optimization AI agents

© 2026. All rights reserved.