Models & Research
-
Mystery AI ‘gpt2-chatbot’ Rivals GPT-4
A new AI model, ‘gpt2-chatchat’, surprises experts by rivaling GPT-4’s abilities, leading to speculation about its origin and impact.
-
Snowflake Unveils Arctic, Cutting-Edge Open LLM
Snowflake launches ‘Arctic’ LLM, boasting competitive performance and efficiency. Available as open-source, it targets enterprise AI needs.
-
Apple’s OpenELM Innovates On-Device AI Capabilities
Apple released OpenELM, a range of on-device LLMs with 270M to 3B parameters, outperforming AI2’s OLMo and broadening AI accessibility.
-
DeepMind’s LLMs Master Tasks with Expanded Contexts
Google DeepMind research allows LLMs to process inputs equivalent to several books, enhancing AI capabilities and attracting investor interest.
-
Microsoft’s Phi-3: Big Power in a Tiny AI Package
Microsoft released Phi-3, a 3 billion parameter model, enabling affordable and responsible AI deployments with significant business impact.
-
AI Agents Evolve: Smarter Planning and Execution
The survey highlights AI agent advancements in reasoning, planning, and tool execution. Single and multi-agent systems excel in different tasks.
-
Dify: Revolutionizing AI Deployment Across Industries
Dify integrates various language models, tools, and deployment options, enhancing AI accessibility, productivity, and market potential for businesses.
-
Reka Core Surpasses GPT-4 in Record Time
Reka Core, developed by a team from DeepMind, Google, and Meta, competes with leading AI models, with implications on markets and innovation.
-
Meta Launches Llama 3: A New Era of Open AI
Meta releases Meta Llama 3, a significant improvement in LLMs with 70B parameters, broad platform integration, and advanced safety features.
-
STORM Enhances Wikipedia-Like Article Creation
The STORM Writing System revolutionizes pre-writing research for Wikipedia-like articles, with potential for improved educational tools and AI investment.
Eksentricity 