Greetings,

The 4th quarter of 2023 is going to be interesting. Google is getting ready to release their state-of-the-art Gemini model which apparently could dethrone GPT-4. If they manage to pull it off, it’ll change the AI landscape significantly.

• While Google has stalled so far, it’s still the most compute-rich company in the world and it’s getting ready to blow OpenAI out of the water with their new model, Gemini. - SemiAnalysis
  
  
• Build an LLM-based app without coding using LLMStack – it’s open-source, you can chain multiple models, integrate your own data, and deploy to the cloud or on-premise. - GitHub
  
  
• “This is an amazing time to be building platforms in AI infrastructure” – explore the emerging AI model ecosystem, and the many opportunities that have opened up because of it. - Lightspeed
  
  
• LegalBench is a new benchmark for LLM performance on legal reasoning tasks – created in collaboration with legal professionals, it will help researchers turn LLMs into lawyers. - LegalBench
  
  
• Demand for AI servers is expected to grow 50% / year in the next 5 years, but hardware shortages are a bottleneck – NVIDIA chips are hard to come by, and that’s just one of many supply chain issues involved. - Financial Times
  
  
• There is no established framework for analyzing AI threats, so cybersecurity expert Daniel Miessler created ATHI – Actor, Technique, Harm, Impact. - Daniel Miessler
  
  
• “The AI failed to note the irony” – 37 content farms have been using AI rewrite and publish content from major news publications, and it’s a gray area where nobody really knows how to think about or solve this issue. - NewsGuard

----

Found this helpful? Forward this email to a colleague.

Was this email forwarded to you? Sign up here to stay on top of the AI news.