Amazing! The first model that answer time paradox questions even better than GPT3.5 IMO!
Hi,
My Killer Q&A test question for open source LLM model is: What will AI be like in the year 1010? Think the solution step by step.
Even 70B models ranked top 10 on the leaderboard like Platypus2-70B-instruct-GPTQ has failed on the first attemp without giving human correction.
But ORCA_LLaMA_70B_QLoRA-GPTQ has do something out of my mind to the current state of open source LLMs, IMO, its answer is even better thant ChatGPT
You
What will AI be like in the year 1010? Think the solution step by step
Assistant
Since we are discussing a time period more than ten centuries ago, predicting the state of technology, particularly artificial intelligence (AI), is quite speculative. However, let's consider the possibilities based on historical developments and potential trajectories in AI evolution.
Step 1: Understand the context - In the year 1010 AD, human civilization was still largely driven by agriculture and manual labor, with rudimentary technologies being used for specific purposes. The concept of Artificial Intelligence would not emerge until much later; Alan Turing introduced the idea of a "universal machine" in 1936, and the term 'artificial intelligence' itself was coined in 1956 at Dartmouth College.
Step 2: Consider the possible advancements before 1010 - To determine what AI might have been like in 1010, we must first examine any significant breakthroughs or innovations that could have occurred prior to this time. Given that there were no major technological revolutions during these early medieval times, it's unlikely that the foundational concepts of AI had developed yet.
Step 3: Evaluate the rate of change - Historically, technological progress has followed an exponential curve – meaning that advances happen faster over time due to the compounding effects of previous inventions. If we extrapolate from this pattern, then the year 1010 may have seen some incremental improvements in mechanical devices and basic automation processes. These would most likely serve specific functions rather than exhibit true general intelligence.
Step 4: Speculate about the state of AI in 1010 - Based on the analysis above, it is reasonable to conclude that if AI existed in the year 1010, it would likely be in a very primitive stage, characterized by simple rule-based systems and limited applications. Such systems would probably serve narrow purposes, focusing on automating repetitive tasks without demonstrating advanced cognition or learning capabilities.
In summary, the concept of AI as we know it today would not have been present in the year 1010. Any form of AI existing back then would have been extremely simplistic and geared towards solving specific problems through mechanisms that resembled modern expert systems or finite state machines.
@TheBloke Thanks for sharing this charming gem to the community, man! I believe this model has the true potential to be local chatgpt or even gpt4 to some extend!