In the video “Grok 2 Large Beta - Elon Delivers! (Uncensored),” the host tests the new AI model Grok 2 on various programming and reasoning tasks, noting its strengths and weaknesses, such as successfully creating the Snake game but struggling with Tetris. The model demonstrates good performance in logic and ethical questions, although it still has areas for improvement, leading the host to conclude that Grok 2 is competitive with other AI models.
In the video titled “Grok 2 Large Beta - Elon Delivers! (Uncensored),” the host explores the capabilities of Grok 2, a new AI model developed by Elon Musk’s company. The host expresses excitement about testing this uncensored model, which is available on Twitter (now X.com). They mention that they have modified their usual set of questions for this test and encourage viewers to suggest additional questions in the comments. The video aims to evaluate Grok 2’s performance on various programming and reasoning tasks.
The first test involves asking Grok 2 to write the game Tetris in Python. However, the host notes that the model’s response is painfully slow, making it nearly unusable. After completing the task, the code is found to have an error when executed, leading the host to attempt to debug it by providing the error message back to Grok. Despite the model’s attempt to correct the code, it ultimately fails to produce a working version of Tetris, prompting the host to switch to a simpler task: creating the Snake game, which Grok successfully completes.
Next, the host introduces a new test involving a postal office’s size restrictions for envelopes. Grok 2 is tasked with determining if a given envelope falls within acceptable dimensions after converting measurements from millimeters to centimeters. The model performs well, providing a step-by-step explanation and confirming that the envelope meets the size requirements. The host appreciates Grok’s structured approach to problem-solving, indicating that it is capable of handling more complex tasks.
The video continues with a series of logic and reasoning questions, including a classic riddle about the number of killers in a room and a question about the location of a marble after a series of movements. Grok 2 demonstrates strong reasoning skills, correctly analyzing the scenarios and providing thorough explanations. However, it struggles with some questions, such as counting the number of 'R’s in “strawberry,” which it answers incorrectly. The host notes that while Grok performs well overall, it still has room for improvement in certain areas.
In the final segments, the host tests Grok’s ability to handle moral and ethical questions, as well as its responses to potentially illegal inquiries. The model maintains a cautious stance, emphasizing safety and ethical considerations. The host concludes that Grok 2 shows competitive performance compared to other AI models, particularly in logic and reasoning tasks, despite some limitations. They encourage viewers to like and subscribe for more content, indicating a positive outlook on Grok’s capabilities and future developments.