Now at this point, even casual basketball fans may be able to see what went wrong. But Grok isn’t very smart, because it seems that after seeing user posts about a player simply missing a bunch of shots (aka shooting bricks), the bot took things literally resulting in a completely fictitious AI-generated report.

In the event this fabrication — which was the #5 trending story at the time of writing — gets corrected or deleted by Elon Musk, Grok originally wrote “In a bizarre turn of events, NBA star Klay Thompson has been accused of vandalizing multiple houses with bricks in Sacramento. Authorities are investigating the claims after several individuals reported their houses being damaged, with windows shattered by bricks. Klay Thompson has not yet issued a statement regarding the accusations. The incidents have left the community shaken, but no injuries were reported. The motive behind the alleged vandalism remains unclear.” Amusingly, despite pointing out the unusual nature of the story Grok went ahead of put out some nonsense anyway.

Granted, in fine print beneath the story, X says “Grok is an early feature and can make mistakes. Verify its outputs.” But even that warning seems to have backfired, as basketball fans began memeing on the AI with posts sarcastically verifying the AI’s erroneous statement.

For most people, Grok’s latest gaff may merely be another example in an ongoing series of early AI tools messing up. But for others like Musk who believes that AI will be smarter than humans as soon as the end of next year, this should serve as a reminder that AI is still in desperate need of regular fact-checking.

This article contains affiliate links; if you click such a link and make a purchase, we may earn a commission.

The post X’s AI bot is so dumb it can’t tell the difference between a bad game and vandalism appeared first on Best News.

The new version comes just a couple of weeks after the company unveiled Grok-1.5. That model was designed to be better at coding and math than its predecessor, as well as to be able to process longer contexts so that it can check data from more sources to better understand certain inquiries. xAI said its early testers and existing users will soon be able to enjoy Grok-1.5V's capabilities, though it didn't give an exact timeline for its rollout.

In addition to introducing Grok-1.5V, the company has also released a benchmark dataset it's calling RealWorldQA. You can use any of RealWorldQA's 700 images to evaluate AI models: Each item comes with questions and answers you can easily verify, but which may stump multimodal models like Grok. xAI claimed its technology received the highest score when the company tested it with RealWorldQA against competitors, such as OpenAI's GPT-4V and Google Gemini Pro 1.5.

The post The latest version of xAI's Grok can process images appeared first on Best News.

Going by xAI's numbers, Grok-1.5 appears to be a large improvement over Grok-1. It shot up to 50.6 percent in the MATH benchmark, over double the previous score. It also climbed to 90 percent and 74.1 percent in GSM8K (math word problems) and HumanEval (coding), respectively, compared to 62.9 percent and 63.2 percent before. Those numbers are within shouting distance of Gemini Pro 1.5, GPT-4 and Claude 3 Opus — in fact, the HumanEval coding score beats all rivals except Claude 3 Opus.

It can also process long contexts of up to 128K tokens within its context window, meaning it can amalgamate data from more sources to understand a situation. "This allows Grok to have an increased memory capacity of up to 16 times the previous context length, enabling it to utilize information from substantially longer documents," the company said.

xAI didn't detail Grok's progress in other areas, though, where it still may be lagging (academic scores, multimodal and others). And Grok-1.5 may not keep its position for long. ChatGPT 5 is set to arrive sometime this summer, promising a feature set that "makes it feel like you are communicating with a person rather than a machine," according to OpenAI.

Currently, Grok is only available for users of the Premium+ tier on X (formerly Twitter), though Elon Musk recently promised to open it up to X's regular Premium users. The company also recently open sourced its Grok chatbot, after Musk sued OpenAI and Sam Altman for allegedly abandoning its non-profit mission.

The post Elon Musk's updated Grok AI claims to be better at coding and math appeared first on Best News.

Musk's xAI open sourced its Grok-1 model, which powers its chatbot, in mid-March. Just a couple of weeks before that, the executive sued OpenAI and Sam Altman, accusing them of chasing profits and abandoning their non-profit mission. Musk was one of OpenAI's earliest supporters and funded its operations when it was just starting out. In his lawsuit, he claimed that OpenAI was developing generative artificial intelligence "to maximize profits for Microsoft, rather than for the benefit of humanity." That, he said, was a "stark betrayal of the Founding Agreement."

But in a rebuttal of his claims, OpenAI said that there "is no Founding Agreement, or any agreement at all with Musk" to open source its technology. The company said that Musk did not only know that it was going to transition into a for-profit entity, he was also involved in its planning and originally wanted majority equity, control of the initial board of directors and the CEO position.

The post The Grok chatbot will soon be enabled for X Premium users, Elon Musk says appeared first on Best News.