xAI launches 'smartest AI in the world'

Quick Note

Sorry for missing your inbox yesterday — I’m still working out some kinks with this newsletter, including the referral program mentioned last time. Everything should be good to go by Tuesday!

xAI launches Grok 4, claims it’s the ‘smartest AI in the world’

Elon Musk’s xAI has launched Grok 4, which the company claims is the most powerful AI model in the world. 

Key background: Grok has made headlines for all the wrong reasons lately. When xAI launched, Musk promised an “anti-woke” and truth-seeking chatbot, but in recent weeks that’s taken the form of antisemitic remarks. 

The rundown:

• The base Grok 4 model reportedly sets new industry records across major benchmarks.* Musk says it’s smarter than almost every graduate student on Earth.

• Demo videos show the model solving advanced math problems, generating an image of two blackholes colliding and predicting the winner of the World Series. 

• xAI also launched a “heavy” version of Grok 4 that allows multiple AI agents to work together on especially complex problems. 

The bigger picture: Grok 4 is xAI’s answer to OpenAI’s upcoming flagship model GPT-5, which is expected to launch this summer. Despite all of the controversies surrounding Musk and xAI, the company continues to put out highly advanced models that compete with industry leaders.

How do we measure AI models’ intelligence?

AI models are evaluated using benchmarks — standardized tests that assess performance on tasks like language understanding, reasoning, coding, math, etc. Models are scored based on accuracy, speed and their ability to solve tasks outside their training data. The best models perform well across many benchmarks and real-world tasks. 

OpenAI is testing a ‘study together’ feature in ChatGPT

Some ChatGPT subscribers have noticed a new tool called “study together” appearing in their drop-down menus. The feature is designed to make the chatbot a better educational tool. 

The details:

• The mode turns ChatGPT into a more interactive study partner. Instead of just telling you the right answer, it will ask questions and help users solve the problem on their own. 

• Study together is OpenAI’s answer to Google’s LearnLM, a similar tool that lets students learn and practice curriculum without receiving direct answers. 

• The name also suggests that multiple users may be able to join the same session, although OpenAI has not disclosed any additional details about the tool. 

Why it matters: AI is a double-edged sword in education. Students can use it like a tutor — or they can use it to write entire papers for them. This is a way for ChatGPT to encourage the “good” uses while discouraging cheating. 

Reality check: As a recent college graduate, I can confidently say that most students who use AI are using it to cheat. This feature looks great on paper, but it won’t solve the AI cheating dilemma. 

More trending news

Elon Musk says Grok will come to Tesla vehicles next week 

OpenAI reportedly plans to release an AI-powered browser to compete with Google Chrome

YouTube prepares to crack down on mass-produced AI-generated content

Thanks for reading! As you can see, this newsletter’s design is still a work in progress. My goal is to make these emails as easy to read as possible — if there’s anything I can do for you to achieve that goal, let me know!

See you next week,

Chain of Thought