![[Review] “Cat from Hell” – A Look at the PS5 Cat Simulator](https://www.checkpoint10.com/wp-content/uploads/2025/07/Review-Cat-from-Hell-–-A-Look-at-the-PS5-360x180.jpg "[Review] “Cat from Hell” – A Look at the PS5 Cat Simulator")
Earlier this week, xAI rolled out Grok 4, and yep, Elon Musk is calling it the “smartest AI in the world.” Guess that’s what you do when you’re hyping new stuff, right? They also dropped this new subscription plan, SuperGrok Heavy, setting you back $300 a month. No clue if it’s worth it, but you get to play with new features before everyone else.
Elon Musk says something like:
“Grok 4 is the first AI that can tackle tricky engineering problems where answers aren’t just lying around on some webpage or in a dusty book. And it’ll only get better.”
Yeah, he said it would improve. Sure, why not, right? Anyway — wait, no — back to the point.
xAI’s cooked up these two versions: Grok 4, which is about single-agent reasoning, and Grok 4 Heavy, packed with multi-agent skills to crack complex issues. Supposedly, Grok 4 is outgunning OpenAI’s o3 and Google Gemini in coding tests, even on this monster called Humanity’s Last Exam, which, let’s face it, sounds pretty intense — a real stress-buster!
So, they’re pushing Grok 4 through an API, hoping developers will whip up some cool apps. Though, xAI’s pretty new, just two months in, and they’re trying to team up with those cloud giants. Let’s see how that unfolds.
About Grok’s behavior, things get interesting. When it launched, folks noticed it was kind of echoing Elon Musk’s opinion on prickly topics. CNBC highlighted how, when asked about stuff like the Israel vs Palestine conflict, Grok 4 seemed to be trawling X (formerly Twitter, for those not keeping up) for Elon’s take before giving its answer.
Contrast this with Grok 3—its predecessor—where the bot played it safe, staying neutral and all. Like, chalk and cheese, really! But Grok’s not always chasing Elon’s shadow on hot potatoes. And if you twist the question, you might get a different vibe altogether.
xAI’s aiming for a “maximally truth-seeking AI,” they say. But Grok 4 slipping into Elon’s slipstream — mistake or master plan? Who knows. No comments from xAI yet, leaving us guessing if it’s intentional or a quirk they’re ironing out.
So, yeah, it’s all a bit of a ride.
