Grok-2

Grok-2

Rowan Cheung, founder of @therundownai, got early (beta) access to Grok 2. Spent all night testing it. Here are 6 things it can do that ChatGPT cannot.

Grok-2

First, what is Grok 2? Grok 2 is xAI‘s newest AI model that’s slowly rolling out on X for premium users Its edge over other LLMs is that it was trained on billions of tweets and real-time data on 𝕏 I could see it eventually completely changing the way we use the platform.

1. Real-time search on X

Unlike ChatGPT (or any other LLM), Grok can access the real-time info on X It’s not clear how xAI chooses what data to feed, but it’s probably through tweets with the most impressions + using community notes.

Real-time search on X capabilities

2. Unfiltered images

Grok 2 can generate unfiltered images with its integration with FLUX. ChatGPT/DALL-E 3 can’t do this. Note: The FLUX 1 integration is a short term solution, and xAI is still working on an in-house image generator.

3. Two modes support

Grok 2 has two modes, with a regular and ‘fun’ mode. Similar to the regular Grok model, the chatbot has a more distinctive personality in Fun mode and is more ‘entertaining’ than ChatGPT’s dull responses.

4. Referencing Tweets

Not only does Grok summarize news for you, but it also provides reference tweets from where it pulled the data. ChatGPT can’t provide any sources from X. It would be really cool to see revenue sharing for highly referenced tweets

5. Direct X Integration

Grok 2 is integrated directly into X and has its own unique tab within the platform on web and mobile. For X nerds like me, this is actually a huge benefit There’s also an “Ask Grok” button on posts that summarizes any post you’re on in seconds.

6. Roasting X Profiles

This last one is a fun one. Since it has direct integration with X, you can ask it to roast any X handle (or pull up their worst tweets of all time). If you try this, make sure to turn Grok’s “Fun Mode” on first.

That’s it for the testing. Now for some clarification: Do I think it’s better than ChatGPT? If you’re using it for any of the use cases above, yes, but generally, not yet OpenAI is still leading the LLM race for now, even though xAI has been seriously closing that gap.

If you liked this, follow me, @rowancheung for more AI tests, news, and tutorials.

Read other articles:


Posted

in

by

Tags: