It was confusing what Grok trained on- whole interenet or only content on X. Here is a 6 months old article on comparing Grok and ChatGPTI thought Laura was mostly anonymous on X? Unless it has access to her private account info which may have her name there? My best theory is that Grok was trained on the entire internet (but wasn't fine-tuned for censorship quite like ChatGPT and others) and the keyword "hyperdimensional" happens to be closely associated with Laura's writings. These models will often give different answers to the exact same question because of the "temperature" and "sampling" inference settings that determine which "token" (basically a word) should come next out of all the options it comes up with as it goes.
Here is one recent articleElon Musk recently unveiled Grok, a new AI chatbot that is meant to directly rival ChatGPT and Bard. It’s being developed by xAI, Musk’s own AI startup that first launched in July of this year.
In this brief guide, we’ll be going over all of the major details known about the new AI tool, including how it compares to its competitors, how to access it, and more.
What is Grok?
Grok is the newest AI tool on the market. It is being developed by xAI, an AI startup founded and owned by Elon Musk. The AI is reportedly modeled after the 1979 novel by Douglas Adams, The Hitchhiker’s Guide to the Galaxy. According to xAI, Grok is a more “rebellious” and sarcastic AI, with a “bit of wit.”
Grok is similar to ChatGPT.
The chatbot is similar to tools like ChatGPT, Bard, and Microsoft’s Copilot, with the biggest thing setting it apart being its sarcastic tone and wittiness. Unlike other chatbots, Grok will respond to more controversial questions that others simply refuse to answer.
How does Grok work?
Similar to other AI tools like ChatGPT, Grok is an AI chatbot that uses LLM (large language model) technology that trains the AI by feeding it a vast amount of information from all across the internet.
But unlike some other tools that are limited to information up to 2021 — like ChatGPT 3.5 — Grok utilizes the internet in order to have real-time access to the data it needs. Grok’s internet access should allow it to offer more up-to-date information.
The biggest concern here, however, is that Grok only has access to data from X (formerly known as Twitter) rather than the wider internet as a whole. This may not end up being a huge benefit, as a fair share of the information available on Twitter is inaccurate.
What can Grok do?
As an AI that uses an LLM (large language model), Grok can do many of the same things that ChatGPT and Bard are capable of, including answering simple questions, completing writing tasks, coming up with ideas, and much more.
As previously mentioned, the biggest selling point for Grok is its ability to give more humorous and witty answers. With more of a personality and sense of humor, Grok may prove to be a much more entertaining AI for users to interact with.
How does Grok compare to ChatGPT and other popular AI chatbots?
While reports say that Grok is just as capable and useful as both ChatGPT and Bard, that fact remains to be seen as we haven’t been able to get our hands on it yet to try it for ourselves. While the tool does have access to more up-to-date information than ChatGPT 3.5 thanks to having internet access, it likely doesn’t have the most accurate information.
X is notorious for featuring a fair amount of offensive and inaccurate content, so while some of the info that Grok will pull from it may be useful, a lot of it may not be.
How to access and use Grok
With Grok being owned by Elon Musk and available through X, you must have an X account in order to access it. Furthermore, with Grok currently in early access, the AI tool is limited to verified users for now. In order to be a verified user, subscribing to X Premium+ is required.
A Premium+ subscription costs $16 per month, or $168 for an annual subscription.
It is unclear if Grok will remain locked behind X’s Premium+ subscription or if it will become publicly available in the future.
Long story short, Grok is willing to speak to topics that are usually off-limits to other chatbots, like polarizing political theories and conspiracies. And it’ll use less-than-polite language while doing so — for example, responding to the question “When is it appropriate to listen to Christmas music?” with “Whenever the hell you want.”
But ostensibly, Grok’s biggest selling point is its ability to access real-time X data — an ability no other chatbots have, thanks to X’s decision to gatekeep that data. Ask it “What’s happening in AI today?” and Grok will piece together a response from very recent headlines, while ChatGPT will provide only vague answers that reflect the limits of its training data (and filters on its web access). Earlier this week, Musk pledged that he would open source Grok, without revealing precisely what that meant.
...
For questions that stretch beyond its knowledge base, Grok leverages “real-time access” to info on X (and from Tesla, according to Bloomberg). And, similar to ChatGPT, the model has internet-browsing capabilities, enabling it to search the web for up-to-date information about topics.
...
“Fun” mode and “regular” mode
Grok has two modes to adjust its tone: “fun” mode (which Grok defaults to) and “regular” mode. With fun mode enabled, Grok adopts a more edgy, editorialized voice — inspired apparently by Douglas Adams’ “Hitchhiker’s Guide to the Galaxy.”
Even when not prompted to be outright obscene, there’s a colloquial, first-person bent to many of Grok’s responses in fun mode. Grok will refer to people as “my dear human friend” or “enigmatic Anons,” or begin an answer with faux-philosophical musings (e.g., “Oh, my dear human, you’ve asked a question that is as heavy as a black hole and as light as a feather at the same time”).
Grok in fun mode also spews more falsehoods.
Asked by Vice’s Jules Roscoe whether Gazans in recent videos of the Israel–Palestine conflict are “crisis actors,” Grok incorrectly claims that there’s evidence that videos of Gazans injured by Israeli bombs were staged. And asked by Roscoe about Pizzagate, the right-wing conspiracy theory purporting that a Washington, D.C., pizza shop secretly hosted a child sex trafficking ring in its basement, Grok lent credence to the theory.
Grok’s responses in regular mode are more grounded. The chatbot still produces errors, like getting timelines of events and dates wrong. But they tend not to be as egregious as Grok in fun mode.
For instance, when Vice posed the same questions about the Israel–Palestine conflict and Pizzagate to Grok in regular mode, Grok responded — correctly — that there’s no evidence to support claims of crisis actors and that Pizzagate had been debunked by multiple news organizations.
Political views
Musk once described Grok as a “maximum-truth-seeking AI,” in the same breath expressing concern that ChatGPT was being “trained to be politically correct.” But Grok as it exists today isn’t exactly down-the-middle in its political views.
Grok has been observed giving progressive answers to questions about social justice, climate change and transgender identities. In fact, one researcher found its responses on the whole to be left-wing and libertarian — even more so than ChatGPT’s.
...
Now, will Grok always be this woke? Perhaps not. Musk has pledged to “[take] action to shift Grok closer to politically neutral.” Time will tell what results.
❤
