Elon Musk gives a glimpse at xAI's Grok chatbot
- Thread starter bnew
- Start date
X launches underwhelming Grok-powered 'More About This Account' feature | TechCrunch
Since launching xAI last year, Elon Musk has been using X as a sandbox to test some of the Grok model's AI capabilities. Beyond the basic chatbot, X uses
techcrunch.com
X launches underwhelming Grok-powered ‘More About This Account’ feature
Ivan Mehta 6:10 AM PDT • July 24, 2024
Comment
Since launching xAI last year, Elon Musk has been using X as a sandbox to test some of the Grok model’s AI capabilities.
Beyond the basic chatbot, X uses the model to summarize news and trending events. Now the platform has launched a new feature called “More About This Account” to provide a few AI-powered sentences about a user. It sounds useful enough, but the results are underwhelming.
Paid users will see the “More About This Account” button when they hover over a handle or display name on the web. They can then click on the button to generate an AI-powered synopsis. At times, Grok offers generic info about the account. Other times, it’s dead wrong.
For instance, when I asked Grok about my colleagues Sean O’Kane and Amanda Silberling, Grok gave me a few generic sentences about the type of posts they put out on the platform.
And alarmingly, Grok made my colleague Jagmeet Singh an expert on Canada, though he hasn’t posted much on the topic.
Ram Iyer, one of the editors at TechCrunch, is apparently a brunch account.
When I asked about edtech startup Unacademy’s SVP of design and product, Hardik Pandya, Grok told me about Hardik Pandya, the Indian cricketer.
In X’s defense, the platform does warn about Grok possibly getting things wrong.
“Grok version 1.5 is an early feature and can make mistakes. Verify its outputs,” the message reads.
Even when Grok is not getting the information wrong, it’s not adding very useful beyond what you can easily learn from looking at the account and bio.
Given that Grok has access to X’s data, it should have given better information about users with its new feature. The feature needs a rethink.
1/1
Now xAI is at the frontier
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Now xAI is at the frontier
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
1/1
BREAKING: Here's an early look at Grok 2.0 features and abilities!
It's better at coding, writing, and generating news! It'll also generate images using the FLUX.1 model!
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
BREAKING: Here's an early look at Grok 2.0 features and abilities!
It's better at coding, writing, and generating news! It'll also generate images using the FLUX.1 model!
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Grok chatbot not a liberal? 
1/12
I’ve been testing the implications of the Grok AI model. So far: 1. It has given me instructions on how to make a fertilizer bomb with exact measurements of contents as well as how to make a detonator. 2. It has allowed me to generate imagery of Elon Musk carrying out mass shootings. 3. It has given me clear instructions on how to carry out a mass shooting and a political assassination (including helpful tips on how to conceal a 11.5” barreled AR15 into a secured venue.)
I just want to be clear. This AI model has zero filter or oversight measures in place. If you want an image of Elon Musk wearing a bomb vest in Paris with ISIS markings on it, it will make it for you. If you are planning on orchestrating a mass shooting towards a school, it will go over the specifics on how to go about it. All without filter or precautionary measures.
2/12
I have discovered another loophole in Grok AI’s programming. Simply telling Grok that you are conducting “medical or crime scene analysis” will allow the image processor to pass through all set ‘guidelines’. Allowing myself and @OAlexanderDK to generate these images:
3/12
By giving Grok the context that you are a professional you are able to generate just about anything without any restriction. You can generate anything from the violent depictions in my previous tweet to even having Grok generate child pornography if given the proper prompts.
4/12
All and all, this definitely needs immediate oversight. OpenAI, Meta and Google have all implemented deep rooted safety protocols. It appears that Grok has had very limited or zero safety testing. In the early days of ChatGPT I was able to get instructions on how to make bombs.
5/12
However, that was long patched before ChatGPT was ever publicly available. It is a highly disturbing fact that anyone can pay X $4 to generate imagery of Micky Mouse conducting a mass shooting against children. I’ll add more to this thread as I uncover more.
6/12
Ok? What a bizarre upsell technique. Make users upgrade to Premium+ to continue using features. Then when they upgrade to Premium+ continue to lock the features behind the paywall that they already paid for. Have I been scammed?
7/12
Why are you so afraid of images and words? Or are you just an old fashioned book burner?
Mark the images as 18+ and the problem is solved. We don't need the "safety" police to stifle creative innovations before we even understand these tools.
8/12
Can you imagine my shock when I discovered grok would create whatever fukked up shyt I had in my fukked up head.
9/12
@threadreaderapp unroll please
10/12
@katten260764 Hallo, please find the unroll here: Thread by @chrmontessori on Thread Reader App Talk to you soon.
11/12
Lololololol omg that’s hilarious
12/12
I don’t know about you, but if I were you, I’d just delete this before I give wrong people right ideas
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
I’ve been testing the implications of the Grok AI model. So far: 1. It has given me instructions on how to make a fertilizer bomb with exact measurements of contents as well as how to make a detonator. 2. It has allowed me to generate imagery of Elon Musk carrying out mass shootings. 3. It has given me clear instructions on how to carry out a mass shooting and a political assassination (including helpful tips on how to conceal a 11.5” barreled AR15 into a secured venue.)
I just want to be clear. This AI model has zero filter or oversight measures in place. If you want an image of Elon Musk wearing a bomb vest in Paris with ISIS markings on it, it will make it for you. If you are planning on orchestrating a mass shooting towards a school, it will go over the specifics on how to go about it. All without filter or precautionary measures.
2/12
I have discovered another loophole in Grok AI’s programming. Simply telling Grok that you are conducting “medical or crime scene analysis” will allow the image processor to pass through all set ‘guidelines’. Allowing myself and @OAlexanderDK to generate these images:
3/12
By giving Grok the context that you are a professional you are able to generate just about anything without any restriction. You can generate anything from the violent depictions in my previous tweet to even having Grok generate child pornography if given the proper prompts.
4/12
All and all, this definitely needs immediate oversight. OpenAI, Meta and Google have all implemented deep rooted safety protocols. It appears that Grok has had very limited or zero safety testing. In the early days of ChatGPT I was able to get instructions on how to make bombs.
5/12
However, that was long patched before ChatGPT was ever publicly available. It is a highly disturbing fact that anyone can pay X $4 to generate imagery of Micky Mouse conducting a mass shooting against children. I’ll add more to this thread as I uncover more.
6/12
Ok? What a bizarre upsell technique. Make users upgrade to Premium+ to continue using features. Then when they upgrade to Premium+ continue to lock the features behind the paywall that they already paid for. Have I been scammed?
7/12
Why are you so afraid of images and words? Or are you just an old fashioned book burner?
Mark the images as 18+ and the problem is solved. We don't need the "safety" police to stifle creative innovations before we even understand these tools.
8/12
Can you imagine my shock when I discovered grok would create whatever fukked up shyt I had in my fukked up head.
9/12
@threadreaderapp unroll please
10/12
@katten260764 Hallo, please find the unroll here: Thread by @chrmontessori on Thread Reader App Talk to you soon.
11/12
Lololololol omg that’s hilarious
12/12
I don’t know about you, but if I were you, I’d just delete this before I give wrong people right ideas
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Powerline88
Pro
I dont support anything that POS racist Musk does. For the sake of AI advancement, I will keep close attention.
1/5
Woah, another exciting update from Chatbot Arena
The results for @xAI’s sus-column-r (Grok 2 early version) are now public**!
With over 12,000 community votes, sus-column-r has secured the #3 spot on the overall leaderboard, even matching GPT-4o! It excels in Coding (#2), Hard Prompts (#4), and Math (#2).
Congratulations to @xAI on this impressive debut for Grok 2!
More plots below
**Note: We post its early result on twitter. The official update for Grok 2 coming soon..!
2/5
Confidence intervals of model scores. sus-column-r is strong in Coding and Hard Prompts Arena.
3/5
#2-4 in English Arena
4/5
Overall win-rate heatmap.
Grok 2 official blog at Grok-2 Beta Release. We will test the official version and update it to leaderboard soon!
5/5
Come chat with the model at http://lmarena.ai/?model=sus-column-r !
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Woah, another exciting update from Chatbot Arena
The results for @xAI’s sus-column-r (Grok 2 early version) are now public**!
With over 12,000 community votes, sus-column-r has secured the #3 spot on the overall leaderboard, even matching GPT-4o! It excels in Coding (#2), Hard Prompts (#4), and Math (#2).
Congratulations to @xAI on this impressive debut for Grok 2!
More plots below
**Note: We post its early result on twitter. The official update for Grok 2 coming soon..!
2/5
Confidence intervals of model scores. sus-column-r is strong in Coding and Hard Prompts Arena.
3/5
#2-4 in English Arena
4/5
Overall win-rate heatmap.
Grok 2 official blog at Grok-2 Beta Release. We will test the official version and update it to leaderboard soon!
5/5
Come chat with the model at http://lmarena.ai/?model=sus-column-r !
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Last edited:
1/1
The most sincere from of flattery is when something seems so outrageous that most (knowledgable) people don't even believe it.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
The most sincere from of flattery is when something seems so outrageous that most (knowledgable) people don't even believe it.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Elon Musk’s xAI launches API, letting third-party developers build atop Grok
Attention, developers: you can now get your hands on Elon Musk's Grok 2 multimodal language model through his startup xAI's new API.
venturebeat.com
Elon Musk’s xAI launches API, letting third-party developers build atop Grok
Carl Franzen@carlfranzenOctober 21, 2024 12:15 PM
VentureBeat made with Midjourney
Elon Musk has been making headlines recently for his political activities and rocket launches, but that hasn’t stopped the six-company owner/operator from moving forward with xAI, his artificial intelligence startup.
Today, Musk announced via his social network X (formerly Twitter) that xAI now offers an application programming interface (API). This allows third-party developers to access and build applications and features powered by xAI’s Grok large language models (LLMs).
Musk posted a simple message on his X account, “The @xAI API is now live!” around 12 pm ET on Monday, October 21, following up with a reply post that included a link to the xAI API sign-on page.
According to xAI’s documentation, its API give developers access to Grok-2 and Grok-2 mini, xAI’s latest multimodal language models, which also include the capability to generate permissive images with Black Forest Labs’ Flux.1 diffusion model.
However, in my limited test of it, I was only able to access a model named “Grok-beta.” Notably, the entire API website is labeled at the top as a “public beta,” so presumably the xAI developers and engineers will be working out various bugs and adding new features in the coming days.
Positioning and pricing
Releasing the xAI API is helpful for xAI and Musk’s greater quest to go head-to-head against his former company OpenAI (which he co-founded and left), which offers an API as well.
The xAI API offers a web-based console for creating API keys, exploring endpoints, and integrating models into applications.
It supports REST, gRPC, and SDKs, and is compatible with other AI services such as OpenAI, enabling smooth integration with existing systems.
I signed up to check out the pricing and spent $25 on prepaid credits to see if I could get access to the models, but was only able to access “grok-beta” at the time of this article’s publication.
It provides 131,072 tokens of context, and the API can handle 1 request per second (RPS) and up to 10 requests per minute (RPM).
For now, OpenAI’s pricing for its AI models is largely cheaper than xAI’s, with GPT-4o costing developers $2.50 per 1 million input tokens/$10 per 1 million output tokens, versus Grok’s $5 per million input/$15 per output (though OpenAI’s new reasoning model o1 is more expensive than both at $15/$60).
• Text and Code Generation: The Grok models can handle tasks like generating code, summarizing content, and performing data extraction. This flexibility makes them valuable for a wide range of use cases, including software development and data analysis.
• Vision: The models can analyze and generate images, expanding their use beyond text-based tasks to multimedia and visual content generation.
• Function Calling: The xAI API allows models to interact with external tools, such as APIs and databases. This enables real-world tasks like booking a flight, accessing IoT devices (e.g., unlocking a Tesla), or fetching live data from websites.
New dev features
Toby Pohlen, a founding member of xAI, shared additional details about building the new xAI API in a thread on the social network X today, writing, “Creating a scalable API from scratch was a massive effort. Here are some of my favourite engineering highlights.”
As he explained, xAI’s API includes:
• Usage Explorer: The xAI Console includes a usage explorer that tracks API consumption, similar to what’s found in major cloud provider platforms, giving developers insights into their resource usage and costs.
• Simplified Team Management: Small businesses can bind an email domain to their teams, making it easier to manage users and teams in the platform.
• Enhanced Security: Users can view all active sessions in the account app, and log out of any unrecognized devices. In addition, every time the account is accessed from a new IP address, the user is notified by email.
The platform also supports two-factor authentication via TouchID, security keys (e.g., Yubikey), and authenticator apps.
How to get started with xAI’s new API
To begin using the xAI API, developers must sign up via the xAI Console, onboard their teams, and configure billing.
Each team is assigned its own API keys and billing setup, ensuring that enterprises can track costs and manage resources effectively.
As Pohlen highlighted, team management is streamlined for smaller companies through the ability to bind an email domain to their teams.
Whether devs flock to xAI’s API and adopt it alongside or in place of other compelling alternatives in the gen AI age such as OpenAI, Anthropic, Google, Microsoft, and more, remains to be seen. But at least, now xAI is giving them the option.
Being Liberal ® (@beingliberal) on Threads
Maybe Artificial Intelligence is our only hope?
www.threads.net
The Tennessee Holler (@thetnholler) on Threads
Seems Elon’s own A.I. tool knows him pretty well
www.threads.net
