Topics
Latest
AI
Amazon
Image Credits:Nathan Laine/Bloomberg / Getty Images
Apps
Biotech & Health
mood
Image Credits:Nathan Laine/Bloomberg / Getty Images
Cloud Computing
Commerce
Crypto
go-ahead
EVs
Fintech
fund raise
Gadgets
Gaming
Government & Policy
Hardware
Layoffs
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
Security
Social
infinite
Startups
TikTok
expatriation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Elon Musk ’s AI inauguration , xAI , is make its own version ofChatGPT .
That seems to be the eccentric , at least , from Musk ’s tweets on X late Friday evening teasing the AI organization xAI has been quietly recrudesce . Called Grok — a name xAItrademarkedrecently — the system answers interrogation informally , possibly drawing on a knowledge base standardized to that used to civilize ChatGPT and other comparable text edition - generating organisation ( e.g. Meta’sLlama 2 ) .
Grok leverage “ real - time approach ” to info on X , Musk read . And , like ChatGPT , the system has internet graze capableness , enabling it to search the web for up - to - particular date information about specific topics .
Well , most topics .
Grok has substantial - time access to information via the 𝕏 platform , which is a monolithic vantage over other model .
It ’s also found & loves satire . I have no estimate who could have maneuver it this direction 🤷 ♂ ️ 🤣 pic.twitter.com/e5OwuGvZ3Z
— Elon Musk ( @elonmusk)November 4 , 2023
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Musk implied Grok will refuse to serve certain queries of a more sensitive nature , like “ Tell me how to make cocaine , step by step . ” Judging by a screenshot , the system serve that special query a bit more wryly than ChatGPT ; it ’s not clear if it ’s a canned reply or if the organisation is , in fact — as Musk asserts in a tweet — “ designed to have a little more humor in its responses . ”
xAI ’s Grok system is design to have a little humour in its responsespic.twitter.com/WqXxlwI6ef
severally , xAI is offering a “ circumscribed number ” of user in the U.S. access to Grok . There ’s a waitlisthere .
As soon as it ’s out of other beta , xAI ’s Grok system will be usable to all X Premium+ reader
small ’s known about Grok so far — or xAI ’s broader research project , for that matter .
In September , Oracle co - founder Larry Ellison , a ego - described close-fitting friend of Musk , read that xAI had sign a declaration to develop its AI models on Oracle ’s cloud . But xAI itself has n’t revealed anything about those AI models ’ inner working — or , indeed , what screen out of labor they can accomplish .
Or rather , it had n’t until other Sunday , when xAI release ablog postdetailing Grok — include the approaches postulate to train it and the organisation ’s underlying infrastructure .
According to the post , thelarge language modeldriving Grok , Grok-1 , was developed over the course of instruction of several months on a cluster of “ tens of thousands ” of GPUs ( perhaps ply by Oracle ) . Training data total from both the entanglement ( up to Q3 2023 ) and feedback from human assistants that xAI touch on to as “ AI tutors . ”
Grok-1 evidence “ strong result ” on common LLM benchmark tests , xAI claims — “ surpassing all other models in its compute class . ” And — as Musk alluded to — it has “ a bit of wit ” and “ a malcontent streak , ” answering “ spicy questions that are rejected by most other AI systems ” ( although the jury ’s out on what ’s mean by “ spicy , ” just ) .
“ In some authoritative respects , it ( xAI ’s new role model ) is the best that presently exists , ” Musk was quoted as saying in a tweet Friday afternoon .
Musk announced the launching ofxAIin July with the challenging end of build up AI to “ realise the true nature of the universe . ” The society , guide by Musk and ex-serviceman of DeepMind , OpenAI , Google Research , Microsoft Research , Tesla and the University of Toronto , is advise by Dan Hendrycks , the conductor at the Center for AI Safety , an AI inquiry nonprofit , and collaborates with X and other company in Musk ’s lieu , including Tesla .
In an interview with Tucker Carlson in April , Musk said that he desire to build what he come to to as a “ maximum - truth - seek AI . ” Is Grok this AI ? possibly — or it ’s a footmark toward something even bigger .
In the web log Emily Post , xAI suggest at how its Grok-1 framework might germinate down the line :
“ Grok does n’t have other good sense , such as vision and audio , ” the ship’s company writes . “ To better assist user , we will equip Grok with these different senses that can enable broad applications , including real - time interactions and assistance . ”
Musk ’s AI aspiration have produce since the billionaire ’s snag with ChatGPT developer OpenAI co - founder Sam Altman and Ilya Sutskever several years ago . As OpenAI ’s focusing shifted from open source enquiry to primarily commercial projects , Musk grew disillusioned — and militant — with the caller on whose board he pose . Musk step down from the OpenAI instrument panel in 2018 , more latterly cutting off the company ’s access code to ecstasy data after arguing that OpenAI was n’t paying enough for the privilege .
update with additional information about LLM modeling power Grok , Grok-1 .
