AI bro fight! AI bro fight! Clash of the TAItans <\/p>\n
Buttons? Pressed. <\/p>\n
Over the weekend, researchers Dylan Patel and Daniel Nishball, who together write a semiconductor blog called SemiAnalysis, published a controversial post declaring that Google's secretive upcoming AI model, dubbed Gemini, is about to blow OpenAI's GPT-4 out of the water. According to the blog brusquely titled \"Google Gemini Eats The World Gemini Smashes GPT-4 By 5X, The GPU-Poors\" Google's expansive infrastructure of advanced GPUs gives the Silicon Valley stalwart and its next-gen model a leg up over the latest iteration of OpenAI's GPT-4. <\/p>\n
Which, as conversations across sites like X-formerly-Twitter and Hacker News made clear, is a contentious take. Does more computing power reallyequal a better AI model? <\/p>\n
It's a fascinating question, and some online debates got heated. But no one, it seems, was more perturbed by the statements in the blog than OpenAI CEO Sam Altman, who took to X-formerly-Twitter on Monday to scoff at the researchers' analysis. <\/p>\n
\"Incredible Google got that SemiAnalysis guy to publish their internal marketing\/recruiting chart,\" Altman wrote in the post, referring to an infrastructure chart included in the blog. His signoff to the message: \"lol.\" <\/p>\n
Patel, one of the blog authors, didn't take Altman's critiques lying down, and on Tuesday hit back with an X post of his own. <\/p>\n
\"Sundar to the GPU-poors,\" the researcher captioned the post, which included an NSFW meme of Google-owning Alphabet CEO Sundar Pichai appearing to force-feed milk to Altman. \"The data is from a supplier of Google,\" Patel added, \"and we made the chart.\" <\/p>\n
It's not surprising that Altman isn't a fan of the SemiAnalysis post, given that it pretty much just says that Google products will absolutely wallop OpenAI's by the end of the year. And to his credit, as a number of netizens pointed out, the Semiconductors Rule All argument might lack some nuance. <\/p>\n
\"Computational power alone is not the only resource. It is also the training process itself... and, obviously, data and its quality,\" one Hacker News user wrote, as caught by Insider. \"I will be convinced only after Google demonstrates that Gemini is better than GPT4 (in some, or all, tasks).\" <\/p>\n
A fair point. Still, it'salsotrue that, thoughOpenAI's ChatGPT release kickstarted the publicview of the AI race, Google which has used its immensely deep pockets to pioneer the world of AI R&D for some time now likely has some powerful stuff up its sleeves. Regardless, at the end of the day, it's always fun to witness Silicon Valley's most powerful bristle at any affront to their perceived superiority. <\/p>\n
More on AI: AI's Dirty Secret: Poor People in the Developing World Are Doing Most of the Work <\/p>\n
<\/p>\n
See the original post here: AI bro fight! AI bro fight! Clash of the TAItans Buttons? Pressed. Over the weekend, researchers Dylan Patel and Daniel Nishball, who together write a semiconductor blog called SemiAnalysis, published a controversial post declaring that Google's secretive upcoming AI model, dubbed Gemini, is about to blow OpenAI's GPT-4 out of the water Continue reading
\nOpenAI Rages at Report That Google's New AI Crushes GPT-4 - Futurism<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"