Google on Wednesday infused its Bard chatbot with a new-generation artificial intelligence (AI) model called Gemini, which it touts as being able to “reason” better than ChatGPT and other rivals.
“This is incredible momentum, and yet, we’re only beginning to scratch the surface of what’s possible,” Google chief executive officer Sundar Pichai said in a release. “This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company.”
It is the first AI model to outperform human experts in certain benchmarks involving problem solving, math, physics, history, law, medicine and ethics, Google DeepMind vice president of product Eli Collins said during a briefing.
A demonstration showed Gemini recognizing what it was shown, from a person acting out a Matrix movie scene to someone drawing a duck and then holding up a rubber duck.
Gemini commented on what it was shown, making comparisons, drawing conclusions and offering suggestions.
Performance of an “Ultra” version of Gemini “far exceeds” that of other state-of-the-art models in 30 benchmark tests measuring capabilities such as image understanding or mathematical reasoning, Collins said.
A “Pro” version of Gemini built into Bard is designed to handle a wide range of tasks. A “Nano” version is tailored for smartphones, coming first to Google’s top-of-the-line Pixel 8 handset.
Bard will use Gemini for more advanced reasoning, planning and understanding capabilities, a demonstration showed.
It will be available in English in more than 170 countries and territories, with more languages added soon, Bard vice president Sissie Hsiao (蕭茜) said.
Gemini-infused Bard would be expanded to be “multimodal,” meaning it will be able to work with auditory and visual input as well as text prompts, Google said.
The Ultra version of Gemini designed to handle highly complex tasks would be released early next year, it said.
Google in September integrated Gmail, YouTube and other tools into its Bard chatbot as tech giants seek to persuade users that generative AI is useful and not dangerous or just a fad.
Those capabilities closely match offerings from Microsoft Corp that infuse its Office 365 apps with AI powers, although those come at an extra cost to customers and are not available through the chatbot on its search engine Bing.
Taiwan Semiconductor Manufacturing’s (TSMC, 台積電) first wafer fab in Kumamoto, Japan is still set to launch commercial production in the fourth quarter of this year as planned, the world’s largest contract chipmaker said on Saturday in response to reports that mass production might begin ahead of schedule. TSMC said the monthly production capacity of the joint venture fab, Japan Advanced Semiconductor Manufacturing (JASM), is expected to hit 55,000 units of 12-inch wafers, using the mature 12-nanometer, 16-nanometer, 22-nanometer and 28-nanometer processes. JASM is owned by TSMC and its Japanese business partners Sony Semiconductor Solutions Corp and Denso Corp, with the Taiwanese company
US President Joe Biden’s administration is in talks to confer more than US$10 billion in subsidies to Intel Corp, people familiar with the matter said, in what would be the largest award yet under a plan to bring semiconductor manufacturing back to US soil. Intel’s award package is expected to include both loans and direct grants, the source said. They stressed that negotiations are still under way. The US Department of Commerce and Intel declined to comment. The incentives would come from the 2022 Creating Helpful Incentives to Produce Semiconductors (CHIPS) and Science Act, which set aside US$39 billion in direct grants as
A new artificial intelligence (AI) tool that promises to create short videos from simple text commands has raised concerns along with questions from artists and media professionals. OpenAI, the creator of ChatGPT and image generator DALL-E, on Thursday said it was testing a text-to-video model called “Sora” that can allow users to create realistic videos with simple prompts. The San Francisco-based start-up said that Sora can “generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background,” but added that it still has limitations, such as possibly “mixing up left and right.” Examples of Sora-created clips
Super Micro Computer Inc’s lengthy rally came to a shuddering halt on Friday, with a selloff that derailed what had looked to be the server maker’s best week on record. Shares fell 20 percent, their biggest one-day percentage drop since August last year. The decline comes in the wake of a nine-session run of gains, the longest such streak for the stock since 2016. However, even with the day’s selloff, the stock rose 8.5 percent for the week. Despite Friday’s drop, recent gains show how Super Micro has become one of the hottest names in artificial intelligence (AI). The stock has risen