ChatGPT will soon be able to see, hear and speak. Tesla’s robot Optimus can do yoga. And the AI competition heats up as Amazon prepares to drop billions into OpenAI competitor Anthropic.
These and more top tech stories on Hashtag Trending
I’m your host Jim Love, CIO of IT World Canada and Tech News Day in the US.
It’s one of those days when even though you look for non-AI stories, the most interesting and relevant stories are all AI. So let’s go with the flow….
Amazon has declared its intention to invest up to $4 billion in Anthropic, a leading competitor to OpenAI, the organization behind ChatGPT. This strategic move places Amazon in a strong alliance with a prominent AI startup, mirroring Microsoft’s partnership with OpenAI.
As part of the collaboration, Anthropic will utilize Amazon Web Services’ Trainium and Inferentia chips for building, training, and deploying its AI models. Amazon’s initial commitment is $1.25 billion for a minority stake in Anthropic, with the potential to increase to $4 billion. Earlier in the year, Google invested over $300 million in Anthropic, but Amazon’s investment significantly surpasses this. Anthropic, established in 2021 by former OpenAI members, aims to develop a safer and more ethically sound AI product.
Source include: Axios
OpenAI has enhanced ChatGPT’s capabilities, allowing it to “see, hear and speak.” Initially, when GPT-4 was launched, it had multimodal capabilities, enabling ChatGPT to process image inputs. Formerly, ChatGPT had some capabilities via a plug and voices through 11Labs but this appears to be something “built in not bolted on.’
Now, ChatGPT can also accept voice inputs and respond in voice conversations. The image input feature assists users in various tasks, such as identifying objects or suggesting recipes based on pantry items. The voice feature transforms ChatGPT into a voice assistant, similar to Amazon’s Alexa. In a demonstration, ChatGPT narrated a bedtime story about a hedgehog on a user’s verbal request. OpenAI has integrated Whisper, its speech recognition system, and a new text-to-speech model to support this voice feature. These features will soon be available for ChatGPT Plus and Enterprise users.
Source include: ZDNET
Google’s experimental AI chatbot, Bard, has introduced three significant features in September 2023. These include a “double-check response with Google” option, extensions to access data from other Google services, and a “share-and-continue chat” feature. Bard now offers the ability to verify its generated content with Google Search results, enhancing user confidence in the information provided. The chatbot can also integrate with Google services like Maps, YouTube, Gmail, and Drive, enhancing its utility. Additionally, Bard chats can now be shared via a link, allowing others to view and continue the conversation. Notably, Bard has expanded its language support, now catering to over 40 languages, and can return images alongside text responses. These updates hint at Google’s potential plans to evolve Bard from an experimental phase to a comprehensive service.
I don’t think that our Canadian audience can get Bard yet, but it’s easy to access if you get a VPN. Check it out.
Source include: Tech Republic
Getty Images has launched a generative AI tool that creates images based solely on licensed content from Getty’s extensive creative collection. This move distinguishes Getty from other text-to-image engines. The company assures that its AI-generated images are safe for commercial use and has even offered to indemnify customers purchasing these images, taking responsibility for any legal issues that might arise.
Listeners might remember that Microsoft made a similar commitment for its software development assistance with AI.
The tool was developed in collaboration with Nvidia and was trained exclusively on Getty’s creative images, excluding editorial images and news photos. Getty plans to charge customers for each AI-generated image, and contributors whose images were used for training will be compensated. Getty CEO Craig Peters emphasized that the tool aims to complement, not replace, professional artists, offering a legally safe and innovative image creation solution.
Source include: Axios
Tesla’s humanoid robot, Optimus, has undergone significant advancements since its prototype unveiling by Elon Musk a year ago. Initially, Optimus could only wave, but now it can pick up and sort objects, perform yoga, and navigate its environment. Notably, it can execute the yoga pose “Vrikshasana.” Unlike other robots, such as those from Boston Dynamics, Optimus operates on neural networks rather than just rules. Jim Fan, a senior AI scientist at NVIDIA, suggests that Optimus’s smooth hand movements are likely a result of imitation learning, where the robot mimics human actions. This learning approach is akin to how game characters are animated. Optimus might also utilize motion capture technology, similar to Hollywood films, and various other methods for its movement and learning. Musk envisions a future where everyone would want an Optimus, signaling a shift towards a world dominated by humanoid robots.
Source include: Analytics India
A comprehensive study, backed by renowned institutions like Harvard Business School and MIT Sloan, involving over 750 BCG consultants, revealed both the promise and pitfalls of GenAI. When applied to creative product innovation, nearly 90 per cent of participants enhanced their performance with GPT-4, achieving results 40 per cent superior to those without the tool. However, the technology’s application in areas outside its core competence, like business problem-solving, led to a 23 per cent performance dip. A concerning observation was the misplaced trust in GenAI: participants often over-relied on it in unsuitable areas while underutilizing it where it could add immense value. Additionally, GenAI’s uniform outputs can diminish group thought diversity by 41 per cent. Leaders are urged to judiciously integrate GenAI, considering its evolving capabilities and potential impact on diverse tasks.
Source included: BCG
That’s the top tech news stories for today. For more fast reads on top stories, check us out at TechNewsDay.com or ITWorldCanada.com on the homepage.
Hashtag Trending goes to air 5 days a week with a special weekend interview show we call “the Weekend Edition.”
You can get us anywhere you get audio podcasts and there is a copy of the show notes at itworldcanada.com/podcasts
I’m back for a couple of days, but off to cover a huge AI conference in Montreal Canada this week called “All In.”
I’m sure you are enjoying the mellow tones of James Roy and the great stories from Ashee Pamma.
I’m your host, Jim Love. Have a Terrific Tuesday!