Alphabet, the parent company of Google, shakes up the A.I. landscape as it merges two of its cutting-edge research groups, DeepMind and Google Brain called Google DeepMind


Read The Internal Memo Alphabet Sent In Merging AI-Focused Groups DeepMind And Google Brain

Alphabet is merging two of its A.I.-focused units, DeepMind and Google Brain, into a single entity called Google DeepMind, which will be led by DeepMind CEO Demis Hassabis. The merger is aimed at bringing together two groups working on A.I. research and accelerating the progress of Google’s A.I. efforts. The move comes amid the rapid developments in A.I. and is a response to OpenAI’s launch of the chatbot ChatGPT last year. Jeff Dean, currently leading Google’s A.I. efforts, will be promoted to Google’s chief scientist and will head up the technical projects related to A.I. for the newly merged entity.


Google Will Offer Customer AI Tools To ‘Remix’ Content For Ad Campaigns

Google plans to integrate generative AI into its advertising business this year, allowing customers to input content like text, images, and videos, which will be “remixed” to create ads based on specific goals. The new tools will likely be integrated into Google’s Performance Max program, which currently uses machine learning for strategic recommendations. However, the use of generative AI introduces risks, as such systems have been known to introduce errors and replicate biases. Despite Google’s assurances, the company may not be able to fully safeguard new AI products, and its rivals are also offering similar AI tools for advertising work.


Microsoft Releases Copilot For Viva, As It Continues To Roll Out Generative AI To Apps

Microsoft has integrated its generative AI-powered 365 Copilot into Viva, the employee engagement and experience platform it launched in February 2021. Copilot in Viva focuses on streamlining workforce communications and improving organizational alignment. It will be available to customers later this year. Copilot combines OpenAI’s GPT-4 with Microsoft Graph data and 365 apps including Teams, Word, Outlook, and Excel. Copilot in Viva Goals can draft OKR recommendations, summarize OKR status, identify blockers and suggest next steps. Viva Engage with Copilot can create post ideas for company intranet pages, while Copilot in Viva Glint can help summarize thousands of employee survey comments.


Meta Unveils DINOv2, An AI Vision Transfer Model: Is It The Next Leap In Image Recognition?

Meta, the parent company of Facebook, has unveiled its new AI model, DINOv2, which is expected to revolutionize the field of computer vision. The model is designed to accurately identify individual objects within images, videos, and other visual inputs. It employs self-supervised learning, allowing it to learn from any collection of images, regardless of whether they are manually labeled or not. The model works on a framework with a teacher network and a student network, both of which learn from data without labels. DINOv2 has a wide range of potential applications, including medical image analysis, security monitoring, and assisting with real-time decision-making for autonomous vehicles.


OpenAI Cofounder Greg Brockman Demos Unreleased ChatGPT Plug-ins

At TED2023, OpenAI co-founder Greg Brockman demonstrated the capabilities of the company’s newest large language model, GPT-4, with a series of unreleased plug-ins for ChatGPT. These plug-ins include generating a recipe, image, tweet, and grocery list without leaving the chatbot, fact-checking with citations, and interpreting spreadsheets with vague instructions. Brockman hopes that AI will align with the user’s intent and create a deep, trustworthy collaboration between humans and AI. Although there are concerns about the potential of AGI, Brockman believes that widespread participation and input from people on what the guardrails for the AI should be will create a better world for everyone. The release of the machine before it’s super powerful allows for feedback from users for incremental improvement. AGI is set to transform almost every aspect of how we use computers, and Brockman suggests that everyone should get literate in this technology.


ChatGPT Costs $700,000 Per Day To Run, Which Is Why Microsoft Wants To Make Its Own AI Chips

ChatGPT, a popular AI language model developed by OpenAI, reportedly costs $700,000 per day to run, which could prompt Microsoft to develop its own AI chips to decrease costs. Building its own chips can reduce Microsoft’s reliance on GPU manufacturers such as NVIDIA, with the potential to cut costs by a third or more. Microsoft’s AI chip, Athena, is already being tested and is set to be distributed among Azure AI services next year. This move is part of Microsoft’s aggressive push into AI, following Google’s botched launch of its AI product, Bard.


Stakes Are ‘Incredibly High’ For AI In Healthcare, Says General Catalyst’s Maloney

During HIMSS23’s Venture Connect, General Catalyst’s Managing Director, Holly Maloney, spoke about responsible AI use in healthcare and what the company looks for when considering investment in startups. She emphasized the importance of responsible innovation to offer data-driven and consumer-centric healthcare services. Maloney said that one reason for slow AI adoption is a lack of trust, and the stakes are high, with private information that can be monetized in many ways. General Catalyst believes that collaborations are needed to go to market and has developed partnerships with health systems. Despite the challenges, many companies establish partnerships to scale operations, and there is a lot of innovation happening in healthcare.


Ways AI Can Help You Be More Productive At Home

AI isn’t just for the workplace anymore, as it can also assist with tasks at home. OpenAI’s GPT4 technology can help with meal planning, generating bedtime stories for children, and even creating books with corresponding images using AI tools like DALL-E. However, AI-generated text may contain errors or inappropriate statements, so fact-checking is necessary. AI can also be useful for generating emails and content creation, as well as assisting computer programmers and coders in their work. While AI is helpful for the broad strokes of navigating life, it doesn’t have the same level of judgment and insight as humans.


Apple Is The One Big Tech Company Without A Clear ChatGPT Strategy

Everyone is scrambling to copy ChatGPT, the AI language model created by OpenAI. Microsoft, Google, and Amazon have already invested in their AI development, but Apple seems to be lagging behind. Despite improving its AI capabilities, Apple has not been prioritizing generative AI, which could be a game-changer in the industry. With the massive computational power and cloud services required for this technology, Apple may be at a disadvantage. However, AI in Apple products today is still essential and helpful for many functions. The company may not need to be in the generative AI fight, but if AI evolves into the post-mobile platform, Apple could find itself at risk. The lesson for Apple may be to focus on innovation and staying ahead of the curve, as it did in the early days of the iPhone.