3.9.2023
The August edition of Frontnow highlights rapid advancements in AI, from breakthroughs in games and speech recognition to its integration into products by tech giants like Apple, Intel, and Google.
Cedric May
Chief Technology Officer
Welcome to Frontnow's August Edition: The pace of AI's evolution is unyielding, and this month has been a testament to its transformative potential. From its undeniable prowess in complex games and businesses to its breakthroughs in image and speech recognition, AI continues to push boundaries. Yet, with immense power comes profound responsibility, and many are pondering the broader implications of these technological leaps.
Dive deeper in this issue:
Sit back and immerse yourself in the dynamic landscape of artificial intelligence with Frontnow. Your monthly compass to the future!
In the past decade, AI has developed rapidly, achieving significant breakthroughs in complex games, surpassing humans in image and speech recognition, passing difficult business school exams, and even acing coding interviews at Amazon. However, recent advances in AI have left some people feeling uneasy, even describing it as "scary". During a Senate Judiciary Committee hearing on regulating AI, expert witnesses emphasized the pace of progress in AI and how the three key factors driving AI progress are compute, data, and algorithms. Compute power has increased steadily since the first artificial neural network was developed in 1957, and researchers have been spending increasingly large amounts of money to train larger scale models. Data is essential to train AI systems, and AI models are being trained on larger and larger amounts of data collected from the internet. Finally, algorithms play a crucial role in determining how AI systems use computational power to model the relationships between variables in the data they are given.
Experts predict that AI progress will continue at a breakneck pace for the next few years, with companies spending more on compute, using remaining data on the internet to train AI models, and finding more efficient ways to train and run AI systems. However, this rapid pace of progress has many experts worried about potential misuse of AI systems in areas such as cybersecurity, nuclear technology, chemistry, and biology.
GenAI startups raised $14.1B in H1 2023, surpassing 2022's year-end total by more than 5x. Generative AI is at the peak of Gartner's hype cycle for emerging technologies in 2023. Despite the decline in ChatGPT traffic, generative AI adoption is predicted to double the adoption rate of smartphones and tablets, reaching 77.8 million users in the next few years. Amazon CEO Andy Jassy has emphasized the importance of GenAI and revealed that every team at Amazon is actively working on generative AI projects.
However, there are major uncertainties in generative AI adoption, such as security, copyright/training data, inaccuracy/model hallucinations, data privacy, and pending regulation. According to a recent McKinsey survey, 79% of all respondents have had some exposure to gen AI, and 22% are regularly using it in their own work. But when it comes to enterprise adoption, many organizations are still reluctant to embrace generative AI at scale. Only 21% have policies about using generative AI at work. The Gartner report also highlights other emerging technologies that will present benefits for early adopters, including Emergent AI, Developer Experience, Pervasive Cloud, and Human-centric security and privacy.
During Intel's Q2 2023 earnings call, Intel CEO Pat Gelsinger expressed his strong belief in AI and announced that Intel plans to incorporate AI into every product they produce. Intel's upcoming Meteor Lake chip, which is set to release later this year, will be the company's first consumer chip to include a built-in neural processor for machine learning. Gelsinger has previously mentioned the "four superpowers" or "five superpowers" of technology companies, which include AI and the cloud. However, he now suggests that AI and the cloud do not necessarily go together. Gelsinger believes that AI will eventually be integrated into every product that Intel sells, including hearing aids. While people are starting to use the cloud for certain tasks such as ChatGPT, Gelsinger notes that becoming AI-enabled must be done on the client for that to occur. Intel sees a wide range of AI enablement across every aspect of consumer, developer, and enterprise efficiency use cases, which will be client-centered and at the edge. Gelsinger emphasizes that you cannot round trip to the cloud for every client, as it will all happen at the edge and at the client. While it is Nvidia, not Intel, that makes the chips that power the AI cloud, Intel is determined to find its own way in the industry.
Apple is expanding the AI capabilities of its mobile devices by hiring for dozens of roles to work on large language models. The company is focused on bringing technologies like LLMs specifically to mobile, whereas competitors like Google have released AI products like chatbots. Apple CEO Tim Cook views AI and machine learning as core fundamental technologies that are integral to virtually every product they build. The company’s Q3 spending on R&D was $3.1 billion larger than the same quarter in 2022, which Cook says is due in part to its generative AI efforts. As companies build LLMs, they’re also facing the challenges of collecting and classifying vast amounts of data, while also understanding the complexities of how models now operate. And with all this development and investment comes regulation, as governments around the world try to get a better handle on AI. University of Pennsylvania Law School Professor Cary Coglianese argued that regulating AI will be a multifaceted activity that changes depending on the type of algorithm and how it is used.
Google is launching new updates to its Search Generative Experience (SGE), an AI-powered conversational mode in Search. The updates include features to see definitions of unfamiliar terms, improve understanding and coding information across languages, and tap into the AI power of SGE while browsing. These improvements aim to help people better understand complicated concepts or complex topics, boost their coding skills, and more. One of the new features allows users to preview definitions of certain words and see related images or diagrams, while another will help those using SGE for programming assistance. SGE while browsing is available as a stand-alone experiment in Search Labs, and will automatically roll out to those who have already opted into SGE.
IBM and NASA are teaming up to develop an innovative open-source AI model that will revolutionize the way we evaluate satellite images and conduct geoscientific observations. This groundbreaking technology will enable relevant entities to quickly and confidently react to changes related to deforestation, wildfires, harvests, climate change, and natural disasters. The Geospatial Foundation Model will serve geo-ecological earth monitoring and meticulously track forest cover, greenhouse gases, crop yields, and natural disasters. IBM's cutting-edge Watsonx.ai AI, Hugging Face's open-source AI platform, and NASA's detailed satellite data are all contributing to this pioneering project. With the AI model, we can analyze massive amounts of data that were previously impossible to evaluate effectively. This collaboration is a crucial step towards improving our planet by providing faster and more effective solutions in critical areas such as climate change.
Google Lab Sessions is a collaboration between AI experts and innovators. In the latest session, they worked with rapper and MIT Visiting Scholar Lupe Fiasco to explore how AI could help expand human creativity. They observed Lupe's creative process and discovered that he deconstructs and reassembles language in innovative ways. Google wanted to find ways to use AI to help Lupe explore even more creative possibilities with text and language.
To do this, they used large language models (LLMs) to generate text and developed a few-shot prompting strategy to fit into Lupe's lyric-writing workflow. They designed ten prompts to explore creative possibilities that may arise from a given word, phrase, or concept using the MakerSuite platform. They then built an app called TextFX, which houses these prompts using Google's PaLM 2 model and the PaLM API.
TextFX is an excellent example of how artists can experiment with the PaLM API and build applications that leverage Google's state-of-the-art large language models.
Google and Universal Music Group (UMG) are in talks to create AI-generated music. The proposed agreement would allow users to create songs using musicians' voices and pay for the licensing fees. The tool is still in its early stages and artists would have the option to opt in. The rise of audio deepfakes is a concern for the music industry as it raises questions about ownership and reputational harm. UMG's partnership with Google would permit the use of their artists' voices, but ethical issues remain. For instance, authors like Jane Friedman have discovered AI-generated books falsely written under their name on Amazon. The SAG-AFTRA strike was partly triggered by a proposal from studios to use AI-generated likenesses of background actors.
August illuminated the AI universe with technological marvels, compelling collaborations, and critical ethical discussions. From tech giants pushing boundaries to the harmonious blend of art and algorithms, we've delved into AI's multifaceted landscape. As we transition to September, remain inquisitive and anticipate even more groundbreaking revelations in AI's ever-evolving saga. Until next time, keep pushing the boundaries of possibility. See you next month!
Welcome to Frontnow's August Edition: The pace of AI's evolution is unyielding, and this month has been a testament to its transformative potential. From its undeniable prowess in complex games and businesses to its breakthroughs in image and speech recognition, AI continues to push boundaries. Yet, with immense power comes profound responsibility, and many are pondering the broader implications of these technological leaps.
Dive deeper in this issue:
Sit back and immerse yourself in the dynamic landscape of artificial intelligence with Frontnow. Your monthly compass to the future!
In the past decade, AI has developed rapidly, achieving significant breakthroughs in complex games, surpassing humans in image and speech recognition, passing difficult business school exams, and even acing coding interviews at Amazon. However, recent advances in AI have left some people feeling uneasy, even describing it as "scary". During a Senate Judiciary Committee hearing on regulating AI, expert witnesses emphasized the pace of progress in AI and how the three key factors driving AI progress are compute, data, and algorithms. Compute power has increased steadily since the first artificial neural network was developed in 1957, and researchers have been spending increasingly large amounts of money to train larger scale models. Data is essential to train AI systems, and AI models are being trained on larger and larger amounts of data collected from the internet. Finally, algorithms play a crucial role in determining how AI systems use computational power to model the relationships between variables in the data they are given.
Experts predict that AI progress will continue at a breakneck pace for the next few years, with companies spending more on compute, using remaining data on the internet to train AI models, and finding more efficient ways to train and run AI systems. However, this rapid pace of progress has many experts worried about potential misuse of AI systems in areas such as cybersecurity, nuclear technology, chemistry, and biology.
GenAI startups raised $14.1B in H1 2023, surpassing 2022's year-end total by more than 5x. Generative AI is at the peak of Gartner's hype cycle for emerging technologies in 2023. Despite the decline in ChatGPT traffic, generative AI adoption is predicted to double the adoption rate of smartphones and tablets, reaching 77.8 million users in the next few years. Amazon CEO Andy Jassy has emphasized the importance of GenAI and revealed that every team at Amazon is actively working on generative AI projects.
However, there are major uncertainties in generative AI adoption, such as security, copyright/training data, inaccuracy/model hallucinations, data privacy, and pending regulation. According to a recent McKinsey survey, 79% of all respondents have had some exposure to gen AI, and 22% are regularly using it in their own work. But when it comes to enterprise adoption, many organizations are still reluctant to embrace generative AI at scale. Only 21% have policies about using generative AI at work. The Gartner report also highlights other emerging technologies that will present benefits for early adopters, including Emergent AI, Developer Experience, Pervasive Cloud, and Human-centric security and privacy.
During Intel's Q2 2023 earnings call, Intel CEO Pat Gelsinger expressed his strong belief in AI and announced that Intel plans to incorporate AI into every product they produce. Intel's upcoming Meteor Lake chip, which is set to release later this year, will be the company's first consumer chip to include a built-in neural processor for machine learning. Gelsinger has previously mentioned the "four superpowers" or "five superpowers" of technology companies, which include AI and the cloud. However, he now suggests that AI and the cloud do not necessarily go together. Gelsinger believes that AI will eventually be integrated into every product that Intel sells, including hearing aids. While people are starting to use the cloud for certain tasks such as ChatGPT, Gelsinger notes that becoming AI-enabled must be done on the client for that to occur. Intel sees a wide range of AI enablement across every aspect of consumer, developer, and enterprise efficiency use cases, which will be client-centered and at the edge. Gelsinger emphasizes that you cannot round trip to the cloud for every client, as it will all happen at the edge and at the client. While it is Nvidia, not Intel, that makes the chips that power the AI cloud, Intel is determined to find its own way in the industry.
Apple is expanding the AI capabilities of its mobile devices by hiring for dozens of roles to work on large language models. The company is focused on bringing technologies like LLMs specifically to mobile, whereas competitors like Google have released AI products like chatbots. Apple CEO Tim Cook views AI and machine learning as core fundamental technologies that are integral to virtually every product they build. The company’s Q3 spending on R&D was $3.1 billion larger than the same quarter in 2022, which Cook says is due in part to its generative AI efforts. As companies build LLMs, they’re also facing the challenges of collecting and classifying vast amounts of data, while also understanding the complexities of how models now operate. And with all this development and investment comes regulation, as governments around the world try to get a better handle on AI. University of Pennsylvania Law School Professor Cary Coglianese argued that regulating AI will be a multifaceted activity that changes depending on the type of algorithm and how it is used.
Google is launching new updates to its Search Generative Experience (SGE), an AI-powered conversational mode in Search. The updates include features to see definitions of unfamiliar terms, improve understanding and coding information across languages, and tap into the AI power of SGE while browsing. These improvements aim to help people better understand complicated concepts or complex topics, boost their coding skills, and more. One of the new features allows users to preview definitions of certain words and see related images or diagrams, while another will help those using SGE for programming assistance. SGE while browsing is available as a stand-alone experiment in Search Labs, and will automatically roll out to those who have already opted into SGE.
IBM and NASA are teaming up to develop an innovative open-source AI model that will revolutionize the way we evaluate satellite images and conduct geoscientific observations. This groundbreaking technology will enable relevant entities to quickly and confidently react to changes related to deforestation, wildfires, harvests, climate change, and natural disasters. The Geospatial Foundation Model will serve geo-ecological earth monitoring and meticulously track forest cover, greenhouse gases, crop yields, and natural disasters. IBM's cutting-edge Watsonx.ai AI, Hugging Face's open-source AI platform, and NASA's detailed satellite data are all contributing to this pioneering project. With the AI model, we can analyze massive amounts of data that were previously impossible to evaluate effectively. This collaboration is a crucial step towards improving our planet by providing faster and more effective solutions in critical areas such as climate change.
Google Lab Sessions is a collaboration between AI experts and innovators. In the latest session, they worked with rapper and MIT Visiting Scholar Lupe Fiasco to explore how AI could help expand human creativity. They observed Lupe's creative process and discovered that he deconstructs and reassembles language in innovative ways. Google wanted to find ways to use AI to help Lupe explore even more creative possibilities with text and language.
To do this, they used large language models (LLMs) to generate text and developed a few-shot prompting strategy to fit into Lupe's lyric-writing workflow. They designed ten prompts to explore creative possibilities that may arise from a given word, phrase, or concept using the MakerSuite platform. They then built an app called TextFX, which houses these prompts using Google's PaLM 2 model and the PaLM API.
TextFX is an excellent example of how artists can experiment with the PaLM API and build applications that leverage Google's state-of-the-art large language models.
Google and Universal Music Group (UMG) are in talks to create AI-generated music. The proposed agreement would allow users to create songs using musicians' voices and pay for the licensing fees. The tool is still in its early stages and artists would have the option to opt in. The rise of audio deepfakes is a concern for the music industry as it raises questions about ownership and reputational harm. UMG's partnership with Google would permit the use of their artists' voices, but ethical issues remain. For instance, authors like Jane Friedman have discovered AI-generated books falsely written under their name on Amazon. The SAG-AFTRA strike was partly triggered by a proposal from studios to use AI-generated likenesses of background actors.
August illuminated the AI universe with technological marvels, compelling collaborations, and critical ethical discussions. From tech giants pushing boundaries to the harmonious blend of art and algorithms, we've delved into AI's multifaceted landscape. As we transition to September, remain inquisitive and anticipate even more groundbreaking revelations in AI's ever-evolving saga. Until next time, keep pushing the boundaries of possibility. See you next month!
Read next
Stay Updated With Frontnow's News & Cutting-Edge AI Innovation.
Your partner in redefining e-commerce, leveraging AI for hyper-personalized online experiences and superior product discovery.
