ByteSize Daily Wrap - 2023/06/16

AI regulations meet high-risk challenges, AMD takes a swing at Nvidia, and AI vision transformers outshine your Excel skills! Today's smorgasbord of tech content also asks: Is the legal system ready for AI-generated content? Dive in for more!

Welcome back, ByteSizers, to another fantastic edition of Atometrix ByteSize! Today, we have a smorgasbord of tech content that's sure to tickle your neurons and spark your imagination. We'll be covering everything from new AI regulations in high-risk areas and AMD's latest GPU, to Microsoft's AI breakthroughs and the ever-evolving world of AI-generated content.

In today's opinion piece, let's talk about how the regulation of AI, especially in high-risk areas, is an essential step towards responsible AI development. As Europe passes new laws to restrict and disclose AI usage in such areas, we will see how a balance can be struck between fostering AI innovation and protecting our human rights. We're also excited to see how the new AMD MI300X might challenge Nvidia's stronghold. And it's worth considering if our legal system is truly prepared for the challenges that AI-generated content will bring. But hey, who isn't excited to see AI vision transformers put our Excel cooking skills to shame? Let's dive into these buzzing topics and more!

Going hands free? Now available as a podcast!

Have you had this forwarded to you? Subscribe now!

News & Articles

  • (Link) European Parliament passes AI regulations for high-risk areas: Excitingly, the European Parliament has taken a big leap forward in regulating AI, as they have passed a draft law which places restrictions on the use of AI in high-risk areas such as the legal system. The law also imposes more stringent disclosure requirements on companies regarding the data they use to generate AI systems. This is a great step forward to ensure that AI is used ethically and responsibly! Looks like our robotic overlords will have to wait a little longer.

  • (Link) AMD unveils new AI GPU to challenge Nvidia: AMD has revealed its latest AI-dedicated GPU, the MI300X, which will be available later this year. The chip will challenge Nvidia's dominance in the AI chip market and is designed to take advantage of the growing AI sector, predicted to be worth over $150 billion by 2027. With its superior memory capacity and support for ROCm software, the MI300X could cut costs of AI applications and make inroads into Nvidia's market share. AMD has long had ambitions in AI and wants to be a leader in the sector.

  • (Link) Bipartisan bill proposes removing AI's legal shield: Senators Richard Blumenthal and Josh Hawley have introduced a bipartisan bill that proposes removing protection under Section 230 for generative AI. This is an attempt to clarify the application of internet's liability law to AI. However, the bill has limitations, as it does not cover machine learning algorithms and does not define what constitutes "generative AI". It remains to be seen if the bill will gain traction in Congress. Let's see if AI's lawsuits will come piling in!

  • (Link) Microsoft dominates the AI arena with breakthroughs!: Microsoft CEO, Satya Nadella, has prioritised AI for the company's long-term growth, leading to impressive breakthroughs in AI technologies such as OpenAI's language model now integrated with Bing and Copilot, an AI tool that automates coding. There were initial bumps in the road, but now Microsoft is becoming a heavyweight in the AI arena. By investing $10 billion in OpenAI, the company has placed a huge bet on AI's potential as a key strategic move. Nadella places importance on aligning AI with core human values to utilise it as an empowerment tool, and reduce any unintended consequences. Nadella's AI vision has Microsoft poised for an AI-powered future. Microsoft won't take over the world with AI, unless Excel starts working on its own and stops needing formulas.

  • (Link) Bard: Google's AI chatbot launch in EU delayed over privacy concerns: Google has been forced to delay the launch of its new artificial intelligence chatbot, Bard, in the European Union due to concerns raised by the bloc's main data regulator over privacy. This is a bump in the road for Google, who have been seeking to diversify their business portfolio, expanding into new industries with an aim to strengthen their market dominance. Let's hope they can iron out the concerns before its launch. Maybe Bard just needs a better privacy agreement, like the rest of us!

Insights & Papers

  • (Link) Experts Warn AI Model Collapse Risk Due to AI-generated Content: Scientists are cautioning that there is a risk of artificial intelligence (AI) model collapse when models are trained using AI-generated content. The phenomenon occurs when these models forget the actual underlying data distribution and generate inaccurate responses. Since the model collapse can result in issues such as discrimination and a distorted perspective, experts are advocating for AI developers to utilise approaches that promote transparency and encourage human oversight in AI. We can now breathe a sigh of relief that AI has yet to take over the world.

  • (Link) Revolutionary AI System Detects Anomalies with Precision.: The Coupled Attention-based Neural Network (CAN) has been developed to improve the accuracy of multivariate time series anomaly detection (MTAD). This system is highly adaptable to different scenarios and utilises a special network to better detect anomalies. CAN has vast real-world applications such as risk management and smart surveillance. The authors highlight its importance in the detection and prevention of anomalies. Overall, this is an exciting development in the field of MTAD. Who knows, perhaps the CAN concept might 'CAN-opener' a new era for AI and anomaly detection!

  • (Link) The $4.4 Trillion Revolution: How AI is Changing the Global Economy for the Better: A recent study has revealed that AI has the potential to add an estimated $4.4 trillion of value to the global economy every year. This value would come from increased productivity and new growth opportunities in various industries including healthcare, retail and finance. This study highlights the transformative impact that AI can have on the world and the importance of investing in it. Looks like robots really will be taking over the world... in a good way!

  • (Link) Revolutionary AI vision transformer optimises computation and performance: Researchers have developed SoViT, a more efficient vision transformer for image classification, visual question answering, and zero-shot transfer. This technology has been optimised for size and shape using advanced scaling methods, resulting in competitive performance compared to larger models, with less computation. SoViT paves a new direction for optimising existing vision models whilst keeping resource consumption low. The 24-minute read provides greater detail on the optimisation process. Now, if only they could optimise my cooking skills, I’d be all set!

Code & Engineering

  • (Link) Extract Events Like a Pro with OmniEvent: The Open-Source Toolkit You Need!: Looking for an open-source toolkit for event extraction? Look no further than OmniEvent! This comprehensive toolkit includes both event detection and argument extraction processes and provides a unified approach for different paradigms. Plus, its modular design means it's easy to extend. Give it a try and extract events like a pro. Just don't forget to thank the toolkit for doing all the hard work!

  • (Link) Revolutionize your coding with Starcoder's Open Models: A new GitHub repository called Starcoder has shown that open code models can outperform open language models on challenging coding tasks. Closed models, which are actually coded models disguised as language models, often outperform open language models because they can utilise reasoning skills from extensive code training. The Starcoder model has achieved a remarkable score of 57 on the HumanEval Benchmarks, using instruction tuning and pretraining on Evol instructions. Both the models and code are available on the repository. It's time to switch to the Starcoder model and give ChatGPT a run for its money!

  • (Link) Introducing LEGO-Net: The Mind-Reading Butler for Your Home!: The article discusses a new innovation on Github called LEGO-Net, which uses machine learning algorithms to learn the regular arrangements of objects in rooms. The software has been trained on a dataset of over 20,000 room configurations and can recognise patterns in object arrangement and predict how objects will be rearranged. One possible application is in robotics, where the software could be used to help robots navigate and interact with their environments. It's like having a virtual mind-reading butler for your home!

  • (Link) Revolutionary Tutoring Tech Understands Your Thoughts!: The article introduces a new technology called Tutor GPT, which allows for more efficient tutoring by using theory of mind, a human-like understanding of others' thoughts and beliefs. Tutor GPT creates a meta-prompt system that adapts to a student's understanding by using natural language processing and machine learning. This technology could revolutionise the education industry by providing more personalised and effective tutoring experiences. Just when we thought machines couldn't get any smarter, they can now understand our thoughts too!

Tools & Products

  • (Link) Make any celebrity say anything with CopyCat: CopyCat is a new app that allows users to make their favourite celebrity say anything. A user needs to select a celebrity from the app's catalogue and then type or record the message they want the celebrity to say. The app uses deep learning algorithms to generate a video of the celebrity saying the message. The app can be used for fun or marketing purposes, such as creating a digital advertisement featuring a celebrity endorser. With CopyCat, the possibilities are endless! Just be careful not to make that secret crush declaration too obvious.

  • (Link) Revolutionise Your Design Workflow with Duply: Duply is an all-in-one visual automation platform that simplifies and speeds up your visual content creation process. With its artificial intelligence capabilities, you can quickly create stunning graphics, videos, and gifs for social media, websites, and other marketing channels. Duply also provides ready-to-use templates and a drag-and-drop editor to help you customise your content. Say goodbye to tedious editing tasks and hello to streamlined design creativity with Duply! It's like having your own personal designer at your fingertips.

Videos

  • (Link) Unlocking the Power of AI: Key Insights: The article highlights the key takeaways from Sam Altman's world tour discussing Artificial Intelligence (AI). He explains that AI's impact will be significant in the coming years, and it is crucial to ensure that its benefits are broadly shared. Additionally, Altman emphasises the importance of maintaining AI's ethical and safety standards. In 16 moments throughout his world tour, he elaborates on the need for continuous learning, strong governance, and cooperation to make AI work for everyone. Who knew AI could be so interesting?

Training & Education

  • Link) Learn to Create Chatbots without Coding with IBM's Free Course: Looking to learn how to create AI chatbots but don't have any programming experience? IBM has just launched a free online course that teaches you how to create chatbots with no programming skills required. The course covers everything from the basics of chatbots to natural language processing and advanced techniques. Plus, it's completely free! Time to say goodbye to the days of being intimidated by programming jargon and coding languages. Sign up now and start your journey to becoming a chatbot expert. And who knows...maybe you'll even create your own personal chatbot to help you with everyday tasks!

  • (Link) Limitless Career Possibilities with an LLM Degree: The article explores the diverse paths that individuals who hold a Master of Laws (LLM) degree can take in their careers. Some LLMs choose to work in academia while others opt for roles in private law firms or multinational companies. Some also go into public service, with a focus on international organisations or government institutions. The article concludes that the flexibility and prestige of an LLM degree allows for a wide range of career possibilities. Who knew that legal eagles had so many options?

Miscellaneous

  • (Link) Unlocking the Benefits of AI in Workflow: The article talks about the opinions people have regarding artificial intelligence (AI) in the workplace. While some believe AI will lead to job loss, others think it will help employees be more productive. Many acknowledge AI's potential to reduce human bias and enable more data-driven decision making. Some experts suggest that teamwork between humans and AI will be the key to success. Overall, the article highlights the different perspectives surrounding AI's impact on work and the need for careful consideration when implementing AI technologies. Remember, the robots are not taking over just yet!

  • (Link) Google Maps: Say Goodbye to Getting Lost!: Google Maps has introduced two new features - glanceable directions and an expanded Immersive View. Users can now get directions that are easier to follow, with key turns and decision points highlighted along the way. Additionally, the immersive view has been expanded to allow users to explore a 3D map of their destination. So, let's say goodbye to the days of getting lost and frantically looking for the nearest petrol station!

  • (Link) AI-Generated Pitches Three Times More Effective: According to a recent experiment conducted by startup research lab OpenAI, pitches generated by GPT-4, an artificial intelligence (AI) language model, are three times more likely to secure funding than those written by humans. The researchers found that investors were impressed by the language used in the AI-generated pitches and gave higher scores to them over human-written ones. Looks like robots are taking over the pitching world! Better start practicing our machinespeak.

  • (Link) Google Lens' New Tool Identifies Skin Conditions: Google Lens has a new feature that allows it to search for skin conditions in photos taken by users. This new tool will provide quick access to reliable information and can help individuals identify potential skin issues. However, this is not a substitute for seeing a doctor or dermatologist. So, don't skip your appointments--even if Google helps!

  • (Link) RoboGrammar: AI Plays with LEGO to Complete Tasks: AI has been trained to learn how to manipulate and rearrange objects like playing with LEGO to complete tasks. The new technology called RoboGrammar uses natural language processing to understand how to assemble different parts as a child would with LEGO bricks. The ability to allow AI to manipulate objects in this way will have a range of applications for use in industries such as manufacturing, assembly lines, and even healthcare. The AI of the future is looking very flexible. Time to hire a new robotic assistant!

Art of the Day

“aboriginal Australian dream time” By @ellephunt

Well, ByteSized buddies, it looks like our tech-tastic journey has come to an end today! Thank you so much for tuning in to Atometrix ByteSize, your daily dose of AI, low code, and tech goodness. We hope you've enjoyed every byte of information we've shared with you today. Remember, folks: stay plugged in, stay curious, and always stay ByteSized! Catch you on the flip side!

p.s. Don’t forget to checkout our platform or our apps!

  • Atometrix: The only low code platform the allows easy deployment of server side React Native. Now you can edit your UI and deploy changes without a developer and see it happen in your production apps!

  • Storyworm: The world’s first server side React Native application. Storyworm is a dynamic platform for all storytellers. Powered by AI, it lets you craft unique narratives, voice them through a text-to-speech feature, and combat writer's block with AI-generated plot suggestions. Beyond being an app, Storyworm is an inclusive community for story-lovers worldwide. This tool enables an immersive narrative experience that adapts and evolves according to users' needs.