ByteSize Daily Wrap - 2023/06/27

Welcome, tech enthusiasts, to another episode of Atometrix ByteSize, the podcast that brings you daily updates in the world of AI, low-code, and all things tech! Today, you're in for a real treat as we'll be discussing Google's next-gen speech technology, AudioPaLM, along with futuristic AI art models, YouTube's AI dubbing feature, and so much more!

With Google's new AudioPaLM offering a groundbreaking blend of text-based and speech-based language models, it's impossible not to be blown away by the potential applications for speech translation and voice-to-text. Imagine the convenience of conversing with our phones in any language we choose! Speaking of exciting developments, Midjourney's "Zoom Out" AI art model has us all rethinking the bounds of what's possible in the realm of AI-generated art. It's a beautiful time to be alive, and even our precious robot companions are learning to handle delicate objects, thanks to SoftGPT!

In today's era of rapid automation, it comes as no surprise that 46% of companies are boosting their automation efforts. After all, who can resist the allure of increased efficiency and productivity? However, we must also ensure that we strike the right balance between technological progress and human values, lest we risk alienating our workforce in the process. On the topic of AI code models that self-repair, it's fascinating to consider that while humans sometimes struggle with basic DIY tasks, AI models like GPT-4 are stepping up to save the day by mending their own errors. What a time to be in tech!

Going hands free? Now available as a podcast!

Have you had this forwarded to you? Subscribe now!

News & Articles

• (Link) Google Launches AudioPaLM: Next-Gen Speech Technology!: Google has launched its new technology called AudioPaLM, which is a huge language model designed for speech comprehension and generation. The technology integrates both text-based and speech-based language models to create a singular system that can analyse and produce speech and text. This invention is ideal for audio and speech projects, including speech translation and voice-to-text applications. With this technology in the market, the world of speech technology is changing drastically. Who knows, we may all soon be chatting with our phones, speaking any language we choose.

• (Link) "Zoom Out" AI Art Model Wows Community: Midjourney has revealed the latest version of its AI-powered image synthesis model, which has impressed the AI art community with its new "zoom out" feature. The addition allows for a larger scene to be built around a central synthesised image, mimicking the effect of zooming out with a camera lens. Midjourney continues to innovate, with even more exciting developments to come in the future. Who needs a human artist when you have AI talent like this?

• (Link) Revolutionary YouTube AI Dubbing Feature Unveiled!: Get ready to see more videos in your native language as YouTube rolls out an AI-powered dubbing feature. The new tool will allow creators to automatically add captions to their videos, as well as select from a range of synthetic voices to dub their content. This will make it easier for international audiences to enjoy videos produced in other languages. The feature is still in early development, but it's an exciting glimpse into the future of video translation. Just don't expect the AI voices to win any Oscars for Best Actor!

Insights & Papers

• (Link) 46% of Companies Boosting Automation: According to a research report by Appfire, 46% of companies are planning to increase their automation efforts within the next six months. The report explores the potential use cases of automation across various business sectors and dispels any misconceptions regarding the benefits of automation. To learn more insights and data from more than 200 industry experts, Appfire is offering a free download of the report. Who knew machines would one day do all our work?

• (Link) Teaching Robots to Handle Delicate Objects: SoftGPT is teaching robots how to handle soft and delicate objects, such as clothes and pillows, found within households. This is an often challenging task for robots, but SoftGPT's model allows robots to understand the movements and shapes of objects and predict the outcome of actions. With extensive exploration data, robots can learn to handle objects safely and with precision. Who knew robots could be so gentle?

• (Link) AI code models that self-repair: The process of program synthesis involves automatically writing computer code. While models such as GPT-4, Turbo, WizardCoder and StarCoder excel at writing code, many fail to self-repair when the code is incorrect. A new paper has measured the self-repair ability of such code models and found that GPT-4 is one of the few to exhibit such repair. The assumption is that this model was explicitly trained for self-repair by OpenAI rather than this being an emergent feature. Isn't it great to know that some AI models can now self-repair, while most of us humans struggle to do the same with even basic DIY tasks?

Code & Engineering

• (Link) Revolutionary PanoHead Creates Hyper-Realistic 3D Heads!: This new project, PanoHead by GitHub, offers geometry-conscious 3D full-head synthesis in 360°. It allows for highly detailed head synthesis and rendering using geometric analysis and careful attention to picture depth. It's a great tool for creating realistic and compelling 3D head models. And just when you thought 360° technology couldn't get any better, Panohead comes along!

• (Link) Get Creative with Drag Your GAN: Official Code Released!: Exciting news for machine learning enthusiasts! The official source code for Drag Your GAN has finally been released on GitHub, allowing users to experiment with the Generative Adversarial Network (GAN) model which creates realistic images. The creators have also provided a paper and a link to try it out. Get ready to unleash your creativity with Drag Your GAN. And who knows, maybe you'll create images of dragons or even a GAN that can teach you how to cook!

Tools & Products

• (Link) AI Solution for Privacy Protection: MOSTLY AI offers a solution to data hurdles by creating synthetic data based on samples of a company's real data. This AI technology learns the granular level details of correlations, distributions, and properties while keeping data privacy in mind. It allows for shorter time-to-data and more machine learning models in production. MOSTLY AI can provide statistically representative data that is fully compliant with the strictest privacy laws for free. The service can readily generate up to 100k rows of data per day, and it is free forever.

• (Link) Testing with Codium's AI Automation!: The Codium product offers a solution to improve accuracy and customer satisfaction by using AI to automate test writing. This means that results are more reliable, and time is saved, allowing for increased productivity. Say goodbye to tedious manual testing! With Codium, you'll have more time to do the things you actually enjoy.

• (Link) Transform into Hyper-Realistic Avatars with Avaturn: Avaturn is a new technology that uses 3D scanning and machine learning algorithms to create hyper-realistic avatars. Users can have their bodies scanned into the system, then use the avatars for virtual and augmented reality experiences. The technology has potential uses in gaming, social media, and even medical simulations. With Avaturn, you'll never have to settle for a generic character model again. Time to say goodbye to your dad bod and hello to your new buff avatar.

Videos

• (Link) Get Ready for the 7th Research and AI Summit!: The 7th Research and Applied AI Summit is coming! This exciting event brings together top AI researchers and practitioners to share their latest breakthroughs and insights. Attendees will have the chance to network with fellow experts and learn about cutting-edge AI applications across a variety of industries. Don't miss this incredible opportunity to stay on the forefront of AI innovation! See you there, or will I be seeing robots instead?

Training & Education

• (Link) Welcome to the Age of Smart Learning: The rise of AI has drastically changed our understanding of knowledge acquisition. In the past, knowledge was hard to get, but now AI has made it accessible to us. With the power to access vast amounts of information faster, AI has created a new "burden of knowledge" where individuals are expected to know more about various subjects. It's like the classic "just Google it" phrase, but on steroids. Looks like we have to keep up with the AI race!

• (Link) Harvard Launches AI-Powered Program to Revolutionise Education: Harvard University has launched an AI-powered initiative aimed at improving education quality by supporting teachers in classrooms. The program relies on AI algorithms to track students' work and progress, delivering feedback to teachers in real-time. The main goal is to help teachers personalise learning for each student and address their individual academic struggles. The project is expected to scale up across the US in the coming years, significantly benefiting students' education journeys. AI is becoming more common in the education sector, with several institutions harnessing its power to help enhance academic outcomes. Looks like we might need to brush up on our coding skills!

Miscellaneous

• (Link) Capcom Town: Your Ultimate Anniversary Gaming Destination!: Capcom, the popular video game company, is celebrating its 40th anniversary by launching a new website called Capcom Town. If you’re bored and looking for something fun to do, the website has a range of exciting games and activities for all to enjoy. It’s an excellent way to spend some leisure time or take a break from work! Just don't let your boss catch you playing!

• (Link) Voice AI Library: ElevenLabs has created a Voice Library that allows users to share and use artificially intelligent voices. This Voice Library could demonstrate the AI capabilities of a particular system. It will help ease the workload and make artificial voice recording more efficient. Using this library would help individuals trying to implement voice technology by decreasing time and cost requirements. With the Voice Library created by ElevenLabs, we are on our way to a more advanced future with AI!

• (Link) Corporate Spying with Military-Grade AI: Invasion of Privacy: This article talks about how companies are using military-grade artificial intelligence (AI) to track down internal leakers, critics, and labor unions. They are misusing the technology which was designed during wartime to locate enemies for their own unfair advantage. This is disturbing because such technology is being used to invade the privacy of the employees which is a breach of their fundamental rights. Looks like corporations would rather bully their workforce with powerful machines than work towards providing better working conditions. Scary stuff!

Art of the Day

_{“funny MAN WITH LONG HAT in Alebrije style 3D tromp l'oeil showing the texture of thick oil paint strokes on the rustic canvas, vibrant colors, basquiat style sharp focus” By @tombarreto}

Well, folks, it looks like our daily summation of AI and tech goodness is coming to an end. Thanks for listening and keeping those neurons firing with us. Don't forget to stay ByteSized, and follow us for more scrumptious daily updates on all things AI and tech. Let's continue to consume knowledge as fiercely as our future robot overlords do energy! Until next time, fellow tech enthusiasts – stay ByteSized!

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.