The AI-ronman 🚀

Hello, Future Thinkers! 🧠💡 Dive into the world of AI where machines are getting smarter, and so are you. Let’s decode the latest together!

Quick Takes ⚡

1. Google rolls out Gemma 2 and upgraded Gemini models

2. Perplexity AI faces AWS investigation over policy breaches

3. Google launches Imagen 3 and two million token context windows in its Vertex AI cloud

4. Major security flaw found in Rabbit R1’s code

5. OpenAI imposes API restrictions on China, Russia, and more

6. Meta AI tests chatbots on Instagram

7. OpenAI partners with TIME magazine for century-old archives

8. Character.AI introduces multilingual AI call avatars

Deep Dive 🔍

• Google rolls out Gemma 2 and upgraded Gemini models 🌟
  
  Google introduces Gemma 2, the latest in their lightweight AI model series. The Gemini 1.5 Pro model now includes a 2M token context window and enhanced coding abilities. Gemma 2 is available in 9B and 27B parameter models, with a future 2.6B version hinted. The 27B model rivals larger models, and the 9B surpasses similar-sized models like Llama 3 8B. Gemini’s new features boost Python code generation and improve math and data tasks.
  
  

  

  
  

  
  
  

• Perplexity AI faces AWS investigation over policy breaches 🚨 
  
  Perplexity AI is under investigation by Amazon Web Services for possible terms of service violations, including copyright infringement and improper data collection. The company is accused of ignoring the Robots Exclusion Protocol by crawling prohibited websites. Controversy surrounds its 'Pages' product, which aggregates content from various sources, competing with original content. The CEO's defense and lack of understanding have sparked criticism. Despite substantial funding, Perplexity's practices and lack of transparency raise concerns.

• Google launches Imagen 3 and two million token context windows in its Vertex AI cloud 📈
  
  Google has launched Imagen 3 and introduced two million token context windows in its Vertex AI Cloud. Gemini 1.5 Flash offers low latency and a one million token context window, ideal for retail chatbots and document processing. Gemini 1.5 Pro supports up to two million tokens for multimodal tasks like code analysis and audio processing. Imagen 3 generates images 40% faster than its predecessor, though not yet matching industry leaders in quality. Google is also enhancing Vertex AI with more third-party models and cost-saving features.
  
  

  

  
  

• Major security flaw found in Rabbit R1’s code 🔐
  
  Rabbitude developers found a major security flaw in Rabbit's R1 AI gadget, with hardcoded API keys in the code. These keys exposed Rabbit's accounts with ElevenLabs and SendGrid, risking unauthorized data access. The breach allowed access to all R1 device responses. Rabbit delayed addressing the issue for over a month, leaving some keys still active. Rabbit acknowledged the breach and is investigating customer data safety on its website and Discord channel.
  
  

  

  
  

• OpenAI imposes API restrictions on China, Russia, and more 🚫
  
  Starting July 9th, OpenAI will tighten API restrictions on unsupported countries, including China, Russia, North Korea, and Iran. This move follows a warning to developers about API use from these regions. Developers struggle to identify traffic from these areas due to limited data. OpenAI found state-backed actors from these countries misusing AI for propaganda, with minimal impact due to frequent errors. The new restrictions aim to prevent such misuse, crucial during an election year.
  
  
  

• Meta AI tests chatbots on Instagram 🚀 
  
  Mark Zuckerberg announced that Meta AI is testing creator-made AI characters on Instagram. Starting in the U.S., these AI chatbots will be clearly marked for user awareness. AI Avatars will help creators and small businesses engage with their communities. Initially, Meta will test the feature with 50 creators and a small user group. The full launch is planned for August after further testing.
  

• OpenAI partners with TIME magazine for century-old archives 📰
  
  OpenAI has formed a multi-year partnership with TIME magazine, gaining access to its 101-year archive to enhance ChatGPT and other products. TIME will leverage OpenAI's technology to develop new offerings. This collaboration follows deals with publishers like Le Monde, Axel Springer, and Associated Press. While OpenAI is securing substantial partnerships, the selection criteria and US market focus raise some concerns.

• Character.AI introduces multilingual AI call avatars 📞
  
  Character.AI now lets users talk to AI characters over calls in multiple languages. With over 3 million users making 20 million calls for various purposes, the feature offers a smooth experience with user-generated AI avatars. It includes reduced latency, seamless switching between calling and texting, and a 'Tap to interrupt' option. Character.AI also launched 'Character Voices,' providing over 1 million voices. Reddit users voiced concerns about AI character personality changes, which the company addressed.
  
  

  

• More updates from OpenAI
  
  OpenAI has introduced CriticGPT, an AI model based on GPT-4 designed to detect errors in language models like ChatGPT. Aimed at improving human trainers' evaluations through RLHF, CriticGPT was preferred over ChatGPT's error detection 63% of the time. Combining human feedback with CriticGPT yielded more thorough criticism and fewer hallucinations. OpenAI views CriticGPT as a step towards scalable oversight for evaluating advanced AI systems.
  
  

  

  
  
  
  OpenAI has delayed the launch of ChatGPT's advanced voice capabilities, initially set for late June, to the fall. The company is refining content moderation, user experience, and infrastructure for a safer rollout, now starting in late July for Plus users.
  
  
  Toys"R"Us has launched its first AI-generated commercial using OpenAI's video generator, Sora, in collaboration with Native Foreign. The ad, featuring mascot Geoffrey the Giraffe and founder Charles Lazarus, tells the brand's story and is viewable on the Toys"R"Us website.
  
  

  

GlazeGPT just announced their industry-leading innovation - text-to-MongoDB queries, a revolutionary feature that translates natural language into precise MongoDB queries.🔥This latest update empowers users to effortlessly interact with complex MongoDB datasets and generates complex queries in seconds while understanding your business glossary.

Level up your skills with this 14-day free trial - don't miss out! ⚡️

That's a wrap, Digital Dynamos! 📡🤓 Keep innovating, stay connected, and remember—AI is only as smart as the questions we ask. Catch you next issue!

Ciao for now!

Author: Poonam 👧 

Karan 😎 🚀