Meta’s 4000 Languages AI, OpenAI AGI Fear

Meta’s 4,000 Languages AI

Meta has released an AI language model called Massively Multilingual Speech (MMS) that can recognize over 4,000 spoken languages and generate speech in over 1,100 languages. To collect audio data, Meta utilized translations of religious texts, which provided publicly available recordings in various languages. They trained the MMS models using wav2vec 2.0, a self-supervised speech representation learning model. The results showed impressive performance, surpassing existing models like OpenAI's Whisper in terms of word error rate and language coverage.

Despite most recordings being read by male speakers, the models performed equally well with both male and female voices. Meta recognizes the models' imperfections and the possibility of mistranscribed words or offensive language, highlighting the importance of collaborative efforts in responsible AI development.

By sharing the MMS models as open-source for research, Meta aims to combat the decline of lesser-supported languages due to technology. They envision a world where technology, including assistive tools and virtual/augmented reality, supports language preservation and enables individuals to access information and communicate in their native languages.

Read More

OpenAI Want’s AI to be treated like Nuclear

OpenAI's cofounders have proposed the need for an AI regulatory body, akin to the International Atomic Energy Agency (IAEA), to address the existential risks associated with AI. This agency would have the authority to conduct audits, inspect systems, impose deployment restrictions, and monitor computer and energy usage to ensure the safe development of superintelligence.

While this proposal aims to mitigate risks, it differs from the EU AI Act, which imposes licensing requirements on all models.

Some skeptics argue that OpenAI's push for regulation may be driven by a desire for market dominance through regulatory capture, potentially stifling competition in the AI startup landscape, but its also worth mentioning that the blog makes it clear that smaller projects should not be subject to the same “burdensome” restrictions, so as to foster innovation.

Read OpenAI Blog

In other News:

Couldn’t decide which of these Articles to feature third so i’ll just give a short summary on each headline then you can decide which captures your interest.

Google’s AI Policy

Google recently published an AI Policy Agenda paper outlining its vision for responsible AI deployment and suggestions for government regulations. The agenda emphasizes the need for a balanced approach that protects against harmful outcomes while promoting innovation. It focuses on three core objectives: opportunity, responsibility, and security.

Apple 30% ChatGPT Tax

Basically the ChatGPT App which is currently live on the App store is being charged 30% of every Plus subscription (as Apple does with every in-app purchase on IOS). The Article highlights how Apple despite not launching their own AI will benefit massively from the AI boom as AI Platforms will build for Mobile (IOS), which is where their 30% fee will help with capturing financial value created.

Fake AI image of Pentagon exploding goes viral on Twitter and causes US markets to plummet

An AI generated Image went viral on twitter and fooled a couple media outlets who posted without fact-checking, which ended up tanking the stock market for a bit.

Bill Gates says A.I. could kill Google Search and Amazon as we know them

Gates is claiming with the advent of AI , people won’t go to google to search or amazon to buy stuff, they’ll just ask their AI’s to do it which will lead to a massive drop in traffic for Google and Amazon, essentially just change how we use the internet.

Twitter accuses Microsoft of misusing its data for AI

Twitter sent a legal letter to Microsoft saying they Misused their Access to twitter API. The major line in the article is “Microsoft’s agreement with Twitter barred it from overuse of the service’s data such as exceeding “reasonable request volume” or “excessive or abusive usage.” Spiro then noted that “despite these limitations,” Microsoft had retrieved more than 26 billion tweets in 2022 alone.”

Request assistance from ChatGPT to help you use it more effectively on a particular topic

Prompt:

Copy/paste the prompts here: https://www.promptvibes.com/view-and-test?recordId=rec5vYCGUZV5VM8lp

CourseCreatorGPT

You can buy that $199 Gumroad Course, or you can get ChatGPT to make it for you.

Prompt:

Copy/paste the prompts here: https://www.promptvibes.com/view-and-test?recordId=recmL1AfadxksnTkN

Generate/Identify new, innovative big ideas in a field of interest.

Prompt:

Copy/paste the prompt here: https://www.promptvibes.com/view-and-test?recordId=recZz9kDSO9PZL5zp

How this Tool Can Boost Your Productivity

Fibery AI can boost productivity by facilitating idea generation, enhancing writing quality, automating repetitive tasks, enabling process experimentation, and providing a seamless work context. By leveraging AI capabilities, the tool aims to streamline and optimize various aspects of your work, ultimately helping you accomplish more in less time.

Realism painting of a bunny smoking a cigarette after a hard day working in a coal mine