♾️ Meta’s open-sourced multi-sensory model

Welcome, humans.☕

What to decode:

• ♾️ Meta’s open-sourced multi-sensory model

• 🤖 IBM launches an AI platform for enterprises

• 🅰️ Anthropic is backing Constitutional AI over RLHF

Read time: 4 minutes

♾️ Meta’s open-sourced multisensory model

Meta is making a comeback (once more!) with an exhilarating open-source venture. Let us introduce ImageBind, a cutting-edge AI research model designed to comprehend and fuse text, audio, visual, movement, thermal, and depth data.

This amalgamation of various data types unleashes a realm of fresh opportunities, including:

• Multimodal search capabilities

• Arithmetic within embedding spaces

• Audio-to-image generation conversion

Currently, ImageBind exists as an open-source experiment, yet it harbors immense potential for multimedia search, virtual reality, and robotics in the foreseeable future. Meta continues to astound the community with their generous act of open-sourcing these remarkable models.

🤖 IBM launches an AI platform for enterprises

During IBM's Think conference, the introduction of Watsonx took center stage—a cutting-edge AI platform developed by the company to empower businesses in building and leveraging AI models.

Amidst fierce competition from fellow tech behemoths and startups, IBM distinguishes Watsonx by providing pre-trained enterprise models on a cost-effective infrastructure.

With IBM's recent decision to replace 7,800 jobs with AI, and now the unveiling of Watsonx, it becomes evident that the company is making substantial investments in AI. The burning question lingers: Can IBM match the pace of industry giants such as Google, Meta, and Microsoft? Only time will unveil the answer.

🅰️ Anthropic is backing Constitutional AI over RLHF

Currently, there appear to be two prevailing viewpoints within the field of artificial intelligence: RLHF versus Constitutional. This divide can now be likened to the competition between OpenAI and Anthropic, a company founded by former OpenAI employees with the goal of developing more ethically-driven AI. RLHF, or Reinforcement Learning from Human Feedback, relies on human involvement to interpret and evaluate the outputs of AI models according to human preferences.

On the other hand, the Constitutional approach eliminates the need for human intervention by employing a "constitution" of predefined values. A language model is then employed to assess the extent to which the responses generated by another language model align with these constitutional values. The Constitutional approach boasts greater scalability and does not rely on inexpensive human labor, making it an appealing prospect for Anthropic and their vision for the future.

For an in-depth exploration of how Constitutional AI can enhance RLHF, please refer to this detailed write-up: [Read More].

FROM OUR SPONSOR

Experience the power of Notion AI, your all-in-one writing assistant. It saves you time, enhances your creativity, and improves your writing in countless ways.

• Effortless first drafts: Let Notion AI handle the challenging task of starting your writing. It generates ideas for you to turn into great content.

• Ignite creativity: Instantly get a list of ideas on any topic, sparking new and unexpected directions for your work.

• Impeccable editing: Notion AI serves as your eagle-eyed editor, catching spelling and grammar mistakes, and even providing translations.

• Concise summaries: Say goodbye to sifting through long meeting notes. Notion AI extracts the most important points and action items for you.

Notion AI is your teammate throughout the writing process, supporting you from start to finish.

Try it now and unleash the full potential of your writing!

FASCINATING FINDS

• On this day - May 10, 1894, “Wireless” is born when Guglielmo Marconi sends a radio wave three-quarters of a mile. Three years later the Marconi Company will successfully communicate “ship to shore” over a distance of twelve miles. Marconi’s work leads to the commercialization and proliferation of most of the radio technologies we know today. (link)

• Weird but interesting - Humans have been performing dentistry since 7000 B.C., which makes dentists one of the oldest professions. (link)

• Mind-boggling statistics - It takes an individual photon several millenia to go from the center to the surface of the sun.

  Then it takes only 8 minutes to reach Earth, and a couple days at most to leave the solar system. (link)

MISCELLANEOUS

(News, podcasts, videos, blogs etc)

• Bessemer Venture Partners is committing $1 billion to invest in AI startups. (link)

• Google Cloud’s generative AI partnership with popular workplace apps. (link)

• Accelerating tech adoption with AI. (link)

• AI is not good software. It is pretty good people. (link)

• Palantir stock soars 21%, says demand for AI unprecedented. (link, without paywall here)

• IBM intros a slew of new AI services, including generative models. (link)

• With seed funding secured, AI detection tool GPTZero launches a new browser plugin. (link)

• Ascend raises $25 million for pre-seed AI startups in the Pacific Northwest. (link)

• Eventbrite integrates GPT capabilities into the platform to aid the event planning process. (link)

• What made Hinton into an AI doomer. (link)

• TidyBot - Personalized robot assistance with large language models. (link)

• FrugalGPT - How to use LLMs while reducing cost by 98% and improving performance. (link)

That's all for now!

As always, thanks for reading, and see you next time. 🫡