Inflection AI: the company that created the ChatGPT analogue

Date:

Pi is not just a chatbot, but an interlocutor: this is how Inflection AI positions its product. The startup, which is only two years old, is already attracting large investments and collaborating with well-known corporations

Who created the company, Inflection AI?

Inflection AI was founded in March 2022. The startup positions itself as “an artificial intelligence studio creating personal AI for everyone” [1].

Inflection AI is led by three entrepreneurs who previously worked in other IT projects.

  • Mustafa Suleiman is the co-founder of the artificial intelligence company DeepMind. The company was subsequently acquired by Google. Suleiman left the corporation at the beginning of 2022 and started thinking about creating his own startup, which would develop algorithms with conversational interfaces.He is the CEO of Inflection AI.
  • Reid Hoffman is the creator of the business social network LinkedIn, one of the first investors in OpenAI. Co-founder and board member of Inflection AI.
  • Karen Simonyan is an Oxford graduate who developed his own startup, which was later acquired by DeepMind. Worked as chief scientist at DeepMind. At Inflection AI, he takes the same position.

Introducing Inflection, Reed Hoffman said, “Our mission was to create a tool that is designed to understand the user, not the other way around. We wanted it to not only answer questions, but also be part of a dialogue that helps you on your journey in life – whether it’s communicating with other people, making decisions, figuring out how to deal with a problem at work . ”

Key milestones in the history of Inflection AI

The startup launched in March 2022 with financial support from Microsoft, Nvidia, as well as billionaires Reed Hoffman (co-founder of LinkedIn), Bill Gates (co-founder of Microsoft) and Eric Schmidt (former CEO of Google). The startup raised $225 million in seed capital [3] .

In May 2023, Inflection AI launched a conversational chatbot called Pi, an analogue of ChatGPT [4] . The main difference from other chatbots is that Pi talks to the user and gradually accumulates knowledge about him based on conversations. Pi itself responded that it is not a chatbot, but “a conversational AI, which means it can have more natural and interesting conversations” [5] .

The Pi chatbot was launched based on the Inflection language model. In November 2023, the company announced that it had completed training of Inflection-2, “the best model in the world for its computing class and the second largest LLM (large language model) in the world today.” .

The company noted that the new model is more capable than the previous version: the developers have improved the style of answers and reasoning capabilities [6]. The latest update to date is the Inflection-2.5 model, which was introduced in March 2024 [7]. In terms of performance, it is as close as possible to the results of GPT-4, while requiring 40% less computing resources.

Work is also underway on the company’s flagship product, the Pi chatbot. So, in March 2024, the company announced that it became possible to send messages to a chatbot via iMessage, a messaging system on Apple devices. Today the company employs several dozen employees. The headquarters is located in the Californian city of Palo Alto.

future of the company

In June 2023, the startup, which was valued at $4 billion, attracted $1.3 billion in investments from Microsoft, Hoffman, Gates, Schmidt, and Nvidia. This further strengthened the startup’s ties with two key partners in the AI ​​race [8].

Microsoft provides cloud computing infrastructure, and Nvidia (a GPU developer) partners with CoreWeave (cloud services for solving AI problems) to integrate GPUs for LLM training.

Together with partners, Inflection AI is creating the world’s largest supercomputer from 22 thousand Nvidia H100 Tensor Core GPUs. Such computing power will make it possible to create a new generation of artificial intelligence models and train them quickly and efficiently.

Now the future supercomputer includes 3.5 thousand processors; it has already proven to be the fastest at training large language models [9]. With its latest major funding in June 2023, Inflection AI plans to continue expanding the supercomputer’s computing capabilities as well as developing its core product, Pi [10].

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Share post:

Subscribe

Popular

More like this
Related

Artificial Intelligence Tools That Can Be Used in E-Export

In the "ChatGPT and Artificial Intelligence Tools in E-Export"...

What are SMART goals, why are they needed and how to set them correctly

In the modern world, where everyone strives to achieve...

How and why the United States is developing a lunar economy

The United States is seriously thinking about developing an...

China faces problem of untreatable gonorrhea

In China, there are a growing number of strains...