Chapter 1: Unveiling GPT-Next

OpenAI has notably slowed its pace of releases this year amidst fierce competition in the realm of large language models. However, an exciting new model is on the horizon.

At the recent KDDI Summit, Tadao Nagasaki, CEO of OpenAI Japan, shared groundbreaking updates regarding the upcoming ChatGPT model. He introduced "GPT-Next," which is anticipated to be a staggering 100 times more powerful than the current GPT-4 model, with a launch expected later this year.

Reports suggest that GPT-Next has been developed using a compact variant of OpenAI's enigmatic "Strawberry" model. The enhancements in this model stem from advancements in its architecture and learning efficiency rather than solely depending on extensive computing resources.

Tadao also disclosed that OpenAI employs fewer than 2,000 staff members globally, with about half dedicated to AI development. The organization’s notable clients include industry giants like Apple, Coca-Cola, and the biotechnology leader Moderna.

In his keynote, Tadao provided an optimistic view of the GPT series' future, contrasting the capabilities of GPT-3 and GPT-4. He highlighted that, unlike traditional software, AI is set for exponential advancements.

Furthermore, various media outlets have reported that OpenAI is in negotiations for a new funding round, which could elevate its valuation to over $100 billion, potentially positioning it as the most valuable AI unicorn globally. Major investors in this round include Thrive Capital, Microsoft, Apple, and Nvidia. Reports from The Information indicate that this funding round may also see OpenAI transition from a nonprofit to a for-profit model.

Additionally, statistics reveal that since November of last year, the number of weekly active users on ChatGPT has doubled, surpassing 200 million.

In connection with OpenAI's fundraising efforts, former chief scientist Ilya Sutskever has launched a new company, SSI, which recently secured $1 billion from investors such as Sequoia Capital and Andreessen Horowitz. This funding will be directed towards acquiring computational power and attracting elite AI talent. Despite being a new startup with only ten employees, SSI has already reached a valuation of $5 billion.

The founding team of SSI includes Ilya, along with Daniel Gross and Daniel Levy. Levy serves as the chief scientist, while Gross is tasked with fundraising and securing computing resources. Before co-founding SSI, Daniel Gross was a prominent investor and led Apple's AI division, gaining recognition for backing significant AI startups, including Keen Technologies, which aims to develop AGI (Artificial General Intelligence).

Public records indicate that Gross's early venture, Greplin, eventually became Cue, which was acquired by Apple in 2013 for about $40 million. At just 23, Gross became Apple's Director of Technology, focusing on integrating machine learning into iOS and macOS.

In an interview with Reuters, Gross emphasized the importance of finding investors who align with their mission of pursuing safe superintelligence. "We plan to dedicate a few years to product development before taking it to market," he stated.

Levy, praised for his collaboration with Ilya at OpenAI, remarked, "The timing is perfect for launching this project. My vision is in sync with Ilya's—building a focused, agile team aimed at achieving safe superintelligence."

Unlike OpenAI's initial nonprofit model, which emphasized AI safety, SSI has adopted a standard for-profit framework. Regarding their recent funding, SSI has indicated plans to partner with cloud and chip companies to meet its computational needs, with Microsoft and Nvidia as potential collaborators.

Ilya Sutskever, a prominent advocate of the Scaling Law, believes that as AI models grow in scale—through parameters, training data, or computational resources—their performance and capabilities significantly improve.

As per Reuters, Ilya aims to pursue scaling in a manner distinct from OpenAI, although specific details remain undisclosed. "While everyone discusses the Scaling Hypothesis, they overlook a crucial question: What are we actually scaling?" he noted. "Some may work harder, but they’re merely accelerating the same approach. We strive for innovative outcomes through diverse methodologies."

OpenAI's GPT-Next is set to revolutionize AI with unprecedented capabilities, outpacing existing models by a significant margin.

Chapter 2: The Competitive Landscape of AI

As the AI landscape evolves, the emergence of new companies like SSI reflects a growing investment in the future of artificial intelligence.

A deep dive into whether ChatGPT-4o outperforms GPT-4, exploring OpenAI's latest flagship model and its impressive features.