SAN FRANCISCO, CALIFORNIA – On the 6th of November, 2023, in San Francisco, OpenAI’s CEO, Sam Altman, took the stage at the inaugural OpenAI DevDay event. Altman headlined the conference with a keynote speech that unveiled OpenAI’s latest advancements and strategic moves in the competitive landscape of artificial intelligence.
OpenAI introduced its enhanced GPT-4 Turbo, a formidable upgrade to its AI lineup, and announced a groundbreaking feature allowing the creation of personalized ChatGPT variants. In a strategic pricing move, OpenAI has also reduced the cost of using its AI services, a change that is likely to resonate well with the developer and corporate community.
The announcements underscore OpenAI’s dynamic progression and its commitment to maintaining a lead in the AI industry, contending with other tech giants such as Anthropic, Google, and Meta. OpenAI’s ChatGPT has seen a meteoric rise in popularity, now boasting around 100 million weekly active users. OpenAI CTO Mira Murati highlighted the platform’s widespread adoption, noting its use by over 92% of Fortune 500 companies, a significant increase from 80% just a few months prior.
A highlight of the event was a guest appearance by Microsoft’s CEO, Satya Nadella, who emphasized the importance of cutting-edge systems and computing power in advancing AI technology—a commitment that Microsoft is ready to support as OpenAI’s strategic partner.
Earlier in the year, Microsoft’s bolstered investment in OpenAI, totaling an additional \(10 billion, stood out as the year’s most significant AI investment. OpenAI’s valuation soared, with recent funding rounds and talks positioning the company at an estimated \)80 billion valuation.
Altman’s address to the audience reflected a year’s worth of dialogue with developers, focusing on their requirements and feedback. Regarding the future, specifically the anticipated GPT-5, Altman expressed OpenAI’s intent without committing to a definitive timeline.
Highlights from OpenAI’s Announcements:
GPT-4 Turbo
The new GPT-4 Turbo model now processes information up to April 2023, surpassing previous models’ knowledge cutoffs. This update means GPT-4 Turbo can provide current and relevant responses to queries about recent events—something its predecessors couldn’t manage.
Altman expressed a shared frustration with the limitations of past models, emphasizing the importance of staying current. GPT-4 Turbo also boasts an impressive input capacity, capable of processing documents up to 300 pages, allowing for comprehensive summaries of extensive texts.
Additionally, GPT-4 Turbo integrates DALL-E 3 for AI-generated imagery and offers text-to-speech capabilities with a selection of six preset voices, providing a more versatile and engaging user experience.
Developers can access GPT-4 Turbo in a preview release, with a full rollout expected in the following weeks. OpenAI also announced significant price reductions for developers, making GPT-4 Turbo’s input tokens three times more affordable and output tokens twice as cheap as the previous GPT-4 model, enabling more cost-effective operations for businesses and developers utilizing the AI models.
Custom Chatbot Creation
Previously, only enterprise and business clients had the privilege of uploading their data to tailor ChatGPT for specific sectors and applications. OpenAI is now democratizing this capability, allowing anyone to craft their own bespoke chatbots.
The concept of AI "agents" has become increasingly popular, with numerous startups competing to deliver personalized AI tools akin to those seen in pop culture, such as Tony Stark’s J.A.R.V.I.S. from the Marvel Cinematic Universe or Pam from the Disney Channel’s "Smart House."
OpenAI has streamlined the process of creating a personalized GPT. According to their announcement, no coding skills are needed. Individuals can design these chatbots for personal use, for internal company operations, or for public access. The creation process is as straightforward as initiating a dialogue, providing the bot with instructions and additional knowledge, and selecting its capabilities, which can range from web searches to generating images or analyzing data.
With over two million developers already utilizing ChatGPT’s API to build their tools, the potential for encountering customized AI chatbots across various apps and websites is set to increase significantly.
Open AI’s Take on the App Store
With the new ability for users and developers to deploy their personalized AI chatbots, OpenAI is launching a platform that could become a significant source of revenue: its own app store equivalent.
The GPT Store is a marketplace where creators can share their custom GPTs for public use. OpenAI has announced that in the near future, creators will have the opportunity to monetize their GPTs based on user engagement.
Creators’ GPTs will be searchable within the store and can ascend the leaderboards. OpenAI plans to feature standout GPTs in various categories such as productivity, education, and entertainment, spotlighting those that offer exceptional utility or enjoyment.
Regarding the revenue model for creators whose chatbots are hosted in the store, OpenAI initially plans to distribute a portion of the overall subscription revenue. Altman mentioned that the payout would be based on active user metrics and category bonuses, with potential future support for specific GPT subscriptions.
Altman emphasized to the press that OpenAI’s core product is intelligence. The company’s trajectory is expected to increasingly focus on selling not just intelligence but intelligent agents, shaping the future of how AI integrates into everyday life.
Integrated AI-Powered Features for Enhanced User Experience
Previously, utilizing the full suite of OpenAI’s capabilities required users to navigate through multiple applications and websites, which presented a steeper learning curve. However, as of Monday, OpenAI has consolidated its various AI tools into a single, user-friendly interface. Now, ChatGPT users can enjoy a seamless experience with integrated features such as image creation with DALL-E, web browsing, data analysis, and the ability to upload and search through documents and PDFs. Before this integration, Anthropic’s Claude was the sole chatbot offering in the market with the ability to search PDFs.
Proactive Legal Protection for Copyright Issues
In response to the increasing number of legal challenges associated with generative AI, Altman announced a new initiative to actively protect OpenAI’s customers. Should any user face legal allegations of copyright infringement, OpenAI has committed to intervening on their behalf and covering any resulting expenses. This move is in line with similar assurances provided by tech giants such as Google, Microsoft, and Adobe, aiming to alleviate the legal concerns of users engaging with AI-generated content.