HomeWinBuzzer NewsOpenAI Dev Day Unveils GPT-4 Turbo, Multi-Modal Capabilities, Assistants API and More

OpenAI Dev Day Unveils GPT-4 Turbo, Multi-Modal Capabilities, Assistants API and More

By making AI more cost-effective and easier to deploy with better and cheaper tools, OpenAI could disrupt the AI marketplace and alter business strategies regarding AI utilization.


OpenAI has unveiled a series of new features for its GPT API at its Developer Day event attended by developers and industry professionals. During the Developer Day held in San Francisco, OpenAI CEO announced the updates which include an expanded context window for an updated and faster version of GPT-4, an up-to-date knowledge cutoff, and copyright protections for developers using the API.

's Chief Executive Officer, Satya Nadella, graced the event, emphasizing the significant investment and collaboration between the two titans. During his keynote, Altman praised the relationship with Microsoft as potentially the “best partnership in tech.” On stage, Nadella underscored the transformation that AI workloads have prompted within Azure's infrastructure, from data center designs to networking solutions tailored to cater to the specific needs of models. The dynamic partnership focuses on creating an optimized environment that allows OpenAI to develop state-of-the-art models while simultaneously providing developers with access to these advancements through Azure's cloud platform.

Custom GPTs with Store and Monetization

The highlight of the event was the introduction of specialized models of , known simply as GPTs. These models can employ custom instructions and incorporate context from uploaded files, thus extending beyond the capabilities of the standard ChatGPT. Demonstrating the power of these customizations, Altman himself crafted a GPT that provided startup advice based on a lecture he had given. Aside from text-based customizations, these AI models can be integrated with third-party platforms. For example, a GPT was showcased that could identify calendar conflicts and communicate about them through a messaging app.

Towards More Autonomous AI Agents

While GPTs currently serve as sophisticated tools that enhance productivity, OpenAI is already paving the way for them to evolve into autonomous agents. These future AI agents could potentially interact with various applications and perform complex tasks without direct human oversight. The implications of such advancements are substantial, raising questions about and the ethical considerations of AI systems with increased agency.

New GPT-4 Turbo With Massive Context Window Increase

With the launch of GPT-4 Turbo, an improved variant of OpenAI's large language model, the organization has effectively doubled the context window capacity to 128,000 tokens—translating into an impressive comprehension span of up to 300 pages of text. This enhancement grants Turbo extended memory and advanced reasoning abilities, enabling it to facilitate deeper, more intricate conversations and execute complex instructions with better nuance. OpenAI further reveals that GPT-4 Turbo comes with a cost reduction—a 3x decrease per token for input and a 2x reduction for output compared to the previous iteration of GPT-4. This update not only promises quicker returns on AI investments for enterprises but also democratizes access to cutting-edge generative AI for startups and smaller development teams.

Addressing Safety Concerns

Altman underscored a cautious approach when it comes to the kinds of actions GPT-4 may perform on the internet. The company currently sanctions simple, direct actions while avoiding tasks that require high-level planning, reflecting a mindful stance on safety. Nevertheless, OpenAI remains forward-thinking, with Altman advocating for AI to become a tool for individual empowerment and societal advancement, equipping people with ‘superpowers on demand'.

New Tools to Foster Innovation

OpenAI also announced the debut of the Assistants API, a suite of devised to create bespoke AI agents tailored to specific applications. These custom agents could range widely in purpose, from coding assistants to vacation planners, or even voice-controlled music selectors. 

The Assistants API enables OpenAI's clientele to craft assistants tailored with precise instructions that draw upon external knowledge. Additionally, it can engage various generative AI models and tools that OpenAI offers, aiming to execute a plethora of tasks. Programmers have the flexibility to create an AI capable of a wide range of functionality, from analyzing data through natural language processing to aiding in software development or orchestrating vacation plans by harnessing AI's predictive power.

One of the core technologies propelling the Assistants API is the Code Interpreter, a system released in March for ChatGPT, which has shown proficiency in writing and executing Python code within a secure and controlled setting. This capability allows the Assistants API to perform iterative code and mathematical problem-solving, as well as visualizing results through generated graphs and charts.

Copyright Protection Efforts

OpenAI has also taken a proactive stance in safeguarding businesses utilizing its AI products against potential copyright infringement claims. As part of a commitment, the company declared that it has initiated Copyright Shield, a program under which it will cover all legal costs encountered by businesses in the wake of IP-related litigation concerning works generated by OpenAI tools. The program specifically caters to the users of the generally accessible features of OpenAI's developer platform as well as ChatGPT Enterprise, the more advanced, subscription-based tier of its AI-chatbot. However, it was noted that the offer does not extend to users of the free and Plus tiers of ChatGPT, and uncertainty remains about whether OpenAI will provide indemnity for claims made regarding the training data used for its AI models.

Additional Updates and Business Prospects

Adding to the array of enhancements, OpenAI has incorporated computer vision and text-to-speech technology into its suite. The integration includes direct API access to DALL-E 3, the startup's highly realistic image generation model. OpenAI is also taking proactive steps to protect creators with the introduction of Copyright Shield, a program designed to guard against infringement claims when utilizing the platform's general-purpose features.

Navigating the Future of AI

The introduction of these innovative models and development tools is set to carry substantial implications for various industries. By making AI more cost-effective and easier to deploy, OpenAI could disrupt the AI marketplace and alter business strategies regarding AI utilization. These sophisticated models and multimodal capabilities could lead to more intricate AI-driven applications and services, resulting in richer and more interactive user experiences.

Markus Kasanmascheff
Markus Kasanmascheff
Markus is the founder of WinBuzzer and has been playing with Windows and technology for more than 25 years. He is holding a Master´s degree in International Economics and previously worked as Lead Windows Expert for Softonic.com.

Recent News