Most Important AI Developments in March 2023
In the world of AI, a lot has been happening lately. The pace of change in AI technology is so significant that it's challenging to keep up with everything. That's why we've decided to publish a monthly summary of AI events.
This is the first post in the series, summarizing AI developments in March. Below is a list of the most important AI events from March:
1. Figure is Building Autonomous AI Robot
Figure, a robotics startup, plans to create Figure 01, the first commercially viable general-purpose humanoid robot to address labor shortages and hazardous jobs. With an experienced team from leading institutions, the robot will be designed for real-world applications in human workspaces, focusing on practicality over superhuman capabilities.
2. AI Piloted Fighter Jet
The U.S. Defense Department announced it has completed a dozen test flights with a real fighter jet over the California desert piloted by AI
3. Meta LLAMA available on Torrent
Meta's sophisticated large language models, known as LLaMA, intended for authorized researchers only, were made publicly available through a link posted on 4chan. This has raised concerns about potential misuse, such as generating spam, marketing material, or disinformation, as the model is now the most powerful publicly accessible large language model.
4. ChatGPT 3.5 turbo API Released, and It is 10x cheaper
The ChatGPT 3.5 Turbo API has been released, offering users enhanced AI capabilities. This updated version is now 10 times more affordable, making it more accessible to a wider range of users.
5. Microsoft just introduced Kosmos-1
Kosmos-1 is a huge AI model that can: -pass visual IQ tests -analyze images for content
Microsoft researchers introduced Kosmos-1, a multimodal AI model capable of analyzing images, solving visual puzzles, and understanding natural language instructions. This advancement in multimodal AI, which combines different input modes, is seen as a crucial step toward achieving artificial general intelligence (AGI).
6. ChatGPT 4 released
OpenAI has released GPT-4, a large multimodal model demonstrating human-level performance on various professional and academic benchmarks. Despite its impressive capabilities, GPT-4 has limitations like hallucinating facts, making reasoning errors, and being unaware of events after September 2021.
7. MidJurney 5 released
Midjourney 5 has been released with incremental improvements, offering more fine-grained control, better photorealism, and faster upscaling for experienced users. The new version renders faces and hands more accurately and supports a wider range of aspect ratios, making it a valuable tool for generating photorealistic images, such as book covers.
8. Microsoft added chatGPT to Microsoft 365
Microsoft has introduced Copilot, integrating ChatGPT into its Microsoft 365 apps, unlocking the potential of large language models and making AI more accessible. Copilot enhances apps like PowerPoint, Excel, Outlook, Word and Skype by generating content based on prompts, providing users with contextual results, suggested interactions, and easier ways to tweak and refine output.
9. ChatGPT 4 API released
For now, it is only available through a request. You need to sign up for the waitlist and wait for access to be granted.
10. Google released Bard
Google is launching Bard, an AI chatbot tool competing with ChatGPT, allowing users to join a waitlist for access. Bard aims to help users with tasks such as essay drafting, event planning, and lunch ideas, and will initially roll out in the United States and United Kingdom.
And 4 Bonus once:
1. Adobe released Firefly AI
Adobe has released Firefly, a family of creative AI models, featuring tools similar to DALL-E and Midjourney for image creation and AI-powered stylized text. The company plans to integrate generative tools into its apps, such as Photoshop, Illustrator, and Premiere, while addressing copyright and bias concerns and compensating artists contributing to training data.
2. NVIDIA launches cloud tools for Generative AI
The age of enterprise AI has witnessed rapid growth, with major tech companies incorporating generative AI features into their products. NVIDIA's AI Foundations allows businesses to develop custom large language and generative AI models for domain-specific tasks, providing unique models like NeMo, BioNemo, and Picasso to supercharge various industries.
3. GitHub unveils CopilotX
GitHub Copilot X integrates AI into software development workflows, offering context-aware conversations, personalized documentation, and assistance with pull requests and automated testing. With support from OpenAI's GPT-4, it helps developers code faster, focus on more satisfying work, and enhance productivity.
4. ChatGPT plugins released
ChatGPT plugins enhance the AI chatbot's capabilities by accessing the internet and integrating with third-party services like Expedia, Klarna, Slack, and more. This update allows ChatGPT to provide up-to-date information, run computations, and interact with various platforms for improved user experience.