OpenAI now allows fine-tuning to GPT-4o with 1Mn Free tokens till Sep 23

Here is the top trending news from the world of technology. News that every tech enthusiast should keep a tab on.  

1)

OpenAI now allows fine-tuning to GPT-4o with 1Mn Free tokens till Sep 23

OpenAI on Tuesday made a significant announcement by allowing fine-tuning to GPT-4o, its latest flagship LLM. For designers and companies alike, this will pave for more proficiency and accuracy. To bring more customers on-board, OpenAI will be providing 1 million free training tokens each day to companies. But this offer will be only for limited time, until September 23. The fine tuning will enable developers to customize GPT-4o as per their unique requirements, paving way for better performance and efficiency.  Even though this will incur cost once the free trail period is completed, it will be still less expensive than training a AI model from ground zero.

2)

Microsoft launches Three new AI models under PHi family

It seems Microsoft does not want to leave any stone unturned in its quest to remain ahead in the cutthroat AI race. On Tuesday, the tech giant announced three new Phi 3.5 models. These three new models are Phi-3.5-mini (suitable for chat formats), Phi-3.5-vision (a multimodal model that is apt for tasks demanding visual reasoning), and Phi-3.5-MoE-instruct (a large model with multilingual capacity). All three models are available for download on Hugging Face under a Microsoft branded MIT license. What’s more important to note that these three new Phi models have managed to outperform many of its competitors on several key benchmarks, including Google’s Gemini 1.5 Flash and Meta’s Llama 3.1.

3)

Luma AI rolls out Dream Machine 1.5 for more Realistic & High-Quality Videos

Luma AI, a prominent AI video generation platform, has rolled out a new AI model. Dream Machine 1.5 promises to offer more realistic videos, enhanced motion tracking and more in-depth as well as intuitive understanding of prompts. However, the feature that is worth note taking is its ability to render text within generated videos, a feature that is missing in other AI models and can prove to be a game changer. It can be especially useful for making dynamic title sequences, animated logos and on-screen graphics for presentations. Additionally, Dream Machine 1.5 has also shown proficiency in understating non-English prompts.

4)

Microsoft finally launches a single unified Teams app  

After years of testing, Microsoft has finally decided to roll out a single unified Teams app, which means now you won’t have to face the inconvenience of switching between personal and work accounts. Furthermore, you won’t even have to switch between Team tenants. Apart from Windows, this single unified Teams app will also be available for MacOS. With this integration, the painful and often confusing double app situation for work and personal accounts will become a thing of past. If you’ve automatically downloaded Teams on your Windows and Mac system then you won’t have to download it again. It will be updated automatically or else you can download it from Microsoft’s site.

5)

OpenAI inks Content Deal with Conde Nast’

In its relentless quest to procure more data for training its AI models, OpenAI has struck a multi-year content deal with Conde Nast’. Conde Nast’ is a media company behind popular news publications like Vogue and Wired. As part of this deal, OpenAI will get access to Conde Nast’ large content repository and in exchange the latter gets access to OpenAI’s advanced many tools and features. Such deals not only help OpenAI in procuring troves of precious data but also spares it from getting sued for copyright infringement. The ChatGPT’s parent company has already signed a similar content deal with Times and Vox media.