Good morning, everyone! It’s Friday, September thirteenth, two thousand twenty-four, and you're tuned into AI Tech News Today. We've got a lineup of exciting news in the world of Artificial Intelligence that you won't want to miss. 

OpenAI's new 'o1' model thinks longer to give smarter answers. OpenAI has released GPT-o1, a new model designed to enhance reasoning capabilities by extending its thought process before responding. Unlike its predecessors, GPT-o1 focuses on prolonged inference. While it doesn't always beat the previous GPT-4o model across the board, it excels in tasks that require deep reasoning. Alongside o1, two companion versions—o1-preview and o1-mini—are available for ChatGPT Plus and Team users. These are aimed at refining specific use cases and offer strong performance in science and technology challenges.

OpenAI reportedly seeking six point five billion dollars investment at one hundred fifty billion dollar valuation. OpenAI is in discussions to secure six point five billion dollars in new funding, which would set its valuation at an eye-popping one hundred fifty billion dollars. Thrive Capital is expected to lead the investment round with a billion dollars, while current investor Microsoft may also chip in. This funding would facilitate the launch of OpenAI’s new o1 model and help build its AI infrastructure. 

Suno releases new "covers" feature to reimagine music you love. The music platform Suno has unveiled its latest feature called "Covers," allowing users to transform their songs into different styles while keeping the original melody intact. This innovative tool supports various audio inputs, making it simple for users to create unique covers of their favorite tracks. The feature is currently in beta with an allocation of one hundred free covers available for subscribers.

Adobe announces Firefly Video model AI video tool. Adobe is stepping into the AI video editing realm with the upcoming Firefly Video model. This tool, set for beta release later this year, allows users to create short video clips from text or image prompts. With custom camera angles, pans, and zoom effects, Adobe claims it provides superior results compared to other competitors in the market.

Three new AI tools worth checking out include Thunderbit, an automation copilot for web tasks; Objective, a search API designed for various applications with advanced language processing; and Lifesight, a unified marketing measurement platform promising smarter insights driven by AI. 

That wraps up today’s episode! Stay curious and keep exploring the evolving world of Artificial Intelligence. For more updates, keep tuning in to AI Tech News Today.

OpenAI Launches New GPT-o1 Model as It Seeks $6.5 Billion Funding; Suno and Adobe Unveil Innovative AI Features

Good morning, everyone! It’s Friday, October Eighteenth, Two Thousand Twenty-Four. Welcome to another exciting edition of AI Tech News Today, where we bring you the latest and greatest in artificial intelligence developments. Let’s dive right into today’s highlights.

OpenAI reveals GPT-O1 model.
Noam Brown from OpenAI has unveiled a new video titled "Learning to Reason with Large Language Models." This presentation introduces the O1 model, which uses reinforcement learning to enhance reasoning in large language models, consistently outperforming previous models in various benchmarks. With increased computational power, this model shows promising implications for advancing reasoning in fields such as mathematics and programming.

Meta and OpenAI face criticism on open-source claims.
Meta has been criticized by the Open Source Initiative for labeling its Llama AI models as “open-source.” According to OSI chief Stefano Maffulli, Meta's approach dilutes the true meaning of open-source, as vital components remain proprietary. In a related note, OpenAI has pledged to avoid using patents offensively, but experts deem this more of a public relations stunt than a genuine commitment to collaboration.

Open Sora releases version 1.3 of video generation model.
Open Sora Plan recently launched version 1.3.0 of its video generation model, which comes with five major features, including a more efficient Wavelet Variational Autoencoder and a Prompt Refiner that enhances short text inputs. Unlike other models, Open Sora will be open-sourced, making it available to the community for further advancements in video generation.

Nvidia drops new AI model surpassing GPT-4.
In the race of AI models, Nvidia has quietly launched the Llama-3.1-Nemotron-70B-Instruct AI model, which boasts benchmark scores that exceed those of industry leaders like OpenAI and Anthropic. This model offers improved alignment capabilities, delivering precise and contextually relevant responses, making it a strong candidate for businesses seeking effective language processing solutions.

AI software corrects eye contact in videos for ten cents per minute.
A startup named Sieve has released a new API that can automatically correct eye contact in videos for just ten cents per minute. The software analyzes the eye region and adjusts gaze direction in real-time, enhancing viewer engagement by creating an illusion of direct eye contact. This innovation holds potential for various applications, from video production to educational content.

That’s all for today! Stay tuned and keep an eye on AI innovations as they continue to reshape our world. Until next time, keep learning and exploring!

New tracks tagged #GPT-o1