Carl Franzen

Meta Introduces Spirit LM open source model that combines text and speech inputs/outputs

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Just in time for Halloween 2024, Meta has unveiled Meta Spirit LM, the company’s first open-source multimodal language model capable of seamlessly integrating text and speech inputs and outputs. As such, it competes directly with OpenAI’s GPT-4o (also natively multimodal) and other multimodal models such as Hume’s EVI 2, as well as dedicated text-to-speech and speech-to-text

Read More »

Why AI is going nuclear

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More If you’re of a certain age, the words “nuclear energy” probably conjure up dystopian images of power plants melting down, glowing radioactive waste, protesters, and other dark scenes ranging from the unfortunate to apocalyptic. The truth is, nuclear power’s reputation has been mostly unfairly blemished since 1970s and ’80s thanks to the Three Mile Island and

Read More »
AR/VR

The one question you need to ask ChatGPT right now

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Do you use ChatGPT regularly? Do you have the “memory” feature turned on — which allows the chatbot from OpenAI to recall important information about you and your preferences? If so, navigate over to it when you have a free moment and enter in the following question: “From all of our interactions what is one thing

Read More »

Reflection 70B saga continues as training data provider releases post-mortem report

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More On September 5th, 2024, Matt Shumer, co-founder and CEO of the startup Hyperwrite AI (also known as OthersideAI) took to the social network X to post the bombshell news that he had fine-tuned a version of Meta’s open source Llama 3.1-70B into an even more performant large language model (LLM) known as Reflection 70B — so

Read More »

Pika 1.5 launches with physics-defying AI special effects

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Pika, also known as Pika Labs, was one of the first startups to emerge with its own AI video generation model, allowing users to simply type in text and get video clips in return, raising $35 million in a Series A in November 2023. But it’s been nearly a year since Pika launched its 1.0 text-to-video

Read More »