Matthias Bastian Online journalist Matthias is the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers. Amazon has unveiled a revamped version of its Alexa AI voice assistant, featuring a large language model (LLM) for more natural and longer conversations with users. The new …
Summary Midjourney version 6 is expected to bring a big leap in quality. It will be released later this year. According to Midjourney CEO David Holz, the leap from Midjourney’s current version 5 to version 6 will be greater than the leap from version 4 to version 5. Holz did not want to comment on …
Midjourney’s upcoming version 6 promises a big leap in quality Read More »
Summary OpenAI’s new DALL-E 3 text-to-image system offers significant improvements over its predecessor and is natively integrated with ChatGPT. According to OpenAI, this integration should eliminate the need for “prompt engineering”. According to OpenAI, DALL-E 3 shows significant advances in its ability to generate images that accurately follow complicated prompts. This includes accurately representing a …
OpenAI announces DALL-E 3 with native ChatGPT integration Read More »
Summary Generative AI is supposed to change our work in many areas, but web statistics show that the sector is driven and dominated by ChatGPT. A16z analyzed traffic to various GenAI-related websites using Similarweb. Among the top 50 “GenAI” web applications, ChatGPT is by far the most popular, accounting for 60 percent of the total …
Web stats show generative AI boom is mostly ChatGPT hype Read More »
Summary Updated September 20, 2023: OpenAI is rolling out a new fine-tuning interface that documents the status of fine-tuning. In the coming months, it will be possible to fine-tune via the UI, which should make it easily accessible to more users. In addition, three GPT 3.5 models can now be fine-tuned with your own data; …
You can now fine-tune three GPT-3.5 models, GPT-4 to follow this fall Read More »
Summary OpenAI has announced an open call for experts across a range of domains to join the OpenAI Red Teaming Network, aimed at improving the safety of their AI models. Red teaming is part of OpenAI’s iterative development process, which includes internal and external adversarial testing. With this formal launch, OpenAI aims to deepen and …
OpenAI launches red teaming network to root out AI risks Read More »
Summary In the race with ChatGPT, Google has added new features to Bard. In particular, the integration into the Google ecosystem may give it a leg up. According to a company blog post, Google has updated its PaLM 2 language model to make Bard more versatile and useful. A few weeks ago, Google added 40 …
Google Bard now understands images, checks facts, and uses your Google data Read More »
Summary Google is about to release Gemini, a multimodal language model that could challenge OpenAI’s GPT-4. But OpenAI is reportedly preparing a fitting response. According to The Information, OpenAI is preparing to roll out GPT-4’s image understanding feature. This multimodal capability was introduced at GPT-4’s unveiling and was a major aspect of the presentation. A …
OpenAI could answer Google’s Gemini with GPT-Vision and “Gobi” Read More »
Summary OpenAI introduces “gpt-3.5-turbo-instruct”, a new instruction language model that is as efficient as the chat-optimized GPT-3.5 Turbo. OpenAI is introducing “gpt-3.5-turbo-instruct” as a replacement for the existing Instruct models, as well as text-ada-001, text-babbage-001, text-curie-001, and the three text-davinci models that will be retired on January 4, 2024. The cost and performance of “gpt-3.5-turbo-instruct” …
OpenAI releases new language model InstructGPT-3.5 Read More »
Summary AI startup Delphi wants to bring one-on-one interactions with celebrities and influencers to the masses. You can’t be everywhere at once – can you? That’s how you could describe Delphi’s business model. The startup aims to model personal chatbots, i.e. make personalities chattable, from multiple data sources such as chats, emails, YouTube videos, audio …
AI startup Delphi wants to create digital clones of influencers Read More »
