GPT-4, Multimodal?

Can't read or see images? View this email in a browser

OpenAI seems to be in a bit of a rush with GPT-5 as it recently filed for a trademark. But, what about GPT-4 and its multimodal promises, where users were expecting to play with images, create websites, etc? When OpenAI introduced GPT-4 a few months ago, it claimed that the advanced capabilities would soon be available to ChatGPT and API. They said they were working with ‘Be My Eyes’ to make image input capabilities widely available. So far, the collaboration is in closed beta, being tested for feedback among a small subset of the users with little to no updates. Also, the multimodal features of GPT-4 are not accessible in the APIs.  OpenAI, in its blog post, states that users can currently make text-only requests to the GPT-4 model, and the capability to input images is still in a limited alpha stage. However, OpenAI assures users that they will automatically update to the recommended stable model as new versions are released over time. In other words, more advanced features and capabilities would be available to users as the model continues to evolve and improve.  Recently, OpenAI introduced Code Interpreter in ChatGPT Plus. Many termed it as a GPT-4.5 moment but sadly it was just old-school OCR from Python libraries and didn’t use multimodal for image generation. Read the full story here.  All Eyes on NVIDIA  The demand for high-performance GPUs, particularly the NVIDIA H100s, has skyrocketed. “‘Who’s getting how many H100s and when’ is the top gossip of the Valley right now,” said OpenAI’s Andrej Karpathy. Also, AWS’ Stephen Balaban, said “Lambda has a few thousand more H100s coming online before the end of this year — if you need 64 H100s or more, DM me.”  Many AI leaders, including Quora’s Adam D’Angelo and OpenAI’s Sam Altman, have voiced their concerns about the GPU shortage. OpenAI has revealed that the limited GPU supply is hindering their short-term plans, including model fine-tuning and dedicated capacity. This explains why it is still stuck with GPT-4 and not able to fulfil its multimodal ambitions.  Read more here.  Freshworks’ Generative AI Plans  Zoho rival Freshworks seems to be going all-in on generative AI. In an exclusive interview with AIM, Prakash Ramamurthy, chief product officer of Freshworks, revealed a slew of new generative AI capabilities offered by the company, alongside touching upon strategic collaborations, challenges and more.  Ramamurthy believes that the success of Freshworks’ AI-driven solutions comes from its strategic partnerships with hyperscalers such as Microsoft Azure, Google Cloud, and AWS. Notably, Freshworks has integrated Microsoft Azure OpenAI Service into their Freddy AI solutions, resulting in a significant boost to their capabilities.  Check out the complete interview here.  Generative AI APIs  Working with open-source generative AI models – the likes of Stable Diffusion, Llama 2, Falcon, GPT-J, etc. – just got a lot faster. Thanks to generative AI APIs such as Monster API, Gooey.AI, Illusion AI, Fantasy.ai, One AI, and Outset.ai, among others, developers can now test generative AI models on the go.  However, the challenges galore. Most of these companies that are developing generative AI APIs are majorly specialising in API calling and are not solving enterprise problems like custom data training, fine-tuning of generative AI models, security, data privacy, etc. It would be interesting to see how the space evolves in the coming months.  TAUSIF ALAM & AMIT RAJA NAIK Thursday, Aug 3, 2023 | Was this email forwarded to you? Sign up here DOWNLOAD OUR MOBILE APP Stay Connected info@analyticsindiamag.com © 2023 Analytics India Magazine

This email was sent by support@analyticsindiamag.com to alexvarboffin.abbb@blogger.com Not interested? Unsubscribe | Manage Preference | Update profile Analytics India Magazine | 280, 2nd floor, 5th Main, 15 A cross, Sector 6, HSR layout Bengaluru, Karnataka 560102