IBL News | New York
Nine weeks after ChatGPT was launched, the company behind it, OpenAI, released this week a brief technical note stating, “we’ve upgraded the ChatGPT model with improved factuality and mathematical capabilities.”
On November 30, 2022, San Francisco-based research lab, now heavily supported by Microsoft, OpenAI described what it released as an “early demo” of a part of the GPT 3.5 series — an interactive, human-like, conversational model. Its dialogue format “makes it possible for ChatGPT to answer follow-up questions, admit its own mistakes, challenge incorrect information, and reject inappropriate requests.”
ChatGPT took the world by storm and made AI the next big thing, according to experts. Since then, tech giants like Google are in “red code,” and every day seems to bring new contenders.
At least four top players are working on “generative” A.I., technologies making moves to challenge ChatGPT:
Google: LaMDA. Launched in 2021, Google said in a launch blog post that LaMDA’s conversational skills “have been years in the making.”
Like ChatGPT, LaMDA is trained in dialogue. It’s built on Transformer, the neural network architecture that Google Research invented and open-sourced in 2017.
The Transformer architecture “produces a model that can be trained to read many words (a sentence or paragraph, for example), pay attention to how those words relate to one another, and then predict what words it thinks will come next.”
Anthropic: Claude. Founded in 2021 by a group of people that included several researchers who left OpenAI, this San Francisco AI start-up has raised $300 million in new funding, in exchange of taking a stake of 10%.
The deal values Anthropic at $3 billion.
According to a report posted at The Financial Times today, Google has already made that investment.
The British paper said that Google confirmed it had made that investment and that it had a large cloud contract with Anthropic to use the Google Cloud infrastructure, but did not provide further details. This deal would echo OpenAI’s agreement with Microsoft’s Azure.
Anthropic developed an AI chatbot, Claude — available in closed beta through a Slack integration — that reports say is similar to ChatGPT and has even demonstrated improvements.
Character AI. This news AI chatbot technology allows users to chat and role-play with anyone, living or dead — as it can impersonate historical figures like Queen Elizabeth and William Shakespeare.
It was launched by two engineers that left Google in October 2022, Noam Shazeer, one of the authors of the original Transformer paper, and Daniel De Freitas.
Now they are trying to raise $250 million in new funding, a striking price for a startup with a product still in beta.
DeepMind: Sparrow. DeepMind, the British-owned subsidiary of Google parent company Alphabet, introduced Sparrow in a paper in September.
For now, Sparrow is a research-based, proof-of-concept model that is not ready to be deployed, according to Geoffrey Irving, a safety researcher at DeepMind and lead author of the paper introducing Sparrow.
DeepMind’s CEO and cofounder Demis Hassabis said in a TIME article two weeks ago that its company is considering releasing its chatbot Sparrow in a “private beta” sometime in 2023.
DeepMind says “Sparrow is a dialogue agent that’s useful and reduces the risk of unsafe and inappropriate answers.” The agent is designed to “talk with a user, answer questions and search the internet using Google when it’s helpful to look up evidence to inform its responses.”
It was hailed as an important step toward creating safer, less-biased machine learning (ML) systems, thanks to its application of reinforcement learning based on input from human research participants for training.
Axios: How ChatGPT became the next big thing
:format(webp)/cdn.vox-cdn.com/uploads/chorus_asset/file/24422832/lTEtt4F.jpeg)
