OpenAI launched fine-tuning for GPT-4o yesterday, allowing developers to customize the structure and tone of responses or follow domain-specific instructions. Fine-tuning was one of the most requested features. It can significantly impact model performance and cost reduction. OpenAI announced it offers 1 million training tokens daily for free for every organization through September 23.GPT-4o fine-tuning training costs $25 per million tokens, and inference is $3.75 per million input tokens and $15 per million output tokens. For GPT-4o mini, OpenAI offers 2 million training tokens daily for free through September 23. One featured example is Genie, an AI software engineering assistant that can autonomously identify and resolve bugs, build features, and refactor code in collaboration with users. It is powered by a fine-tuned GPT-4o model trained on examples of real software engineers at work, enabling the model to learn to respond in a specific way. The model is also trained to output in specific formats, such as patches that could be easily committed to codebases. The San Francisco-based research lab ensured that these fine-tuned models remain entirely under the customer's control. The customer has full ownership of his business data, including all inputs and outputs. This ensures that your data is never shared or used to train other models.
In May this year, NASA named its first AI officer, David Salvagnini. This role expanded Salvagnini’s current role as chief data officer. "AI can accelerate the pace of discovery," said NASA Administrator Bill Nelson. The chief AI Officer is a new role within companies intended to keep up with everything related to AI and implement it in certain operations. In addition to NASA, other brands such as Dell, LVMH, Morgan Stanley, and the Dubai government recently hired managers for this role. Experts say that major organizations are trying to fall behind when using AI tools more effectively. "For example, this person might find a ChatGPT prompt that makes the Finance team work 10% faster, get the Marketing team using Midjourney to work 25% faster, or implement AI tools to streamline HR processes and save teams hours a day," wrote Rowan Cheung, founder of the popular newsletter "The Rundown AI." It's this new role where someone is hired to specifically keep up with everything happening in AI and implement it into certain operations within the company. For example, this person might: – Find a ChatGPT… pic.twitter.com/qoHU7MP4PB — Rowan Cheung (@rowancheung) July 2, 2024
Google is rolling out new Gemini-powered features for Chrome, such as Lens for desktop, tab compare for shopping, and AI-powered history search, as an opt-in feature for U.S. users this month. Lens allows the user to ask questions about what is seen —a video, item, or image—while browsing. It will live in the address bar and the three-dot menu. After clicking, the user selects a part of a page and asks more questions to get search results. With Lens—which is located in the address bar and the three-dot menu—users can tap on objects and ask questions through multi-search to find a similar item in different colors or brands. Depending on the question, users might also get AI Overviews in answers.
The U.S. Department of Education released this month "Designing for Education with Artificial Intelligence: An Essential Guide for Developers." This 49-page guide describes how developers and educators can work together to strengthen trust. It also provides insights into how educational organizations or development teams can strengthen their approach to responsible AI in education. The guide builds on the Department’s prior report, Artificial Intelligence and the Future of Teaching and Learning: Insights and Recommendations (2023 AI Report). • PDF Download • Webinar: The Power of a Clean Slate – Removing Barriers to Career-Connected Education
Groq Inc. — a startup based in Mountain View, California, that designs semiconductors and software for AI systems — raised $640 million in new funding at a valuation of $2.8 billion. The founding round was led by BlackRock Inc. funds and included backing from the investment arms of Cisco Systems and Samsung Electronics. This Series D round almost triples its valuation from $1 billion in a funding round in 2021. The company has entered into competition against incumbents such as the leader Nvidia, as well as Intel Corp., Advanced Micro Devices Inc. Chief Executive Officer Jonathan Ross said in an interview on Bloomberg Television that Groq plans to use the funding to build about 108,000 language processing units, hire significantly, and consider some acquisitions. The company said that former Intel executive Stuart Pann is joining Groq as its chief operating officer. Additionally, Yann LeCun, Meta’s chief AI scientist, will become an adviser to the startup. Ross said LeCun’s push for open-source models, such as Meta's Llama, allowed Groq to grow. "Groq would not exist today if not for open source," Ross said. We built the best chips, but if we didn’t also have the software, we would not be able to demonstrate that." Open-source models allow people to get into the AI business without having to train models. What are we doing with this capital? Originally we intended to raise $300M which was going to allow us to deploy 108,000 LPUs into production by end of Q1 2025. We raised 2x that, so we're also expanding our cloud and core engineering teams. We're hiring! https://t.co/rnk6PGzfXy — Jonathan Ross (@JonathanRoss321) August 5, 2024 Check out this new app powered by Groq! 🤩 https://t.co/7YXL3OvUH9 — Groq Inc (@GroqInc) August 5, 2024 Very, very fast voice bots. Llama 3.1 running on @GroqInc. 🚀 500ms voice-to-voice response times pic.twitter.com/l9u5GNPhCo — kwindla (@kwindla) July 23, 2024
