Paris Olympics broadcasters diverge on AI approach
Paris Olympics broadcasters diverge on AI approach
US markets suffer worst day(7/24) since 2022 as Tesla and AI stocks fall
Anthropic named ‘Best Performing’ AI as LLM arms race intensifies
An entirely new way to present ideas
Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.
***
Paris Olympics broadcasters diverge on AI approach
Paris Olympics will serve as a prominent platform for showcasing generative AI to American audiences. While European viewers will not experience a similar technological approach. The Olympic Broadcasting Services, responsible for producing neutral coverage for media outlets worldwide, is leveraging AI to efficiently edit extensive footage into concise highlight reels.
In the United States, NBCUniversal is fully embracing AI for its broadcast of the Games, including the innovative re-creation of the voice of legendary sportscaster Al Michaels. In contrast, Warner Bros. Discovery's sports division in Europe has expressed reservations, deeming the technology too immature for critical roles such as sports commentary.
"In every part of their demos, it feels like yes, you've translated the words correctly, but you haven't translated or narrated the feeling," said a representative from Warner Bros. Discovery. For example, during the Tokyo Olympics, Italian sprinter Marcell Jacobs' gold medal victory in the men's 100 meters elicited ecstatic reactions from Italian commentators, showcasing the authentic connection of "experts sitting side-by-side, really living out that story." Such genuine moments are difficult to replicate through automated systems now.
In contrast, U.S. audiences will benefit from AI enhancements while watching the Games on NBC and the streaming service Peacock, thanks to a new collaboration between NBCUniversal, Google, and Team USA. AI-enhanced Google Map images will provide viewers with a virtual tour of the Olympic venues in Paris, while NBC hosts will demonstrate how Google AI search can address questions about the competitions.
Additionally, NBCUniversal plans to utilize generative AI to produce personalized daily briefings of Olympic events, narrated by an AI-generated version of Al Michaels' voice. The network estimates that nearly seven million unique variations of these daily recaps could be created throughout the Paris Olympics….
Originally by Sheila Dang in Paris of Reuters
***
US markets suffer worst day(7/24) since 2022 as Tesla and AI stocks fall
“We’ve got some of the Magnificent Seven digging their heels in, in terms of spending on AI,” said Kevin Gordon, a senior investment strategist at Charles Schwab, “and if they’re not wildly exceeding [earnings] expectations, that’s when you get some profit-taking.”
The US stock market experienced a significant decline last Wednesday, marking its worst day in over 18 months. This drop was largely due to underwhelming results from key index contributors Tesla and Alphabet(Google), which further fueled a technology sector sell-off that had already begun.
Notably, the losses were led by big tech stocks such as Nvidia, Microsoft, Apple, and Tesla. These technology and AI stocks have been primarily responsible for the majority of market gains seen this year.
Nvidia, a leading chipmaker specializing in AI chips, was the biggest drag on the S&P 500, with its stock price dropping 6.8% on Wednesday. Super Micro Computer and ASML Holdings, both semiconductor stocks, were the worst performers in the Nasdaq.….
Originally by George Steer and Jennifer Hughes of the FT
***
Anthropic named ‘Best Performing’ AI as LLM arms race intensifies
Which AI Performed the Best: Current benchmarks are often based on academic use-cases, rather than real-world applications. Our new Index seeks to address this by testing AI models in real-world use cases that require the LLMs to retrieve data, a common practice in business AI use cases.
Best Overall Performing Model: Anthropic's Claude 3.5 Sonnet. The closed-source model outpaced competing models across short, medium, and long context scenarios. Anthropic's Claude 3.5 Sonnet and Claude 3 Opus consistently scored close to perfect scores across categories, beating out last year's winners, GPT-4o and GPT-3.5, especially in shorter context scenarios.
Best Performing Model on Cost: Google's Gemini 1.5 Flash. The Google model ranked the best performing for the cost due to its great performance on all tasks.
Best Open Source Model: Alibaba's Qwen2-72B-Instruct. The open source model performed best with top scores in the short and medium context.
Key Findings and Trends:
Open-Source Closing the Gap: Closed-source models like Claude-3.5 Sonnet and Gemini 1.5 Flash remain the top performers thanks to proprietary training data, but open-source models, such as Qwen1.5-32B-Chat and Llama-3-70b-chat, are rapidly closing the gap with improvements in performance and lower-cost barriers than their closed-source counterparts.
Large Models Are Not Always Better: In certain cases, smaller models outperform larger models. For example, Gemini-1.5-flash-001 outperformed larger models, which suggests that efficiency in model design can sometimes outweigh scale….