AI Top-of-Mind for Feb 9
There have been many rumors on the status of GPT-5, expected to include 10x the parameters of GPT-4. For perspective:
The latest predictions are by Rohan Balkondekar, and his guess is that it will be released this coming November.
Also on the model front, Azhar looks at the Eagle 7B model and its multilingual capabilities. Its power is derived from a combination of Recurrent Neural Networks (RNNs) and transformer technologies. Stats from the article:
Turning to the global scene, I’m sure you’ve read about TSMC, the global semiconductor supply chain, and how the US spending hundreds of billions to re-home production. Manual Brenner looks at this fragility.
A few days back I spoke of Bing’s image generator and its inclusion of watermarks. Next on the block is OpenAI with both ChatGPT and DALL-E3. The movement is based on the requirement to adhere to the new ‘Coalition for Content Provenance and Authenticity (C2PA)’ standards and images can be verified at the site ‘Content Credentials Verify.’ More at ‘The Verge’ article. And speaking of creative, a viewpoint by Jeff Hayward in ‘Counter Arts’ where he looks at the downside of tools such as Nightshade and Glaze, which are designed to prevent faithful copying of AI-generated images.
On the retail front, Microsoft’s Super Bowl Co-Pilot ad. More from the ‘SJ Mercury’ on Super Bowl ads and AI.
Lastly, on the corporate front, good ways to leverage Gen AI for teambuilding. The ‘SmartBrief’ articleidentifies ten top skills as identified by the World Economic Forum, and then describes how GPT can be used to reenforce them. The list, with ‘soft skills’ marked with an *:
· Analytical thinking.
· Creative thinking.*
· Resilience, flexibility, and agility.*
· Motivation and self-awareness.*
· Curiosity and lifelong learning.*
· Technological literacy.
· Dependability and attention to detail.*
· Empathy and active listening.*
· Leadership and social influence.*
· Quality control.
