Call us now:
A great article on AI.
It’s interesting the amount of communication one hears about AI and how it will transform everything we do. There is no doubt that AI has the opportunity to make real difference in our lives. But as this article points out, we are dealing with Generative AI based on Large Language Models, we’re not dealing with Machine learning. One was involved in AI in machine Learning 20 years ago, working at Rolls-Royce and Warwick University on machine learning for the manufacture of Turbine blades. Yep an academic and an grease monkey engineer.
What we are seeing today is not the true power of AI. We are seeing systems that take data from what has already been created and reproducing it. It ain’t going to solve complex problems like turbine blade manufacturing just yet.
‘Text generation produced by large language models depends for the moment on the data used to train those models. LLMs produce better results when they reflect recurring concepts. They struggle with new scenarios and tasks outside that envelope.’ AKA they give an answer based on what everyone else has said, even if that is not correct.
‘LLMs meanwhile struggle to identify their own mistakes. Requesting a correction does not produce a more accurate answer. In a study of LLMs, Originality. AI found that every single one produced errors. OpenAI’s ChatGPT-4 offered inaccuracies in nearly a third of responses’ – As KPMG and an Attorney found out recently.
As AI is based on LLMs it is not ready to go do something complex. Where it has to learn from a real world, in real time situations, and make a specific repeatable decisions, that’s machine learning, and we ain’t there yet. That’s not to say it is not useful, it will find its place.
Why does one feel that we are heading towards a crypto, NFT investment bubble.
Ted Dziekanowski Brian D. McCarthy Rachel V. Rose, JD, MBA Bob Dix
#paravasecuritysolutions #ai #largelanguagemodels
View Original Post
https://www.linkedin.com/feed/update/urn%3Ali%3Ashare%3A7131195250768769024