5G is still being rolled out across geographies. It was supposed to herald the revolution in healthcare. Even in those areas where 5G has established its presence and the market has “matured”, I haven’t seen or heard about any “killer application”. Unless, of course, you have a blind faith in the McKinsey’s of the world.
ChatGPT has joined the ranks since it got hyped up. Massively.
This paragraph addresses the concerns in one go:
ChatGPT is a bullshit generator. But it can still be amazingly useful
The philosopher Harry Frankfurt defined bullshit as speech that is intended to persuade without regard for the truth. By this measure, OpenAI’s new chatbot ChatGPT is the greatest bullshitter ever. Large Language Models (LLMs) are trained to produce plausible text, not true statements. ChatGPT is shockingly good at sounding convincing on any conceivable topic. But OpenAI is clear that there is no source of truth during training. That means that using ChatGPT in its current form would be a bad idea for applications like education or answering health questions. Even though the bot often gives excellent answers, sometimes it fails badly. And it’s always convincing, so it’s hard to tell the difference.
The choice of words is confusing, but nuanced. If it gives excellent answers, why is it failing? Possibly you need to ask the “right questions”. It took me a little time (and then generated output) from specific queries. I used it once to generate a response to insurance denial, and it gave me a factually correct answer with attached references. I got the insurance clearance immediately, thereafter.
Will I use it? Possibly for low-end grunting work. Higher order of thinking requires me to write on my blog on my own steam. In routine work, I hate distractions, but need to get work done quickly. I will create templates (with PDF’s) to be shared immediately with the team, without impeding their workflows.
ChatGPT (and 5G) are here. They are solutions to a problem. I am sure everyone’s convinced it will require massive infusions of capital (in these times of recession) to drum up “interest”. Virtual reality/mixed reality are again additional investments which no one requires; however, I won’t be surprised to see “excitement” building up.
The authors ended with this excellent conclusion:
Generative AI releases tend to look impressive based on cherry-picked examples that go viral. But that’s not the full picture. For many applications, even a 10% failure rate is too high. There seem to be a fairly limited set of use cases where the lack of a source of truth isn’t a deal breaker. While these uses are still tremendously exciting, it’s far from obvious that people will soon be using chatbots in their everyday lives — for learning, or as a search replacement, or as a conversation partner.
RadOncNotes 