- Powered By AI Newsletter
- Posts
- Testing ChatGPT Plugins to their limits... and beyond?
Testing ChatGPT Plugins to their limits... and beyond?
This week saw more AI-related job cuts, US Senators grilling (kinda) OpenAI’s CEO, and I tested ChatGPT plugins and found them to be… hmm
ChatGPT Plugins
ChatGPT plugins promised a lot: ChatGPT’s natural language and reasoning, with added functionality, real world data, and new capabilities. Plugins would see us:
Booking hotels and flights, and even arranging our whole holidays
Getting product recommendations, finding the best prices, and making purchases
Pulling all sorts of data from plugins: census, weather, stock market etc
So what’s the reality? You can watch the video above to see my full tests, but so far I’ve been fairly underwhelmed. It’s slow (as slow as 1-2 minutes per answer), and I think starkly demonstrates the limitations of the text interface.
Agree? Disagree? Let me know in the YouTube comments!
AI job cuts
The latest company to predict a shrinking of their workforce due to AI was BT. These will be customer service and diagnosis roles, so following the pattern we’ve seen with IBM and others who are already forecasting reduced need for back office, customer service and some technical roles.
It’s interesting to see companies so quickly chopping out huge chunks of their workforce, despite some of the generative AI tools they’ll be replaced with being relatively new. And not just scrappy hi-risk startups, but large ‘slow’ corporations…
US Senate vs Sam Altman. No contest
Open AI’s CEO appeared in front of the US senate, along with IBM’s chief privacy officer and an NYU professor to try and help determine what should be done about regulation.
There were some entertaining and frankly bizarre moments, but basically everyone agrees about the need for regulation and the Senators clearly had no idea who should be in charge of this (one questioned whether Sam should do it, which a Twitter commentator likened to allowing the foxes to choose who should protect the chickens).
They touched on job displacement, the dangers of AI, copyright and IP and national security - the standard stuff.
The scariest part of the whole hearing was how soft the questions were.
There’s a BBC page with video highlights of key segments here. Or you could watch the whole 2 hour 50 minute hearing here.
If you have any feedback, just hit reply and let us know what you think of the newsletter!