BBC Radio 4 – Word of Mouth, Chatbots 1
Like lots of other folk I’ve been reading plenty about Large Language Models, AI & Chatbots and playing with some of the toys.
I really liked Professor Bender’s approach and method. I also found this a very easy listen. My mind has tended to wander off when reading blogs post about AI. Very clear on the “not intelligent” and the risks associated with chatbots trained on large piles of language.
And specifically the things that they’re predicting is what would be a plausible next word given all the preceding words here and then again and then again and again.
…
And so that’s linguistically interesting that once you get to billions of words of text, there’s enough information in there just in the distribution of words to stick with things that are both grammatical and seemingly coherent.
…
So that’s a cool observation and it’s dangerous because we tend to react to grammatical, fluent, coherent, seeming text as authoritative and reliable and valuable.
…
So instead of talking about automatic speech recognition, I prefer to talk about automatic transcription because that describes what we’re using it for and doesn’t attribute any cognition to the system that is doing the task for us.2
- I subscribe to the RSS feed of this BBC radio program as a podcast, pity you can’t find the feed on the webpage. ↩
- Ironically I used Aiko to get the text of the podcast for the quotes: “transcription is powered by OpenAI’s Whisper model running locally on your device” ↩
*** Although I think it’s essential to watch and listen to the clearly evolving experts regarding how these AI (neural net based architectured)systems do things differently than we do with our brains…
*** THE ESSENTIAL THING to watch and listen for IS WHAT THESE neural net based architectured SYSTEMS OF AI … CAN DO.
9 – Imperfect PERCEPTION DECEPTION
It’s good to remember how imperfect WE ARE – how fuzzy OUR BRAINS work.
and the evolving present imperfect fuzzy state of our world
Here’s One thing every one of our public schools in America should teach our students… and here’s the perfect tool… to watch and discuss… watch and discuss…
*** call it biology or social studies or civics
Your Brain: Perception Deception NOVA Full Documentary | NOVA | PBS – YouTube
Your Brain: Perception Deception | Full Documentary | NOVA | PBS – YouTube
I continue to pay most attention to and to rely on Geoffrey Hinton for my fundamental starting point in my ongoing process of trying to understand the systems and issues and our rapidly changing situation; I watch and listen and re-watch and re-listen to everything Geoffrey Hinton says and slowly merge that with other phrasings of the situation, other views of the situation – going back to re-think about what Geoffrey Hinton says so clearly.
Food for thought Geoffrey Hinton
MAY 2023
1 – Interviewer Hari Sreenavasin
*** On (Christiane) Amanpour & Co.
*** May 9, 2023 – 211 Thousand views
“Godfather of AI” Warns of the “Existential Threat” of AI – YouTube
*** GEOFFREY HINTON
2 – Interviewer: Brook Silva-Braga
*** On CBS Mornings
*** March 1, 2023 – 1.2 Million Views
(at Toronto Vector Institute)
Full interview: “Godfather of artificial intelligence” talks impact and potential of AI – YouTube
*** GEOFFREY HINTON
OR
3 – Interviewer Geoff Bennett
*** On PBS NEWSHOUR
*** May 5, 2023 – 292 Thousand Views
‘Godfather of AI’ discusses dangers the developing technologies pose to society – YouTube
*** GEOFFREY HINTON
Great food for thought.
*** Although I think it’s essential to watch and listen to the clearly evolving experts regarding how these AI (neural net based architectured)systems do things differently than we do with our brains…
*** THE ESSENTIAL THING to watch and listen for IS WHAT THESE neural net based architectured SYSTEMS OF AI … CAN DO.
8 – Emily M Bender AI on ChatBots
BBC Radio 4 – Word of Mouth, Chatbots
Via John Johnston’s
World Wide Wall Display
Blog 5-16-23
https://www.bbc.co.uk/programmes/m001l97m
Hi Dan,
Thanks for all the recommendations for more information. I’ll have a listen. I’ve started to put them through huffduffer so I can listen while commuting.