well-temperedforum.groupee.net
Articles on AI
12 April 2023, 01:37 PM
wtgArticles on AI
Seeing a lot of it on the news, so thought I'd start a thread where we can throw any interesting articles about how AI may affect us in the future, in a positive or a negative way.
First up, Eric Topol on medical AI:
quote:
In the journal Nature today, my colleagues and I published an article on the future directions of generative A.I. (aka Large Language or Foundation models) for the practice of medicine. These new AI models have generated a multitude of new and exciting opportunities in healthcare that we didn’t have before, along with many challenges and liabilities. I’ll briefly explain how we got here and what’s in store.
https://erictopol.substack.com...7473&isFreemail=true
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
12 April 2023, 03:45 PM
wtgA bit of history.
quote:
The Thrill — And The Mystery — Of A 1970s Bell Labs AI Chatbot Known As ‘Red Father’
https://www.forbes.com/sites/a...her/?sh=67ed44ff2277
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
12 April 2023, 03:45 PM
wtgThe people who love AI.
quote:
AI is flooding the workplace, and workers love it
Who’s afraid of ChatGPT? Not these workers.
https://www.vox.com/technology...-bing-bard-work-jobs
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
12 April 2023, 07:51 PM
dolmansaxlilI used ChatGPT last week to help me come up with an interesting title for a workshop I’m doing. Then I used it to help me revise my bio for the same event. It’s not going to give you original or groundbreaking content, but it can be a really helpful tool.
--------------------------------
"Your first 10,000 photographs are your worst." ~ Henri Cartier-Bresson
21 April 2023, 06:11 AM
Axtremus"Worse than useless"
"Pathological liar"
Some of the more colorful comments reportedly made by Google employees who internally trialed the Bard AI.
Not sure how much weight to give them, though, most large enough organizations tend to have a few people that are super cautious or super critical when reviewing something, anything.
25 April 2023, 10:27 PM
AxtremusArticle on a large set of open source images used to train image generation AI.
This is a big deal with many big unresolved issues. Copyright ownership, whether AI company can use the images to train their AI models without permission, whether the AI companies should be compelled to disclose what data sets they use to train their AI models, etc. All these can impact the development and use/commercialization of AI.
quote:
A man widely seen as the godfather of artificial intelligence (AI) has quit his job, warning about the growing dangers from developments in the field.
Geoffrey Hinton, 75, announced his resignation from Google in a statement to the New York Times, saying he now regretted his work.
He told the BBC some of the dangers of AI chatbots were "quite scary".
"Right now, they're not more intelligent than us, as far as I can tell. But I think they soon may be."
Dr Hinton also accepted that his age had played into his decision to leave the tech giant, telling the BBC: "I'm 75, so it's time to retire."
Dr Hinton's pioneering research on neural networks and deep learning has paved the way for current AI systems like ChatGPT.
https://www.bbc.com/news/world-us-canada-65452940
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
04 May 2023, 03:45 AM
AmandaI quake for those whose livelihoods depend on carrying out functions which chatbots can do as well or better.
Among them I include pathologists (hitherto one of the most lucrative fields in medicine).
Tssk, you just never know when you're choosing your major or medical specialty especially if remuneration is a key factor!
--------------------------------
The most dangerous word in the language is "obvious"
04 May 2023, 03:47 AM
AmandaAnd as far as my personal Turing test goes, I wait for a chatbot that can come up with really funny jokes (including the timing!). Sense of humor is so different from language to language, from culture to culture. Likewise, metaphors and folk references.
I still remember with a blush even to myself, the (few!) times I attempted to tell a joke (foolproof ones, I thought) to friends in a foreign language. That ghastly silence following my delivering the (hilarious) punchline still haunts me.
Then too (to illustrate my point) I remember the night I was crossing the border into West Germany- ages ago, of course. I handed a border guard a tourist guide ("Europe on Five Dollars a Day"?) to inquire about a few inexpensive hotels recommended.
He shook his head discouragingly, remarking enigmatically "Paper is patient" ("Papier ist geduldig").
Dead silence from me. He repeated his informative statement again, louder. "PAPER IS PATIENT" [you fool!].
What th-? Since when does paper have personality traits?
Giving up, he translated the expression from its metaphoric meaning. To him, it had been straightforward information but - not really.
"You can lie to paper and it won't complain!" AHA! I finally got it. The hotels were a lot more expensive than the author had claimed. But I doubt even an advanced chatbot could have done better than me. It was one of many times I realized why the Geneva Institute for Simultaneous Interpretation (for the UN) is so exclusive in acceptance, and so time-consuming in training. (And no wonder the salaries are so high, especially for certain language combinations)! You're NOT just absorbing a Google Translation function, but centuries of culture and history!
I did not pursue that ambition.
--------------------------------
The most dangerous word in the language is "obvious"
05 May 2023, 08:57 AM
AxtremusHat tip George K
https://www.semianalysis.com/p...-no-moat-and-neitherArticle purported to be a leaked memo by a Google employee. But regardless of its provenance, it's a good article with lots of good information in there.
As a matter of opinion, it argues that open source is the way of the future for AI, and it is rather bullish of the prospect of a certain class of model training technique called Low Rank Adaptation (LoRA).
The article also has a nice timeline showing very rapid improvements/innovations in the space starting from Feb. 2023 (yes, this year) when Meta (Facebook) released a consequential set of AI code.
Google introduces text-to-music AI.
quote:
Google today released MusicLM, a new experimental AI tool that can turn text descriptions into music. Available in the AI Test Kitchen app on the web, Android or iOS, MusicLM lets users type in a prompt like “soulful jazz for a dinner party” or “create an industrial techno sound that is hypnotic” and have the tool create several versions of the song.
Users can specify instruments like “electronic” or “classical,” as well as the “vibe, mood, or emotion” they’re aiming for, as they refine their MusicLM-generated creations.
https://techcrunch.com/2023/05...placement=newsletter
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
quote:
How AI Knows Things No One Told It
Researchers are still struggling to understand how AI models trained to parrot internet text can perform advanced tasks such as running code, playing games and trying to break up a marriage
https://www.scientificamerican...ings-no-one-told-it/
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier
16 May 2023, 06:39 AM
ShiroKuroHere’s one from this morning, MS claims the ChatGPT model not available to the public (I.e., the more powerful one), showed artificial generalized intelligence:
Microsoft Says New A.I. Shows Signs of Human Reasoning28 May 2023, 03:55 PM
AxtremusArticle with sample outputs from Google's MusicLM AI that generates music from text prompts. The reviewers largely panned the results.
quote:
We asked Google’s new AI music bot to write us a song. We instantly regretted it
After listening to the samples they included in the article, I can see why.
--------------------------------
When the world wearies and society ceases to satisfy, there is always the garden - Minnie Aumônier