• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Master AI (Artificial Intelligence) Discussion/News Thread

Where is it now? I'm predicting (guessing) large databases will be soon locked from common free access.
I think this whole kerfuffle started a few decades ago:

Google Books (previously known as Google Book Search, Google Print, and by its code-name Project Ocean)
Google's original vision (circa 2002) was to create a digital library of all books
This initiative was later expanded (in 2004) with the announcement of the Library Project, which involved partnerships with major university and public libraries...
Google publicly stated its intention to scan all 129,864,880 known books within a decade (in 2010), reaffirming its commitment to digitizing as much printed material as possible.
Then, the project got ugly and quickly!

A 2017 media proclamation summarized "What Happened to Google's Effort to Scan Millions of University Library Books", like this:
...It’s a pillar of the humanities’ growing engagement with Big Data...
... yet the promised library of everything hasn’t come into being...
...An epic legal battle was ultimately dismissed the case (in 2013), handing Google a victory that allowed it to keep on scanning. Yet, the dream of easy and full access to all those works remains just that...
TL&DR: Hidden in the bowels of Google's data centers, there is a database containing >25 million books but nobody is allowed to read them.... Google helped create this db and uses it as a dataset they can query, even if they can’t consume full texts.

I had suspected [not being a history fan] that this Google db (e.g. Google Books) may have made it the universe's gold-standard text corpora; for Google's LLM training.
Gemini answers my suspicion thusly:
...is highly likely to have been used in training Google's large language models (LLMs), though the company has not confirmed this explicitly.
:confused:
 
I think this whole kerfuffle started a few decades ago:

Google Books (previously known as Google Book Search, Google Print, and by its code-name Project Ocean)
Google's original vision (circa 2002) was to create a digital library of all books
This initiative was later expanded (in 2004) with the announcement of the Library Project, which involved partnerships with major university and public libraries...
Google publicly stated its intention to scan all 129,864,880 known books within a decade (in 2010), reaffirming its commitment to digitizing as much printed material as possible.
Then, the project got ugly and quickly!

A 2017 media proclamation summarized "What Happened to Google's Effort to Scan Millions of University Library Books", like this:

TL&DR: Hidden in the bowels of Google's data centers, there is a database containing >25 million books but nobody is allowed to read them.... Google helped create this db and uses it as a dataset they can query, even if they can’t consume full texts.

I had suspected [not being a history fan] that this Google db (e.g. Google Books) may have made it the universe's gold-standard text corpora; for Google's LLM training.
Gemini answers my suspicion thusly:

:confused:
Yes that pertains to Google. The "where is it now?" was about data that DOGE took from departments of the US Government and would add what are doing with it? Deleting? Changing? Is it still available to US Gov. employees who normally access it? Are copies floating around?
 
Or better not, because of that Adam and Eve thing - thrown out of paradise because they knew too much :cool:
 
Does any "AI" really know what time it is? (Apologies to Robert Lamm.)


I really like the graphic they used to head that article (as an illustration of how 'AI' would assemble 'time').

Vrg_illo_Kristen_Radtke_why_cant_chatgpt_tell_time.jpg.webp
 
Last edited:
Meanwhile, on the industrial relations front:

More than 1,000 Amazon workers warn rapid AI rollout threatens jobs and climate

Reflecting broader AI concerns across the industry, the letter was also supported by more than 2,400 workers from companies including Meta, Google, Apple and Microsoft. ... Workers emphasized they are not against AI outright, rather they want it to be developed sustainably and with input from the people building and using it. “I see Amazon using AI to justify a power grab over community resources like water and energy, but also over its own workers, who are increasingly subject to surveillance, work speedups, and implicit threats of layoffs,” said the senior software engineer. “There is a culture of fear around openly discussing the drawbacks of AI at work, and one thing the letter is setting out to accomplish is to show our colleagues that many of us feel this way and that another path is possible.”

In the eternal struggle of capital vs labour, the latter are usually more reasonable, the former more avaricious. That hasn't changed in this phase, tragically.
 
If there is any technology that needs usable AI, it is photoshop. Alas, the few times I have tried their AI features, they have been terrible.
I use photoshop AI often, for expanding images, replacing unwanted items, etc. and have generally found it amazingly good. The prompts and sequences of operations do take some messing around with ,and I'm using the Beta versions, but my experience is that it's magical. What doesn't work at all is images of people, Adobe has very tight censorship limitations to the point of being ridiculous.
 
I'm looking forward to being able to zoom into an image indefinitely.
Ala the "Esper" Machine in Blade Runner.
 
An AP News story on how generative and agentic A.I. are changing how shopping is done in the U.S.

Google has introduced an 'agent' that can call local stores to ask if a product is in stock.

 
A lot of what's now been labeled "AI" has been around for a bit. Now everything that remotely touches a GPU is "AI".

You could say something like Photoshop has been "AI" for many, many years - just looked it up and it was introduced in... 1990!!!! Never had a clue, I started using it when I got into digital photography probably 2000 or so, and it helped my ex-wife worked at Adobe and I got it for free, but it was mindblowing... but a complicated pig to use. Over time it adopted and streamlined the most used workflows and crunched the hell out of GPUs. Now that linear, evolutionary path is suddenly all about AI, supposedly :-D. And note that AI to a certain degree can *cost* companies some revenues - with all he "AI" features, now Photoshop Elements does the job for me perfectly, at a fracton of the cost of the full product suite.

I personally find it silly to call anything that happens in my smartphone "AI", but it may leverage functions in an app that sends stuff into a *truly* AI powered data center, meaning the multi-billion stuff leveraging top Nvidia GPUs that set you back over 30k a pop and you'll need hundreds or thousands of them to do what I'd call "real AI", the more disruptive aspect of it.
 
'OOFSkate' - an app that uses the camera of phone or tablet and A.I. to measure and assess the technical form of a figure skater.

The unfurled URL, arguably, begs the question: Does Olympic-level figure skating need to be revolutionized?
I mean... I'd say... no...
and/or I'd wonder why it needs to be revolutionized?
EDIT: Seems to be there are better things to revolutionize... maybe I am too naive.
 
The unfurled URL, arguably, begs the question: Does Olympic-level figure skating need to be revolutionized?
I mean... I'd say... no...
and/or I'd wonder why it needs to be revolutionized?
EDIT: Seems to be there are better things to revolutionize... maybe I am too naive.

Some may regard the aim of separation of the assessment of technical form by an app and the assessment of artistic form by human judges to be revolutionary in the sport of figure skating.
 
The unfurled URL, arguably, begs the question: Does Olympic-level figure skating need to be revolutionized?
I mean... I'd say... no...
and/or I'd wonder why it needs to be revolutionized?
EDIT: Seems to be there are better things to revolutionize... maybe I am too naive.

The big issue there is that AI would not be trained to rate creative, new complicated moves when they sudddenly appear, which would be utterly unfair to innovators and would potentially kill attempts to grow beyond the same old moves. AI's strength is to accelerate cutting through complexity within the *established* parameters provided to its algorithms.
 
The big issue there is that AI would not be trained to rate creative, new complicated moves when they sudddenly appear, which would be utterly unfair to innovators and would potentially kill attempts to grow beyond the same old moves. AI's strength is to accelerate cutting through complexity within the *established* parameters provided to its algorithms.

I am unfamiliar with most of the sport. Could human judges assess new moves where A.I. does not have sufficient training to be effective in measuring technical form ?
 
The current state of the AI playing field puts me to mind of one Dr. George Davis, who worked for me in my early days in biotech.
George was fond of saying
If the only tool you have is a hammer, all of your problems look like nails.
 
I am unfamiliar with most of the sport. Could human judges assess new moves where A.I. does not have sufficient training to be effective in measuring technical form ?
It has happened before, many times, in gymnastics. The judges with many years of background may go "Wow, I could never do that and that's so awesome - definitely a 10!", while AI with its rules would most likely go "I would have preferred to see the classic complications in several routines formerly ranked as 10s, that's what I was trained on". And that's guaranteed. What AI hasn't been trained on, it adds 0% value on. In fact that's the source of many so called AI hallucinations.

In a nutshell, we're trying to fit "AI" into may use cases it's not a fit for. AI is crap when i comes to interpretative shades of grey (no references to the book please :-D).
 
It has happened before, many times, in gymnastics. The judges with many years of background may go "Wow, I could never do that and that's so awesome - definitely a 10!", while AI with its rules would most likely go "I would have preferred to see the classic complications in several routines formerly ranked as 10s, that's what I was trained on". And that's guaranteed. What AI hasn't been trained on, it adds 0% value on. In fact that's the source of many so called AI hallucinations.

In a nutshell, we're trying to fit "AI" into may use cases it's not a fit for. AI is crap when i comes to interpretative shades of grey (no references to the book please :-D).
I disagree. AI does not need to be trained on everything possible to evaluate technique.

Maybe artistry.
 
Back
Top Bottom