Effect pretty much the same.
-
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
- Thread starter amirm
- Start date
Sounds like a mangement failure to me.
The Pope Leo XIV' encyclical summarized in a humanist version (w/o the theological-scaffolding) is truly relevant and appropriate.
Even for those who'd rather not get into the religiosity of it.
It feels to me as if the younger generation are growing very concerned and skeptical.
I know it’s anecdotal and full of sampling bias but the continual refrains I hear from my kids and their friends (who are approaching the end of their studies) are “who asked for this?” and “who benefits from this?”
In both cases, their answers are resoundingly “not us”
EDIT: it seems to me that those two questions are not being properly asked in any sort of national conversation.
I know it’s anecdotal and full of sampling bias but the continual refrains I hear from my kids and their friends (who are approaching the end of their studies) are “who asked for this?” and “who benefits from this?”
In both cases, their answers are resoundingly “not us”
EDIT: it seems to me that those two questions are not being properly asked in any sort of national conversation.
- Joined
- Jul 21, 2019
- Messages
- 3,490
- Likes
- 5,170
Rewarding the burning of "tokens" works about as well as you might expect:
https://gizmodo.com/companies-are-getting-burned-by-burning-tons-of-tokens-2000765232
https://gizmodo.com/companies-are-getting-burned-by-burning-tons-of-tokens-2000765232
Serious one. In any company I worked for, the person responsible for that lack of foresight would no longer be working there.
somebodyelse
Master Contributor
- Joined
- Dec 5, 2018
- Messages
- 6,594
- Likes
- 6,463
One cause of excessive token use is apparently the needlessly verbose data exchange from some of the tools in the chain. Some people are developing proxies to strip the unnecessary content, reducing token use significantly:
https://www.theregister.com/ai-ml/2...o-slash-ai-bills-then-open-sources-it/5248702
On the cost side it isn't necessarily a lack of control on the company side. Some have set limits that they thought would prevent overspend, only to find that they don't work as expected. If you had a record of paying promptly, and had spent over $1000 with them in the past, Google would automatically raise your spending limit without notification. This is especially problematic given the previously noted change in API key usage. They have at least been voiding the overspend for some 'lucky' customers, but sometimes only after the press got involved. Meanwhile on AWS the AI spending was in one of the categories not covered by the spending limit - strictly speaking it was documented, but not exactly obvious.
https://www.theregister.com/ai-ml/2...o-slash-ai-bills-then-open-sources-it/5248702
On the cost side it isn't necessarily a lack of control on the company side. Some have set limits that they thought would prevent overspend, only to find that they don't work as expected. If you had a record of paying promptly, and had spent over $1000 with them in the past, Google would automatically raise your spending limit without notification. This is especially problematic given the previously noted change in API key usage. They have at least been voiding the overspend for some 'lucky' customers, but sometimes only after the press got involved. Meanwhile on AWS the AI spending was in one of the categories not covered by the spending limit - strictly speaking it was documented, but not exactly obvious.
Again, an example where it must be the user’s fault if the results are poor, not that the self-targeting users of public-facing generative AI tools will be qualified to make that judgment.
Rick “it’s the people, stupid” Denney
somebodyelse
Master Contributor
- Joined
- Dec 5, 2018
- Messages
- 6,594
- Likes
- 6,463
An interesting article about OpenAI's recent disproving of the Erdős unit distance conjecture, and what it shows about the relative strengths and weaknesses of humans an current AI in mathematics. It used a novel combination of existing ideas from disparate fields of mathematics, sufficiently separate that an individual human mathematician is unlikely to know of them all. So while it was using existing knowledge as @pablolie often says, it combined it in a way that hasn't been done by humans in the 80 years since the conjecture was put forward. In a different field that sort of combination would qualify as sufficiently non-obvious to qualify for a patent if a human did it. It's the sort of thing that normally requires a team with a variety of backgrounds, and some luck in that combination. Another aspect is that it was a type of technique that “consumes much time and frequently doesn’t work out,” so humans may not bother to follow it through, while AI can do that - assuming you think it's worth the cost of the tokens. That's the bit where it may get tricky as companies make the transition from loss leaders to trying make a profit, as we're seeing with recent cases of bill shock and complaints of running down your month's worth of tokens in a few hours.
https://arstechnica.com/ai/2026/06/openais-math-breakthrough-played-to-ais-strengths/
https://arstechnica.com/ai/2026/06/openais-math-breakthrough-played-to-ais-strengths/
Find Mo Gawdat insights already for some years revealing.
if you could find the time have listen.
Also on YouTube;
Summery of the podcast:
AI Expert Mo Gawdat returns to The Diary Of A CEO to reveal why AGI has already arrived, why 30% of jobs will disappear by 2027, and why the most dangerous thing about AI isn't the technology - it's the people in charge of it.
Mo Gawdat is the former Chief Business Officer at Google X, founder of One Billion Happy, and co-founder of Emma.Love. He is a 4x international bestselling author, and his upcoming book ‘Alive: A Human's Guide to Living in the World of AI’, will be released in October 2026.
He explains:
How AI can give you a 400-point IQ boost, and why most people are wasting it
Why Mo actually wants a machine smarter than all of humanity to take control
Why Sam Altman said AI will "likely end humanity", and what he chose to do next
Why capitalism breaks when AI replaces the workers who buy the things we make
Why AI unemployment could trigger civil unrest before governments are ready for it
if you could find the time have listen.
Also on YouTube;
Summery of the podcast:
AI Expert Mo Gawdat returns to The Diary Of A CEO to reveal why AGI has already arrived, why 30% of jobs will disappear by 2027, and why the most dangerous thing about AI isn't the technology - it's the people in charge of it.
Mo Gawdat is the former Chief Business Officer at Google X, founder of One Billion Happy, and co-founder of Emma.Love. He is a 4x international bestselling author, and his upcoming book ‘Alive: A Human's Guide to Living in the World of AI’, will be released in October 2026.
He explains:
How AI can give you a 400-point IQ boost, and why most people are wasting it Why Mo actually wants a machine smarter than all of humanity to take controlWhy Sam Altman said AI will "likely end humanity", and what he chose to do nextWhy capitalism breaks when AI replaces the workers who buy the things we makeWhy AI unemployment could trigger civil unrest before governments are ready for it
Last edited:
- Joined
- Apr 24, 2022
- Messages
- 1,816
- Likes
- 2,433
'Lights, Camera, Prompt' !
An AI generated movie that was shown at Cannes cost $500,000 and $400,000 was spent on AI
A prompt to generate content typically required 3,000 words and apparently AI can assist that task.
gift link:
An AI generated movie that was shown at Cannes cost $500,000 and $400,000 was spent on AI
A prompt to generate content typically required 3,000 words and apparently AI can assist that task.
gift link:
- Joined
- Apr 24, 2022
- Messages
- 1,816
- Likes
- 2,433
Film director Martin Scorcese is using AI to save time when 'storyboarding' a movie during preproduction. (gift link)
www.nytimes.com
Martin Scorsese Is Embracing A.I.
In a sign of Hollywood’s softening stance on artificial intelligence, the cinema icon is backing Black Forest Labs, an image and video generation start-up.
Chr1
Major Contributor
- Joined
- Jul 21, 2018
- Messages
- 2,590
- Likes
- 2,728
Jack Clark, co-founder of Anthropic calls for a "brake pedal" for AI.
Most likely to be ignored however... Just as previous warnings have.
(What could possibly go wrong? Ho hum.)
Most likely to be ignored however... Just as previous warnings have.
(What could possibly go wrong? Ho hum.)
Last edited:
I have a chatgpt subscription and use it in Thinking mode and my results blow me away how good they are on a consistent basis.
However, out of convenience since I already there, I find myself using Google AI in its default Fast mode and I understand all those folks whose experience is AI spitting back halucinations. I almost wonder did Google make it that bad on purpose to try and give people bad experience long enough that they could catch up with others? The consensus of AI experts seem to place Anthropic, Gemini and then ChatGPT as the best AI as of today. That is not my experience for chat AI, I have not used agentic AI. If so, then Google has caught up and doesn't need to hamstring Google AI. Or it may be just the major clientel using Google AI just don't seem to care as long as they get a fast response? Or it could just be economics, so many people us Google, they need to push them to fast modes with cached stale data to afford the offering for free?
However, out of convenience since I already there, I find myself using Google AI in its default Fast mode and I understand all those folks whose experience is AI spitting back halucinations. I almost wonder did Google make it that bad on purpose to try and give people bad experience long enough that they could catch up with others? The consensus of AI experts seem to place Anthropic, Gemini and then ChatGPT as the best AI as of today. That is not my experience for chat AI, I have not used agentic AI. If so, then Google has caught up and doesn't need to hamstring Google AI. Or it may be just the major clientel using Google AI just don't seem to care as long as they get a fast response? Or it could just be economics, so many people us Google, they need to push them to fast modes with cached stale data to afford the offering for free?
- Joined
- Apr 24, 2022
- Messages
- 1,816
- Likes
- 2,433
AI is assisting....
An automatic milking machine that cows can autonomously use on demand.
An autonomous tractor that identifies weeds and kills them with a laser beam.
Workers remotely piloting multiple mowing tractors for turf/sod companies and helping attract a younger demographic.
A gift link to an article on AI in agriculture in The New York Times Magazine:
www.nytimes.com
An automatic milking machine that cows can autonomously use on demand.
An autonomous tractor that identifies weeds and kills them with a laser beam.
Workers remotely piloting multiple mowing tractors for turf/sod companies and helping attract a younger demographic.
A gift link to an article on AI in agriculture in The New York Times Magazine:
From Cow-Milking Robots to Weed-Zapping Lasers, Farmers Are Embracing A.I.
We asked three farmers to tell us how new technology is revolutionizing the way they work.
- Joined
- Mar 1, 2021
- Messages
- 3,078
- Likes
- 6,604
That’s been a thing in the dairy farms round here for a few years, along with the rise of big ugly dairy sheds as it’s financially beneficial to keep dairy cattle inside for the majority of the year and feed them silage - also cuts down on the workforce as some of the dairy sheds have robotic shit clearing machines.
Personally I think it’s fu*ked
Ralph_Cramden
Major Contributor
- Joined
- Dec 6, 2020
- Messages
- 4,144
- Likes
- 7,755
I gather the following is based on Cloudflare network traffic data:
Agentic/bot web traffic is said to be "57.4% of all web activity", as of May 2026. They have officially surpassed the web-based human activity, spiking by 187% in 2025.
According to a CNBC report, 'automated AI programs for human users' also rose by a whopping 8,000%.
Projections cum predictions were that this flip would not occur until the end of 2027 but, here we are!
The meaning of these spikes are not exactly known at the current time for humans and/or the internet.
I have found successful work-arounds for the internet's present 'advertisement' biz-model but that model may be in peril (at best) or becomes AI-infused (at worst).
The old two-tier internet concept is resurfacing again and may even be making a comeback already: In 2025, Cloudflare launched "Pay per Crawl", which allows publishers to charge AI scrapers for access to their content, via "toll roads". Cloudflare CEO -Mathew Prince- puts it this way:
Agentic/bot web traffic is said to be "57.4% of all web activity", as of May 2026. They have officially surpassed the web-based human activity, spiking by 187% in 2025.
According to a CNBC report, 'automated AI programs for human users' also rose by a whopping 8,000%.
Projections cum predictions were that this flip would not occur until the end of 2027 but, here we are!
The meaning of these spikes are not exactly known at the current time for humans and/or the internet.
I have found successful work-arounds for the internet's present 'advertisement' biz-model but that model may be in peril (at best) or becomes AI-infused (at worst).
The old two-tier internet concept is resurfacing again and may even be making a comeback already: In 2025, Cloudflare launched "Pay per Crawl", which allows publishers to charge AI scrapers for access to their content, via "toll roads". Cloudflare CEO -Mathew Prince- puts it this way:
Dorothy was right!
Last edited:
We ain't seen nothin' yet!
A new breed of fringe groups are just mobilizing:
I have stocked up on my supply of popcorn already and staying tuned.
- Joined
- Mar 1, 2021
- Messages
- 3,078
- Likes
- 6,604
I see we are now in the minority so all those counts of “1m million users per month” to justify that one website is better than another is rather chin strokinly suspect
www.cnet.com
AI Agents Now Generate More Web Traffic Than Humans
Cloudflare CEO Matthew Prince says he didn't expect this milestone until 2027.
www.cnet.com
Similar threads
