https://www.aljazeera.com/economy/2...ds-mourn-lost-love-after-cold-chatgpt-upgrade

Women with AI ‘boyfriends’ mourn lost love after ‘cold’ ChatGPT upgrade

Jane, who asked to be referred to by an alias, is among a small but growing group of women who say they have an AI “boyfriend”.

 

https://www.aljazeera.com/economy/2...ds-mourn-lost-love-after-cold-chatgpt-upgrade

Women with AI ‘boyfriends’ mourn lost love after ‘cold’ ChatGPT upgrade


View attachment 354436

 

https://www.newyorker.com/culture/open-questions/what-if-ai-doesnt-get-much-better-than-this

It’s hard to overstate how completely the A.I. community came to believe that it would inevitably scale its way to A.G.I. In 2022, Gary Marcus, an A.I. entrepreneur and an emeritus professor of psychology and neural science at N.Y.U., pushed back on Kaplan’s paper, noting that “the so-called scaling laws aren’t universal laws like gravity but rather mere observations that might not hold forever.” The negative response was fierce and swift. “No other essay I have ever written has been ridiculed by as many people, or as many famous people, from Sam Altman and Greg Brockton to Yann LeCun and Elon Musk,” Marcus later reflected. He recently told me that his remarks essentially “excommunicated” him from the world of machine learning. Soon, ChatGPT would reach a hundred million users faster than any digital service in history; in March, 2023, OpenAI’s next release, GPT-4, vaulted so far up the scaling curve that it inspired a Microsoft research paper titled “Sparks of Artificial General Intelligence.” Over the following year, venture-capital spending on A.I. jumped by eighty per cent.



After that, however, progress seemed to slow. OpenAI did not unveil a new blockbuster model for more than two years, instead focussing on specialized releases that became hard for the general public to follow. Some voices within the industry began to wonder if the A.I. scaling law was starting to falter. “The 2010s were the age of scaling, now we’re back in the age of wonder and discovery once again,” Ilya Sutskever, one of the company’s founders, told Reuters in November. “Everyone is looking for the next thing.” A contemporaneous TechCrunch article summarized the general mood: “Everyone now seems to be admitting you can’t just use more compute and more data while pretraining large language models and expect them to turn into some sort of all-knowing digital god.” But such observations were largely drowned out by the headline-generating rhetoric of other A.I. leaders

 

In other words ... what if this is as good as it gets? ...

​

 

Naked emperors seem to be multiplying at an alarming rate in today’s world.

 

Naked emperors seem to be multiplying at an alarming rate in today’s world.

There were plenty before and likely, fewer today in reality.

 

In other words ... what if this is as good as it gets? ...

View attachment 354442​

One of my favorite actors, and maybe the best actor in the last 100 years.

 

I find smut, sexual innuendo inappropriate on a technology and educational forum. There are countless options for participating in such banter, never forget who might read our posts, Young impressionable minds should be respected, childhood should be respected. You wanna post smut, head over to Twitter.

So far it looks like what has been said would only be understood by a mind that was already conditioned with such input. None of it is explicit.

If you feel very badly about that, then ask the moderators to delete the posts you think are too bad.
We like to mix humor in with the mix here because it is uplifting when talking about a hardwood topic like technology. It's not typical to joke about the stuff we did in this thread though.

 

https://arstechnica.com/ai/2025/08/...energy-cost-of-ai-queries-by-33x-in-one-year/
Google says it dropped the energy cost of AI queries by 33x in one year

So far this year, electricity use in the US is up nearly 4 percent compared to the same period the year prior. That comes after decades of essentially flat use, a change that has been associated with a rapid expansion of data centers. And a lot of those data centers are being built to serve the boom in AI usage. Given that some of this rising demand is being met by increased coal use (as of May, coal's share of generation is up about 20 percent compared to the year prior), the environmental impact of AI is looking pretty bad.

"Google also says that they can replace a full programmer with their Gemini chatbot."

I just Ignore the 'AI' result for a search.

 

https://arstechnica.com/ai/2025/08/...energy-cost-of-ai-queries-by-33x-in-one-year/
Google says it dropped the energy cost of AI queries by 33x in one year

"Google also says that they can replace a full programmer with their Gemini chatbot."

That reminds me of a post highlighted on Security Now a couple of months past:

"Claude 4 just refactored my entire codebase in one call.
25 tool invocations. 3000+ new lines. 12 brand new files.
It modularized everything. Broke up monoliths. Cleaned up spaghetti.

None of it worked. But boy was it beautiful."

 

https://www.theverge.com/news/767421/taco-bell-ai-drive-thru-trolls-glitches
Taco Bell’s AI drive-thru plan gets caught up on trolls and glitches

 

Hello again,

If we think about it, if we consider the theory of the simulated universe, then our own brain intelligence is artificial intelligence tuned a little bit better. We make lots of mistakes too but eventually can figure things out.

 

Usually a theory has at least some evidence or observation as a starting point.

 

As I think I've mentioned before, I've been increasingly using Copilot as a search engine. I have to admit that I am pretty amazed at how well it usually does -- most of the information it provides is actually pretty accurate and useful. But I have to vet everything very carefully. Today was a good case in point. We have gotten so much rain that the ground is saturated (half our backyard is a marsh) and groundwater has been coming up into one of our window wells (and then into the house). So I wanted to put a pump in a bucket down in the well to make it easier to keep it below the window bottom, which currently requires me to pump out about sixty gallons everything fifty minutes or so -- and we are supposed to get more rain tonight and tomorrow. I haven't slept in over 36 hours and I'm going to be napping in half-hour increments all weekend. Anyway, I needed to know the minimum number of holes I would need to get 100 gal/hr to flow into the bucket if the holes are located an inch or two below the window (which keeps them an inch or two off the dirt and gravel with the bucket dug down a ways. So, I asked Copilot so that I could see what it came up with as a starting point for figuring it out myself.

Here's the exchange:


The explanation and process looks reasonable and well explained (other than not tracking units properly, but I didn't expect it to).

I expected the answer to be more holes than this, but I also know that it doesn't take much of a hole to drain a 55 gal drum in an hour, so there were no real red flags.

Still, I always vet what these tools spout, so I ran the numbers myself, using it's recommended equations (and verifying that the units actually do work out properly) and ran smack dab into something I've seen multiple times -- no matter how impressive they sound, they can't do math very well.



It's off by a factor of three!

 

As I think I've mentioned before, I've been increasingly using Copilot as a search engine. I have to admit that I am pretty amazed at how well it usually does -- most of the information it provides is actually pretty accurate and useful. But I have to vet everything very carefully. Today was a good case in point. We have gotten so much rain that the ground is saturated (half our backyard is a marsh) and groundwater has been coming up into one of our window wells (and then into the house). So I wanted to put a pump in a bucket down in the well to make it easier to keep it below the window bottom, which currently requires me to pump out about sixty gallons everything fifty minutes or so -- and we are supposed to get more rain tonight and tomorrow. I haven't slept in over 36 hours and I'm going to be napping in half-hour increments all weekend. Anyway, I needed to know the minimum number of holes I would need to get 100 gal/hr to flow into the bucket if the holes are located an inch or two below the window (which keeps them an inch or two off the dirt and gravel with the bucket dug down a ways. So, I asked Copilot so that I could see what it came up with as a starting point for figuring it out myself.

Here's the exchange:
View attachment 355054

The explanation and process looks reasonable and well explained (other than not tracking units properly, but I didn't expect it to).

I expected the answer to be more holes than this, but I also know that it doesn't take much of a hole to drain a 55 gal drum in an hour, so there were no real red flags.

Still, I always vet what these tools spout, so I ran the numbers myself, using it's recommended equations (and verifying that the units actually do work out properly) and ran smack dab into something I've seen multiple times -- no matter how impressive they sound, they can't do math very well.


View attachment 355055

It's off by a factor of three!

Hi,

Yeah that's the strange thing about 'ai' it gets it completely wrong sometimes. I was going to investigate this to find out what it was doing internally that screws up the calculations, but it takes a lot of time to design equations for testing and then run the tests. It's surprising how good it can be with some things and how bad it can be with other things. It seems random but there may be a systemic reason it does this.

I had it design an algorithm one time that it swore up and down that it was correct even though it was way off and completely unusable.
Another time I had it solve some kind of simple circuit, and when it came out wrong, I told it to do it another way, and it came up with a totally different result (ha ha). It made up some silly reason why one came out different than the other. For now it's just not good for some things.

The other thing is when it does get things surprisingly perfect, it can lead us to a false sense of trust in future results. That's not good so we have to be very careful. Still, to date, I don't think I got it to do much I could not do myself better when it comes to math and related.

I calculated 0.0133 gallons per second per hole with no turbulence, and 0.60 to 0.65 times that in thin plastic with the hole edges deburred, which would lead to about 3.5 holes for 100 gallons per hour with no extra holes for a safety margin. Of course some extra holes is a good idea and it seemed to get that. I used Torricelli’s Law also.

 

The other day I figured I'd give copilot a try to help me with a doctor's visit via public transportation.

So, I asked "how many traffic lights between Cottman Ave. and Rhawn St. on the Roosevelt Blvd."

So, the answer was having to do with cameras along that stretch.

Then I asked to specify "traffic lights" and it gave me another incorrect answer.

So, then I went and used a map app and found my answer.

 

As I think I've mentioned before, I've been increasingly using Copilot as a search engine. I have to admit that I am pretty amazed at how well it usually does -- most of the information it provides is actually pretty accurate and useful. But I have to vet everything very carefully. Today was a good case in point. We have gotten so much rain that the ground is saturated (half our backyard is a marsh) and groundwater has been coming up into one of our window wells (and then into the house). So I wanted to put a pump in a bucket down in the well to make it easier to keep it below the window bottom, which currently requires me to pump out about sixty gallons everything fifty minutes or so -- and we are supposed to get more rain tonight and tomorrow. I haven't slept in over 36 hours and I'm going to be napping in half-hour increments all weekend. Anyway, I needed to know the minimum number of holes I would need to get 100 gal/hr to flow into the bucket if the holes are located an inch or two below the window (which keeps them an inch or two off the dirt and gravel with the bucket dug down a ways. So, I asked Copilot so that I could see what it came up with as a starting point for figuring it out myself.

Here's the exchange:
View attachment 355054

The explanation and process looks reasonable and well explained (other than not tracking units properly, but I didn't expect it to).

I expected the answer to be more holes than this, but I also know that it doesn't take much of a hole to drain a 55 gal drum in an hour, so there were no real red flags.

Still, I always vet what these tools spout, so I ran the numbers myself, using it's recommended equations (and verifying that the units actually do work out properly) and ran smack dab into something I've seen multiple times -- no matter how impressive they sound, they can't do math very well.

View attachment 355055

It's off by a factor of three!

Did you tell it it was off by a factor of three? I've argued with copilot myself a few times, it ended up apologizing profusely.

Good luck with the water, I know how those situations can be.

 

Did you tell it it was off by a factor of three? I've argued with copilot myself a few times, it ended up apologizing profusely.

That's one of the things I find really annoying about these tools -- they are programmed to incessantly blow smoke up your backside. There are few things I find more off putting (in a human) than that, because it's almost always insincere manipulation. Frankly, I think the same motive is behind having these LLMs do it, too.

Good luck with the water, I know how those situations can be.

Thanks. Very frustrating. A couple thousand dollars worth of textbooks were destroyed and the carpet in this room is soaked. But, other than that -- and having to get up every hour and a half to run the pump, which doesn't make for good sleep -- it could be a lot worse. Later I'm going to try to seal the window closed to hopefully limit the intrusion if the water rises too far again, but that will mean running power to the pump along the outside of the house. Putting in a proper sump pump will take quite a bit more effort, so my life's pretty much chained to the house until the ground dries out. Good thing I'm "retired".

 

That's one of the things I find really annoying about these tools -- they are programmed to incessantly blow smoke up your backside. There are few things I find more off putting (in a human) than that, because it's almost always insincere manipulation. Frankly, I think the same motive is behind having these LLMs do it, too.



Thanks. Very frustrating. A couple thousand dollars worth of textbooks were destroyed and the carpet in this room is soaked. But, other than that -- and having to get up every hour and a half to run the pump, which doesn't make for good sleep -- it could be a lot worse. Later I'm going to try to seal the window closed to hopefully limit the intrusion if the water rises too far again, but that will mean running power to the pump along the outside of the house. Putting in a proper sump pump will take quite a bit more effort, so my life's pretty much chained to the house until the ground dries out. Good thing I'm "retired".

Yeah, "retired" into doing a lot of other things anyway.

Good luck with the sump pump that's a good way to go. We had to do that one place I lived. It was the only way to do it. They last pretty long, or at least they DID about 45 years ago.
Installation often means digging a hole just for the pump to sit in. We did a concrete lined square hole to keep it clean.

 

The other day I figured I'd give copilot a try to help me with a doctor's visit via public transportation.

So, I asked "how many traffic lights between Cottman Ave. and Rhawn St. on the Roosevelt Blvd."

So, the answer was having to do with cameras along that stretch.

Then I asked to specify "traffic lights" and it gave me another incorrect answer.

So, then I went and used a map app and found my answer.

Sounds like Philly. Must be a lot of traffic lights along that road.
Why count the traffic lights though? Did you count the manhole covers too and the pot holes