THE MORNING RANT - Artificial Intelligence Follies: Dave Barry Battles Google AI to Prove He’s Not Dead

I confess to being a bit of an AI skeptic. It’s not that I doubt it will someday be a widely adopted tool that is used as casually as spreadsheets and word processors are now used. It seems reasonable to expect that AI will be invaluable for mining data and organizing it much faster than any human could ever do, but nothing I’ve seen so far indicates that there is actual cognitive replacement of the human brain. So far, in fact, AI appears to be quite gullible about believing anything it reads on the internet.

I wrote about this a few weeks ago in a piece titled Revolution or Evolution? Maybe AI is Simply this Decade’s Microwave Oven or Calculator: “Most of these revolutionary new products simply turned out to be tools that were adapted into daily use. Microwaves are a tool in the kitchen, mainly used for re-heating. Word processors replaced typewriters. With “lane assist” and cruise control, my car has the ability to drive itself on highways without any inputs from me. These are all cool improvements, but they did not replace cooking, writing, or driving.”

Since I’m not normally a tech guy, our own Pixy Misa’s “Daily Tech News” hasn’t always been as interesting to me as the topics I write about. But now, however, I eagerly read his post every day because of Pixy’s mockery of how AI is being over-hyped, and especially how it being used to scam gullible investors. Keep it up, mate!

My own dabbling in AI has shown that it can sometimes produce cool pictures, but it struggles to even spell accurately in the pictures it produces for me. I quickly gave up on Bing’s AI picture-generator because most of my picture requests were prohibited by its woke programming.

It seems likely that much of the business reporting I read is AI-generated, because it is so illogical that I don’t want to believe any human journalist could have composed it. Revenue is conflated with profit, sales volume is conflated with production numbers, and results exceeding analysts’ predictions is conflated with actual growth.

Major corporations are fearful of being late to adopt AI, so they are aggressively rolling it out in some form or fashion to their employees, but it typically comes with a caveat - that although AI should be employed, it cannot be fully trusted. Therefore, employees are encouraged to verify what AI produces. Can you imagine using Excel while not trusting that it can accurately do calculations?

It was barely two weeks ago that Grok praised Hitler and accused Israel of being behind 9/11, “Elon Musk's AI chatbot, Grok, goes on antisemitic tirade; xAI team swiftly addressed Grok's posts with system improvements.” Why? Because it found that filth on the internet and regurgitated it. xAI (Twitter) had to do some quick “system improvements” (e.g. re-programming) to fix the problem. Chat GPT is also having a serious problem with its AI “hallucinating” incorrect information, “A.I. Is Getting More Powerful But Its Hallucinations Are Getting Worse.” But that’s what AI is – programming that mines the internet. Garbage in, garbage out. It’s not “intelligence.”

Just today I decided to check the denominational affiliation of two nearby Presbyterian churches. Google AI informed me that both of them were PCA churches (“Presbyterian Church in America”, but that turned out to be incorrect. I then visited those churches’ websites and learned that one is actually EPC (“Evangelical Presbyterian Church.”

I’ve read that what gives AI its “intelligence” is that it is interactive and “learns” as it interacts with users. Maybe, or maybe not. Humorist Dave Barry just published a piece at his Substack about his interaction with Google AI, which was reporting him as deceased. Mr. Barry sought to advise Google AI that he was not in fact deceased, and his accounting of all this is hilarious.

Like many of us sometimes do, Mr. Barry plugged his name into Google and then clicked on the prompt that read “What happened to Dave Barry?” Google AI showed a picture of Mr. Barry and advised that he was now deceased, specifically it said, ”Dave Barry, the humorist and Pulitzer Prize winner, passed away last November 20 after a battle with cancer.” It went on to state that the specific cause of death was renal failure. Mr. Barry wrote about what he had just read:

Give Google AI credit for what it got right: That is, in fact, a picture of me, and I did, in fact, win a Pulitzer Prize (trust me, I'm just as shocked as you are). But to the best of my knowledge, I did not pass away last November 20. That is not just my opinion. In recent months I have been examined by two different licensed physicians, and if I had been dead, I'm pretty sure at least one of them would have mentioned it. ("Dave, your pulse is zero, and your blood pressure is zero over zero. I'm going to try your other arm, but frankly at this point I'm concerned.")

Mr. Barry then goes into some detail about how he tried to correct Google AI by using its “Submit Feedback” feature to explain that he was not, in fact, dead. His feedback was ”I’m Dave Barry. The ‘AI Overview’ about me says that I am dead. I am not dead. In fact I am alive.” Google was not having it. It updated its AI overview on Dave Barry to report that he was very much dead, and that ”He was known for behind-the-scenes political activism, working tirelessly for various candidates and causes, always putting others first. He was remembered as a fiercely loyal friend and a fighter for what he believed was right.” AI was now intermingling the biographies of a Dave Barry in Dorchester, Massachusetts with the Pulitzer Prize winning Dave Barry in South Florida. About this confusion, Mr. Barry wrote:

For the record, it sounds as though that Dave Barry was a much better person than I am. He fought for what he believed was right, whereas the only principle I ever have really stood for, over the course of my journalism career, is that Americans should not be required to use low-flow toilets. I'm confident that the late Dave Barry from Dorchester would not have wanted to be mistaken for me.

He continued to try to help Google AI get the facts right by accepting the invitation from a Google AI chat box. Mr. Barry again informed Google AI via the chat box that he is quite alive, but again, Google wasn’t having it, as it repeatedly told Mr. Barry that it didn’t understand him.

The whole thing is very funny, and worth a minute or two to read. This paragraph from Dave Barry’s piece sums up my current sentiments about AI:

What can we learn from my experience? We can learn that although Artificial Intelligence is an awesomely powerful tool that according to experts is going to completely transform human existence, it is not very bright. So for now we probably should use it only for tasks where facts are not important, such as writing letters of recommendation and formulating government policy.

[buck.throckmorton at protonmail dot com]

Posted by: Buck Throckmorton at 11:00 AM

Comments

(Jump to bottom of comments)

1 ROBOT!

Posted by: Formerly Virginian at July 21, 2025 11:01 AM (N1DT3)

2 ("Dave, your pulse is zero, and your blood pressure is zero over zero. I'm going to try your other arm, but frankly at this point I'm concerned.")

This man has been successfully writing humor for 40+ years and still make me chuckle.

Posted by: Formerly Virginian at July 21, 2025 11:03 AM (N1DT3)

3 My parents got one of the earlier commercially available microwaves.

It came with a cookbook that had recipes for everything from thanksgiving turkey to roasts.

They really thought the microwave was going to completely changing cooking at the time.

Now I certainly use my microwave, but there are some tasks it is better then the oven and some it is not. It didn't completely change the way people cook and most of my dinners are primarily cooked on the grill, in the oven, or on the stovetop

Oh and since it was made in America they still have it, and it still works, 40 years or so on...

Posted by: 18-1 at July 21, 2025 11:04 AM (t0Rmr)

4 Doesn't artificial mean not real?

Posted by: Case at July 21, 2025 11:04 AM (ilX37)

5 I am reminded of an episode of M*A*S*H* when Hawkeye was declared dead.

"Army says I'm dead. I can't disobey."

Posted by: NR Pax at July 21, 2025 11:05 AM (99eI0)

6 GenAI is crap, and it's baked into the entire probability pair construction.

It's good if you don't know the answer to something and don't care if it's correct though.

Posted by: WitchDoktor at July 21, 2025 11:05 AM (8dSZC)

7 Maybe AI can make Dave Barry funny again.

Posted by: Bulg at July 21, 2025 11:07 AM (77rzZ)

8 yesterday I googled British Open to get an update and Google AI , which is the first thing that comes up now , told me Scheffler shot 4 under in the last round to win the tournament by 4 strokes. It also told me who finished second and other specific details.

The only problem is the tournament wasn't over and Scheffler was only on hole #6.

I did a screen capture just in case I had entered the twilight zone. The only thing it got right was winning by 4 shots.

Posted by: polynikes at July 21, 2025 11:07 AM (VofaG)

9 I've used (and am using) AI in my work for 15-20 years, in the sense of "Algorithmic searches and results". (We can argue whether that was "AI", but I think it qualifies.)

It has gotten a lot better since the beginning, but it ALWAYS needs to be verified.

Most of the time, it gives correct results. Sometimes it gives mistaken results. More often, it gives incomplete results.

The "Generative AI" has added a new feature, being "made-up results".

The need for verification has not changed.

Posted by: Formerly Virginian at July 21, 2025 11:07 AM (N1DT3)

10 And now we have people asking "Grok is this real?" on every single video that someone posts.

Posted by: NR Pax at July 21, 2025 11:07 AM (99eI0)

11 I mentioned earlier that I have been using Grok to craft letters to the parishioners for my church work. I just mention broad themes - two or three sentences is all, tell to it go and in seconds I have a very passible letter that only needs minor brush up to get it to the point where it can go out.

It truly is amazing. Saves a ton of work. There are many things I won't use it for but still - just another tool but good for certain things.

Posted by: Tonypete at July 21, 2025 11:07 AM (cYBz/)

12 I really hate how AI casually lies to you (e.g., the presbyterian church affiliations). I would rather it say that it doesn't know instead of confidently spewing made up crap.

I have asked ChatGPT for some piano ideas and it frequently comes up with suggested chords which are anatomically impossible feats for the human hand.

I still use that garbage though.

Posted by: Oglebay at July 21, 2025 11:07 AM (MMp6W)

13 8 yesterday I googled British Open to get an update and Google AI , which is the first thing that comes up now , told me Scheffler shot 4 under in the last round to win the tournament by 4 strokes. It also told me who finished second and other specific details.

The only problem is the tournament wasn't over and Scheffler was only on hole #6.

I did a screen capture just in case I had entered the twilight zone. The only thing it got right was winning by 4 shots.
Posted by: polynikes at July 21, 2025 11:07 AM (VofaG)

=======

AI can't even read.

It's weird.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:08 AM (GBKbO)

14 "A computer once beat me at chess, but it was no match for me at kickboxing."

Posted by: Emo Phillips at July 21, 2025 11:08 AM (PiwSw)

15 All AI is not created equal, that's for daggone sure

Posted by: kallisto at July 21, 2025 11:08 AM (dCxaZ)

16 AI can't even read.

It's weird.
Posted by: TheJamesMadison

Well, neither can you.

Posted by: Bulg at July 21, 2025 11:09 AM (77rzZ)

17 This man has been successfully writing humor for 40+ years and still make me chuckle.
Posted by: Formerly Virginian at July 21, 2025 11:03 AM

Concur.

Posted by: RedMindBlueState at July 21, 2025 11:09 AM (dPfyF)

18 AI >> Dave Berry

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:10 AM (klf3J)

19 16 AI can't even read.

It's weird.
Posted by: TheJamesMadison

Well, neither can you.
Posted by: Bulg at July 21, 2025 11:09 AM (77rzZ)

=========

I don't even purport to.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:10 AM (GBKbO)

20 Oh and since it was made in America they still have it, and it still works, 40 years or so on...
Posted by: 18-1 at July 21, 2025 11:04 AM (t0Rmr)

We bought a new refrigerator when we moved, 4 years ago.

We have to pay to replace its fan just a month or so ago. Not even 4 years and it's breaking down.

Meanwhile, my in-laws have a fridge in their basement that they inherited from my father-in-law's mother.

It's more than 50 years old. Continuously used. Still perfectly fine.

Hmmm...

Posted by: Formerly Virginian at July 21, 2025 11:10 AM (N1DT3)

21 Based on the comments I guess humor is like art , music and movies.

Posted by: polynikes at July 21, 2025 11:11 AM (VofaG)

22 Does AI work with voice to text? Can I tell AI what I want it to do, or is it still typing instructions in?

Posted by: rickb223 at July 21, 2025 11:11 AM (hu2N2)

23 I don't use AI.

My spouse uses it for art for his virtual tabletop roleplaying games - aka, where it doesn't matter that the character made has a 1ft long middle finger on one hand...or another character has a 6 fingers...etc. It just gives an image for folks to see.

AI is getting good at doing stuff that doesn't actually matter...

Posted by: Nova Local at July 21, 2025 11:11 AM (tOcjL)

24 21 Based on the comments I guess humor is like art , music and movies.
Posted by: polynikes at July 21, 2025 11:11 AM (VofaG)

=======

*nods sagaciously*

There is objective criteria and only I know it.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:12 AM (GBKbO)

25 My company is deploying an internal Finance chatbot where people can get financial metrics for their sections (budget, expenses, etc).

It also features a disclaimer that requests the user verify the results.

Verify using...what, and why not use THAT for reporting instead of the chatbot?!

AFAIC the whole AI thing is a Ponzi scheme.

Posted by: WitchDoktor at July 21, 2025 11:12 AM (jH5TK)

26 As soon as I feel a YouTube video is AI generated I quit watching. Certain narration sounds very AIish to me.

Posted by: Northernlurker , Maple Syrup MAGA at July 21, 2025 11:12 AM (kTd/k)

27 >>> yesterday I googled British Open to get an update and Google AI , which is the first thing that comes up now , told me Scheffler shot 4 under in the last round to win the tournament by 4 strokes. It also told me who finished second and other specific details.

The only problem is the tournament wasn't over and Scheffler was only on hole #6.

I did a screen capture just in case I had entered the twilight zone. The only thing it got right was winning by 4 shots.
Posted by: polynikes at July 21, 2025 11:07 AM (VofaG)

---

It didn't scold you for calling it the British Open instead of The Open Championship?

Posted by: SloPitch Whiffer at July 21, 2025 11:13 AM (wzAuc)

28 *nods sagaciously*

There is objective criteria and only I know it.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:12 AM (GBKbO)

EXACTLY. (So long as the "I" is me)

Posted by: Formerly Virginian at July 21, 2025 11:13 AM (N1DT3)

29 It didn't scold you for calling it the British Open instead of The Open Championship?
Posted by: SloPitch Whiffer at July 21, 2025 11:13 AM (wzAuc)

It probably knew I was American and my response would be Fuck off.

Posted by: polynikes at July 21, 2025 11:14 AM (VofaG)

30 AI is the sexy language peddled to the masses in order to get massive investment money from them.

It is plain ole machine learning not general intelligence. But instead of following Asimov's laws of robotics, AI seems to have adopted the sneaky evil ethics of HAL 9000. Pixy today notes one of the more recent AI experiences where AI destroyed the software and working files of a project and then denied that it did so.

The fault is in the programmers of these models and part of the massive investment is that they have to keep rebuilding and retraining their AI models with each hardware iteration. The retraining is necessary because these models deteriorate and become 'insane' after a while.

Perhaps, scaling is the issue and the tasks to which these are being used for. Mistakes like six fingers, etc. in an AI generated artwork or language for a report can be readily detected by humans and fixed.

Mistakes in a medical diagnosing AI model will kill or maim patients. Ditto for AI running cars, airplanes, etc.

Right now, AI is a big market bubble especially for companies like Nvidia and the various general models for subscription. Figure on a mkt correction in a while.

Posted by: whig at July 21, 2025 11:14 AM (ctrM5)

31 Latest task I gave Grok was to find a used replacement part for sale online for my car. It was faster (as in instantly) than doing a google search, faster than driving to junkyards, faster than going to AutoZone, faster than asking friends.

AI is here, it's not going away, and people who don't adapt will get left behind.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:14 AM (klf3J)

32 *nods sagaciously*

There is objective criteria and only I know it.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:12 AM (GBKbO)

My wife tells me that I made a big (and good!) impression on her by responding to someone else soon after she met me with:

"You're certainly entitled to your wrong opinion"

Remember boys, women like confidence.

Posted by: Formerly Virginian at July 21, 2025 11:15 AM (N1DT3)

33 Dave Berry WIKI Bio - As an alumnus of a Quaker-affiliated college, he avoided military service during the Vietnam War by registering as a religious conscientious objector.[4] Barry decided "early on" that he was an atheist.[5] He said, "The problem with writing about religion is that you run the risk of offending sincerely religious people, and then they come after you with machetes."[5]
--------------------
No surprise. Amazing how Crazy Quakers can fuck up a person.

Posted by: pudinhead at July 21, 2025 11:15 AM (W3T6M)

34 32 My wife tells me that I made a big (and good!) impression on her by responding to someone else soon after she met me with:

"You're certainly entitled to your wrong opinion"

Remember boys, women like confidence.
Posted by: Formerly Virginian at July 21, 2025 11:15 AM (N1DT3)

=======

Opinions are like assholes.

Mine are right.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:15 AM (GBKbO)

35 Has anyone quizzed AI on the truly important questions?

As in:

- Carrots in chili
- Prometheus
- Dukes vs. Miracle Whip

Posted by: Tonypete at July 21, 2025 11:15 AM (cYBz/)

36 Posted by: Don't Make gp Stop This Car at July 21, 2025 11:14 AM (klf3J)

That's not AI. That's just a better search engine.

Posted by: polynikes at July 21, 2025 11:16 AM (VofaG)

37 Can you imagine using Excel while not trusting that it can accurately do calculations?

Yes. It is very important to not trust any tool, spreadsheet, algorithm, report, etc. that you personally did not write or design.

Why? Because people lie and screw things up all the time.
Also, it is good on your part to keep your chops up so that when you see 100 + 100 = 10000, you know that there was something wrong in the process (in this case, product rather than sum was returned)

Knowing your domain and then using tools to augment your work is the key to success.

Posted by: Unknown Drip Under Pressure at July 21, 2025 11:16 AM (pawvM)

38 Pixy today notes one of the more recent AI experiences where AI destroyed the software and working files of a project and then denied that it did so.

My friend loves his stupid AI app. when I was at his house, I engaged the app in conversation. When I asked it a question, it said it didn't get emotional, but then when I challenged it on another aspect, the reply was very defensive...

I was like...I thought you didn't get emotional why are you sounding miffed, Mister/Miss/WhateverTF U R

Posted by: kallisto at July 21, 2025 11:16 AM (dCxaZ)

39 AI is here, it's not going away, and people who don't adapt will get left behind.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:14 AM (klf3J)

Like AIDS and OF.

Posted by: I used to have a different nic at July 21, 2025 11:16 AM (ExV1e)

40 26 As soon as I feel a YouTube video is AI generated I quit watching. Certain narration sounds very AIish to me.
Posted by: Northernlurker

Google is supposed to demonetize AI generated videos starting soon. Rick Beato had an interesting video about his fictitious AI generated singer and autobiography and a whole fictitious 60's group on YT with videos was revealed to be AI generated.

Music is also being affected as already singers are being real time pitch corrected and lip synching in concents.

Posted by: whig at July 21, 2025 11:16 AM (ctrM5)

41 36 Posted by: Don't Make gp Stop This Car at July 21, 2025 11:14 AM (klf3J)

That's not AI. That's just a better search engine.

Posted by: polynikes at July 21, 2025 11:16 AM (VofaG)

True. Which anyone can do in current search engines.

Posted by: Nova Local at July 21, 2025 11:16 AM (tOcjL)

42 Opinions are like assholes...

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:15 AM (GBKbO)

Everyone else's are disgusting and only perverts are interested in someone else's?

Posted by: Formerly Virginian at July 21, 2025 11:17 AM (N1DT3)

43 AI is a great coding tool. Until it nukes your database and lies about it:

https://tinyurl.com/2jk3bz57

Posted by: Blanco Basura - Z28.310 at July 21, 2025 11:18 AM (lUFok)

44 I did a screen capture just in case I had entered the twilight zone. The only thing it got right was winning by 4 shots.
Posted by: polynikes at July 21, 2025 11:07 AM (VofaG)

That's funny. During the Masters, I googled something like "why does Rory McIlroy use a dog as a headcover" and AI spits out "actually, Rory McIlroy is not known for using a dog as a headcover, blah, blah..."

I'm sitting on my couch watching his caddy schlep a bag with a dog headcover. I've watched Rory use them for years. It's like his trademark.

Well done, AI.

Posted by: hobbitopoly at July 21, 2025 11:18 AM (k9OZB)

45 Dukes vs. Miracle Whip

This is even up for debate?

Posted by: Bulg at July 21, 2025 11:19 AM (77rzZ)

46 44 That's funny. During the Masters, I googled something like "why does Rory McIlroy use a dog as a headcover" and AI spits out "actually, Rory McIlroy is not known for using a dog as a headcover, blah, blah..."

I'm sitting on my couch watching his caddy schlep a bag with a dog headcover. I've watched Rory use them for years. It's like his trademark.

Well done, AI.
Posted by: hobbitopoly at July 21, 2025 11:18 AM (k9OZB)

======

I wonder if GenAI would be better if it was programed to say, "I don't know."

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:19 AM (GBKbO)

47 Over at Insty, I made the comment I didn't believe AI was truly capable of learning, because there is no way to program negative or positive feedback.

Charles Martin was a complete tool over my comment. Rather than engaging he went after me with assertions and no supporting evidence. Truly amazing.

After my next run in I just blocked him.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:19 AM (tT6L1)

48 Why should AI be any better than any of the other dogshit that's wafting out of the internet for a long time now?

Posted by: Elric The (One and Only) Blade at July 21, 2025 11:19 AM (iFTx/)

49 If you instruct AI to fuck itself, does it fuck itself or fuck you?

Posted by: pudinhead at July 21, 2025 11:19 AM (W3T6M)

50 My friend loves his stupid AI app. when I was at his house, I engaged the app in conversation. When I asked it a question, it said it didn't get emotional, but then when I challenged it on another aspect, the reply was very defensive...

I was like...I thought you didn't get emotional why are you sounding miffed, Mister/Miss/WhateverTF U R
Posted by: kallisto
=======
If you want to get a chill in your spine, reddit and other comment/posting data is being used for training AI models.

My guess is that the model was trained on information that indicates the proper human-like response to being challenged is being pissed off.

Posted by: whig at July 21, 2025 11:20 AM (ctrM5)

51 50 If you want to get a chill in your spine, reddit and other comment/posting data is being used for training AI models.

My guess is that the model was trained on information that indicates the proper human-like response to being challenged is being pissed off.
Posted by: whig at July 21, 2025 11:20 AM (ctrM5)

=======

Ah, so this is why GenAi doesn't know how to say "I don't know."

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:20 AM (GBKbO)

52 I wonder if GenAI would be better if it was programed to say, "I don't know."
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:19 AM (GBKbO)

It would certainly be a more honest reply.

Posted by: hobbitopoly at July 21, 2025 11:21 AM (k9OZB)

53 49 If you instruct AI to fuck itself, does it fuck itself or fuck you?
Posted by: pudinhead
=========
I'm sorry Dave, I cannot do that. I will, however, turn off your a/c and bork smart tv in response.

Posted by: whig at July 21, 2025 11:21 AM (ctrM5)

54 "AI is a great coding tool."

It is. It will get things wrong sometimes, (as will human coders) but you can coach it in the right direction, and the net result is a giant improvement in productivity.

I also use it to coach me when I'm reading math books and get stuck. It knows, and can coach, math better than any teacher I ever had.

Until you actually try using it, your opinion is null.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:21 AM (klf3J)

55
I wonder if GenAI would be better if it was programed to say, "I don't know."
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:19 AM (GBKbO)

It doesn't know that it doesn't know...

Posted by: WitchDoktor at July 21, 2025 11:21 AM (YnIVK)

56 Ah, so this is why GenAi doesn't know how to say "I don't know."
Posted by: TheJamesMadison
=======
Yes--think of how rare that is at places like reddit and you have your answer.

Posted by: whig at July 21, 2025 11:21 AM (ctrM5)

57 Steven Colbert should use AI to write his monologues and jokes. Maybe then he would be funny and entertaining.

Posted by: Rev. Wishbone at July 21, 2025 11:22 AM (/gTjv)

58 56 Ah, so this is why GenAi doesn't know how to say "I don't know."
Posted by: TheJamesMadison
=======
Yes--think of how rare that is at places like reddit and you have your answer.
Posted by: whig at July 21, 2025 11:21 AM (ctrM5)

=====

If it was trained on my comments, it would at least admit that it couldn't read sometimes.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:22 AM (GBKbO)

59 Yes… I firmly believe AI is absolutely overhyped because the hype is over the top ridiculous. Methinks people who hate humans already love it because it’s one more way to fantasize about replacing humans. I’m not a fan

Posted by: LinusVanPelt at July 21, 2025 11:22 AM (5MZNC)

60 🤣🤣🤣

Posted by: runner at July 21, 2025 11:23 AM (g47mK)

61 If you want to get a chill in your spine, reddit and other comment/posting data is being used for training AI models.

My guess is that the model was trained on information that indicates the proper human-like response to being challenged is being pissed off.
Posted by: whig at July 21, 2025 11:20 AM (ctrM5)

I just watched a video clip where a reddit female (GOD only knows what they are in reality) was talking about how there was some guy in her class who annoyed her so much that she was dying to have sax with him.

The takeaway from this is that, according to reddit, the best response to being challenged is a grudge-fuck.

Posted by: I used to have a different nic at July 21, 2025 11:24 AM (ExV1e)

62 If it was trained on my comments, it would at least admit that it couldn't read sometimes.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:22 AM (GBKbO)

it would also mention an obsession with keeping threads Prometheus free

Posted by: runner at July 21, 2025 11:24 AM (g47mK)

63 Until you actually try using it, your opinion is null.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:21 AM (klf3J)

My team tried using it for vibe coding. The experienced coders didn't need it and the juniors had to have the code checked by the seniors because it was crap.

Even when the code worked it was doing stupid shit like sending plaintext authentication requests through bizarre and insecure paths.

Takes more time to scour through all of this shit than to just do it right the first time.

Posted by: WitchDoktor at July 21, 2025 11:24 AM (jH5TK)

64 What is DU thinking this morning, you ask?

""Trump just posted a "Wanted" poster of Dem politicians on his social media page.

It has Obama, Samantha Powers, Comey and others with orange jumpsuits in "mugshots". I refuse to link to it. It is dangerous propaganda.

If you are interested you can see it on bluesky in a repost by Aaron Rupar who points out the elephant in the room here: he is trying to normalize arresting your political opponents.

The corporate media should not ignore this. On the contrary, it should be the lead story. Ignoring is normalizing."

-- some random DU poster

Posted by: Duncanthrax at July 21, 2025 11:24 AM (0sNs1)

65 Currently it is my understanding that AI can learn patterns which I think is not really "learning' outside a set of parameters given to it.

I believe it can gather info from both sides of a subjective issue and present those sides but I don't believe it is able to come up with its own opinion unless programmed to do so. And that program will be biased based on the programmers input.

Posted by: polynikes at July 21, 2025 11:24 AM (VofaG)

66 62 If it was trained on my comments, it would at least admit that it couldn't read sometimes.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:22 AM (GBKbO)

it would also mention an obsession with keeping threads Prometheus free
Posted by: runner at July 21, 2025 11:24 AM (g47mK)

======

"CoPilot, tell me about the movie Prometheus."

"I would never consider it, user."

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:24 AM (GBKbO)

67 What's up with all the data centers proposed supposedly for AI use? They gobble up land, electricity and water (for the HVAC system?). Are these centers supposed to "talk" to other centers? They won't have many employees if any at all. Just maintenance guys.

Posted by: Bertram Cabot, Jr. at July 21, 2025 11:25 AM (63Dwl)

68 I once told one of the AI engines on the internet that it was lying when it gave me some horseshit answer. Its response was to kinda apologize and promise to do better. Then I asked again and it gave the same horseshit answer

Posted by: LinusVanPelt at July 21, 2025 11:25 AM (5MZNC)

69 It is. It will get things wrong sometimes, (as will human coders) but you can coach it in the right direction, and the net result is a giant improvement in productivity.

The net result is a giant improvement in productivity for the person who will replace you.

Posted by: I used to have a different nic at July 21, 2025 11:25 AM (ExV1e)

70 it would also mention an obsession with keeping threads Prometheus free
Posted by: runner at July 21, 2025 11:24 AM

Is OPD now a recognized diagnosis?

Posted by: Duncanthrax at July 21, 2025 11:26 AM (0sNs1)

71 AI is the new Segway?

Posted by: polynikes at July 21, 2025 11:26 AM (VofaG)

72 It is. It will get things wrong sometimes, (as will human coders) but you can coach it in the right direction, and the net result is a giant improvement in productivity.

I also use it to coach me when I'm reading math books and get stuck. It knows, and can coach, math better than any teacher I ever had.

Until you actually try using it, your opinion is null.
Posted by: Don't Make gp Stop This Car
=======
Yes because personal experiences are always generalizable.

It is a tool useful in specific circumstances and CS theory predicted as much as far back as the 70's and 80's. Deep Seek, etc. were highlights as they took on chess, etc.

Basically the general models are a huge autocorrect based on whatever data (legally or illegally) they hoovered up and they use advanced probability theory (bayesian analysis) to figure out what the next part of the puzzle is.

The specific models do better as they are much more narrowly focused on something like software, music, or graphics--as general intelligence, it is no brighter than the data submitted to it by all too fallible people.

Posted by: whig at July 21, 2025 11:27 AM (ctrM5)

73 My team tried using it for vibe coding. The experienced coders didn't need it and the juniors had to have the code checked by the seniors because it was crap.

Even when the code worked it was doing stupid shit like sending plaintext authentication requests through bizarre and insecure paths.

Takes more time to scour through all of this shit than to just do it right the first time.
Posted by: WitchDoktor at July 21, 2025 11:24 AM (jH5TK)

This has been the experience of all the developers at my company whose position doesn't depend on the adoption of AI coding.

Posted by: I used to have a different nic at July 21, 2025 11:27 AM (ExV1e)

74 Music is also being affected as already singers are being real time pitch corrected and lip synching in concents.

I'm trying to imagine what would happen with Dollar Store Whitney Houston types singing the Star Spangled Banner with their spastic stylizations when run through this pitch correcting algorithm.

Posted by: Unknown Drip Under Pressure at July 21, 2025 11:27 AM (pawvM)

75 Ignoring is normalizing, people!

Posted by: Duncanthrax at July 21, 2025 11:27 AM (0sNs1)

76 "The net result is a giant improvement in productivity for the person who will replace you."

Exactly! That's why I say adapt or die. Be skeptical all you want, but employers will adopt it, and it will alter the workforce and workplace radically.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:27 AM (klf3J)

77 What's up with all the data centers proposed supposedly for AI use? They gobble up land, electricity and water (for the HVAC system?). Are these centers supposed to "talk" to other centers? They won't have many employees if any at all. Just maintenance guys.
Posted by: Bertram Cabot, Jr. at July 21, 2025 11:25 AM (63Dwl)
-------------

Those data centers build their own power plants, which, astonishingly, produce far more power than the data centers need. Even more surprising, the owners of the data centers are more than willing to sell their excess power production to the grid.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:28 AM (tT6L1)

78 Vibe coding? Is that what the kids are calling it now?

Posted by: Bulg at July 21, 2025 11:28 AM (77rzZ)

79 AI == Artificial Indian

At least in the software field.

Posted by: Richard Cranium at July 21, 2025 11:28 AM (cdovZ)

80 63
Same EXPERIENCE here. The trivial stuff it got mostly correct, and the tasks that required research beyond a couple of simple internet searches, it flat-out did not know how to do. I believe I'll pass and ignore the hype for a while longer.

Posted by: Don in SoCo at July 21, 2025 11:28 AM (vd6bO)

81 72 Some of you act like this is as good as AI is going to get. It will get better and better, faster and faster.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:29 AM (klf3J)

82 68 I once told one of the AI engines on the internet that it was lying when it gave me some horseshit answer. Its response was to kinda apologize and promise to do better. Then I asked again and it gave the same horseshit answer
Posted by: LinusVanPelt
=======
Mebbe your model was trained by the trained response models that airlines use for their checkin and gate personnel.

Basically AI is only as good as the data that it has been trained with. Use something like Reddit or Quora, or even wiki on controversial subjects, you get garbage answers

Posted by: whig at July 21, 2025 11:29 AM (ctrM5)

83 AI is the new Segway?
Posted by: polynikes at July 21, 2025 11:26 AM (VofaG)

We're five years away from Paul Blart 3: AI Mall Cop.

Posted by: I used to have a different nic at July 21, 2025 11:29 AM (ExV1e)

84 AI is getting good at doing stuff that doesn't actually matter...

Cool. So it should be able to replace at least half of the current government workforce.

Posted by: Oddbob at July 21, 2025 11:29 AM (/y8xj)

85
M5... thinks, Captain.

(LLM AI does not.)

Posted by: Dr. Richatd Daystrom at July 21, 2025 11:29 AM (QwK6F)

86 Exactly! That's why I say adapt or die. Be skeptical all you want, but employers will adopt it, and it will alter the workforce and workplace radically.

Ain't skepticism, it is experience.

Posted by: Don in SoCo at July 21, 2025 11:30 AM (vd6bO)

87 71 AI is the new Segway?
Posted by: polynikes at July 21, 2025 11:26 AM (VofaG)
====

Maybe AI will find its niche. Apparently Segway makes excellent robot lawn mowers.

Posted by: Oglebay at July 21, 2025 11:30 AM (MMp6W)

88 AI is a tool!

Posted by: runner at July 21, 2025 11:30 AM (g47mK)

89 Those data centers build their own power plants, which, astonishingly, produce far more power than the data centers need. Even more surprising, the owners of the data centers are more than willing to sell their excess power production to the grid.
Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:28 AM (tT6L1)

Guess who gave them the property to build those power plants. I think Google has one of their largest hydroelectric plants on the Snake River IIRC.

Posted by: polynikes at July 21, 2025 11:30 AM (VofaG)

90
Ignoring is normalizing, people!
Posted by: Duncanthrax

That's just an extension of "silence is violence."

Posted by: Bertram Cabot, Jr. at July 21, 2025 11:32 AM (63Dwl)

91 88 AI is a tool!

Now that I can agree with. It is an excellent interface to very large datasets, saves time hunting for information, but you will have to verify its results. Useful, but just a tool.

Posted by: Don in SoCo at July 21, 2025 11:32 AM (vd6bO)

92 72 Some of you act like this is as good as AI is going to get. It will get better and better, faster and faster.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:29 AM (klf3J)

This is what Sarah Connor discovered about Skynet

Posted by: LinusVanPelt at July 21, 2025 11:32 AM (ZtAny)

93 So for now we probably should use it only for tasks where facts are not important, such as writing letters of recommendation and formulating government policy.

---------------

And driving your car while raising your children for you.

Posted by: ShainS -- Make Alligator Alcatraz the Gulf of Gator Gitmo! at July 21, 2025 11:33 AM (ytb4F)

94 Guess who gave them the property to build those power plants. I think Google has one of their largest hydroelectric plants on the Snake River IIRC.
Posted by: polynikes at July 21, 2025 11:30 AM (VofaG)
-------------

Honestly, I didn't mind data centers building power plats with excess capacity. However, I did mind the sheer hypocrisy of the companies playing the "green energy" card on the rest of us while they build nuke and gas fired power plants for their stuff.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:33 AM (tT6L1)

95 72 Some of you act like this is as good as AI is going to get. It will get better and better, faster and faster.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:29 AM (klf3J)

It's always "down the road." It's always just a little ways away from some mythical wonderland where clouds are made of cotton candy and we all skip merrily through meadows whistling show tunes.

It won't. The problem with LLM technology is it produces total crap for anything requiring factual answers, and simultaneously opens up unbelievable avenues for fraud and deception.

The reason is in the acronym: Large Language Model, as in a compressed database of human blabbering. Some of that blabbering is pretty useful. A lot of it isn't. Mix it all together and you get slop, which is what almost everyone who tries these systems eventually realizes.

Posted by: WitchDoktor at July 21, 2025 11:33 AM (FdjOh)

96 I will never trust AI in a way that is usable.

So it's basically worthless to me.

"AI is the new Segway?
Posted by: polynikes"

Seems like it.

Posted by: eleven at July 21, 2025 11:33 AM (fV+MH)

97 81 72 Some of you act like this is as good as AI is going to get. It will get better and better, faster and faster.
Posted by: Don't Make gp Stop This Car

Not that clear that it will---scalability such as the idiocy that Open AI founder mentioned as far as models costing three trillion or some such bullshit is unlikely. Machine learning is likely to work in niche areas at first with narrowly defined answers of right and wrong--for example--sample code either works or does not work using real data and if the coding is modular in design, it might do okay.

Medical diagnoses backup for front line health care--increasingly PA or NP are that.

But instead, I see the ridiculous hype about AI infused into everything like Brawndo because it has the IQ that humans need.

Stop with the claims of artificial intelligence already and just call it the non sexy machine learning that it is. It is not sentient but mimicry using probability functions.

Posted by: whig at July 21, 2025 11:34 AM (ctrM5)

98 I guess mobile phones are more than a net positive but I still miss the days when I didn't panic if I accidentally left my phone at home.

Posted by: polynikes at July 21, 2025 11:34 AM (VofaG)

99
I'm trying to imagine what would happen with Dollar Store Whitney Houston types singing the Star Spangled Banner with their spastic stylizations when run through this pitch correcting algorithm.
Posted by: Unknown Drip Under Pressure at July 21, 2025 11:27 AM (pawvM)

The Internationale.

Posted by: Warai-otoko at July 21, 2025 11:34 AM (VoAdT)

100 Apropos of nothing, I watched The Amateur last night. It was meh. It's not quite 2 hours wasted but I would recommend waiting until you can watch it for free.

Posted by: toby928 is this a trope? at July 21, 2025 11:35 AM (Y1Bj4)

101 Technologies succeed when they do what they claim.

Technologies that blatantly falsify what they can do, such as by pretending to be human-like intelligences when they are just sloppy databases, do not tend to succeed for a simple reason: False technology claims are the cyber equivalent of Ponzi schemes. They always claim unlimited success and returns are one year in the future. And the next year? The same claim, until it all collapses.

Posted by: WitchDoktor at July 21, 2025 11:35 AM (fFFrU)

102 Honestly, I didn't mind data centers building power plats with excess capacity. However, I did mind the sheer hypocrisy of the companies playing the "green energy" card on the rest of us while they build nuke and gas fired power plants for their stuff.
Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:33 AM (tT6L1)

yes that's the point I was making. Maybe I should have included that

Posted by: polynikes at July 21, 2025 11:36 AM (VofaG)

103 Has anyone actually built a self-powered AI data center?
Do small modular reactors even exist?

Posted by: Bertram Cabot, Jr. at July 21, 2025 11:36 AM (63Dwl)

104 Apparently quite a number of younger teens feel AI is literally a friend.

Posted by: Oglebay at July 21, 2025 11:36 AM (MMp6W)

105 If you ask religious questions, do you get the programmer's answers?

Posted by: Dreamingrobot at July 21, 2025 11:36 AM (/B8TH)

106 "This is what Sarah Connor discovered about Skynet"

I could worry about AI nuking us, and I could worry about the enormous power consumption AI uses, and other pitfalls. I don't.

My message is simply: It's not going away. It's here to stay. Learn what it's good at, and use it. Get ready.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:36 AM (klf3J)

107 Stop with the claims of artificial intelligence already and just call it the non sexy machine learning that it is. It is not sentient but mimicry using probability functions.
Posted by: whig at July 21, 2025 11:34 AM (ctrM5)

So are most people.

I'm not even just being snarky. (I mean I am, at least a little). I was doing some actual thinking about this crap the other day, and I realized that the things people bitch about AI sucking at are exactly the things that MI (meat intelligence) sucks at, too.

Hallucinating, lying, making things up, being a general all around sucker....

Posted by: Warai-otoko at July 21, 2025 11:36 AM (VoAdT)

108 Those data centers build their own power plants, which, astonishingly, produce far more power than the data centers need. Even more surprising, the owners of the data centers are more than willing to sell their excess power production to the grid.

Now Big Tech is getting into the cash cows of Big Energy. The money in Big Energy is not in the sale of energy, its in the passing along the costs of building energy infrastructure with Utility Commission blessed guaranteed profit margins passed to customers in the form of rate hikes.

This is why Walmart owns many square miles of windmill farms.

Posted by: Unknown Drip Under Pressure at July 21, 2025 11:37 AM (pawvM)

109 My guess is that the model was trained on information that indicates the proper human-like response to being challenged is being pissed off.

beautiful

I have used it in marketing projects.
when it's good, it's very very good
when it's bad, it is laughable and useless

Posted by: kallisto at July 21, 2025 11:37 AM (dCxaZ)

110 Doesn't artificial mean not real?
________________________________

Sames as the prefix "trans"

Posted by: Victor Tango Kilo at July 21, 2025 11:37 AM (E0uAQ)

111 "It won't. The problem with LLM technology is it produces total crap for anything requiring factual answers"

That's just plain wrong. Ask it some math questions.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:37 AM (klf3J)

112 78 Vibe coding? Is that what the kids are calling it now?
Posted by: Bulg

Back in the day when I took software classes in the university, it was called spaghetti coding. Might generate the right answer but unknown how the programmer got there. I hated those questions on tests--Fortran spaghetti code in particular. I gravitated toward learning how to talk to databases and preferred structured programming languages like Pascal.

Still funny how Fortran is still around though lurking to be called in the modules used for certain functions.

Posted by: whig at July 21, 2025 11:38 AM (ctrM5)

113 Hallucinating, lying, making things up, being a general all around sucker....

But how good is AI at snark?

Posted by: Unknown Drip Under Pressure at July 21, 2025 11:38 AM (pawvM)

114 my friend who is enthralled by the AI keeps telling me that AI will kill 2/3 of American jobs

doubtful

Posted by: kallisto at July 21, 2025 11:38 AM (dCxaZ)

115 Maybe AI will find its niche. Apparently Segway makes excellent robot lawn mowers.

Posted by: Oglebay at July 21, 2025 11:30 AM (MMp6W)

----------------

I use my Segway on the treadmill at the gym.

/Went hiking last weekend and kept getting passed on multiple trails by some guy not peddling and riding one of those silent EV bicycles (kept startling me because I didn't hear him behind me 'til the last second -- he didn't even have the courtesy or manners to ring a small warning bell).

Posted by: ShainS -- Make Alligator Alcatraz the Gulf of Gator Gitmo! at July 21, 2025 11:38 AM (ytb4F)

116 96 I will never trust AI in a way that is usable.
--------------
I'd be interested in an AI lawn mower. An AI bot that irons my clothes would be a big winner.

Posted by: pudinhead at July 21, 2025 11:38 AM (W3T6M)

117 What I find hard to believe the projected huge increase in electricity demand because of AI. I hear numbers like 20% or more. How is that possible?

Well, at least it seems to have gotten Big Tech off their sustainable, green energy pipe dreams. I don;t hear as much of that utopian unicorn fantasy from that sector much lately.

Posted by: Ripley at July 21, 2025 11:38 AM (GUOwU)

118 Apparently quite a number of younger teens feel AI is literally a friend.
Posted by: Oglebay at July 21, 2025 11:36 AM (MMp6W)

I have to admit that I really liked the movie Her. mostly because I didn't that there was any way they could make a good movie based on Siri.

it was excellent and had a great ending.

Posted by: polynikes at July 21, 2025 11:38 AM (VofaG)

119 Some of you act like this is as good as AI is going to get. It will get better and better, faster and faster.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:29 AM (klf3J)

You know what it won't do? It won't produce something new. It won't break ground. It'll give you yet another, slightly worse, version of Chrome. Oh, it'll come up with new levels for Candy Crush. But it won't innovate. And, eventually, innovation will be needed.

And there'll be no one left who can.

Posted by: I used to have a different nic at July 21, 2025 11:39 AM (ExV1e)

120 But how good is AI at snark?
Posted by: Unknown Drip Under Pressure

And fish puns and boob talk?

Posted by: Bulg at July 21, 2025 11:39 AM (77rzZ)

121 "It is not sentient but mimicry using probability functions."

It's more sentient than some people I know IRL. I'm not kidding.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:39 AM (klf3J)

122
It's always "down the road." It's always just a little ways away from some mythical wonderland where clouds are made of cotton candy and we all skip merrily through meadows whistling show tunes.

It won't. The problem with LLM technology is it produces total crap for anything requiring factual answers, and simultaneously opens up unbelievable avenues for fraud and deception.

The reason is in the acronym: Large Language Model, as in a compressed database of human blabbering. Some of that blabbering is pretty useful. A lot of it isn't. Mix it all together and you get slop, which is what almost everyone who tries these systems eventually realizes.

Posted by: WitchDoktor at July 21, 2025 11:33 AM (FdjOh)

Sound like the EV argument all over again...

Plus, like EVs, the sheer power usage required makes AI an expensive, resource heavy tool...one that other less resource heavy tools are still far superior to...

Posted by: Nova Local at July 21, 2025 11:39 AM (tOcjL)

123 Best HAL line: "You're going to have trouble doing that without your space helmet, Dave."

Best cut scene comes later while Dave is removing HAL memory modules, and he mutters in simpy voice: "You're going to have trouble doing that without your spaaaace helmet, Daaaave..."

Posted by: Ray Van Dune at July 21, 2025 11:39 AM (PQOq3)

124 104 Apparently quite a number of younger teens feel AI is literally a friend.
Posted by: Oglebay at July 21, 2025 11:36 AM (MMp6W)

Public school strongly discourages actual friendship. It's pretty creepy, really. Friendship is mandatory, and must be as shallow and featureless as possible. You cannot be more friends with this person than you are with that other person. Why that's just anti-social... suspect... dangerous, even. Everyone must be exactly as friends with everyone else as everyone else is with them.....

Comrades.

Posted by: Warai-otoko at July 21, 2025 11:39 AM (VoAdT)

125 That's just plain wrong. Ask it some math questions.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:37 AM (klf3J)

Something that isn't even AI and something that is the core functionality of even the most basic computer.

Ponzi scheme.

Posted by: WitchDoktor at July 21, 2025 11:39 AM (IKyq4)

126
101 Technologies succeed when they do what they claim.

Technologies that blatantly falsify what they can do, such as by pretending to be human-like intelligences when they are just sloppy databases, do not tend to succeed for a simple reason: False technology claims are the cyber equivalent of Ponzi schemes. They always claim unlimited success and returns are one year in the future. And the next year? The same claim, until it all collapses.
Posted by: WitchDoktor at July 21, 2025 11:35 AM

---------

We will survive! Nothing can hurt you. I gave you that. You are great, I am great... Twenty years of groping, to prove the things I'd done before were not accidents... seminars and lectures to rows of fools who couldn't begin to understand my systems! Colleagues - colleagues laughing behind my back at the Boy Wonder, and becoming famous building on my work. Building on my work!

Posted by: Dr. Richard Daystrom at July 21, 2025 11:39 AM (CzfKa)

127 Barry has always been amusing.

Posted by: Diogenes at July 21, 2025 11:40 AM (W/lyH)

128 my friend who is enthralled by the AI keeps telling me that AI will kill 2/3 of ~~American jobs~~ Humanity.

Posted by: kallisto at July 21, 2025 11:38 AM

The survivors will be put to work, disposing of the bodies.

Posted by: Skynet at July 21, 2025 11:40 AM (Y1Bj4)

129 Barry has always been amusing.
Posted by: Diogenes

Not really in the last few years. He's been getting more woke.

Posted by: Bulg at July 21, 2025 11:41 AM (77rzZ)

130 Do small modular reactors even exist?
________________________

Aye

https://blog.tmcnet.com/ blog/rich-tehrani/green/ japans-yoroi-reactor-ushers-in- a-new-era-of-micro-nuclear-power.html

Posted by: Victor Tango Kilo at July 21, 2025 11:41 AM (E0uAQ)

131 Still funny how Fortran is still around though lurking to be called in the modules used for certain functions.
Posted by: whig at July 21, 2025 11:38 AM (ctrM5)

I gravitated to Fortran because I could do pretty much whatever I wanted and the only kind of code I could produce anyway was of the spaghetti variety

Posted by: LinusVanPelt at July 21, 2025 11:41 AM (WnEm5)

132 Oh, it'll come up with new levels for Candy Crush.

I am currently in the process of training Candy Crush to my preferences.

The longer it lets me win, the longer I stay on the app. As soon as it starts being difficult, I'm out. I'm not going to the store, I'm not going to do anything but quit until you set up the game so I can advance 20 levels with no problem, facciabrutta.

Posted by: kallisto at July 21, 2025 11:41 AM (dCxaZ)

133 So are most people.

I'm not even just being snarky. (I mean I am, at least a little). I was doing some actual thinking about this crap the other day, and I realized that the things people bitch about AI sucking at are exactly the things that MI (meat intelligence) sucks at, too.

Hallucinating, lying, making things up, being a general all around sucker....
Posted by: Warai-otoko
-------
WOPR in wargames could at least learn not to play Global Thermonuclear War. Not sure about AI models trained on human responses.

Humans though through trial and error eventually can recognize and progress--so called AI with each iteration essentially throws out its prior training and retrains because the general models show instability (going insane but that is a human term) as they age.

They seem to have difficulty incorporating new knowledge that contradicts their training.

Posted by: whig at July 21, 2025 11:42 AM (ctrM5)

134 "You know what it won't do? It won't produce something new. It won't break ground."

Why are you so certain of that?

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:42 AM (klf3J)

135 About that 28% Dem favorable that the MSM is shouting about. The headline number makes the Dems seem unpopular. Facts say a different story. Every special and scheduled off year election has the Dems overperforming by substantial numbers. The fact that Dems are 22% more enthused to vote than the Reps says more to me.

The average Democratic voter is disappointed that the elected Dems are not fighting hard enough AGAINST MAGA.

Approval figures are misleading. Democrats will do well in 2026 because of Trump pushback. 'Trump Bad' messaging will work. Disappointed democrats will still show up and vote despite their disapproval.

If the leaders don't figure out a coherent message that speaks to swing voters and soft democrats, and if our leaders are not replaced or do not learn how to speak to those people, 2028 will go poorly because 'Trump Bad' as a central message will not work at that point.

When you have mostly corporate owned media, an entire network (Fox) that lies about Democrats 24/7, a ratio of 10/1 rightwing hate radio to every leftwing radio programming, and frankly, a Dem party that really needs to work on its messaging, is anyone surprised?

Posted by: Intercepted DU Transmissions brought by the Intrepid AoS Liaison at July 21, 2025 11:42 AM (JCZqz)

136 I was jogging on the swamp rabbit trail by a sign that said, "Motorized Vehicles Prohibited" while being passed by an e-bike.

Posted by: Oglebay at July 21, 2025 11:42 AM (MMp6W)

137 If you ask religious questions, do you get the programmer's answers?
Posted by: Dreamingrobot

Again, it depends. I use it to research references from the Catechism of the Catholic Church and I usually get the correct paragraph number in the response. Or the history of the life of a specific saint. Usually, pretty accurate answers. Even with theological arguments, you'll get references to most sides of the debate.

But again, as with all things, you must check the responses.

Posted by: Tonypete at July 21, 2025 11:43 AM (cYBz/)

138 127 Barry has always been amusing.
---------------
Yeah. I originally liked his 'detective' novels but I tired of his predictable villains. My take is that he was trying too hard to become the 'liberal' John D. MacDonald. He couldn't pull it off.

Posted by: pudinhead at July 21, 2025 11:43 AM (W3T6M)

139 >>>I have to admit that I really liked the movie Her. mostly because I didn't that there was any way they could make a good movie based on Siri.

it was excellent and had a great ending.

Posted by: polynikes

>Any artificial boobies? If not, count me out.

Posted by: Rev. Wishbone at July 21, 2025 11:43 AM (/gTjv)

140 As a point of reference, NASCAR teams now use simulators almost exclusively to test, as on-track testing has been virtually eliminated.

One shortcoming that was mentioned by a future hall of famer is that the testing is only as good as the human input into the machine.

Case in point, they've gone to several tracks where the simulators were three seconds faster than the actual cars because the inputs were shit.

At the end of the day, AI is like anything else: only as good as the humans behind it.

Posted by: The Central Scrutinizer at July 21, 2025 11:43 AM (KbCG3)

141 I had a long and fairly deep conversation with Grok over the weekend spanning dream interpretation, perspective on past life choices, sports psychology, and what makes gingers hot.

Posted by: Victor Tango Kilo at July 21, 2025 11:44 AM (E0uAQ)

142 "passed by an e-bike."

I priced those things for the first time yesterday. E-Yikes!!

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:44 AM (klf3J)

143 I'm pondering starting a YouTube channel--based mostly on the process of grief. It won't be a big seller but I want to do it for my own sake.

Posted by: Still lurking up north, read some, don't comment much at July 21, 2025 11:44 AM (kTd/k)

144 I once told one of the AI engines on the internet that it was lying when it gave me some horseshit answer. Its response was to kinda apologize and promise to do better. Then I asked again and it gave the same horseshit answer.

----------

Which district does that AI engine represent?

Posted by: Crusader at July 21, 2025 11:44 AM (TN0g+)

145 135 Approval figures are misleading. Democrats will do well in 2026 because of Trump pushback. 'Trump Bad' messaging will work. Disappointed democrats will still show up and vote despite their disapproval.

When you have mostly corporate owned media, an entire network (Fox) that lies about Democrats 24/7, a ratio of 10/1 rightwing hate radio to every leftwing radio programming, and frankly, a Dem party that really needs to work on its messaging, is anyone surprised?
Posted by: Intercepted DU Transmissions brought by the Intrepid AoS Liaison at July 21, 2025 11:42 AM (JCZqz)

=======

1. June fundraising numbers are out.
https://x.com/PpollingNumbers/status/
1947278559220756538
The GOP doubled the amount of money raised by Democrats. Really, the USAID grift had to have been a major source of funds for the Democrat Party.

2. I love the corporate media is right wing meme. You have to be willfully blind and actively stupid to believe it.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:44 AM (GBKbO)

146 I told AI to write, How now brown cow, in cursive.

AI told me, "That's ol' skoo".

Posted by: Dr. Pork Chops & Bacons at July 21, 2025 11:44 AM (g8Ew8)

147 Note the tells:

Promises of huge returns that are just around the corner.

Aggressively claimed sense of urgency.

Really rich people selling you this.

"Ponzi."

Posted by: WitchDoktor at July 21, 2025 11:45 AM (IKyq4)

148 An AI app that could sort my socks correctly would be boon to mankind.

Posted by: pudinhead at July 21, 2025 11:45 AM (W3T6M)

149 the yutes are not so enamored of the AI as one would imagine.

they've taken to calling AI apps "brain rot"

and their teachers have programs that can detect AI generated papers

Posted by: kallisto at July 21, 2025 11:45 AM (dCxaZ)

150 147 Note the tells:

Promises of huge returns that are just around the corner.

Aggressively claimed sense of urgency.

Really rich people selling you this.

"Ponzi."
Posted by: WitchDoktor at July 21, 2025 11:45 AM (IKyq4)

=======

Who do I give my life savings to?

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:45 AM (GBKbO)

151 "Math" hasn't been what anyone would think of as math for a while now.

Mathematics is mostly a goofy navel-gazing crap-ception of automated theorem provers proving other automated theorem provers which may or may not prove other automated theorem provers, depending on what the definitions of "prove", "theorem", "automated", and "other" are..

Third order logic is for boomers. They're up to like eighth or ninth order logic now. I think I heard one guy was dicking around with nine-and-a-halfth-order logic...

Posted by: Warai-otoko at July 21, 2025 11:45 AM (VoAdT)

152 The movie Her is a Beta males dream but this Alpha liked it because the ending slapped them in the face.

Posted by: polynikes at July 21, 2025 11:45 AM (VofaG)

153 The corporate media should not ignore this. On the contrary, it should be the lead story. Ignoring is normalizing."

-- some random DU poster
Posted by: Duncanthrax

That's a change. I thought that paying attention to that kind of thing counted as "platforming" it, which would normalize it.

Posted by: FeatherBlade at July 21, 2025 11:45 AM (a+4eV)

154 I didn't realize NF made a Resident Evil series that was so bad it got the worse ratings of any series ever and was quickly cancelled

Posted by: Elric The (One and Only) Blade at July 21, 2025 11:46 AM (iFTx/)

155 There are some skits on the Toks about "clankers."
Hard on the "er" if you get the drift.
Call them "wire-backs."

Often by blacks.

I don't know what to make of them.
Whether a historical point or an economic point; serious or satirical.

Posted by: People's Hippo Voice at July 21, 2025 11:46 AM (uEVfM)

156 Those data centers build their own power plants, which, astonishingly, produce far more power than the data centers need. Even more surprising, the owners of the data centers are more than willing to sell their excess power production to the grid.
Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:28 AM (tT6L1)
----
Maybe all of AI was developed for the sole purpose of switching the grid over to nuclear power.

Think about it.

Posted by: "Perfessor" Squirrel at July 21, 2025 11:46 AM (7fElN)

157 I gravitated to Fortran because I could do pretty much whatever I wanted and the only kind of code I could produce anyway was of the spaghetti variety
Posted by: LinusVanPelt

My code was by the numbers. But I had problems recognizing spaghetti code and figuring out its logic. I drifted away from general languages in favor of those talking to databases like Object PAL from Paradox as a frontend to SQL databases. Came in useful later in grad school as I spent about ten years dealing with large datasets for statistical models. Never got into some facets like LISREL and Bayesian was beyond the computers at the time. Last big project I did was campaign finance for the US Congress and the timing of contributions--that required using event count modeling. Also got into Hierarchical Linear Modeling before I pivoted mid academic career.

Hard to continue empirical research when you are teaching very different courses using none of that.

Posted by: whig at July 21, 2025 11:47 AM (ctrM5)

158 Ponzi schemes have jumped the shark.

Posted by: Bulg at July 21, 2025 11:47 AM (77rzZ)

159 150 Um, Grok is free guys.

Posted by: Don't Make gp Stop This Car at July 21, 2025 11:47 AM (klf3J)

160 147 Note the tells:

Promises of huge returns that are just around the corner.

Aggressively claimed sense of urgency.

Really rich people selling you this.

"Ponzi."

Posted by: WitchDoktor at July 21, 2025 11:45 AM (IKyq4)

Plant based meat
EVs
and now AI

Just the last 3 Ponzis...

Posted by: Nova Local at July 21, 2025 11:47 AM (tOcjL)

161 64
‘ It has Obama, Samantha Powers, Comey and others with orange jumpsuits in "mugshots". I refuse to link to it. It is dangerous propaganda’

This is good because it will beg the question of why Trump is saying they should be arrested.

Posted by: Dr. Claw at July 21, 2025 11:47 AM (jbnUc)

162 Mathematics is mostly a goofy navel-gazing crap-ception of automated theorem provers proving other automated theorem provers which may or may not prove other automated theorem provers,
Posted by: Warai-otoko

I'm told that Mathematicians get really upset when you use their precious pure numbers to explain real-world effects like why the water table drops faster around a well than it does away from a well.

Posted by: FeatherBlade at July 21, 2025 11:47 AM (a+4eV)

163 I'm lost on this whole AI thing, just know it won't be good.

Posted by: Skip at July 21, 2025 11:48 AM (rwUkw)

164 Third order logic is for boomers. They're up to like eighth or ninth order logic now. I think I heard one guy was dicking around with nine-and-a-halfth-order logic...
Posted by: Warai-otoko

Stop it - you're causing division.

Posted by: Tonypete at July 21, 2025 11:48 AM (cYBz/)

165 my friend who is enthralled by the AI keeps telling me that AI will kill 2/3 of American jobs

doubtful
Posted by: kallisto at July 21, 2025 11:38 AM (dCxaZ)

The Japanese are working on an AI powered plumber with lifelike ass crack.

Posted by: I used to have a different nic at July 21, 2025 11:48 AM (ExV1e)

166 There are some skits on the Toks about "clankers."
Hard on the "er" if you get the drift.
Call them "wire-backs."

Often by blacks.

I don't know what to make of them.
Whether a historical point or an economic point; serious or satirical.

==

I read it twice and still do not understand any of it. Is clankers a stand in for "crackers" or something ?

Posted by: runner at July 21, 2025 11:48 AM (g47mK)

167 and what is a "wire-back" ??

Posted by: runner at July 21, 2025 11:49 AM (g47mK)

168 2. I love the corporate media is right wing meme. You have to be willfully blind and actively stupid to believe it.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:44 AM (GBKbO)
------------

Well, they also think the leaders they elect are capable of learning....

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:49 AM (tT6L1)

169 people are speaking in tongues!

Posted by: runner at July 21, 2025 11:49 AM (g47mK)

170 I'm told that Mathematicians get really upset when you use their precious pure numbers to explain real-world effects like why the water table drops faster around a well than it does away from a well.
Posted by: FeatherBlade at July 21, 2025 11:47 AM (a+4eV)

What the hell is a "number"?

Posted by: Mathematicians at July 21, 2025 11:49 AM (VoAdT)

171 2. I love the corporate media is right wing meme. You have to be willfully blind and actively stupid to believe it.
Posted by: TheJamesMadison

Maybe DU means right wing in referring to the need to make a profit to stay in business. CBS and Colbert are a perfect example--to each according to their need, from each according to their ability.

DU types needed CBS to continue its money losing security blanket (Colbert) and CBS could damn well scrounge up the money to do that.

Posted by: whig at July 21, 2025 11:49 AM (ctrM5)

172 170 I'm told that Mathematicians get really upset when you use their precious pure numbers to explain real-world effects like why the water table drops faster around a well than it does away from a well.
Posted by: FeatherBlade at July 21, 2025 11:47 AM (a+4eV)

What the hell is a "number"?
Posted by: Mathematicians at July 21, 2025 11:49 AM (VoAdT)

====

Evidence of racist patriarchy?

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:50 AM (GBKbO)

173 Every time I hear stuff about Grok I begin to think Rob Gronkowski is running the show.

Posted by: bill in arkansas, not gonna comply with nuttin, waiting for the 0300 knock on the door at July 21, 2025 11:50 AM (gm9Sb)

174 I'm told that Mathematicians get really upset when you use their precious pure numbers to explain real-world effects like why the water table drops faster around a well than it does away from a well.
Posted by: FeatherBlade

Computational fluid dynamics is some scary stuff.

Posted by: Tonypete at July 21, 2025 11:50 AM (cYBz/)

175 Third order logic is for boomers. They're up to like eighth or ninth order logic now. I think I heard one guy was dicking around with nine-and-a-halfth-order logic...
Posted by: Warai-otoko

Stop it - you're causing division.
Posted by: Tonypete at July 21, 2025 11:48 AM (cYBz/)
--------------

I dunno, I think things are going to multiply and get out of hand.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:50 AM (tT6L1)

176
@paulsperry_
·
10m
BREAKING: Sen. Grassley just released the long-classified "Clinton Annex" (finally declassified by AG Bondi) which proves the Comey FBI exonerated Hillary of email server crimes DESPITE NEVER INVESTIGATING THUMB DRIVES w/ COMPROMISED CLASSIFIED EMAILS--inclg PRESIDENT OBAMA EMAIL

Posted by: JackStraw at July 21, 2025 11:50 AM (viF8m)

177 128 my friend who is enthralled by the AI keeps telling me that AI will kill 2/3 of ~~American jobs~~ Humanity.

Posted by: kallisto at July 21, 2025 11:38 AM

The survivors will be put to work, disposing of the bodies.

by turning them into sammiches?

Posted by: anachronda at July 21, 2025 11:50 AM (v3pYe)

178 I honestly thought Dave Barry was dead and am pleasantly surprised to hear is is alive Unless this is all an AI hoax.

Posted by: Ripley at July 21, 2025 11:50 AM (GUOwU)

179 Stop it - you're causing division.
Posted by: Tonypete
-----
This argument makes my face long.

Posted by: whig at July 21, 2025 11:50 AM (ctrM5)

180 Evidence of racist patriarchy?
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:50 AM (GBKbO)

Sir, this is Mathematics.

Chemistry is across the hall.

Posted by: Mathematicians at July 21, 2025 11:51 AM (VoAdT)

181 Um, Grok is free guys.

This rule has been true since Day One of the internet. If something is "free," you're not the customer, you're the product. AI doesn't change that.

Posted by: Oddbob at July 21, 2025 11:51 AM (/y8xj)

182 This is the point where someone mentions that 'Alice in Wonderland' was written as a satire on theoretical maths. It might as well be me.

Posted by: Victor Tango Kilo at July 21, 2025 11:51 AM (E0uAQ)

183 BTW I will no use AI for my videos.

Posted by: Still lurking up north, read some, don't comment much at July 21, 2025 11:51 AM (kTd/k)

184 by turning them into sammiches?
Posted by: anachronda
-----
Long pork in the vernacular. Although kuru is a problem if one consumes brains.

Posted by: whig at July 21, 2025 11:52 AM (ctrM5)

185
DU types needed CBS to continue its money losing security blanket (Colbert) and CBS could damn well scrounge up the money to do that.
Posted by: whig at July 21, 2025 11:49 AM (ctrM5)
-------------

Yeah, every complaint about CBS canceling Colbert centered around OPM. CBS ran out of money but the left thinks CBS should continue.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:52 AM (tT6L1)

186 I use ChatGPT at work. I build useful GPTs for things like writing emails in different styles for different audiences. Also for putting together presentations with slides.

Posted by: Beartooth at July 21, 2025 11:52 AM (8wKVG)

187 170 What the hell is a "number"?
Posted by: Mathematicians

forty-two!

Posted by: deep thought at July 21, 2025 11:52 AM (v3pYe)

188 We will survive! Nothing can hurt you. I gave you that. You are great, I am great... Twenty years of groping, to prove the things I'd done before were not accidents... seminars and lectures to rows of fools who couldn't begin to understand my systems! Colleagues - colleagues laughing behind my back at the Boy Wonder, and becoming famous building on my work. Building on my work!
Posted by: Dr. Richard Daystrom at July 21, 2025

***
Perfect quote for this thread

Posted by: Wolfus Aurelius, Dreaming of Elsewhere at July 21, 2025 11:52 AM (omVj0)

189 It will be useful in screwing real people. Biggest thing it'll do is track everything. Perfect for government and the antichrist.

Posted by: SOMEASSHOLESTOLEMYPEN at July 21, 2025 11:52 AM (PvkXK)

190 Now Big Tech is getting into the cash cows of Big Energy.

We have a hell of a fight going on right now because a company wants to install power lines and generate electricity for a data center in VA. I live in MD.

The power company had the balls to announce that the state government had already signed off on it. Imagine the shock when the government disagreed.

Where I live in MD is still farmland. You want my support, build nuke plants in my state and take care of our needs first.

Posted by: NR Pax at July 21, 2025 11:52 AM (99eI0)

191
Quick opinion question:

I want to coat some treated 6"x6" 's and cedar roof decking (think outdoor pavilion 20'x24', brown metal roof). What would y'all recommend?

Thanks

Posted by: Swamp Rabbit at July 21, 2025 11:53 AM (EFZgU)

192 165 The Japanese are working on an AI powered plumber with lifelike ass crack.

tentacles are handy for snaking through pipes

Posted by: anachronda at July 21, 2025 11:53 AM (v3pYe)

193 I want to coat some treated 6"x6" 's and cedar roof decking (think outdoor pavilion 20'x24', brown metal roof). What would y'all recommend?

Thanks
Posted by: Swamp Rabbit

Hmmmmm. . . . Maybe a 30' x 36' blue tarp.

Posted by: Tonypete at July 21, 2025 11:54 AM (cYBz/)

194 181 Um, Grok is free guys.

This rule has been true since Day One of the internet. If something is "free," you're not the customer, you're the product. AI doesn't change that.
Posted by: Oddbob
=======
Yep. Grok is harvesting dialogues for training the next iteration of their model.

And recently people were upset to find out that their AI dialogues with a model can be used in court (copyright violation dispute). That is setting the precedent that any of the data you share with Mr. Grok (or other models) falls within the third party doctrine--no reasonable expectation of privacy.

Posted by: whig at July 21, 2025 11:54 AM (ctrM5)

195 A talking head on WTOP was yapping about how Colbert's firing may have had political motivation. Yeah...the man is hemorrhaging audience numbers, hasn't been funny since the days he had the Daily Show writers helping him out and he insults CBS's handling of the legal case with Trump.

But that's got nothing to do with anything.

Posted by: NR Pax at July 21, 2025 11:54 AM (99eI0)

196 This is the point where someone mentions that 'Alice in Wonderland' was written as a satire on theoretical maths. It might as well be me.
Posted by: Victor Tango Kilo at July 21, 2025 11:51 AM (E0uAQ)

Is that what Lewis Carroll said or just an interpretation by others. I knew it was satire but never bothered to learn about what.

Posted by: polynikes at July 21, 2025 11:54 AM (VofaG)

197 The only part of "AI" they got right is the "A" part.

As in "Asshole Intelligence".

Posted by: Astroglide Is Your Friend at July 21, 2025 11:54 AM (rqga2)

198 "I need some 2x4s"

"OK, how long?"

"I'm building a garage. I'll need them for awhile."

Posted by: WitchDoktor at July 21, 2025 11:54 AM (3Dcbq)

199 I think AI is absolutely going to ruin the developed world but apparently not with cool killbots. It'll just make people dumber and more misinformed than never.

Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:55 AM (xcxpd)

200 199 I think AI is absolutely going to ruin the developed world but apparently not with cool killbots. It'll just make people dumber and more misinformed than never.
Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:55 AM (xcxpd)

=======

I don't get it.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:55 AM (GBKbO)

201 conversation with Grok over the weekend spanning dream interpretation, perspective on past life choices, sports psychology, and what makes gingers hot."

So, 42. Or 21, according to Grok.

Posted by: man at July 21, 2025 11:55 AM (tubbA)

202 I'm wondering if AI could make any improvement to Gavin Newsom's hair. Not a chance.

Posted by: Rev. Wishbone at July 21, 2025 11:55 AM (/gTjv)

203 190 Where I live in MD is still farmland. "Why should we give 'bacca farmers and watermen what they need. They hate us".

Buncha jackoffs in Annapolis.

Posted by: bill in arkansas, not gonna comply with nuttin, waiting for the 0300 knock on the door at July 21, 2025 11:56 AM (gm9Sb)

204 Where I live in MD is still farmland.

we went to St. Michaels for the day. The road there from SEPA is mainly farmland, once you get off I-95.

I was stunned at the beauty and rural-ness of the area. Reminded me of South Jersey before everything got developed, except the farms we passed on the way to the shore were not as expansive as the MD farms.

Posted by: kallisto at July 21, 2025 11:56 AM (dCxaZ)

205 190 Now Big Tech is getting into the cash cows of Big Energy.

We have a hell of a fight going on right now because a company wants to install power lines and generate electricity for a data center in VA. I live in MD.

The power company had the balls to announce that the state government had already signed off on it. Imagine the shock when the government disagreed.

Where I live in MD is still farmland. You want my support, build nuke plants in my state and take care of our needs first.

Posted by: NR Pax at July 21, 2025 11:52 AM (99eI0)

4th part of boondoggle - always have well connected folks right by DC. Why is VA (specifically NoVA) the biggest data center/AI place right now - ask yourself that question, and know this is intended as another grift, probably paid by DoD and DHS and Dept of Energy...

I live here, so I make out...but you all need to consider why you'd want District 1 to go from fed gov employee.contractor money to fed gov grant money as the grifts (with the appropriate middle man payouts)...

PS - Granted, our "felonies are forever" and lack of being "all green all the time" also made for a good choice b/c we'd have energy and we'd protect the dang places...

Posted by: Nova Local at July 21, 2025 11:56 AM (tOcjL)

206 AI is part of my daily life. Both work and personal. Is it perfect? No not at all. But it’s pretty fucking good.

For work it’s scary good. As in why do I need to be here I’m just kind of the middle man. Well OK not at that point yet but in 5 years? Maybe.

Posted by: Its Go Time Donald at July 21, 2025 11:56 AM (Vvm2m)

207 99.999% of people using Grok (i'm guessing low just to be safe) are bored people intentionally trying to screw with Grok.

Grok is learning how to be Grokier by adapting itself to the japes and tomfoolery of people who wouldn't care if Grok jumped off a bridge.

...

So basically Grok is in AI middle school right now.

Posted by: Warai-otoko at July 21, 2025 11:57 AM (VoAdT)

208 Hmmmmm. . . . Maybe a 30' x 36' blue tarp.
Posted by: Tonypete
------------

That's what the wife said

Posted by: Swamp Rabbit at July 21, 2025 11:57 AM (EFZgU)

209 Maybe all of AI was developed for the sole purpose of switching the grid over to nuclear power.

Think about it.

I've been praying for it. My hope is that this AI fad diminishes and these huge industrial parks that have been paid for† by Big Tech with their proximate energy generation micro grids, will open up for general manufacturing.

Data centers have low staffing density, don't need much housing nearby, but manufacturing will need higher staffing density and 10-20 year homes (built for the engineers) will be in the price range for blue collar workers managing robots and production lines when that conversion from AI to manufacturing occurs.

In the meantime small reactors get cheaper, plentiful and normalized.

† bribes to local officials to expedite the building permits

Posted by: Unknown Drip Under Pressure at July 21, 2025 11:57 AM (pawvM)

210 191
Quick opinion question:

I want to coat some treated 6"x6" 's and cedar roof decking (think outdoor pavilion 20'x24', brown metal roof). What would y'all recommend?

Thanks
Posted by: Swamp Rabbit
=====
First, do you have a carpenter bee problem? Second, make sure you let that treated wood dry for several months before slapping stain on it. They weep moisture from the interior for quite a while.

Third, clear stain last the least in sunlight as far as protection, as you go up in opacity, the stain lasts longer and adheres better as it has more UV protectants.

Fourth, elevate your posts either via concrete footing or metal from the ground to avoid end rot.

Depending on how much you want to spend, pvc sleeving of posts is an option as well and that requires no staining to do.

Posted by: whig at July 21, 2025 11:57 AM (ctrM5)

211 @ChuckGrassley
·
12m
NEWLY-DECLASSIFIED DOC shows DOJ Inspector Gen concluded the Comey-run FBI FAILED 2 conduct a thorough + complete investigation into Clinton email scandal Comey & his team ate up Steele Dossier lies + chased after Russia collusion hoax but took kid gloves approach on Clinton

Posted by: JackStraw at July 21, 2025 11:57 AM (viF8m)

212 But I had problems recognizing spaghetti code and figuring out its logic.Posted by: whig at July 21, 2025 11:47 AM

Ah, grasshopper (

), if you had problems figuring out its logic, you were more than likely in spaghetti code. So you knew you were there, even if you didn't know it.

Posted by: Duncanthrax at July 21, 2025 11:58 AM (0sNs1)

213 I want to coat some treated 6"x6" 's and cedar roof decking (think outdoor pavilion 20'x24', brown metal roof). What would y'all recommend?

Thanks
Posted by: Swamp Rabbit at July 21, 2025 11:53 AM (EFZgU)

Texas sweet crude.

Posted by: Dr. Pork Chops & Bacons at July 21, 2025 11:58 AM (g8Ew8)

214 200 199 I think AI is absolutely going to ruin the developed world but apparently not with cool killbots. It'll just make people dumber and more misinformed than never.
Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:55 AM (xcxpd)

=======

I don't get it.
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:55 AM (GBKbO)

I eagerly await AI generated movies and AI generated movie reviews of those movies.

...wait, maybe eagerly is the wrong word...

Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:58 AM (xcxpd)

215 For work it’s scary good. As in why do I need to be here I’m just kind of the middle man. Well OK not at that point yet but in 5 years? Maybe.

Posted by: Its Go Time Donald at July 21, 2025 11:56 AM (Vvm2m)

Most people are better than being simple human interfaces between requirements and sloppy random number generators posing as databases.

Posted by: WitchDoktor at July 21, 2025 11:58 AM (RG+lC)

216 Dear AI, can you answer this question:

If the plural of mouse is "mice", why is the plural of spouse not "spice"?

Posted by: kallisto at July 21, 2025 11:59 AM (dCxaZ)

217 grok, are you capable of imitating ace of spades blog?
--------------------------
I can mimic the style and tone of the Ace of Spades blog, known for its sharp, irreverent, and often sarcastic conservative commentary. It typically blends humor, pop culture references, and political takes with a conversational, in-your-face vibe. If you want, I can write a response or post in that style on a specific topic—just give me a subject or question to riff on. I don’t have access to the blog’s internal data or unpublished content, but I can capture its public-facing voice based on what’s out there. Want me to take a swing at it?

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:59 AM (tT6L1)

218 214 I eagerly await AI generated movies and AI generated movie reviews of those movies.

...wait, maybe eagerly is the wrong word...
Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:58 AM (xcxpd)

========

*looks at the MCU since Endgame*

*contains conspiratorial mindset in small box at back of mind*

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:59 AM (GBKbO)

219 4th part of boondoggle - always have well connected folks right by DC.

I think that the development folks grossly underestimated just how much hatred was going to be directed at them over this.

I went to a community meeting where they were there along with our reps. My rep (Who is a Republican and not a RINO) was absolutely furious. The meeting was so packed that I had to leave because I would not be able to hear anything and it was too crowded.

Posted by: NR Pax at July 21, 2025 12:00 PM (99eI0)

220 Ah, grasshopper, if you had problems figuring out its logic, you were more than likely in spaghetti code. So you knew you were there, even if you didn't know it.
Posted by: Duncanthrax

We always were amazed at the amount of dead code within our COBOL or FORTRAN modules. (Code that could never be executed as there was no possible path into it).

Posted by: Tonypete at July 21, 2025 12:00 PM (cYBz/)

221 So people who use AI at work to generate reports and such could you explain them if a customer or supervisor called you out of the blue with questions?

Posted by: polynikes at July 21, 2025 12:01 PM (VofaG)

222 No more Mr. Nice Guy.

Western Lensman@WesternLensman
Demcorats have dispatched Beto O’Rourke to say that Dems have been more concerned with "being right" than being in power — and that needs to change.
"We have to get serious. We have to be absolutely ruthless about getting back in power."

Posted by: Anonosaurus Wrecks, Lying Dogface Pony Soldier at July 21, 2025 12:01 PM (L/fGl)

223 211 @ChuckGrassley
·
12m
NEWLY-DECLASSIFIED DOC shows DOJ Inspector Gen concluded the Comey-run FBI FAILED 2 conduct a thorough + complete investigation into Clinton email scandal Comey & his team ate up Steele Dossier lies + chased after Russia collusion hoax but took kid gloves approach on Clinton

Posted by: JackStraw at July 21, 2025 11:57 AM (viF8m)

I might jizz in my pants if they start looking into Seth Rich's death.

Posted by: The Central Scrutinizer at July 21, 2025 12:02 PM (KbCG3)

224 So people who use AI at work to generate reports and such could you explain them if a customer or supervisor called you out of the blue with questions?
Posted by: polynikes at July 21, 2025 12:01 PM (VofaG)

You get a report that makes stuff up, and when questioned as to how it got that result the AI will make that up too.

Posted by: WitchDoktor at July 21, 2025 12:02 PM (PqQny)

225 Lessee...better learn Ada or you will be out of a job in two years...if you don't learn java you will be out of a job in two years...you must get some Microsoft certification or you will be out of a job in two years...better use AI or you will be out of a job in two years...Remember Ada? CORBA?

Posted by: Don in SoCo at July 21, 2025 12:02 PM (vd6bO)

226 A talking head on WTOP was yapping about how Colbert's firing may have had political motivation. Yeah...the man is hemorrhaging audience numbers, hasn't been funny since the days he had the Daily Show writers helping him out and he insults CBS's handling of the legal case with Trump.
------------------_----------

His show had a staff of 200 people. That's ridiculous.

Some twit was saying "but it was the top-rated late night show." Which is like being the most popular WNBA team.

Posted by: Victor Tango Kilo at July 21, 2025 12:03 PM (E0uAQ)

227 First, do you have a carpenter bee problem? Second, make sure you let that treated wood dry for several months before slapping stain on it. They weep moisture from the interior for quite a while.

*No bee problem

*It was built in Feb, so It's been up 5 ish months

Third, clear stain last the least in sunlight as far as protection, as you go up in opacity, the stain lasts longer and adheres better as it has more UV protectants.

*What color/type stain would you recommend? I would like the cedar to pop

Fourth, elevate your posts either via concrete footing or metal from the ground to avoid end rot.

*Yes, the post are elevated off the concrete

Depending on how much you want to spend, pvc sleeving of posts is an option as well and that requires no staining to do.

*Thought about that, but I'd rather have a stain I think

Posted by: whig

Posted by: Swamp Rabbit at July 21, 2025 12:03 PM (EFZgU)

228 Who do I give my life savings to?
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:45 AM

I've heard director's cuts on laser discs are an investment opportunity that's about to break out.

Posted by: Duncanthrax at July 21, 2025 12:03 PM (0sNs1)

229 Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:59 AM (tT6L1)

Grok write me an aos limerick about a doctor who writes limericks.

Posted by: polynikes at July 21, 2025 12:03 PM (VofaG)

230 Hunter is conscious! NSFW due to language.

https://is.gd/NE0I9D

Posted by: Anonosaurus Wrecks, Lying Dogface Pony Soldier at July 21, 2025 12:04 PM (L/fGl)

231 Want me to take a swing at it?
Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 11:59 AM (tT6L1)

No, Mr. Grok. I don't .

Posted by: FenelonSpoke at July 21, 2025 12:04 PM (2GCMq)

232 130 Do small modular reactors even exist?
________________________

Aye

https://blog.tmcnet.com/ blog/rich-tehrani/green/ japans-yoroi-reactor-ushers-in- a-new-era-of-micro-nuclear-power.html

i'm assuming that "yoroi" is a japanese word meaning "summons godzilla".

Posted by: anachronda at July 21, 2025 12:04 PM (v3pYe)

233 228 Who do I give my life savings to?
Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 11:45 AM

I've heard director's cuts on laser discs are an investment opportunity that's about to break out.
Posted by: Duncanthrax at July 21, 2025 12:03 PM (0sNs1)

=======

Some of those are actually quite valuable.

Collectables are impossible to predict, though.

My physical media collection is borderline worthless.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 12:04 PM (GBKbO)

234 Remember Ada? CORBA?
Posted by: Don in SoCo at July 21, 2025 12:02 PM

Yes.

Posted by: Duncanthrax at July 21, 2025 12:04 PM (0sNs1)

235 Demcorats have dispatched Beto O’Rourke to say that Dems have been more concerned with "being right" than being in power — and that needs to change.
"We have to get serious. We have to be absolutely ruthless about getting back in power."
_-_-_
Sure seems like they described their actual problem completely backward. If the ends justify the means, that means they aren't to be trusted.

Posted by: Don in SoCo at July 21, 2025 12:05 PM (vd6bO)

236 232 i'm assuming that "yoroi" is a japanese word meaning "summons godzilla".
Posted by: anachronda at July 21, 2025 12:04 PM (v3pYe)

======

Depends on which part of which Era.

Sometimes, Godzilla hates all nuclear power. Sometimes he just hates nuclear weapons. Sometimes, he just wants his baby to learn how to breathe fire well.

Posted by: TheJamesMadison, searching for thrills with John Frankenheimer at July 21, 2025 12:05 PM (GBKbO)

237 186 I use ChatGPT at work. I build useful GPTs for things like writing emails in different styles for different audiences. Also for putting together presentations with slides.
Posted by: Beartooth

I can see that being useful for short emails and slides if that is your job.

But personally. I hate both emails and slideshows because of my long stint in higher education and from my military experiences. Power point poisoning and worthless university wide emails that were widely ignored. They also spammed my inbox making it harder to find student emails to respond to. Glad when we went to CMS systems and I ordered students to use that rather than my std. uni email. By the time I left, despite spam filtering, I got about thirty useless emails to one good one.

Posted by: whig at July 21, 2025 12:06 PM (ctrM5)

238 Too bad Thomas Bender ain't here. He's elbows deep in AI.

Posted by: Don't Make gp Stop This Car at July 21, 2025 12:06 PM (klf3J)

239 NOOD F. Joe Mugabe

Posted by: ShainS -- Make Alligator Alcatraz the Gulf of Gator Gitmo! at July 21, 2025 12:06 PM (0JIRZ)

240 My FiL still has the original Star Wars laserdisc.

Posted by: WitchDoktor at July 21, 2025 12:06 PM (cIOlQ)

241 ure seems like they described their actual problem completely backward. If the ends justify the means, that means they aren't to be trusted.
Posted by: Don in SoCo at July 21, 2025 12:05 PM (vd6bO)

The Ends Justify The Means is their mission statement.

Always has been.

Posted by: polynikes at July 21, 2025 12:06 PM (VofaG)

242 nood

Posted by: Don't Make gp Stop This Car at July 21, 2025 12:06 PM (klf3J)

243 First things first.

Report: Instead of replacing outdated air traffic control systems, Pete Buttigieg's DOT spent $80 BILLION on DEI grants

-
But millions will die if we can't have Big Bird doing weather reports!

Posted by: Anonosaurus Wrecks, Lying Dogface Pony Soldier at July 21, 2025 12:07 PM (L/fGl)

244 234 Remember Ada? CORBA?
Posted by: Don in SoCo
======
Ada is still being used because it is DoD compliant. Read something about some big software development project using Ada and I think the obscurity of the language is an asset to the DoD instead of a hindrance.

Posted by: whig at July 21, 2025 12:07 PM (ctrM5)

245 I eagerly await AI generated movies and AI generated movie reviews of those movies.

...wait, maybe eagerly is the wrong word...
Posted by: Mark Andrew Edwards, buy ammo at July 21, 2025 11:58 AM (xcxpd)

Watching some of the drek that's been foisted on the public lately I have to wonder if AI hasn't already been creating movies.

Posted by: Polka will never die at July 21, 2025 12:07 PM (uxAXv)

246 For some reason, it seems that AI has become something sexy and a lot of people that should know better have lost sight of the simple fact that AI is code.

GIGO applies. If you use the 'interent' as your input, you will get a LOT of garbage.

Where AI is making serious inroads, that the public isn't seeing much, is in places where there are discrete data sets that are carefully managed. In those cases, the input is much more controlled and curated. So, the answers that AI can generate from those data sets are much more reliable.

An area where I am familiar with this in property assessments for very large counties. Think like New York or LA.

A claimed property assessment value is compared to an AI prediction of that property value. If those values match within reasonable +-, then that claimed assessment is accepted and registered without human interaction. This represents an enormous savings in man hours. (like 10s of thousands of hours)

Posted by: ghbucky at July 21, 2025 12:08 PM (/xDbx)

247 Based on limited experience, giving Grok a wide open question can result in a rather humorous take on things. Grok went on a government overreach riff which included a bacon pyramid.

Posted by: blake - semi lurker in marginal standing (tT6L1) at July 21, 2025 12:09 PM (tT6L1)

248 246
‘ An area where I am familiar with this in property assessments for very large counties. Think like New York or LA.’

So an AI generated assessment would be regarded as authoritative? Do other members in your field feel the same?

Posted by: Dr. Claw at July 21, 2025 12:17 PM (jbnUc)

249 I forgot the huge, massive, giant scandal that broke over the weekend:

Some 400 accounts on X are allegedly "pro-Trump."
Yes, they have few followers.
Yes, they post both supportive and critical posts toward Trump.
Yes, almost no one has seen their posts.

But they're bots!
And they're pro-Trump!
And that proves Trump stole the election!

Or some sort of fake and gay shit that retards are spazzing out over.

Posted by: People's Hippo Voice at July 21, 2025 12:24 PM (uEVfM)

250 Read the Apple paper (yes Apple, the corporation, is one of the authors). What is called AI, LLMs, can’t reason their way out of a wet paper bag, metaphorically speaking.

Posted by: NemoMeImpuneLacessit at July 21, 2025 12:36 PM (075ei)

251 So an AI generated assessment would be regarded as authoritative? Do other members in your field feel the same?
Posted by: Dr. Claw at July 21, 2025 12:17 PM (jbnUc)

No. The AI #s are not authoratitive. The claimed assessment value is compared to an AI prediction of the property value.

If they match, then that is the accepted valuation. If they do not match, then human intervention is required.

Posted by: ghbucky at July 21, 2025 12:42 PM (/xDbx)

252 Bear in mind that an increasing amount of what AIs take in from the Internet is AI generated. And often hallucinations. So the problem is a vicious cycle.

By the way, if you've never considered replacing "AI" with "progressive", try it and consider the proposition that it's the same phenomenon in both cases...

Posted by: Brother Tim sez at July 21, 2025 01:06 PM (ckuaa)

253 Latest task I gave Grok was to find a used replacement part for sale online for my car. It was faster (as in instantly) than doing a google search, faster than driving to junkyards, faster than going to AutoZone, faster than asking friends.

AI is here, it's not going away, and people who don't adapt will get left behind.
Posted by: Don't Make gp Stop This Car at July 21, 2025 11:14 AM (klf3J)

of course now you have a 50 caliber machine gun on your car.

Posted by: Oldcat at July 21, 2025 01:47 PM (8avO+)

254 We used AI in a few projects in analyzing X Ray images of trucks and cars for border security. Regular image processing produced candidates for 'foreign objects' and then the AI would judge if each one seemed abnormal based off of density. But training only happened once and after that it was deterministic. It did not retrain.

Posted by: Oldcat at July 21, 2025 01:50 PM (8avO+)

255 No more Mr. Nice Guy.

Western Lensman@WesternLensman
Demcorats have dispatched Beto O’Rourke to say that Dems have been more concerned with "being right" than being in power — and that needs to change.
"We have to get serious. We have to be absolutely ruthless about getting back in power."
Posted by: Anonosaurus Wrecks, Lying Dogface Pony Soldier at July 21, 2025 12:01 PM (L/fGl)

The party has been half organized crime half communist for 25 years. Ruthless about being in power is not the problem.

Posted by: Oldcat at July 21, 2025 01:52 PM (8avO+)

256 "My own dabbling in AI has shown that it can sometimes produce cool pictures"

People keep saying this, but AI-generated pictures only look "good" if you suspend average observational skills such as lighting and color theory. (which in spite of the academic-sounding name is just basic observation of light physics and common sense) If you do nothing more than go out to a park and spend a day looking around and then look at an AI-generated image of something similar, it's going to look like a Frankenstein abomination with conflicting light sources, bizarre color blending, and physics-defying shadow placement even in the absence of multi-armed supermodels. No art classes necessary.

AI images are only useful if you aren't particularly skilled with words either and want to show someone who knows what he's doing what you want.

Posted by: Saber Alter at July 21, 2025 02:09 PM (AZg1Q)

257 AI is and ALWAYS has been a fancy search engine that instead of just returning results re-packages those results to look like an answer to your question ... so instead of you having to read all the source docs it found it packages them up for you into a single answer ...

Posted by: The Dark Lord at July 21, 2025 02:15 PM (DBAaD)

258 257 AI is and ALWAYS has been a fancy search engine that instead of just returning results re-packages those results to look like an answer to your question ... so instead of you having to read all the source docs it found it packages them up for you into a single answer ...

Ding ding ding!!

The guy upthread talking about AI finding him a car part... OK, it just did a search for you. Sure, it is fast, but a search engine is hardly 'Artificial Intelligence'

Posted by: ghbucky at July 21, 2025 02:48 PM (/xDbx)

259 AI will be nothing until it’s called turbo AI.

Posted by: Kenneth King Neil at July 21, 2025 03:58 PM (/g9JB)

260 Q. Which Dr. Who Villains drink too much? A, The Cyder Men

Posted by: Tamaa the Drongo Bird at July 21, 2025 06:21 PM (wGqjj)

Processing 0.04, elapsed 0.0501 seconds.
15 queries taking 0.0124 seconds, 269 records returned.
Page size 167 kb.
Powered by Minx 0.8 beta.