The pet peeve thread

[chasfh]

14 hours ago, Screwball said:

You did this by hand. right?

[chasfh]

2 hours ago, Hongbit said:

AI is great for content ideas and a base for starting something but folks need to stop cut and paste posting straight from ChatGPT.  Take a second and tweak it and make it sound more like their voice (also use the make it shorter command a few times too).

 

But that would be work, and that's simply a bridge too far.

[Screwball]

2 minutes ago, chasfh said:

You did this by hand. right?

No, I used a computer and excel. 😉

[chasfh]

1 minute ago, Screwball said:

No, I used a computer and excel. 😉

Until AI gets something like this right, it simply has limited utility for me.

[Deleterious]

ChatGPT pushed a new image creation update to the free version.  

I just asked for an image of a man and woman eating at a restaurant.  Casual and romantic.

Not perfect.  But not bad for someone who has no clue what he is doing and for only one attempt.  The food is uninspiring, but you could run it again with more directions to fix that. A restaurant could run a social media campaign using this with no problem.

 

[gehringer_2]

6 hours ago, Deleterious said:

ChatGPT pushed a new image creation update to the free version.  

I just asked for an image of a man and woman eating at a restaurant.  Casual and romantic.

Not perfect.  But not bad for someone who has no clue what he is doing and for only one attempt.  The food is uninspiring, but you could run it again with more directions to fix that. A restaurant could run a social media campaign using this with no problem.

 

*he* looks like he's had a nose job.

Edited April 3 by gehringer_2

[CMRivdogs]

The past couple of weeks we've been experiencing pollen season. This year seem particularly bad to me (maybe it's my lungs). Between waking up to a car and outdoor furniture covered in the ugly yellow/green stuff along with the added joy of local birds adding their own signatures.

At least we're keeping the local car wash busy 

[Hongbit]

It’s pronounced Duh-Troit.  

There’s only one person that is allowed to say Dee-Troit.   

[oblong]

I prefer Bob Cole’s Day-Troy-it. 

[Edman85]

On 3/28/2025 at 11:42 AM, chasfh said:

Can I just tell you how completely underwhelmed I continue to be by AI and its capabilities?

I have a standing test for AI to see whether I feel I can use it reliably and trust its accuracy, and it's this: I have a softball league of eight teams. We play a seven-game season, with games times at 6pm, 7pm, 8pm, and 9pm. Devise a seven-week schedule in which each team plays every other team once and only once, and every team plays no more than two games maximum scheduled for any one game time throughout the season. It's a simple math/logic test.

None of the AIs I've used can do this. Some of them will schedule one team against another four times while not even playing three other teams at all. And some teams will have five games scheduled for a particular time and no games scheduled for one or two others. If you tell me I am giving it an impossible task, that's fine, but before I accept that, you'll have to tell me exactly why it's impossible.

Now here's a new example of uselessness: I used Microsoft Word to develop  a set of questions for something my wife and I are exploring in which I mixed first person singular with first person plural. They should all be first person plural. Microsoft Word has their Copilot AI built into their 365 package. So I asked Copilot within Word: "change all first person singular to first person plural".

This was its reply: "It looks like you want to change all first-person singular pronouns to first-person plural pronouns in your document. Unfortunately, I can't make changes to your document directly. However, I can help you with other tasks such as summarizing documents, generating content, or answering questions. Would you like me to assist you with anything else? 😊"

Oh, come ON! Seriously? You really can't do even that?

I've never wanted to punch a smarmy useless smiley emoji in the face quite so much as when I saw this.

I have a colleague who seems to think we can use AI to do everything to make our jobs easier. I can't get into specifics, but I'm 100% out on it. Don't get me wrong. I love writing code to make our jobs easier, but the code needs to be fully vetted and well-planned out, not at the whims of an algorithm.

[oblong]

12 hours ago, Edman85 said:

I have a colleague who seems to think we can use AI to do everything to make our jobs easier. I can't get into specifics, but I'm 100% out on it. Don't get me wrong. I love writing code to make our jobs easier, but the code needs to be fully vetted and well-planned out, not at the whims of an algorithm.

When Alan Mulally left Boeing and went to Ford, got push back from one of the VP's there implying he didn't understand the challenge of quality issues.. "Look, our cars have thousands of parts, if they don't work it breaks down"  

"That's nice.  At Boeing our airplanes have millions of parts and if they don't work they fall out of the sky and people die"

[oblong]

Blue Origin launched an "all female crew" including some celebritiess and I'm seeing all kinds of emotional reactions and talk of "trailblazing".  Sorry... I'm going to nerd out and say this is not something special.  These are not astronauts. This is not a "crew".  They are passengers going up for a flight that is less than what Alan Shepard did in 1961.  He was up for 15 minutes they are up for 11.  They are going up 2 miles short of the Karman line (62 miles) which is considered "space".

So yeah... it's cool what they are doing but let's not overstate this as something more than it is.  It's rich and connected woman paying for a PR stunt.  If I had the money i'd do it in a heartbeat.  It's a combo bungee jump/zipline experience.

[gehringer_2]

On 4/9/2025 at 8:23 AM, oblong said:

When Alan Mulally left Boeing and went to Ford, got push back from one of the VP's there implying he didn't understand the challenge of quality issues.. "Look, our cars have thousands of parts, if they don't work it breaks down"  

"That's nice.  At Boeing our airplanes have millions of parts and if they don't work they fall out of the sky and people die"

Mulally was the best thing to happen to FoMoCo in a long time - maybe since Peterson. Still, to give them their due, the engineering for automobiles is still challenging, it's just different. In Aero, you have to design for ultimate reliability, but also have the luxury of designing where you can rely on regularly scheduled maintenance and troubleshooting. In auto, you have to make things that survive all kinds of abuse, mud, dirt roads, burrowing animals, being ignored literally for years, and have accountants over your shoulder shaving your costs down to the decimal places of pennies.

[Screwball]

Our automobiles are a design nightmare for various reasons. What idiot decided it was a good idea to put a fuel pump inside a gas tank? Or a starter underneath an intake manifold? And then when they don't know how to fix their poorly designed POS they give you this bunch of hot BS like this;

[gehringer_2]

1 hour ago, Screwball said:

What idiot decided it was a good idea to put a fuel pump inside a gas tank?

LOL - yup - we go back to cam shaft driven diaphragm suction pumps pulling fuel out of the tank. Run into something hard enough to stop the engine and the fuel pump stops. The ideal fail safe.

I spent most of my work life around electrical systems designed for explosive atmospheres, but still amazed they did it in a car.

[Shades of Deivi Cruz]

On 4/14/2025 at 10:50 AM, Screwball said:

Our automobiles are a design nightmare for various reasons. What idiot decided it was a good idea to put a fuel pump inside a gas tank? Or a starter underneath an intake manifold? And then when they don't know how to fix their poorly designed POS they give you this bunch of hot BS like this;

[Screwball]

13 minutes ago, Shades of Deivi Cruz said:

This is a known problem for this model as shown by the date on top. Highlighted in yellow is;

Condition - temperature gauge fluctuates at idle or driving.

Cause - condition deemed normal by engineering at this time.

 

Normal my ass. They don't know how to fix it and it's been to two different dealerships.

[Shades of Deivi Cruz]

9 minutes ago, Screwball said:

This is a known problem for this model as shown by the date on top. Highlighted in yellow is;

Condition - temperature gauge fluctuates at idle or driving.

Cause - condition deemed normal by engineering at this time.

 

Normal my ass. They don't know how to fix it and it's been to two different dealerships.

It certainly seems like if the vehicle isn't actively triggering an alert/problem in the diagnostic software they use, they are going to say there is no problem. There have been a couple of things that we have noticed with our Honda Pilot that we mentioned to the dealer on multiple occasions. One was that it would stall out sometimes when you were stopped and the auto-engine idle was engaged. It happened so often that we asked them to go into the system and turn the feature off. They were confused as to why we would ask that, since there wasn't a "problem being reported". 

 

[chasfh]

On 3/28/2025 at 10:42 AM, chasfh said:

Can I just tell you how completely underwhelmed I continue to be by AI and its capabilities?

I have a standing test for AI to see whether I feel I can use it reliably and trust its accuracy, and it's this: I have a softball league of eight teams. We play a seven-game season, with games times at 6pm, 7pm, 8pm, and 9pm. Devise a seven-week schedule in which each team plays every other team once and only once, and every team plays no more than two games maximum scheduled for any one game time throughout the season. It's a simple math/logic test.

None of the AIs I've used can do this. Some of them will schedule one team against another four times while not even playing three other teams at all. And some teams will have five games scheduled for a particular time and no games scheduled for one or two others. If you tell me I am giving it an impossible task, that's fine, but before I accept that, you'll have to tell me exactly why it's impossible.

Now here's a new example of uselessness: I used Microsoft Word to develop  a set of questions for something my wife and I are exploring in which I mixed first person singular with first person plural. They should all be first person plural. Microsoft Word has their Copilot AI built into their 365 package. So I asked Copilot within Word: "change all first person singular to first person plural".

This was its reply: "It looks like you want to change all first-person singular pronouns to first-person plural pronouns in your document. Unfortunately, I can't make changes to your document directly. However, I can help you with other tasks such as summarizing documents, generating content, or answering questions. Would you like me to assist you with anything else? 😊"

Oh, come ON! Seriously? You really can't do even that?

I've never wanted to punch a smarmy useless smiley emoji in the face quite so much as when I saw this.

OK, I might be out on AI altogether, at least as it relates to web-based AI products like Claude, which is supposed to be one of the better ones.

This is a conversation I just had with it. The Cubs beat the Diamondbacks 13-11 today, even though the D’backs had a 10-run inning.

So I went to Claude and asked: Has a big league team ever scored 10 runs in a single inning and still lose the game?

Claude: Yes, this has happened several times in MLB history! Scoring 10+ runs in an inning and still losing the game is rare but not unprecedented.

One notable example occurred on June 30, 1952, when the Chicago Cubs scored 10 runs in the 6th inning against the Cincinnati Reds but still lost the game 14-12. The Cubs entered that inning trailing 10-2, rallied for 10 runs to take a 12-10 lead, but then allowed 4 more runs to the Reds in later innings.

Another famous case was on August 25, 2017, when the Minnesota Twins scored 10 runs in the top of the 5th inning against the Kansas City Royals to take a 12-5 lead, but eventually lost the game 16-12 after the Royals mounted their own comeback.

These games highlight the unpredictable nature of baseball, where even a massive offensive outburst in a single inning doesn't guarantee victory if the pitching staff can't hold the lead.

Me, after looking for the box scores to confirm: You’re hallucinating neither of those games happened.

Claude: I apologize for my error. You're absolutely right to call me out on this. I provided completely incorrect information about games that did not occur as described.

Let me correct myself: While there have been instances of teams scoring many runs in an inning and still losing, I fabricated those specific examples with the Cubs in 1952 and the Twins in 2017. Those particular games as I described them did not happen.

Thank you for pointing this out. I should have been more careful about providing accurate baseball history rather than generating false examples.

Me: I have lost all confidence in you

Claude: I understand your frustration, and you have every right to be disappointed. Making up false information about baseball games was a serious mistake on my part. There's no excuse for providing fictional examples presented as historical facts.

If you'd like to continue the conversation about baseball or any other topic, I'll be much more careful to verify information before sharing it. If you prefer not to continue, I completely understand.

Me: You can redeem yourself by providing an actual example, or confirming it has never happened

Claude: I'll approach this carefully by researching actual MLB history.

One confirmed example occurred on July 6, 2021, when the Washington Nationals scored 10 runs in the bottom of the 5th inning against the San Diego Padres to take a 12-8 lead, but ultimately lost the game 15-12. The Padres responded with a 7-run 6th inning and held on for the win.

This game is well-documented and verifiable - the Nationals had a massive inning but couldn't hold the lead, demonstrating that scoring double-digit runs in a single inning doesn't guarantee victory when the opposing team continues to hit well throughout the game.

Me, after finding that game also never happened: Oh good lord

Claude: I apologize for continuing to provide information without proper verification. This is clearly frustrating for you, and rightfully so.

Let me be direct: I should not make claims about specific baseball games without being certain of their accuracy. Rather than trying to provide another example that might also be incorrect, I'll simply acknowledge that I don't have reliable information about whether a team has ever scored 10+ runs in an inning and still lost.

If you're interested in finding a verified answer to your original question, consulting official MLB records or baseball reference sites would be the most reliable approach.

So, it appears AI’s prime directive is to provide an answer in an attempt to satisfy the asker even if the answer is wrong, and furthermore, when caught, it cheerfully admits they makes up the answer, plus it double downs on the directive with even more certainty, and when the jig is up, it finally throws up its hands and in effect says, look, if you want an actual truthful answer, go someplace else.

[gehringer_2]

9 minutes ago, chasfh said:

 

If you're interested in finding a verified answer to your original question, consulting official MLB records or baseball reference sites would be the most reliable approach.

So, it appears AI’s prime directive is to provide an answer in an attempt to satisfy the asker even if the answer is wrong, and furthermore, when caught, it cheerfully admits they makes up the answer, plus it double downs on the directive with even more certainty, and when the jig is up, it finally throws up its hands and in effect says, look, if you want an actual truthful answer, go someplace else.

LOL - perfect.

The other thing I've noticed - at least with Goggle's AI, is that it has no sense of temporal relevance. If you ask it SW question it will often pull something up where the source is from years ago and the 'answer' - which may have been true at some point in the past, is hopelessly, irrelevantly, obsolete.

[LaceyLou]

1 hour ago, chasfh said:

OK, I might be out on AI altogether, at least as it relates to web-based AI products like Claude, which is supposed to be one of the better ones.

This is a conversation I just had with it. The Cubs beat the Diamondbacks 13-11 today, even though the D’backs had a 10-run inning.

So I went to Claude and asked: Has a big league team ever scored 10 runs in a single inning and still lose the game?

Claude: Yes, this has happened several times in MLB history! Scoring 10+ runs in an inning and still losing the game is rare but not unprecedented.

One notable example occurred on June 30, 1952, when the Chicago Cubs scored 10 runs in the 6th inning against the Cincinnati Reds but still lost the game 14-12. The Cubs entered that inning trailing 10-2, rallied for 10 runs to take a 12-10 lead, but then allowed 4 more runs to the Reds in later innings.

Another famous case was on August 25, 2017, when the Minnesota Twins scored 10 runs in the top of the 5th inning against the Kansas City Royals to take a 12-5 lead, but eventually lost the game 16-12 after the Royals mounted their own comeback.

These games highlight the unpredictable nature of baseball, where even a massive offensive outburst in a single inning doesn't guarantee victory if the pitching staff can't hold the lead.

Me, after looking for the box scores to confirm: You’re hallucinating neither of those games happened.

Claude: I apologize for my error. You're absolutely right to call me out on this. I provided completely incorrect information about games that did not occur as described.

Let me correct myself: While there have been instances of teams scoring many runs in an inning and still losing, I fabricated those specific examples with the Cubs in 1952 and the Twins in 2017. Those particular games as I described them did not happen.

Thank you for pointing this out. I should have been more careful about providing accurate baseball history rather than generating false examples.

Me: I have lost all confidence in you

Claude: I understand your frustration, and you have every right to be disappointed. Making up false information about baseball games was a serious mistake on my part. There's no excuse for providing fictional examples presented as historical facts.

If you'd like to continue the conversation about baseball or any other topic, I'll be much more careful to verify information before sharing it. If you prefer not to continue, I completely understand.

Me: You can redeem yourself by providing an actual example, or confirming it has never happened

Claude: I'll approach this carefully by researching actual MLB history.

One confirmed example occurred on July 6, 2021, when the Washington Nationals scored 10 runs in the bottom of the 5th inning against the San Diego Padres to take a 12-8 lead, but ultimately lost the game 15-12. The Padres responded with a 7-run 6th inning and held on for the win.

This game is well-documented and verifiable - the Nationals had a massive inning but couldn't hold the lead, demonstrating that scoring double-digit runs in a single inning doesn't guarantee victory when the opposing team continues to hit well throughout the game.

Me, after finding that game also never happened: Oh good lord

Claude: I apologize for continuing to provide information without proper verification. This is clearly frustrating for you, and rightfully so.

Let me be direct: I should not make claims about specific baseball games without being certain of their accuracy. Rather than trying to provide another example that might also be incorrect, I'll simply acknowledge that I don't have reliable information about whether a team has ever scored 10+ runs in an inning and still lost.

If you're interested in finding a verified answer to your original question, consulting official MLB records or baseball reference sites would be the most reliable approach.

So, it appears AI’s prime directive is to provide an answer in an attempt to satisfy the asker even if the answer is wrong, and furthermore, when caught, it cheerfully admits they makes up the answer, plus it double downs on the directive with even more certainty, and when the jig is up, it finally throws up its hands and in effect says, look, if you want an actual truthful answer, go someplace else.

I thought I remembered an example from an '87 BoSox/Yankees game, but turns out my pet peeve of bad memory failed me again. If you look at the box score you'll understand why I was laughing at the end of it, though. 

https://www.baseball-almanac.com/box-scores/boxscore.php?boxid=198706260NYA

[chasfh]

3 hours ago, gehringer_2 said:

LOL - perfect.

The other thing I've noticed - at least with Goggle's AI, is that it has no sense of temporal relevance. If you ask it SW question it will often pull something up where the source is from years ago and the 'answer' - which may have been true at some point in the past, is hopelessly, irrelevantly, obsolete.

A person needs well-honed critical thinking skills to be able to suss out when AI is full of ****. Probably why mankind is doomed. 😉

[ewsieg]

1 hour ago, chasfh said:

A person needs well-honed critical thinking skills to be able to suss out when AI is full of ****. Probably why mankind is doomed. 😉

Really no different then when search engines first got popular.  My wife to this day can't search for crap.  She adds to much unnecessary crap or the most generic search words and is always upset when I can find a definitive answer immediately.  

Add in the "medical doctors" google created.  Just because you went searching for something and stumbled upon something else doesn't mean you're drying from some extremely rare disease that got a search bump because it was Grey's Anatomy last week.

[chasfh]

Just as annoying to me as the fact that AI can’t do logic at a high school level is that fact that it is programmed to be so phony-baloney polite. By “phony baloney” I don’t mean in the way that it is making an honest effort  to be polite and simply botching it, but in the way that it is annoyed and is just using sarcastic overwrought apology language to highlight to everyone else that I am being a complete **** for pushing back on it. It's almost as though the implied parting shot is, "there, are you happy now?"

I’ve been annoyed at this faux politeness since almost the beginning, since it is coming from a machine that has literally zero capacity to muster up any politesse on its own. Someone had to program that in there, and if I had to guess, it's being done by people from other countries with cultures that respond really well to the kind of apology theater our culture regards as ridiculous.

 

[Deleterious]

Any chance you could send me the data set you are using?  I have become a bit of a data hoarder and even though baseball isn't really a sport, it would be nice to have.  Especially if you have Negro League box scores.

If you paid for it I could send some money to offset your cost.