I believe that Amazon’s Prime Assistance for Students and Low Income is being gatekept by an AI. I have a non-conforming confirmation of Public Assistance, but they have no flexibility to send cover letters or explanations to an email address for a human being to see and discuss…considering how wealthy Bezos is, doing shitty gate keeping on one of your few charitable actions is kind of “on brand”…
I've found that AI has been great for vastly scaling up recipes, say "Mexican rice for 2000" and rough drafting the handwasher device my other LLM suggested might be well received by convenience enthusiasts with 6 fingers on one hand.
My experience with AI tells me a lot about it. Amazon uses it in "customer service." A package they said had been delivered, was not. I used their chat to address the problem. At first it seemed fine, but I began to notice that it responded twice at the beginning of every dialogue, as in yes, yes or thank you, thank you. If you get out of its programmed responses it begins to fail. I asked if I should wait a couple of days to see if one of my neighbors received the delivery and it said sure, sure, but it then behaved like I'd never asked that question. It then asks you if it was helpful and I replied yes. But when I received the notice that it would send me the new book as soon as it received my return, I realized it was not helpful. Can't return an item I never received. When I went back, I got a real person who quickly resolved it. Which would have happened in my original contact if I'd received service from a real person.
AI on Google is just kind of a joke. It seems to read Google responses to provide the answer to your question. I asked Google what Dr. Pepper ads were trying to say about Dr. Pepper. Well, because large corporate monstrosities "clean" the internet so any answers about their product will come directly from them or from a friendly source, AI's response sounded like it was spoken by the CEO of Dr. Pepper. "Dr. Pepper uses a fun and interactive ad campaign that engages with people in a positive way." Garbage.
"It’s taken over government functions: Nevada is spending $1 million to use Google AI to deny process unemployment claims, even though it’s only right 76 percent of the time at best. At that price, why not just add more employees? Don’t be silly!"
Watch them having to create a whole new department for when shit goes south, and it will.
The Mild Jrs are Middle School teachers (Jr teaches 7th grade Ancient Civ, Mrs Jr teaches 5th grade Language Arts). The are both big into AI, in and out of the classroom. I love them both, but I’m also disappointed.
Both use ChatGPT as a “backup” in the classroom. They claim that they can tell when their student’s work is AI generated, and that they are teaching the kids how to use AI as a resource, rather than a crutch. My old Boomer brain and cynicism tell me that the Mild Jrs will eventually need to accept more AI produced work from their sutdents, lowering the kids Critical Thinking (there’s that term) skills, and lead to a less educated generation.
I spent over 40 years writing and testing software, mostly in the defense industry. AI is fundamentally software, clever computer programs written to gather any and all information. It has no judgement about right or wrong, truth or falsehood. Any software written by humans is only as good as the humans who developed, wrote, tested, and integrated it.
I have very limited direct experience with AI. However, recently I used the word “sone” in a Scrabble game, and one of my opponents asked one of those programs if it was a legitimate word. It said it was NOT, but I was sure it was OK so I looked it up in the Scrabble dictionary, and it WAS. So fuck you AI.
(Not so) interestingly, AI search also prefers "Pablum" (a brand of cereal) to pabulum (a word meaning tasteless basic nourishment and insipid writing).
White Xmas then Donner and Blitzen for NYE! Right now down the hill from me they’re going nut with fireworks. It’s been all boom boom since midnight, no letup
I get phone messages from an entity related to my 401k that always sound so relaxed about calling at 9am on Sunday to ask me if there's anything they can do for me. And conclude by saying they'll try again in about 10 minutes. On Sunday morning.
Thankfully, I'm not actively practicing law these days. But damn, I've gotten a lot of freaked out emails over the last year from malpractice insurers and the Board of Bar Overseers.
The very idea of this infecting my profession wigs me right the fuck out.
I believe that Amazon’s Prime Assistance for Students and Low Income is being gatekept by an AI. I have a non-conforming confirmation of Public Assistance, but they have no flexibility to send cover letters or explanations to an email address for a human being to see and discuss…considering how wealthy Bezos is, doing shitty gate keeping on one of your few charitable actions is kind of “on brand”…
I've found that AI has been great for vastly scaling up recipes, say "Mexican rice for 2000" and rough drafting the handwasher device my other LLM suggested might be well received by convenience enthusiasts with 6 fingers on one hand.
Am I live or am I Memorax
You am.
My experience with AI tells me a lot about it. Amazon uses it in "customer service." A package they said had been delivered, was not. I used their chat to address the problem. At first it seemed fine, but I began to notice that it responded twice at the beginning of every dialogue, as in yes, yes or thank you, thank you. If you get out of its programmed responses it begins to fail. I asked if I should wait a couple of days to see if one of my neighbors received the delivery and it said sure, sure, but it then behaved like I'd never asked that question. It then asks you if it was helpful and I replied yes. But when I received the notice that it would send me the new book as soon as it received my return, I realized it was not helpful. Can't return an item I never received. When I went back, I got a real person who quickly resolved it. Which would have happened in my original contact if I'd received service from a real person.
AI on Google is just kind of a joke. It seems to read Google responses to provide the answer to your question. I asked Google what Dr. Pepper ads were trying to say about Dr. Pepper. Well, because large corporate monstrosities "clean" the internet so any answers about their product will come directly from them or from a friendly source, AI's response sounded like it was spoken by the CEO of Dr. Pepper. "Dr. Pepper uses a fun and interactive ad campaign that engages with people in a positive way." Garbage.
There are good use cases for generative AI, just as much as there are good use cases for blockchain.
One of the real issues is that tech is being shoehorned into places where it has no business.
"It’s taken over government functions: Nevada is spending $1 million to use Google AI to deny process unemployment claims, even though it’s only right 76 percent of the time at best. At that price, why not just add more employees? Don’t be silly!"
Watch them having to create a whole new department for when shit goes south, and it will.
I do use the "regular" AI somewhat, and find it helpful.
But holy crap, God forbid we let AI ever have control over weapons' systems and targets - although apparently Israel is already doing this. Shudder.
The Mild Jrs are Middle School teachers (Jr teaches 7th grade Ancient Civ, Mrs Jr teaches 5th grade Language Arts). The are both big into AI, in and out of the classroom. I love them both, but I’m also disappointed.
Both use ChatGPT as a “backup” in the classroom. They claim that they can tell when their student’s work is AI generated, and that they are teaching the kids how to use AI as a resource, rather than a crutch. My old Boomer brain and cynicism tell me that the Mild Jrs will eventually need to accept more AI produced work from their sutdents, lowering the kids Critical Thinking (there’s that term) skills, and lead to a less educated generation.
I spent over 40 years writing and testing software, mostly in the defense industry. AI is fundamentally software, clever computer programs written to gather any and all information. It has no judgement about right or wrong, truth or falsehood. Any software written by humans is only as good as the humans who developed, wrote, tested, and integrated it.
Flawed humans will write flawed software and AI.
Just dawned on me that generative AI may be the TechBros equivalent of "flooding the zone with shit" . . .
I have very limited direct experience with AI. However, recently I used the word “sone” in a Scrabble game, and one of my opponents asked one of those programs if it was a legitimate word. It said it was NOT, but I was sure it was OK so I looked it up in the Scrabble dictionary, and it WAS. So fuck you AI.
(Not so) interestingly, AI search also prefers "Pablum" (a brand of cereal) to pabulum (a word meaning tasteless basic nourishment and insipid writing).
https://www.merriam-webster.com/dictionary/pabulum
White Xmas then Donner and Blitzen for NYE! Right now down the hill from me they’re going nut with fireworks. It’s been all boom boom since midnight, no letup
I get phone messages from an entity related to my 401k that always sound so relaxed about calling at 9am on Sunday to ask me if there's anything they can do for me. And conclude by saying they'll try again in about 10 minutes. On Sunday morning.
I can’t figure out this Substack app. Do new comments load?
Why would they load? Comments are not allowed, silly.
No.
You can’t figure it out because it sucks rocks and should be shot into the sun. It’s not you. Stick with a browser, any browser.
Thankfully, I'm not actively practicing law these days. But damn, I've gotten a lot of freaked out emails over the last year from malpractice insurers and the Board of Bar Overseers.
The very idea of this infecting my profession wigs me right the fuck out.
There *have* been lawyers sanctioned for using ChatGPT to write their arguments, citing to non-existent cases.
Like Michael Cohen
https://www.npr.org/2023/12/30/1222273745/michael-cohen-ai-fake-legal-cases
AI is some bullshit so far. It's like getting random consulting from the INTERNET. There is no intelligence.