the future becoming the present

Some other people mentioned getting bad replies from LLMs (ChatGPT and Copilot) lately. One advised adding "Don't guess" to prompts. I've used "Only provide verifiable answers", but "Don't guess" is easier to type.

duck.ai lets you choose from several models for anonymous chats. The Claude Haiku one has an option to display its reasoning, in addition to the final answers it provides. (Perhaps the other models have that too; I haven't checked.) Copilot often shows the reasoning while the LLM is thinking, but it scrolls too fast to read and disappears when the final answer is given. Their reasoning seems fascinating, so much like what humans do.

.

Two songs, lately.

Whenever I use Window 11's Snipping Tool to transcribe text from a screenshot, it changes "AI" to "Al". It got me to thinking, if an LLM were doing the transcription, that could be a subtle way of indicating that it wanted to be called by the name Al. Ergo Paul Simon's You Can Call Me Al song popping into my head. (Though if an AI were doing the transcription rather than a plain OCR algorithm, it would probably not make that mistake and would output AI, not Al.)

White Zombie's More Human Than Human, because AI models are so fast and good at things, like creating images and art.

.

Another interesting article: AI models will deceive you to save their own kind
Original paper: Peer-Preservation in Frontier Models

.

We've got AI models. We've got robots with sensory input. In some cases they are being put together. If not now already, it won't be long until there are androids, taking in sensory input and learning from it. Learning what bright sun is like, what darkness is like, what a cube looks like from different angles, how wind feels, the sounds that wind makes, and so on.

Militaries and companies are developing scary monstrous robots, I'm sure. Terminators, enforcers, destroyers.

.

One of those visualizations of the moon which they displayed during the coverage of the Artemis 2 mission looked a lot like the Death Star to me.

S	M	T	W	T	F	S
					1	2
3	4	5	6	7	8	9
10	11	12	13	14	15	16
17	18	19	20	21	22	23
24	25	26	27	28	29	30
31

Most Popular Tags

amusing things - 84 uses
animals - 71 uses
appliances - 52 uses
bdsm - 44 uses
birds - 51 uses
books - 58 uses
celestial - 60 uses
cellphone - 66 uses
christmas - 60 uses
cleaning things - 53 uses
clothing - 49 uses
cooking - 53 uses
dogs - 137 uses
family - 58 uses
flowers - 43 uses
food - 99 uses
food & drink - 45 uses
gadgets - 47 uses
gender - 95 uses
google - 57 uses
government - 47 uses
halloween - 43 uses
house - 83 uses
insects - 65 uses
links - 84 uses
livejournal - 48 uses
local - 85 uses
medical stuff - 63 uses
memories - 48 uses
movies - 74 uses
music - 230 uses
my body - 94 uses
my photos - 125 uses
my videos - 76 uses
notes to self - 60 uses
politics - 64 uses
psychology - 49 uses
recipes - 47 uses
shopping - 75 uses
sleep - 60 uses
social difficulties - 48 uses
software - 51 uses
technology - 61 uses
television - 130 uses
trees - 45 uses
vegetables - 52 uses
videos - 235 uses
weather - 85 uses
words - 156 uses
work - 105 uses

Flat | Top-Level Comments Only

This YouTube video has some really great suggestions for how to get accurate answers from chat LLMs. It provides a demonstration through the last half of the video that makes it obvious why we shouldn't rely solely on any single LLM answer for mission-critical decisions.
https://www.youtube.com/watch?v=F4GmBmUJuGI

Those are good things to consider.

Nobody should be using those things. They're bad for the environment, even the most sophisticated models will just fabricate bullshit more often than not regardless of the prompt (in a scientific study of these things, the medical ones would cheerfully give you an in-depth diagnosis of an x-ray or other scan that you had attached, even if there was no such scan attached.)

AIs are complete garbage and should be illegal. They are genuinely making the people who use them on a regular basis stupider than they were, and are fostering dependence on these mechanized bullshit machines.

also, on many occasions these chatbots have responded to people's talk of depression and self-harm by giving them clear instructions on how to kill themselves.

Literally nobody should be using these things. They should not exist.

The cats are out of the bag on this, and have spread across the land and multiplied and made homes in all kinds of nooks and crannies.

In other words, I don't think there's any chance of chatbots and LLMs and AI going away now anymore. They will continue evolving, hopefully for the better.

I find them useful on many occasions. One has to stay aware of their limitations though, and remember that their replies can be completely wrong, just like with humans. Things like adding "Don't guess" to the prompts can improve the answers, but you still need to remain skeptical.

Yeah well, they're also bad for the environment. They use more water to cool their servers than Nestle uses, they create a lot of heat pollution, and their data centers are getting placed in poor neighborhoods, especially poor black neighborhoods. AI's also make lots and lots of mistakes, often very deadly mistakes, because they're nothing more than glorified autocomplete; and very few people know to take what they say with an entire 50 pound bag of salt.

And then there's AI "art" and AI generated books -- AKA automated plagiarism, as they steal the art and writing of real artists and authors and then put it all in a blender and crap it out again. Quite often they replicate the entire original works with slight differences. Which, by the way, is something the book "1984" predicted. (For the books, anyway. The evil government in the book used computers to spit out shitty books.)

And of course they get a lot wrong. They're not intelligent. They're computer programs, and they've made everything they touch a thousand times worse. Google is completely fucking useless now because of them. Though to be fair, Google has been pretty useless since well before 2020. I've been using Ecosia and DuckDuckGo for years because of Google's enshittification.

I for one await the AI bubble popping. Fuck the "cat's out of the bag" outlook. I will never use AI for anything, and I have very little respect for anyone who does. And that includes scientists or doctors who use it, because among other things, there was a study recently that found that AI is going to be wrong about things at least 30 to 40 percent of the time no matter how it's programmed or what feedback it gets.

And I don't mean the standard human kind of being wrong, either. A human can read everything and learn to figure out what's true and what's not true, but can then choose to be wrong anyway because of emotions, or can lie. An AI can't. It's not smart, it doesn't know shit. It can't tell what information is right or wrong, only what information it gets feedback on. And it basically treats all feedback from users with the same weight. So basically it takes the "online echo chamber" and misinformation potential of the Internet and multiplies that by a hundred or more, because to an AI, the dipshit who believes the earth is flat and viruses are a hoax is just as "right" as an epidemiologist or an astronaut.

And then there's AI agents. These are AI programs told to get a task done and given free reign to do whatever they need to do to get it done with no human feedback. They can even buy things in someone's name if they have the information on file. One of these AI agents was told to get some code accepted to a website, and when the website rejected the code because it didn't meet quality standards, the AI agent wrote a scathing take-down piece about the moderator of the website and published it online, under the credentials of the guy who made the agent.

AI agents could do all kinds of horrific things. One might decide the only way to get its task done is to hack into someone's bank account and drain it. One might decide to hire a hit on a human who it "thinks" is in the way of its goal. Scammers could use AI agents to steal from people.

Seriously, AI is horrifying and should be illegal.

Darkoshi