AI Is Out To Destroy Your Pipes!
- Thread starter sablebrush52
- Start date
PipesMagazine Approved Sponsor
Site Sponsors
PipesMagazine Approved Sponsor
PipesMagazine Approved Sponsor
PipesMagazine Approved Sponsor
PipesMagazine Approved Sponsor
I can’t think of a worse marketing picture for a GPS than the one above, “Chester, this way”.I’m not really older than the internet but when I was a kid only the dad blasted gubbermint and the evil godless huge multinational corporations had it.
In the early ninties we heard about the “information superhighway” and really couldn’t imagine what the World Wide Web was.
Y2K threatened our national existence until —-January 1, 2000 came and went without our cars not starting or much of anything happening bad.
My experiences with artificial intelligence so far have all been very pleasant.
View attachment 409949
You need a GPS with AI that won’t even tell you how to get there.
“Mr. GPS, I’d like to go to Chester”.
GPS replies, “Did you bump your head when you got out of bed?”
I ran across an alcohol retort cleaning of a pipe.
it involved a stock pot filled with ever clear, a pipe dropped in. and the whole kaboodle brought to a roiling boil.
it involved a stock pot filled with ever clear, a pipe dropped in. and the whole kaboodle brought to a roiling boil.
I’ve used the boiling alcohol retort a number of times and it has never involved dropping a pipe in boiling alcohol. The retort is attached to the stem using a rubber tube. The chamber is filled with cotton, and you use a rocking motion to move the boiling alcohol back and forth between the chamber and the retort. Works great for removing tars and oils from the briar, but total immersion? No way.
Listen to sablebrush52!
What you "ran across" has BAD IDEA written all over it. In the first place Everclear grain alcohol is HIGHLY flammable. To fill a stock pot with it and bring it to a rolling boil is to risk (at a minimum) a fire. In the second place, this would probably strip every bit of finish and patina off of the exterior of the pipe.
I can't wait for all the zoomers who graduated by copy/pasting AI to start serving as judges and doctors and airline pilots.
My favourite use of AI is when folks selling estate pipes use Chat GPT to craft a product description. I’ve seen dozens of variants of "This briar smoking pipe comes in a sleek design"...
Still, I much prefer that to when they actually try to use it to research the brand and value. "Your pipe likely dates to XX-YY and is a valuable collectors item." Yeah, right.
Still, I much prefer that to when they actually try to use it to research the brand and value. "Your pipe likely dates to XX-YY and is a valuable collectors item." Yeah, right.
Legislation becomes, at once, both a limit and a source. Much of the information we pour into the internet does not arise from free choice, but from the obligation or coercion imposed by governments in their relentless drive to control the population. The use of digital money, the mandatory registration of photographs and biometric data for every citizen, the proliferation of cameras and tracking systems—all of this decisively feeds the databases that sustain artificial intelligence algorithms.
To this we must add compulsory online procedures and digitized records that are freely shared with government agencies that have nothing to do with the original purpose of the process. Artificial intelligence is nourished, to some extent, by the voracious appetite for control displayed by modern states, which turn every piece of data into another cog in their machinery of surveillance.
Let's keep the PMSS 2025 Participants out of this, shall we?
Taking this warning seriously, I would say (wild guess) there is a direct link between the dramatic drop in overall human creativity and the increase of IT use in general.
I would also be curious to consult recent global analysis of individuals average IQ now and, let's say, just 50 years ago. To see how it compares. Probably scary.
I would also be curious to consult recent global analysis of individuals average IQ now and, let's say, just 50 years ago. To see how it compares. Probably scary.
It's not impossible for it to be a useful tool but more as a search engine and needs quite lengthy training and constant supervision, so it's practically useless for people who don't already know at least 80% of the information they are looking for.
I, for one, am eagerly anticipating the AI bubble to burst. Hopefully spectacularly.
Oh sure, with AI you absolutely get out what you put in.
Actual human output from someone knowledgeable is always going to be at least a little better if for no other reason than because of the human element. For people who are experts and can already produce any one specific thing, AI can be useful as a way to seed up the process, or shorten the boring bits, etc. because they know what a good result looks like and can guide the AI to reach that point.
But that’s usually turning an 8 hour process down to say 2 hours, not a 5-minute job. AI isn’t really a replacement for anything if you have no clue about a topic and one-shot your way to some output. Sadly that’s practically 80% of the visible use of AI on the web today.
Last edited:
I couldn't have said it better. AI is an incredible tool for almost anything, but it's been over hyped and people are expecting it to do impossible that NGS. As soon as you spend a bit of time with the tool, get familiarized and lower your expectations, then and only then it turns into a great and useful tool.
One thing that often happens even to reasonably experienced people is that they sort of lose focus. People are rigorous with detailed prompts and monitoring what the AI is doing for the first few days or weeks after a new model is out. Later they relax and get complacent, start to expect that the AI should somehow "understand" them better because of their previous conversation history. They stop investing the time in properly prompting the AI and monitoring it, course correcting etc. And then surprise, they get worse results.
All of a sudden you get these reports on Twitter how this or that model "used to be great but was throttled/dumbed down/diluted" while in fact people’s behaviours and expectations of that model shifted during that time period and they altered their process. Sure, sometimes the free services in particular get throttled and their capabilities diminished, but this isn’t true for the vast majority of cases.
Sadly the truth is quite boring - it’s about building a process and then optimising it. There’s no magic, just trial and error. I feel that if AI models weren’t so prone to alter or insert random numbers and couldn’t be trusted with that sort of data, people like accountants or statisticians or analysts of all sorts would have figured out and publicised many of these optimised processes, and then maybe more people would have figured it out, and we would have less of this magical thinking.
My personal opinion regarding AI and LLMs is that they do not offer significant value to work that justifies these large tech companies spending billions on the technology and putting up data centers left and right without oversight. But because big tech has gambled so big on this they're force feeding AI to everyone. Now I cannot control what my employer wants us to use, nor can I control whether Apple chooses to force their AI onto my phone, but I did jettison Windows for Linux and I havdnt been happier.