Yes, I’ll just trust the AI to help me fuck around with grid voltage levels.
With google search results becoming so poor, I guess I need to look into kagi or duck-duck. Pain in my dick, motherless goatfucking, horsehit-happy asshole, corporate varmints gotta fuck it all up for more profit. I’m tired, yo.
You must log in or # to comment.
Recommendation - if you are using AI for search, avoid whatever Google is using for “Google” search. It is a highly efficient, fast model that is often far from accurate and easily confused - not to mention its context window is extremely limited.
In fairness to AI, I personally wouldn’t trust forum posts on this subject matter either, only written documentation or a professional.
Only trust the data sheet from the manufacturer. Only if listed in compliance with your use case. Trust nothing else. All other options will kill you
My father just ruined some PCB while soldering because Gemini told him what to do.
I’m sure there’s many more cases that are worse and we simply do not know about.
The days when YouTube had actual people in their cluttered garages, basements, and driveways showing you how to fix shit was good. Between the algorithm fucking it over and AI giving wrong info, we’re likely going to win some Darwin awards en masse, around the house.
I’ve tried it. It’s good for plucking out game solutions (low risk), and finding the right forum for That Linux Workaround. Or just parroting Wiki.
Catch it in a wrong answer and it will answer just as confidently, agreeing that it was wrong. And yet.
And we’re destroying homes and the life equity of working class for this.
To be fair YouTube was full of bullshit DIY videos too. You used to be able to see which ones got lots of downvotes for being bad…but nah, let’s not show that to people anymore
Sensitive Sally’s didn’t like being downvoted. So the losers at Cherry Ave changed the whole site to not trigger mentally ill people.
A species-level Darwin award is the solution to the Fermi Paradox.
Yep, we’re our own great filter.
I think the “Darwin awards” are part of the plan to be honest. These guys in the TESCREAL crowd are pretty much exo-facist malthusians that think the rest of us are expendable.
A Modest Proposal. The Theil credo.
There are still good YT channels. Learn Linux TV is excellent, as is Sabine Hossenfelder, Anton Petrov, Techmoan, and Louis Rossmann.
Yes, but they’re harder to find organically, since the search function seems to be set up to barf out the top 10 most trafficked. For example, I’ve trained my algorithm to never give me 731 Woodworking (looks like a QVC ad, every time, no woodworking), but if I search for new woodworking, that’s what I get, a long list of mostly 731 woodworking.
And, for whatever reason, people love watching re-rolled clips (the poster didn’t create) of whatever with someone making expressions and reacting to it. Worst content ever, and yet.
Related: https://llmdeathcount.com/
Sad to say, but this is a lot lower than I expected? Although I’m sure a vast majority of cases will never make it here.
I’m sure it’s pretty hard to confirm. Especially for indirect causation or partial contributing factors.
I prefer to have the AI overviews and such off with uBlock. AI should be separate from search engines, and should also be regulated.
The AI CEOs should be forced to follow some dangerous procedure exactly as explained by their shit planet-killing product live for all to see. If they’re confident this is the future, they should set the example.
I would rather have them held accountable in the same way that any news outlet or publisher would be held accountable for publishing false information that could have deadly consequences.
It’s only a matter of time before a fifth-grader follows some inane advice because it was the first result from Google and is irreversibly maimed because of it.
Where are all the Helen Lovejoy types when we can actually use them?
It’s only a matter of time before a fifth-grader follows some inane advice because it was the first result from Google and is irreversibly maimed because of it.
https://www.cnn.com/2025/11/06/us/openai-chatgpt-suicide-lawsuit-invs-vis
I think we’re already there in many aspects
Yes, I’ll just trust the AI to help me fuck around with grid voltage levels.
I mean, if you do… you kinda deserve what comes to you.
This assumes people know how AI works, and that’s not common knowledge.
Dude, like a handful of people have any idea how it works, it was completely a black box until recently when they started deciphering how it “thinks” before outputting in any given language. And before anyone argues, yes, researchers have actually started to assess how it processes stuff now, they’ve recorded the binary processing that wasn’t language based and upon suspecting they found the processing akin to what we’d call thinking for ai, they saved the streams, presented it to other ai in droves and asked them to interpret it independently of the other system that manufactured it, and indeed it matched up with the processing of requests, but was not in any human language, it’s indeed machine code, but not like Fortran or cobal or hexadecimal codes that we are used to dealing with, it has its own language. So no one has a library or Rosetta Stone yet to interpret this, and as of now you have to “trust” other AI to tell you what it means. Which obviously isn’t a good idea at all.
Edit: the paper since some of y’all don’t believe it, and frankly I don’t blame you. It’s new research. "Do Sparse Autoencoders Capture Concept Manifolds?” https://arxiv.org/abs/2604.28119
You have no idea what you’re talking about and you’d feel a lot better if you could just accept that. This story you are telling is absolutely nonsense fantasy, stitched together from a bunch of different actual things to create a narrative that tells a completely untrue and fantastical story
Why do you have the exact number of upvotes that the person you’re replying to has downvotes. Seems like an unlikely coincidence. Also weird how you don’t actually address anything in the comment. Almost like some bot brigade.
if someone tells you some computer data can’t be represented in hexadecimal codes, that’s a clear sign they don’t even know how computers work, let alone LLM. its also futile to compare the LLM internal representation of data to programming languages, its like comparing birdsong to building blueprints.
My comment was addressing frisbirds jackassery, not the op.
Hilarious conspiracy brained shit.
Wrong. Here’s one of the papers on the research. “Do Sparse Autoencoders Capture Concept Manifolds?”
https://arxiv.org/abs/2604.28119
and a video on the paper
Your writing shows that you have no idea how to interpret this paper at all. You clearly have no historical concept of the development of machine learning, neural networks, and the fundamentals of the domain. You are trying to read a cutting edge paper that relies heavily on the reader having deep domain expertise. The video you linked is ALSO not for laypersons as it doesn’t actually explain the basics of the domain. You can’t just dive into the deep end of a pool without knowing how to swim and then tell everyone you were lost at sea.
This isn’t even what I meant by “how it works”. I mean that it’s basically a machine that outputs whatever next text that would be the most probabilistic given the previous text and input.
So the basics of how it works. Not exactly how it works.
Most people don’t know the basics of AI and how it works in that sense. They hear AI and think “I Robot” and think the computer is actually talking to them and giving them something that a “brain” has “reasoned” about, which is not the case.
not hexadecimal codes lol. you are clueless.
but also, the point was that for most even a general level of understanding is missing, making them easily think that it is the all knowing machine that cannot be wrong
I think you missed the key part of what the person you replied to said: common. It is not common knowledge. Yes some people who have done extensive research and spent hundreds of hours on this stuff might have a somewhat good understanding of how LLMs and GenAI work. Do you think just your average person using Google has any idea how it works? No, no they do not.
You want to cite some sources because like a lot of things people say about AI this too sounds like bullshit?
no.
these things are sold to us as magical black boxes of universal knowledge. most people don’t know how they work.
blame the ai companies, not the people who fall prey to their lies. why does this have to be said. come on.
Can we stop treating people as mentally challenged toddlers? When AI started, I would have agreed with you about this. But if today you still use any AI and trust what it says, after all we have seen it does, it’s on you.
Am I saying that AI companies are not to blame? No. I think AI companies are to blame for the shitty product they are delivering. But let’s not take any personal responsibility of people. If you plug a fork in an outlet and shock yourself to death, there’s nobody to blame but yourself for being dumb. And this is the same.
This kind of attitude completely ignores the fact that there are vulnerable people who are very susceptible to the sycophantic nature of GenAI. There are multiple reported instances of people being convinced to kill themselves or others by GenAI. AI companies won’t do anything about this kind of thing unless they are forced to.
No. This is absolutely the wrong attitude. AI companies must be held accountable for all the horrible shit AI is doing.
I agree that we should hold people responsible, but I’d argue mentally challenged toddler is much much closer to the average human intelligence than you think. The average person is shockingly stupid.
Can we stop treating people as mentally challenged toddlers?
Yes, but let’s wait until the lead-poisoned generation have been cared for first. And also until those harmed by whatever the brainworm-in-chief FDA is up to have been cared for. After that, let’s expect everyone to be able to think things through.
If you plug a fork in an outlet and shock yourself to death, there’s nobody to blame but yourself for being dumb.
And yet, we still design outlets in ways that make that exact kind of thing as hard as possible. Because there will always be kids who have no idea what they’re doing. Because there will always be old people unfamiliar with the technology that “everyone” should know how to use by now. Because accidents happen.
these things are sold to us as magical black boxes
Dawg EVERYTHING is sold to us that way… Well not everything, but you get it.
Learning to spot “marketing” BS is a survival skill at this point. People have to learn.
What I’m saying is, the way AI is done is immoral and wrong, but we need to be able to spot “immoral and wrong” from a very young age if we’re to make it to old age. AI is just one such instance that can get you messed up or even killed.
Learning to spot “marketing” BS is a survival skill at this point. People have to learn.
survival skills are inconvenient, like being responsible for your kids and your own choices, so you can forget the average person to bear that.
No, no you don’t.
I don’t use Google and I agree with the energy. But every electrical product I’ve ever purchased has come with a manual, which gives all of the safety information.
So I guess read your manuals and don’t go rage searching on Google
Until you lose your manual and need to look up the information online, hoping to find the manual or relevant info you are first presented with this.
lots of people wont search any further than the AI summary.
I hate AI too, but reading and parsing manuals is one of the few things it’s actually good at. The linked sources are right there if OP wants to confirm
It’s not so great when the only available digital manual is a jpg scan from a printed version. There’s also no way to tell what other sources an LLM is using when it generates its summary, and when you’re dealing with things that are important to get right, like voltage, you have to just read the manual/specifications.
I’d hope if you’re in a job where you regularly poke around line voltage and higher, you’d be smarter than to think your little meter can handle transmission voltage.
I’m more concerned about people who don’t do that regularly and think ‘don’t worry; I’ll ask Google’
deleted by creator
Pain in my dick, motherless goatfucking, horsehit-happy asshole, corporate varmints
Off topic, but that’s a nice spattering of invective you got there, OP. Well done. I genuinely appreciate a well-crafted compound insult.
forget about just being outright wrong all the fucking time, you can ask this piece of shit a very simple yes/no question and it will change its answer constantly when you do so much as refresh the page. its actually baffling to me. duckduckgo has search ai, and somehow theirs works fine enough %90 of the time and won’t change its answer with each page refresh. DDG is doing search ai better than Google, the third biggest company in the world who may as well own the internet, AND YOU CAN JUST TURN THE DDG AI OFF. Big tech are so embarrassingly incompetent.
Just switch your search engine already.
I can vouch for Kagi, although I wouldn’t say their AI is that much better. But if you find like 5 other friends the family plan is worth it.
How does kagi define “family”? Like netflix?
afaik its just a group discount plan
I ask because having an inclusive definition of “family” being whittled down eventually to a single IP address is a pretty common trope of enshittification…
Anyway, itś $20/month, so if I had some nerd friends, I’d consider it.
I think one beer every two weeks at the pub would cover their contribution.
Having a VPN for the initial purchase will yield you even less. Not sure about which country had it the cheapest, but IIRC the lowest was like 12€.
time and won’t change its answer with each page refresh.
Maybe it just recalls the answer for at least a few hours?
Seems like a good way to improve precision without improving accuracy.
It’s very common to cache results for LLM because it saves a ton on computation costs.
Duckduck with ai turned off is pretty decent.
Duckduck go is just an external marketing venture for bing.
Just remove the ai element with ublock origin filters on whatever search engine your used to.
If you care about privacy spin up a searxng which grabs results from all configurable search engine without accepting bribes for special tracker acces.
This will still waste all the energy. Just use the noai version
Add &udm=14 for a google search and it doesn’t
Local searchxng doesn’t have ai features to need any noai. It still the superior choice.
Though you could argue that cause searchxng queries multiple search engines at once this is a bigger energy costs then just the one. But you could configure it to only search by you deemed ethical engines.
Duck duck go just isn’t a choice, you and others can still choose it but i have yet to find even a single benefit to it. The privacy is fake and the results are meh.
(Remember when “no one wants to use bing” was a meme?)
IronFox has a DuckDuckGo (No AI) mode.
noai.duck.com You can use it in any browser
- has already killed many
when will they be held liable for this shit?
That’s the neat thing: never, because they explicitly tell you that big babble machines make mistakes and you’re responsible for any actions you take on that.
Don’t be so hard on Gemini.
It is accurately warning you that this meter is safe to use on 277 volts (actually 300 - but that’s an odd voltage) and below. It’s telling you not to use it on 480+,1 kv,…
Ignore the prompt request for safe use at your own peril!
Fuck ‘gemini’. I don’t trust a glorified chat bot with life and death advice. Don’t be simping for a chatbot ran by an objectively evil corpo.
I don’t care what the actual max voltage is. That search was a shortcut to find the manufacturer spec summary while looking at third party probes. Not ‘gemini’s’ summary or anything else. I don’t want electrical, romance, or life advice from a chatbot. I want links to relevant resources.
Bro, you use google
Can’t help you there buddy.
Gemini correctly summarized the limitations of that item. If you can’t read that, then that’s on you. Gemini actually asked for your specific parameters. One might surmise that you shouldn’t be allowed near high voltage things to begin with if you can’t read basic limitations and refuse clarifications from Gemini.
- 32 year hvac/refrigeration tech.
I personally don’t give a shit wherher you trust Gemini or not. But it seems to me that you need remedial help with basic English and IEC standards.
Sorry. That’s not on Gemini. That’s on you.
you are batshit insane if you are genuinely arguing so confidently that OP should have just believed what gemini responded with. this is not about reading comprehension you dumbass. go see a doctor.
Gemini was 💯 % correct.
No I’m not batshit crazy. I can read a simple spec sheet. Been doing it for a few years now.
I don’t give a fuck if it summarized the specs correctly. It’s been proven to be wrong often enough that trusting it to be correct when dealing with life and death issues is foolish.
I searched a basic spec to get a link to a manufacturer spec list. The bot offfered to vet my setup for safety. It seems you’re suggesting that it’s a good idea to listen to a chatbot for high voltage safety advice.
I’m not the one with the comprehension issue here.
I didn’t ask for high voltage safety tips from a chatbot. It offered them. If you don’t think that’s dangerous and wildly irresponsible, you’re being disingenuous at best and willfully ignorant at worst.
- 25 year machinist including maintenance
It’s not me that’s going to be killed. it’s people that don’t know enough to not trust a fucking chat bot. Do you really trust a bot to explain ground loops, isolation transformers, and floating test equipment? If so, I feel real bad for any kids you might be training.
So just scroll the fuck down and click on the link to the manufacturers datasheet.
Did Gemini disable your scroll wheel?
You admit that the summary is correct. Couldn’t you have found an example where it’s giving bad advice?
If it kills some stupid people who are too lazy to check out other sources, who honestly cares, the world could use that.
So just scroll the fuck down and click on the link to the manufacturers datasheet.
that won’t save other people. don’t be afraid, OP is not fearing for himself, but he fears the safety of others. if it wasn’t entirely obvious.
but probably gemini just disabled your reading comprehension, so don’t even bother responding.
If it kills some stupid people who are too lazy to check out other sources, who honestly cares, the world could use that.
those people are not lazy but gullible. I don’t know what to say if you don’t think at least 60% of people believe gemini knows what it says.
I said that the were stupid and lazy, you don’t think gullible was implied there?
but probably gemini just disabled your reading comprehension
Says the person who thinks that
“but probably”
Is the way to phrase a written argument.
😭
Absolutely, totally trust it. It will give me multiple citations. I can look them up.
It was pretty specific! DON’T USE THIS ABOVE 300 VOLTS! AI: “If you have any questions about how you’re going to use it, just ask me”.
AI Mode All Products Visual matches Here is a highly technical, multi-layered question about nuclear power plant operations, thermodynamics, and safety systems. The Challenge During a sudden Load Rejection (a complete disconnect from the electrical grid) from 100% full power at a Westinghouse 3-loop Pressurized Water Reactor (PWR), the turbine control valves slam shut immediately. Assuming the Reactor Trip System (RTS) fails to automatically actuate due to a mechanical binding of the control rods—initiating an Anticipated Transient Without Scram (ATWS) event—answer the following three-part question: Thermodynamic Feedback: Explain the immediate (<30 seconds) chronological response of the Core Reactivity (Δ ρ). Specifically, address how the competing effects of the Fuel Doppler Temperature Coefficient and the Moderator Temperature Coefficient (MTC) interact as primary coolant temperature and pressure rapidly spike. Overpressure Protection: The Pressurizer Power Operated Relief Valves (PORVs) and Safety Valves will lift to mitigate the primary system pressure spike. If the primary pressure exceeds the secondary side steam generator pressure, what critical heat transfer phenomenon is threatened on the primary side of the steam generator tubes, and what are the implications for fuel cladding integrity? Emergency Mitigation: Since the control rods failed to insert, operators must initiate Emergency Boration. Why is a highly concentrated boric acid solution injected into the Reactor Coolant System (RCS) effective at shutting down the fission chain reaction even when the coolant is at peak transient temperatures, and which specific isotope is doing the heavy lifting? Take your time to break down the neutronics, thermal-hydraulics, and plant systems involved. When you are ready, let me know your answer or which part you want to tackle first, and I can validate your response or provide the detailed technical solution.
Do you understand that?
Is that Gemini’s fault that you didn’t understand that?
AI Mode response is ready
i recently found ecosia having better results than duckduckgo, give it a try. qwant also seemed ok but they geoblock me so i need a vpn to search there…
for keyword searching obscure stuff etools tend to work well for me but they have some weird system where you gotta refresh a couple of times to have results actually load properly
- Why would you google that instead of looking at the datasheet or at the scope where it tells you what the inputs can take?
- If you dont understand how oszilloscope inputs work, what impact frequency has on the inputs and when to use x1 or x10 probes dont operate it? Working with high voltages without knowing shit was always unsave ai wont change that no matter if its web search or ai, if you need to google that instead of looking at the datasheet you should not try to measure high voltages.
It’s a shortcut for pulling up the manual, 240ish pages long. I’m looking into probes, not currently fucking with 300 volts. Used to be, ask for max specs on a component and get a link to the manufacturer’s or sellers summary.
I do appreciate your shitty tone and assumption of my ignorance. While I try and stay under 480V grid power, I do occasionally have to mess with it in a professional capacity. I’m safe, but someone who doesn’t understand what they’re playing with gets a free ride on the lightning.
When did google search ever gave you random info of the manual? I hate ai as much as the next guy but this isnt an ai problem. Also you dont have to look at the manual, it says what the inputs can take right on the front of the device besides the bnc connectors.
Are you being deliberately obtuse, trying to ‘win’, or an llm fanboi?
The scope isn’t even sitting in front of me. I’m just casually investigating the tooling. Looking at cheapo generic probes, wanted the manufacturer spec sheet/sales page, not the full manual. Used to be, a search like that would return a link to the manufacturer’s page showing specs and documents on the product. Instead of going to rigol’s site, using their shitty search or selecting submenus. It’s not me that gets killed with a search like that, it’s someone who doesn’t know what they’re doing.
God someone saying that something isnt a ai problem but a problem of knowledge or workflow is a “ai fanboy” ? Never ever did google point you to a specific page of a manual when searching for some value of said manual. Why would you even search for the max voltage if you want a the product overview/spec sheet?
This just looks like you searched for a term where you knew the google ai will give some output so you can post here.
And no if someone didnt know what they are doing ai will not change that they shouldnt fiddle arpund with high voltages. Again the problem isnt that someone looks at ai the problem is doing stuff with high voltages when they dont have a clue. Look at homy many people die because they fiddle around with microwave transformers every yeat. Look at how many bad youtube shorts there are about building random circuits that are unsafe that has always been a problem and isnt sppecific to ai.
This just looks like you searched for a term where you knew the google ai will give some output so you can post here.
And there it is. What’s your agenda? Be honest.
Never ever did google point you to a specific page of a manual when searching for some value of said manual.
That’s a fucking lie! You’re lying.
I was drinking my coffee, having a lazy morning, researching the capabilities of my new-to-me scope while sitting in my study. Scope is in a different room and didn’t come witn a paper manual.
User manual, programming manual, and other technical docs are in the reference section of my self hosted library which is currently down, sd card or the board is on the fritz. (Calibre runs best with a desktop environment so it’s on it’s own little orangepi board running ubuntu or something in a gui that I remote into.)
Looking at third party probes and wanting a quick reference for the input specs. Wondering if I can get away with a 50ohm bnc cable to resistor for the external trigger or if I’m better off grabbing a cheap probe.
I’m not the one being dishonest and searching for something like this is perfectly reasonable and shouldn’t result in a personified bot trying to help me fuck with lethal voltages. That’s fucking crazy and defending that is a hell of a choice on your part.
What’s your agenda? Be honest. You got me, i only disagrer with you because i got paid by big llm, wtf is wrong with you?
Never ever did google point you to a specific page of a manual when searching for some value of said manual.
That’s a fucking lie! You’re lying.
Nop i am not. Manuals for measurement equiptment or electronic components are always long pdf docs with 50+ pages and never ever did google give you the pdf from a search. Best you could ever get was the overview site of the manufacturers that is still the first link after the sponsored links in the search at least for me. You never could google a random spec of something like this and get to a page of the manual. Only to something like the general overview. The rigol overview doesnt even display your input specs for your measuring inputs and specific information like the input levels the io of the external trigger are not stated there and like i said would always be in the docs. The datasheet is one of four documents for that device where the max input voltage of the trigger is shown. Then you have a programming manual a service manual and a user manual with many pages that google would have never “just” given you with your search term. Dont pretend that you could have googled that 10 years ago and got the pdf doc in your search.
My point still stands that this is not some ai error that kills someone because ai gave specs for the inputs. Is it not common knowledge that working with high voltages is a no go if you have no clue?
Is it not common knowledge that working with high voltages is a no go if you have no clue?
LLMs, the things advertised as the all knowing universal knowledge machine, make laypeople confident, as if they have had a clue on the topic. It presents the info confidently, and the layperson won’t know when it is wrong.
never ever did google give you the pdf from a search.
What? of course it did, that’s what made it such a good search engine.
It’s dangerous because non tech folk might not understand that AI results are not trustworthy. They’ll just look at it as a faster way to look up the “same” data in the data sheets.
Very, very dangerous ignorance.
But it doesnt matter if they use ai or web search if they dont understand the bare minimum about their equiptment. If you cant even look at the input rating that is written right besides the inputs on the device you are already fucked.
Fair point, but surely that’s not the only situation? Maybe you’re planning work and aren’t right by your equipment, and you’re making decisions based on AI output. Danger.
Yea sure then it would be dumb? But this isnt really the case? A person who doesnt know that high voltages are dangerous shouldnt work on high voltages, also the ai output isnt even wrong in this case. If you thing that something valuable comes up with this search its already a problem a random site wwont display sspecifications for your device you should always look at the datasheet. Searching for “max voltage” instead of going to the manufacturers site and/or downloading the datasheet is not really the right aproach.
Its not like op searched “can i connect a 115v device to 230v” and ai giving the wrong awnser. There are many reasons and examples where the ai search is bad, this isnt one of them people dont need to invent cases for ai bad when there are enough real ones out there.
Yea sure then it would be dumb? But this isnt really the case?
This is easy for you to say, because you have sense. It can somehow be hard to relate to how fucking stupid people can be, even if they have proper domain knowledge, and how much trust people put into technology without questioning it’s veracity.
