Anyone else SICK & TIRED of hearing about "AI"??
- Thread starter Stress_Test
- Start date
Help Support Candle Power Flashlight Forum
knucklegary
Flashlight Enthusiast
So, I've been having fun talking with "Rufus" on Amazon. When you ask questions AI cannot answer you'll get same reply every time "sorry but not enough input to give an educated answer" Actually the AI has saved me time.
Asked who is Marjorie Taylor Green. Great unbiased response 😂
Asked who is Marjorie Taylor Green. Great unbiased response 😂
PhotonWrangler
Flashaholic
Sometimes Google's AI takes a joke question seriously...
PhotonWrangler
Flashaholic
Found another one. Yum...
I've found that when I ask a question in google chrome that the first response ia an AI generated synopsis response.
Sometimes it is good... other times it would be a waste of time to read it to the end.
Unfortunately at a meeting I attended, the presenter said that there are times that the AI will generate references for what appear to be scholarly articles, and it will present the artificial references as fact. The presenter stated that it is imperative that you fact check the references before you present them in a paper.
The presenter was presenting at a medico-legal conference, so one can imagine how important it is to remain factually correct.
Sometimes it is good... other times it would be a waste of time to read it to the end.
Unfortunately at a meeting I attended, the presenter said that there are times that the AI will generate references for what appear to be scholarly articles, and it will present the artificial references as fact. The presenter stated that it is imperative that you fact check the references before you present them in a paper.
The presenter was presenting at a medico-legal conference, so one can imagine how important it is to remain factually correct.
PhotonWrangler
Flashaholic
Yes! And there have been several cases where attorneys used AI to come up with citations of case law, only to discover (while in court) that the "references" were completely fabricated.
AI tools are fine, and the underlying technology is sound, but rather than investing in proper education for the neural networks we are relying too much IMO on automated training - basically locking a three-year-old in a room with a web browser for 18 years and wondering why they come out with a skewed understanding of the world.
I am oversimplifying a bit but not that much. We don't necessarily need humans in the loop forever (properly trained networks should be able to do a lot of the work eventually) but if we don't have really smart people probing around in the middle layers of networks looking for misunderstandings (or revelations) this is going to be a rocky ride.
It's only 2024 and I am already totally sick of AI-generated artwork to the point where I get a nails-on-blackboard reaction whenever I see it...
... and this is from someone who has been looking forward to practical AI and working around the edges of it since I first stumbled across Frank Rosenblatt's book in the local bookmobile when I was 8 or 9. Finally retired but the last project I worked on was the AMD MI300:
I am oversimplifying a bit but not that much. We don't necessarily need humans in the loop forever (properly trained networks should be able to do a lot of the work eventually) but if we don't have really smart people probing around in the middle layers of networks looking for misunderstandings (or revelations) this is going to be a rocky ride.
It's only 2024 and I am already totally sick of AI-generated artwork to the point where I get a nails-on-blackboard reaction whenever I see it...
... and this is from someone who has been looking forward to practical AI and working around the edges of it since I first stumbled across Frank Rosenblatt's book in the local bookmobile when I was 8 or 9. Finally retired but the last project I worked on was the AMD MI300:
I won't use Windows 11 because of AI, not that I use much Windows since 2013. I don't have a need for AI and I certainly refuse to pay for any AI.
The internet is nearly useless as a learning resource at this point. I don't see AI helping any and bad actors will certainly sabotage it for trolling or other purposes.
The internet is nearly useless as a learning resource at this point. I don't see AI helping any and bad actors will certainly sabotage it for trolling or other purposes.
Have to disagree. You'd be amazed how many legit tutorials you can find on YouTube, instead of just cat videos or ASMR channels of young women showing off 95% of their clev, making 20 videos and having 60K subscribers. Meanwhile on my ASMR channel, not like I can have part of one of my "stones" hanging out and get tens of thousands of new subs.
Okay getting distracted. I learned how to maintain, take care of, and take apart a chainsaw thanks to watching vids. made by professional arborists. Heck, I've taught some of what I know on my main channel regarding firearms use and safety.
Yeah, you've got to dig through a massive mountain of garbage, filth, and manure to get to the truly useful content on YouTube. But it is there.
Yes, those are annoying. Unfortunately content creators used to have some control over Ads. and when they'd appear in a video. But YT took those controls away awhile back.
Up until about a month ago, I was successfully using an ad blocker. Youtube put an end to that. It's understandable.
I use the Opera browser with Ublock Origin plug-in on Windows 11. No ads on YT. Don't do AI.
Yeah, and working with IT, I am already VERY tired of hearing about AI this and AI that.
Worst thing is, it's still only in its infant stages, and we're already being massively bombarded with it. >-<
"AI will very confidently be wrong" is a phrase I have come to embrace. And I fear, when AI technology is actually mature, how can we trust it then, if we can't trust it now, when it's being pushed so heavily on us. But I guess the various marketing departments, as they always do, will very smoothly gloss over any prior small incidents involving deaths, "beause the owner of the self-driving car at the moment of impact had not updated his or her AI software".
My current potential AI interactions:
- Upscaling and frame generation in games. Not interested. Like, not interested to the point where the words aren't invented yet.
When my GPU can't brute force the desired resolution and the pixels I want, it's time for an upgrade.
- Local data and web searches. I'm still on Windows 10, so fortunately I'm not voluntold to use MacroSuck's AI tools whether I want to or not. Not yet, anyway. But my experiences with any AI online searches is they are completely off the rails in the wrongness.
Example: I googled a Kärcher window vacuum cleaner that I am considering, and I wanted to see an overview between 2 similar models.
With a quick search, the first answer was from an AI, telling me the difference was one model came with one battery, and the other came with 2 batteries. That was total BS when I checked the manufacturer page - both models came with 2 batteries, the AI only said one model came with a single battery because that was how it was mostly depicted online.
- Fake and AI manufactures scam E-mails. We are seeing more and more of them at work.
Fortunately, this is not much an issue for me directly, as I am completely paranoid of anything and anyone across the vastness of the human-created interstellar medium called the internet. The paranoia has its downsides, but it is also pretty much a requirement to safely navigate the interwebs, and not to mention not being a security risk at work.
Worst thing is, it's still only in its infant stages, and we're already being massively bombarded with it. >-<
"AI will very confidently be wrong" is a phrase I have come to embrace. And I fear, when AI technology is actually mature, how can we trust it then, if we can't trust it now, when it's being pushed so heavily on us. But I guess the various marketing departments, as they always do, will very smoothly gloss over any prior small incidents involving deaths, "beause the owner of the self-driving car at the moment of impact had not updated his or her AI software".
My current potential AI interactions:
- Upscaling and frame generation in games. Not interested. Like, not interested to the point where the words aren't invented yet.
When my GPU can't brute force the desired resolution and the pixels I want, it's time for an upgrade.
- Local data and web searches. I'm still on Windows 10, so fortunately I'm not voluntold to use MacroSuck's AI tools whether I want to or not. Not yet, anyway. But my experiences with any AI online searches is they are completely off the rails in the wrongness.
Example: I googled a Kärcher window vacuum cleaner that I am considering, and I wanted to see an overview between 2 similar models.
With a quick search, the first answer was from an AI, telling me the difference was one model came with one battery, and the other came with 2 batteries. That was total BS when I checked the manufacturer page - both models came with 2 batteries, the AI only said one model came with a single battery because that was how it was mostly depicted online.
- Fake and AI manufactures scam E-mails. We are seeing more and more of them at work.
Fortunately, this is not much an issue for me directly, as I am completely paranoid of anything and anyone across the vastness of the human-created interstellar medium called the internet. The paranoia has its downsides, but it is also pretty much a requirement to safely navigate the interwebs, and not to mention not being a security risk at work.
If you are concerned about security use some open wifi with VPN and Tor Browser and use a linux such as Tails or run Ubuntu as a live disk. If anything happens power down. You won't stay infected for long. Anything that doesn't like your connection or needs to know who you are won't have your best interests in mind and will most likely be hacked in the future and lose your data anyways.
