Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Artificial Intelligence In Our Lives
- Thread starter Huntn
- Start date
- Sort by reaction score
Abydos
macrumors 6502
I'm retired so AI serves little if any purpose. I used neural nets back in the day, so I realize they have some uses - with little data in an area you will probably get a SWAG (sh*tty wild a** guess) maybe better than a ouija board maybe not
so they are useful in areas where we have sufficiently dense information and potentially dangerous in areas where we don't - the trick is knowing which is which -- lots of luck guys
so they are useful in areas where we have sufficiently dense information and potentially dangerous in areas where we don't - the trick is knowing which is which -- lots of luck guys
BSDnostalgia
macrumors 6502
turbineseaplane
macrumors Penryn
What's wrong with the link?
Loading fine for me here (I don't pay or subscribe to Fortune or anything)
'It's like we don't exist': Nearly 50,000 Lake Tahoe residents face power loss as utility redirects lines to data centers | Fortune
Roughly 49,000 Lake Tahoe residents could lose 75% of their power after their energy provider said it's directing energy to neighboring data centers.
Also, and I apologize for singling you out, but a simple google search with "lake tahoe power loss" will yield dozens of articles. It really frustrates me when folks don't do a simple search on their own. It takes less time than asking someone else for links.
Again, I apologize to you. I'm not trying to be rude, I promise. 🙏
🔎 lake tahoe power loss - Google Search
www.google.com
Pacific1972
macrumors regular
Last edited:
baconcat_8964
macrumors 6502
skodises
Contributor
Speaking strictly for myself: I refuse to have anything whatsoever to do with LLMs, either at work or in my personal life. Luckily, I work for a company that is in agreement. They are attempting to put a blanket ban in place, to the point of downgrading PC users back to Windows 10. I’m lucky in that my old OCLP’d 5.1 on Sequoia makes this avoidance fairly easy, and my Parallels Windows VM never got upgraded to W11 to begin with.
Getting rid of LLM interactions on my iThings is harder, since the damned things upgraded themselves to iOS 26 before I realized the extent of this exposure- but I’ll figure something out.
The company’s rationale is simple: since no LLM provider can credibly guarantee that our proprietary information will not be quietly sponged up, essentially stolen, and then used for training, they will not permit said information to be exposed to that hazard. I can’t disagree.
We know that avoiding this hazard is becoming harder and harder, especially as e.g. Microsoft jams their LLM hooks into every piece of software they market, but we will still do our level best to try. As a result, our brief dalliance with cloud storage is coming to an end, even as we speak. The marketing people aren’t happy about it, but the engineering and product development teams are fully onboard: engineering refused to put our design data into the cloud from the outset.
I fully realize that I’m an extreme case, and that few will agree with my position. So be it. Frankly, the inevitable LLM-is-God collapse can’t come soon enough. I’m well past retirement age already, and I personally want to see it die before I do. But your mileage will most certainly vary. Be safe out there!
Getting rid of LLM interactions on my iThings is harder, since the damned things upgraded themselves to iOS 26 before I realized the extent of this exposure- but I’ll figure something out.
The company’s rationale is simple: since no LLM provider can credibly guarantee that our proprietary information will not be quietly sponged up, essentially stolen, and then used for training, they will not permit said information to be exposed to that hazard. I can’t disagree.
We know that avoiding this hazard is becoming harder and harder, especially as e.g. Microsoft jams their LLM hooks into every piece of software they market, but we will still do our level best to try. As a result, our brief dalliance with cloud storage is coming to an end, even as we speak. The marketing people aren’t happy about it, but the engineering and product development teams are fully onboard: engineering refused to put our design data into the cloud from the outset.
I fully realize that I’m an extreme case, and that few will agree with my position. So be it. Frankly, the inevitable LLM-is-God collapse can’t come soon enough. I’m well past retirement age already, and I personally want to see it die before I do. But your mileage will most certainly vary. Be safe out there!
Apple fanboy
macrumors Ivy Bridge
That’s an interesting point. Our MD did a report yesterday (more to show us the ease of doing said report in ChatGPT). But the sales data he shared with the system to create said report would not be something we would share outside the business. Is that data going to be available (at a price) to our competitors at some stage? Either explicitly or not?
Interesting times.
one more
macrumors 603
This reminds me of introducing computers on a mass-scale some 40 years ago. The ideal back then was that computers would be able to do handle tasks much faster, thus greatly improving human efficiency and giving people for more leisure time. This worked well in some cases, yet as the tech kept on developing, people ended up spending even longer hours working. As for AI, I find it very useful for tasks as translation, proof-reading or generating images to the given specs, however, imo, it cannot be truly creative and come out with something “out of the box”, as all it currently does is reusing already available knowledge.
one more
macrumors 603
You can disable Apple Intelligence in Settings > Apple Intelligence and Siri, also saving some 10GB of storage space while doing so. 😉
FeliApple
macrumors 601
I agree, completely, with most of the criticism that has been uttered in this thread, and I think you captured the sentiment well. In fact, I have nothing to add to this, it’s simply perfect.
Having said that… the people in charge of AI don’t help their cause either. I have never (not even with divisive politicians, to be honest) seen such a viscerally negative reaction to a person like when Altm*n said what he said about how many resources humans “consume” vs AI.
That level of sociopathy, psychopathy, and all of the “pathies” together generated such an overwhelmingly negative reaction that I had never seen before with any topic. I don’t think I saw a single statement of support.
In any case, the technology may cause (or is already causing, to be honest) a massive loss in the quality of interpersonal relationships (social media has had a horrific effect on this too), an enormous loss in emotion, awareness, inspiration, etc, like you said.
And in my honest opinion… that is the exact opposite of life. If you take away discovery, inspiration, emotional awareness and connection, you take away tenets of life.
cjsuk
macrumors 68020
The thing that gets me is the LLMs amplify two of the worst characteristics of humans.
Firstly, starting with a solution and not a problem. I see this so often. People don't rationalise their problems and then work out the most appropriate solution. They start with an LLM as a solution and find problems that fit it. This inevitably leads to both not solving the problems they actually have and creating more expensive ones in the process.
Secondly, a little bit of knowledge is dangerous. I am a domain expert in a couple of quite distinct areas. I've seen people try and encroach on them "amplified" by an LLM. They are quite embarrassing really because the LLM does a half-arsed job of them at best, sometimes abhorrently bad. Because the results are quite convincing sounding, the people doing it don't actually understand enough to be able to qualify that they aren't good. And then they get pissy with me when I tear the answers to bits. I have to assume that is the case for many other domains. You really can't beat experts with years of experience. And you aren't going to have any of those in the future if you give people these tools because they stop people from actually learning and play on our instant reward traits as a species.
I don't think I've found a use case I'm happy with yet. Thus I shall continue to not use them.
Firstly, starting with a solution and not a problem. I see this so often. People don't rationalise their problems and then work out the most appropriate solution. They start with an LLM as a solution and find problems that fit it. This inevitably leads to both not solving the problems they actually have and creating more expensive ones in the process.
Secondly, a little bit of knowledge is dangerous. I am a domain expert in a couple of quite distinct areas. I've seen people try and encroach on them "amplified" by an LLM. They are quite embarrassing really because the LLM does a half-arsed job of them at best, sometimes abhorrently bad. Because the results are quite convincing sounding, the people doing it don't actually understand enough to be able to qualify that they aren't good. And then they get pissy with me when I tear the answers to bits. I have to assume that is the case for many other domains. You really can't beat experts with years of experience. And you aren't going to have any of those in the future if you give people these tools because they stop people from actually learning and play on our instant reward traits as a species.
I don't think I've found a use case I'm happy with yet. Thus I shall continue to not use them.
skodises
Contributor
Already done, of course. But even that workaround still leaves open questions, questions that hearken back to the early days of Siri, Alexa, and the various other similar mechanisms that quietly ship one’s casual data pff to some waiting mothership for “interpretation”, presumably for one’s convenience: how do you know and trust that “off” really means off?
With the various providers rushing headlong into this brave and foolish new world, I find that I can no longer use the word “trust” in this conversation. The LLM solution-without-a-problem is growing unchecked and unbalanced- and I’m not aware of any packet-logging/filtering mechanism for the iThings, akin to Little Snitch and the like for MacOS. If there is such a thing, I’d like to know about it.
I can’t recall a case of so little forethought having been applied to a new and potentially dangerous technology at any previous point in my career. Perhaps I’ve forgotten it, or blocked it out. Even the onset of ubiquitous computing, and the resulting September That Never Ended, wasn’t this disturbing…
Last edited:
turbineseaplane
macrumors Penryn
And sadly, I’m not sure we can even trust Apple here.
After all, they just settled the lawsuit about their false advertising related to Apple Intelligence and the release of the iPhone 16.
They have some work to do to win trust back.
Partnering with Google, I’m not sure is going to help.
skodises
Contributor
It certainly has been instructive to watch Google’s wholesale and downright gleeful abandonment of their old “Don’t Be Evil” motto, hasn't it? Greed is a powerful force.
You are right to be worried, search on "llm replicating verbatim text from book"
The Augmented Idiot itself admits it's a plagiarism machine. "Recent studies show large language models like GPT-4 and Claude 3 can reproduce significant, near-verbatim portions of copyrighted books from their training data. This ability to recall substantial content, sometimes up to 96% of a book,"
One of the links is https://ai.stanford.edu/blog/verbatim-memorization/
Since your MD sent the super-secret sales data to the LLM, a competitor who queries the LLM to ask for the estimates of your sales will likely get a direct dump from your MD's inputs. After all, what data could be better?"
There is an idea for a new job field. AI Data Poisoner; Intentionally poisoning LLMs with false data to mislead the competition.
Abydos
macrumors 6502
it's an old idea, first documented use at the battle of kadesh (there may be a better spelling) in ca 1276 BCE - I think the general term is disinformation (from a soviet conference ca 1924)
FeliApple
macrumors 601
Russia famously uses the so-called Firehose of Falsehoods (whose origin was conceptually coined in the Soviet Union).
cjsuk
macrumors 68020
Correct. You should not trust any entity in this space at all.
Not because who they are. But who they might become. And when they have all your data from who they were, you're screwed when they become something else.
Absolutely correct. The problem (I think, anyway) with AI lies far beyond what it is doing with the environment & automation of work.
AI reliance is fundamentally affecting human to human relationships, interactions, and understanding. The profound effects of AI on youth development may not be fully realized for several more decades. It is altering fundamental views of reality.
Idiocracy was a comedic warning as to what human society can devolve into when pure thought and creativity are willingly given up for convenience. AI is a wonderful tool in the right context, but it is also extremely dangerous if it takes the place of human intelligence, which it seems primed to do (with younger crowds… perhaps not so much these forums).
My opinion, of course.
thejadedmonkey
macrumors G3
Not just AI, but tech in general. I made the decision long ago that I do trust Microsoft enough, and Apple too.
Microsoft, for all of the crap they've done, I still trust to safeguard my information. Their legal and governance teams seem to be relatively the same, and due to their hands in so many HIPAA and Military contracts, will always need to have the capability to keep things secure. They never promised absolute privacy, but have always been straight forward with me about how they use my data, and how they don't. I can at least respect that.
Apple, on the other hand, I absolutely do not trust anymore. They were making all of the correct decisions, but at some point some MBA somewhere realized that privacy is a part of their product, and any product can be sold to the highest bidder. So now, because e.g. McDonald's is willing to pay more than I am for maps, Apple can sell my privacy to McDonald's, and I am not OK with that.
There is the idea of The Great filter, associated with The Fermi Paradox, that proposes that civilizations destroy themselves before reaching the stars. I propose that management of ourselves is the issue. Technological advancement has great advantage as well as a great dangers. The killer, this is multiplied exponentially when individualistic, short sighted, selfish humans, use automation for profits at the expense of other humans. We are not using our gifts for the group, but to compete against and disenfranchise others, to gain advantages for ourselves, apparantly blind to the civilization big picture. Our team is no longer our nation, but hundreds of competing corporations trying to outdo others.
You think you are part of the team, until technology eradicates your livelihood, or your corporate masters outright disenfranchise you by exporting your jobs, or turning them over to AI.
The key here is our collective mindset. Are things going great or are they not? Pensions gone, cost of living rising, cost of education, unaffordable housing and grocery bills, neglected infrastructure, degraded social safety nets, environmental degradation, destruction. Why? $$$ competition. It’s like being part of a family where most members are hoarding for themselves. Now that we no longer have abundant room to expand, as Capitalists, we are consuming our markets and each other. I may have made an argument for socialism, although, frankly, because of our intelligence and selfish nature, we may not be built for group success, not when 400m or billions are involved. We just as likely prefer conflict to get our way, and gain advantage regardless of the damage we do our planet. It seems to be beyond our consideration.
#TheCapitalistEndGame.
The Great Filter: A possible solution to the Fermi Paradox
There are many major hurdles to becoming an interplanetary species, but one might be tougher than the rest.
www.astronomy.com
yeah it's going great so far 🤣
BSDnostalgia
macrumors 6502
In a different arena, these complications have already arisen. Companies go bankrupt and the assets get sold. A "pledge" or even a real contract is just one more claim in a sea of claims. Personal privacy itself has no legal standing (in the US). So, it gets complicated:
23andMe