Every question you ask, every comment you make, I'll be recording you Recently, OpenAI ChatGPT users were shocked – shocked, I tell you! – to discover that their searches were appearing in Google search. You morons! What do you think AI chatbots are doing? Doing all your homework for free or a mere $20 a month? I think not! When you ask an AI chatbot for an answer, whether it's about the role of …
Let me be clear: every time you use an AI to assist you, you're not delegating. You're donating. You're feeding it your workflow, your decision tree, your intellectual property—and in doing so, you're essentially handing over your job in a neat little JSON file.
And renting AI to run your business? That’s not innovation. That’s corporate self-cannibalization. You're not just outsourcing labor, you're outsourcing legacy. You're teaching the AI how to replicate your business model, optimize it, and—brace yourself—eliminate you from it. Congratulations, you've become the beta test for your own obsolescence.
Even Google is starting to cut its own throat as most people are only reading the AI summary and not clicking on the search links that Google is selling to business.
Why would you buy sponsored search positioning from them if it is only going to feed the front end summary and not get you clicks?
From Google's perspective, if they can improve the targeting and present an AI summary together with one advert. They can charge 10 times as much for that single advert instead of presenting 10 less targeted adverts. Therefore Google saves energy and the competition for the top spot becomes more intense. It's a win-win for Google.
Unfortunately, not using AI isn’t going to save you if it has any value at all. If you’re doing useful work in a sensible way, you almost certainly won’t be alone in how you’re doing it, and some of the others will be feeding the beast if there’s any short term real (or perceived) productivity uplift to be had. Standard tragedy of the commons stuff I’m afraid.
Of course, that “real” is doing a lot of work. If you’re a business owner and others are currently net costing themselves money by ingesting AI kool aid, carry on as you were and profit.
My bet is that regardless of the hype, it’s going to be very hard for AI to completely eliminate most types of job. However some, maybe even many, will eventually (probably over a longer timescale than predicted) see some level of productivity boost that will result in some job losses (unless there’s untapped demand to fill). All that will happen if you refuse to use genuinely advantageous AI is that you’ll lose your job/business earlier; it won’t change the direction of travel.
If you are genuinely a unicorn with a truly unique and genius business model, probably best not to type it into a chatbot. But someone else will likely have imitated you in the old fashioned way before a LLM is trained and released anyway, unless you’ve built some other moats.
Productivity gains should not lead to job losses overall. They should lead to more wealth. Unfortunately, the wealth gain doesn't seem to be getting distributed reasonably, this is leading to people thinking communism is a good idea which will be an even bigger disaster. To enable proper wealth distribution you need to eliminate corruption by increasing transparency of government and corporations. Businesses should not be allowed to hide what they are doing with your data in 200 pages of legal small print. Most of all the world of science needs a clear out of coersion and pay for preferred answers.
@AC
Nobody knows if communism will be a disaster as we have never seen it.
Russia, China etc. were never really communist - a few people involved in the revolutions had some romantic ideals about improving the life of the downtrodden, but soon end up being led by the power / status / wealth obsessed "leaders"* that we see in virtually all countries & political systems.
* as happens in virtually all political systems, due to the flaws of humans, so many of the people that determinedly aim for those positions of power are narcissistic back stabbers with little moral compass. There are lots of nice humans, but proportionately less nice humans high up in politics.
Productivity gains may not lead to job losses *overall* but they certainly can in specific jobs. Industrialisation of much of farming means the majority of agricultural jobs from 200 years ago have been lost and aren’t coming back. This has largely (in the long term, after often painful adjustment) been to society’s benefit as new jobs created have often been better or at least less dangerous than the old ones.
The difference this time around may be in the type of jobs eliminated and the new ones that look likely to replace them. There will likely be a minimum wage job in a nursing home to replace your comfortable one eliminated due to increased productivity for your knowledge worker peers. There will also be completely new, interesting, highly paid jobs - though likely far fewer of those. But unless you’re happy with dealing with actual rather than just metaphorical shit, or are feeling lucky, you may want to try to keep yourself in the employed rump (dwindling as AI improves) as long as possible.
This certainly won’t help with income inequality, particularly as most of the wealth from AI will accrue to capital (e.g. the wealthy) as opposed to labour, because AI is capital intensive.
A stupid person doesn’t need to intend harm - their inability to grasp consequences is the harm. It’s like handing a toddler a loaded gun and saying, “Don’t worry, he’s not malicious.” Malice with intent and malice through incompetence both end up with a hole in the wall and someone bleeding on the floor. Pretending one is somehow excusable is just self-delusion.
You reminded me of a quote from - IIRC - Graham Greene: "Innocence always calls mutely for protection when we would be so much wiser to guard ourselves against it: innocence is like a dumb leper who has lost his bell, wandering the world, meaning no harm."
Not quite the same thing, but a good quote nonetheless.
Ah! That's why water in a bowl sometime explodes when heated on full power. The devil was being summoned and trapped within the water. Never knew that. This new learning is the wonder of our age.
I suppose giving the oven less welly you could expect a lukewarm demon. Crowley?
Searching for this phrase is a revelation in itself. Doesn't give one much confidence in the Intelligence or sanity of one's fellow man.
The village of Lower Wattage is tired of being (mis)used in this manner, and their great rivals up the hill in Higher Wattage are likewise not impressed.
ITYM "power", not "wattage". Note that using "power" means you save having to type three fewer characters, or pronounce only two instead of three syllables, and even avoid some of those tiresome remarks by annoying pedants. It's a win-win-win all round! :-)
/mode pedant off
Well, actually, you get the same demon but for less of the time.
Microwaves reduce their power by running intermittently.
At full power the demon is summoned, and then banished by the "bing" at the end.
At lower power settings the demon is summoned then banished a few times a second. This annoys the demon considerably and is therefore a much greater risk.
I'm sorry, this is 2025.
How many times do you need to be warned that social media is just milking you to get the message ?
If you can't understand by now, then there's only a cluebat to the face - or a great big hole in your bank account - that will get the message across.
And I'm sorry if I do not sympathize.
[...does anyone actually read the terms and conditions?
So many lines there to view
To make sure no-one dares to sue
But does anybody read
All of it? Don't know 'bout you
But my ass avoids it ev'ry time
Quick-scrolling down to find out where I
Can make it all go away, the little box that they've placed
To get on with this program's use
And I know that some tiny part of this will go
And bite me right in the ass, it blows
But it remains the truth, can't change it now
Reading the terms? Won't happen, no doubt]
Original Song Title: "The Words"
Original Performer: Christina Perri
Parody Song Title: "The Terms"
Parody Written by: the_conqueror_of_parodies
This post has been deleted by its author
So if OpenAI is being forced to store data on non-US citizens, does that not invalidate EU Safe Harbour? and according to ChatGPT the answer is Yes,
EU Safe Harbor and Its Replacement (Privacy Shield):
The EU-US Safe Harbor Framework was initially a mechanism for ensuring that data could flow between the European Union and the United States while still complying with EU data protection standards.
The Safe Harbor framework was invalidated by the European Court of Justice in 2015, largely because it did not adequately protect EU citizens' privacy rights, especially with respect to U.S. government surveillance practices.
Following that, the EU-U.S. Privacy Shield was introduced as a replacement, which aimed to address those concerns by providing stronger protections for EU citizens' data. However, in 2020, the European Court of Justice again invalidated the Privacy Shield, citing concerns over U.S. surveillance laws and insufficient protections for EU data subjects.
The Current Situation:
The primary legal mechanism for data transfers between the EU and the U.S. is now based on Standard Contractual Clauses (SCCs), which are used to ensure that personal data is adequately protected when transferred outside of the EU.
However, SCCs still face scrutiny, particularly in light of the Court of Justice ruling that certain aspects of U.S. law (like government surveillance) could make such transfers non-compliant with EU law.
OpenAI and Non-U.S. Citizens' Data:
If OpenAI is indeed storing data in ways that involve non-U.S. citizens, there are key considerations:
Where the data is stored: If data about EU citizens is stored on servers outside of the EU, including in the U.S., the transfer and storage need to comply with EU data protection laws, including the General Data Protection Regulation (GDPR).
Access by U.S. entities: If OpenAI or other U.S.-based entities can access this data, there might be concerns over the extraterritorial application of U.S. surveillance laws (such as the Foreign Intelligence Surveillance Act), which could invalidate EU-based protections, even if data is stored on non-U.S. servers.
So, if OpenAI (or any U.S.-based company) is storing data on non-U.S. citizens and that data is subject to access by U.S. authorities, this could undermine the protections intended by the Safe Harbor or Privacy Shield frameworks, and likely complicates compliance with the EU’s data protection laws.
Ultimately, the legal landscape around data transfers is in flux, and companies are actively working to ensure compliance through measures like SCCs, but as the rulings show, the legality of such data flows may continue to be challenged.
According to that response, Safe Harbour was already invalid, so of course "Yes" was the answer... When posting ChatGPT or other AI responses, please clearly mark the beginning and end of the response. Better yet, search for yourself, and ignore ChatGPT, because it is so frequently wrong, or inaccurate about details.
That response states in several hundred words, what you could have said in less than 100.
I didn't Agree. ChatGPT told me it forgot the whole session once I was done. Well, technically, when it asked about persistent storage I clicked "Block". It started spouting stupid Internet myths, I spent some time challenging it on these and it got things about right in the end. Not seen a whiff of my interactions anywhere since, and its replies now include the same dumb shit it began with last time.
Is this actually a case of too stupid to "Just say No"?
They trained it on "the internet", of course it decides "the internet" is right. It recommends things that don't exist, makes things up out of thin air... The occasional useful result is useful only because paying a human to create a meaningful summary of "everything on Amazon" is too expensive.
recalled me asking about how to make 3D-printed guns
This is a big issue and will fuel the Minority Report desires of the government bureaucrats. When I was young I used to look up how to make explosives, how nuclear weapons worked and read about guns. I was fascinated as a youth. Still have some interest. I never ever had any intention to hurt or damage anyone and never have. I was just a typical boy. But you cannot be a normal boy anymore, that's not allowed.
It appears Google, Bing, et. al. are encouraging LLM chatbot use for web-searching by degrading ordinary web-searching methods.
Two or three times so far, I DuckDuckGo-searched for something, got no relevant hits, though I should have, then searched via Copilot* and got what I needed.
* Not on my own computer, of course -- I used a computer at the library ... which required I sign in via my library card # and password ... for which the library system has my real name and address ... which they surely would not be "sharing" with outside organisations ... except they are using some 3rd-party software to manage (MS-Windows) PC access and printing ... whose TOS I have no idea of, but which may include 3rd-party-use of library patrons' personal info ... because the library's managers are privacy-ignorant/unconcerned morons and accepted those terms ...
D'OH!
