Jerusalem Demsas on AI as centralizing technology; Mark McNeilly on holding AI to a higher standard; Steve Newman on the state of AI; Dean Ball on regulating AI
We usually forgive people for providing bad information because we have the expectation they've done their best to sift through the information readily available to them before formulating an answer, and we usually adjust our expectations for the situation we put them in. If I randomly ask my wife when her next appointment is I'll more than likely be satisfied with a vague answer like 'next month' or even a wrong answer like 'next Tuesday' when it's actually on Wednesday but if she's looking at her calendar I would expect a precise and correct answer. I'd say we're using a similar heuristic for the accuracy of AI answers but our expectations are far higher because we tend to imagine that AI is sifting through mountains of information looking for the nugget that we want, not acting like a giant random number generator creating an answer that we'll find most pleasing.
My experience with LLMs is that they often provide bad - or, at least, misleading - answers at first. If you know enough to push back with facts and logic they’ll provide much better results. The problem is knowing enough to be able to push back or at least to ask clarifying questions. I can do that in some areas, but not at all in others.
They are like news media in that regard. When I know an area or subject, it grinds my teeth how inaccurate they are, their exaggerations, omissions, and cherry picking. But it's incredibly easy to forget that when I read of topics I know little of.
If you’re asking a question about an area about which you know little or nothing - especially one that is likely to be controversial - the best strategy is to assume that the first answer is biased. Question every assertion. For example, if the LLM states that A caused B, ask it to explain the mechanism unless the mechanism is obvious.
Another option is to ask it to critique its own answer or to provide the best arguments against its own answer.
One of the great things about LLMs is that they will admit error and dig deeper for better answers.
But if you ask them to summarise a large or medium sized document you’d better know exactly what’s in it. ChatGPT just leaves out anything it doesn’t like the look of. Just, eg, anything in an education document that favours the teaching of actual disciplinary knowledge, as against ‘critical thinking’, which it adores (and misunderstands). Any lawyer or professional using these platforms to condense anything had better be well in charge of the material in the first place.
I asked my chatbot whether the phrase "13/50" was considered a racist trope. It cited me evidence from the ADL that it is indeed a "numeric hate symbol." Can basic FBI statistics be considered hate symbols? I don't know, maybe?
John Stossel pointed out like thirty years ago we have a tendency to ask if a new thing is perfect instead of better than the current alternatives. We don’t evaluate the new vs the existing, we evaluate it against the perfect. I don’t remember the new tech he was reporting on, but he pointed out natural gas furnaces might be a difficult sale (colorless, odorless, highly flammable even explosive gas piped into your house! People will die!) as something brand new, but no one gives it a second thought because we accept it as existing tech.
The difference is a large one though - these platforms are (or not) delivering ‘meaning’… they are faux interlocutors, spitting out a facsimile of that most human of markers… language.
I guess I’m alone in that I do give it a second thought. There are a half dozen houses in my county that will have to be demolished after one of them blew up the other day, putting people in the hospital. Happens at least once a year here.
Of course, maybe because home heating where I live is not terribly critical nor as expensive as it is other places, I am always happy to find any house I’m interested in is all electric, rather than having gas piped in or a propane tank in the yard.
It’s true that some people are passionately attached to cooking with gas. I’ve noticed this tends to be folks who might be characterized as enjoying a sort of showy element to their cooking, a la TV chefs - more than people who cook for more prosaic reasons, as a daily task.
I know I will be pillared for this take. One is not supposed to have any fears or concerns- in the same way that in Republicans circles it is considered unpatriotic two fear of stock market correction.
The problem isn't with all electric per se, but rather, all electric combined with the insane climate change ideology that provides the justification for shifting from reliable, dispatchable electricity generated from energy-dense fossil fuels (i.e. gas) to electricity generated from intermittent, low-density renewables (solar and wind), plus storage using magic batteries that have yet to be developed (not to mention the mining and processing that goes into batteries and the issue of disposal of dead batteries). I'm not Joe cook, but at least with a gas stove, if there is an electric power blackout (which is increasingly likely with renewables), I can use a match to light the gas burner. With an electric stove, you are screwed.
Given the percentage of power we are now getting in my state - the state, it should be noted most associated with fossil fuels in the entire union - from Wind and solar, I’m not sure how we’re going to instantly double our power output (as we’ve just been told we have to do, so much for happy equilibrium) solely for AI chip manufacture/data centers/y’all’s tiktoks and AI creations - By getting rid of them.
I would ask the legislators, but they’ve gotten so much cash from tech associated with this stuff, that I don’t imagine they care one way or another. They’ve got theirs, and the bright bulbs among them are few anyway.
Is there such a thing as dumb tech? Or even dumb use of it? Have we forfeited the right to ponder that?
I think of the gas lanterns that are fashionable when they spec builders demo and build new houses in my childhood neighborhood.
I’m pretty sure old timers would not have ever wanted those kind of lights.
For starters, or rather for non-starting. There have always been gas post lights In this particular Neighbourhood. They gave a pleasant glow Especially back when the houses were still surrounded by forest.
But something would happen with the filament and they would burn out pretty frequently, and most of the HOA dues went to sending a guy around to fix them. At this point, some of them have been out forever, just spewing gas into the air; some have been Disconnected ; Some have been replaced with an electric light.
It was just a very thin stream of gas, with a little sort of net in the fixture
But these new house builds are often covered with gas lights. They are not operable of course - they just burn all the time. It is 105° outside and you’ve got your gaslights burning away in the middle of the day. And all night like Christmas.
I think this is pretty stupid. I find it hard to imagine how it would be defended as a smart use of tech.
Meanwhile, these larger flames Have no such little net and blow wildly in the wind. In fact, they go out pretty frequently, only now at greater volume. People don’t go outside much and I’m sure that half the people would never notice that they were just spewing gas into the air.
We mentioned it to the next-door neighbor recently, and she said it goes out all the time and her husband will light it again when he’s home. There’s a persistent gas smell around so numerous are these fixtures.
You might say, I delight that they’re spewing gas in the air Because I love climate change or whatever and I think that’s hilarious.
I guess I just don’t think it’s that clever.
It just seems wasteful no matter what your view of climate change is, or had you never heard of it.
I was in a historic house the other day and noted that they had Purchase some lightbulbs that very closely mimicked the look of the flames. Obviously, they did not want flames in their structure.
But maybe that is the one thing that Republicans might decide was a dumb tech 😀.
I’m pretty sure that doubling our energy needs at a stroke *completely unlooked fpr and unplanned*, to say nothing of water, because of the AI industry and the grifting legislators eager for their $ - is not going to work out well, regardless of the energy source lol.
So far there seems to be no indication that there was homeowner error involved. There has been some question about whether there was sufficient of the chemical that fragrances the gas. I forgot the name.
It is being investigated.
Of course homes are only part of it.
I do not think one would call either an idiot or careless.for instance, the child who ran down to fish in the creek behind his house. Not sure what the spark was. There’s often a natural gas line running through creeks. Well, there’s not anywhere that there’s not a pipeline in the vicinity of people down here.
I always remember that particular story out of the many such incidents Because of the last words the child said to his friend as the friend ran to get the little fellow’s mother.
don’t let her see me
Children evince the weirdest nobility at extreme times.
AI should take a page from Karl Popper and for every answer it provides, it should also provide the evidence we need with which to falsify that answer. That might help to offset purely consensus answers.
Also, I may be wrong, but you seem to be saying that total surveillance is a good thing. I think that it is probably inevitable but is a very, very bad thing.
"Catastrophic risks come from bad humans." Reminds me of the dialog from the movie "Shane" where he says to the farmer, "A gun is a tool, Marian; no better or no worse than any other tool: an axe, a shovel or anything. A gun is as good or as bad as the man using it."
"…That’s why I support the AI regulation I support, which, in brief summary, involves the creation of private institutions to sit between the state and the frontier labs precisely so that they can mediate between the inevitable power-seeking impulses of the state and the private business of the frontier AI industry."
I'm glad Dean Ball supports what he supports, that his summary is brief, and that he wants the state to create private institutions sitting between the state and private institutions to slow down the power-seeking impulses of the state.
I'd almost think that was some kind of hallucination, AI or human.
Wow! Your “David Brin's Transparent Society Revisited” is one of the most helpful essays I’ve read of yours. I was so bent out of shape in 2013 about all that. Would have saved me a lot of grief to read this review back then. Thank you.
So I wonder how Brin’s ideas on that topic have manifested into your overall worldview (libertarian strategy)?
The Digital Utilities Commission should be created for AI auditing & all uses of digital info.
It should include explicit commission appointments from both Reps & Dems, at around 45% of each from a Congressional party caucus (Dems choose Dems, Reps choose Reps).
All minutes recorded & transcribed & publicly available, along with all official letters.
Some 5-10% chosen by the Joint Chiefs of Staff of the military, for military secrets NOT available to the public, usually.
China's govt & business are choosing, & investing, & trying to win the AI race.
We usually forgive people for providing bad information because we have the expectation they've done their best to sift through the information readily available to them before formulating an answer, and we usually adjust our expectations for the situation we put them in. If I randomly ask my wife when her next appointment is I'll more than likely be satisfied with a vague answer like 'next month' or even a wrong answer like 'next Tuesday' when it's actually on Wednesday but if she's looking at her calendar I would expect a precise and correct answer. I'd say we're using a similar heuristic for the accuracy of AI answers but our expectations are far higher because we tend to imagine that AI is sifting through mountains of information looking for the nugget that we want, not acting like a giant random number generator creating an answer that we'll find most pleasing.
My experience with LLMs is that they often provide bad - or, at least, misleading - answers at first. If you know enough to push back with facts and logic they’ll provide much better results. The problem is knowing enough to be able to push back or at least to ask clarifying questions. I can do that in some areas, but not at all in others.
They are like news media in that regard. When I know an area or subject, it grinds my teeth how inaccurate they are, their exaggerations, omissions, and cherry picking. But it's incredibly easy to forget that when I read of topics I know little of.
You need to be very knowledgeable to use these Chat platforms. Noticing what they omit or distort - let alone what they flatten and level’
If you’re asking a question about an area about which you know little or nothing - especially one that is likely to be controversial - the best strategy is to assume that the first answer is biased. Question every assertion. For example, if the LLM states that A caused B, ask it to explain the mechanism unless the mechanism is obvious.
Another option is to ask it to critique its own answer or to provide the best arguments against its own answer.
One of the great things about LLMs is that they will admit error and dig deeper for better answers.
But if you ask them to summarise a large or medium sized document you’d better know exactly what’s in it. ChatGPT just leaves out anything it doesn’t like the look of. Just, eg, anything in an education document that favours the teaching of actual disciplinary knowledge, as against ‘critical thinking’, which it adores (and misunderstands). Any lawyer or professional using these platforms to condense anything had better be well in charge of the material in the first place.
I asked my chatbot whether the phrase "13/50" was considered a racist trope. It cited me evidence from the ADL that it is indeed a "numeric hate symbol." Can basic FBI statistics be considered hate symbols? I don't know, maybe?
…is a full time job!
John Stossel pointed out like thirty years ago we have a tendency to ask if a new thing is perfect instead of better than the current alternatives. We don’t evaluate the new vs the existing, we evaluate it against the perfect. I don’t remember the new tech he was reporting on, but he pointed out natural gas furnaces might be a difficult sale (colorless, odorless, highly flammable even explosive gas piped into your house! People will die!) as something brand new, but no one gives it a second thought because we accept it as existing tech.
The difference is a large one though - these platforms are (or not) delivering ‘meaning’… they are faux interlocutors, spitting out a facsimile of that most human of markers… language.
I guess I’m alone in that I do give it a second thought. There are a half dozen houses in my county that will have to be demolished after one of them blew up the other day, putting people in the hospital. Happens at least once a year here.
Of course, maybe because home heating where I live is not terribly critical nor as expensive as it is other places, I am always happy to find any house I’m interested in is all electric, rather than having gas piped in or a propane tank in the yard.
It’s true that some people are passionately attached to cooking with gas. I’ve noticed this tends to be folks who might be characterized as enjoying a sort of showy element to their cooking, a la TV chefs - more than people who cook for more prosaic reasons, as a daily task.
I know I will be pillared for this take. One is not supposed to have any fears or concerns- in the same way that in Republicans circles it is considered unpatriotic two fear of stock market correction.
The problem isn't with all electric per se, but rather, all electric combined with the insane climate change ideology that provides the justification for shifting from reliable, dispatchable electricity generated from energy-dense fossil fuels (i.e. gas) to electricity generated from intermittent, low-density renewables (solar and wind), plus storage using magic batteries that have yet to be developed (not to mention the mining and processing that goes into batteries and the issue of disposal of dead batteries). I'm not Joe cook, but at least with a gas stove, if there is an electric power blackout (which is increasingly likely with renewables), I can use a match to light the gas burner. With an electric stove, you are screwed.
Given the percentage of power we are now getting in my state - the state, it should be noted most associated with fossil fuels in the entire union - from Wind and solar, I’m not sure how we’re going to instantly double our power output (as we’ve just been told we have to do, so much for happy equilibrium) solely for AI chip manufacture/data centers/y’all’s tiktoks and AI creations - By getting rid of them.
I would ask the legislators, but they’ve gotten so much cash from tech associated with this stuff, that I don’t imagine they care one way or another. They’ve got theirs, and the bright bulbs among them are few anyway.
Is there such a thing as dumb tech? Or even dumb use of it? Have we forfeited the right to ponder that?
I think of the gas lanterns that are fashionable when they spec builders demo and build new houses in my childhood neighborhood.
I’m pretty sure old timers would not have ever wanted those kind of lights.
For starters, or rather for non-starting. There have always been gas post lights In this particular Neighbourhood. They gave a pleasant glow Especially back when the houses were still surrounded by forest.
But something would happen with the filament and they would burn out pretty frequently, and most of the HOA dues went to sending a guy around to fix them. At this point, some of them have been out forever, just spewing gas into the air; some have been Disconnected ; Some have been replaced with an electric light.
It was just a very thin stream of gas, with a little sort of net in the fixture
But these new house builds are often covered with gas lights. They are not operable of course - they just burn all the time. It is 105° outside and you’ve got your gaslights burning away in the middle of the day. And all night like Christmas.
I think this is pretty stupid. I find it hard to imagine how it would be defended as a smart use of tech.
Meanwhile, these larger flames Have no such little net and blow wildly in the wind. In fact, they go out pretty frequently, only now at greater volume. People don’t go outside much and I’m sure that half the people would never notice that they were just spewing gas into the air.
We mentioned it to the next-door neighbor recently, and she said it goes out all the time and her husband will light it again when he’s home. There’s a persistent gas smell around so numerous are these fixtures.
You might say, I delight that they’re spewing gas in the air Because I love climate change or whatever and I think that’s hilarious.
I guess I just don’t think it’s that clever.
It just seems wasteful no matter what your view of climate change is, or had you never heard of it.
I was in a historic house the other day and noted that they had Purchase some lightbulbs that very closely mimicked the look of the flames. Obviously, they did not want flames in their structure.
But maybe that is the one thing that Republicans might decide was a dumb tech 😀.
Right, electricity policy in Texas is a mess, with excessive reliance on wind and associated problems: https://wattsupwiththat.com/2026/01/26/storm-ferm-remember-uri-centrally-planned-electricity-transition-in-texas/. Thankfully, their grid is isolated from those of other states. Maybe the proliferation of data centers will force the idiots running your state to fix the system and shift to gas and nuclear.
I’m pretty sure that doubling our energy needs at a stroke *completely unlooked fpr and unplanned*, to say nothing of water, because of the AI industry and the grifting legislators eager for their $ - is not going to work out well, regardless of the energy source lol.
"We made this idiot-proof, and found out there were bigger idiots."
And even non-idiots can get careless if things usually work without a problem.
So far there seems to be no indication that there was homeowner error involved. There has been some question about whether there was sufficient of the chemical that fragrances the gas. I forgot the name.
It is being investigated.
Of course homes are only part of it.
I do not think one would call either an idiot or careless.for instance, the child who ran down to fish in the creek behind his house. Not sure what the spark was. There’s often a natural gas line running through creeks. Well, there’s not anywhere that there’s not a pipeline in the vicinity of people down here.
I always remember that particular story out of the many such incidents Because of the last words the child said to his friend as the friend ran to get the little fellow’s mother.
don’t let her see me
Children evince the weirdest nobility at extreme times.
Meraptan. The gas company where I once lived sent out little "scratch and sniffs" with the smell so you'd know there was a gas leak if you smelled it.
How about this suggestion.
AI should take a page from Karl Popper and for every answer it provides, it should also provide the evidence we need with which to falsify that answer. That might help to offset purely consensus answers.
Also, I may be wrong, but you seem to be saying that total surveillance is a good thing. I think that it is probably inevitable but is a very, very bad thing.
"Catastrophic risks come from bad humans." Reminds me of the dialog from the movie "Shane" where he says to the farmer, "A gun is a tool, Marian; no better or no worse than any other tool: an axe, a shovel or anything. A gun is as good or as bad as the man using it."
Unfortunately, Shane ain't running the country.
"…That’s why I support the AI regulation I support, which, in brief summary, involves the creation of private institutions to sit between the state and the frontier labs precisely so that they can mediate between the inevitable power-seeking impulses of the state and the private business of the frontier AI industry."
I'm glad Dean Ball supports what he supports, that his summary is brief, and that he wants the state to create private institutions sitting between the state and private institutions to slow down the power-seeking impulses of the state.
I'd almost think that was some kind of hallucination, AI or human.
Cannot give this a big enough “like.”
Why on earth would you expect the state to do that? The entirety of human history says 100% otherwise.
They'll get it right this time, honest!
Oppenheimer (The State) should be required reading.
So-to get to the point,are models improving exponentially fast?
probably, but from a very low baseline
Depends on which point they are trying to get to.
AI will become the ultimate propaganda machine.
Wow! Your “David Brin's Transparent Society Revisited” is one of the most helpful essays I’ve read of yours. I was so bent out of shape in 2013 about all that. Would have saved me a lot of grief to read this review back then. Thank you.
So I wonder how Brin’s ideas on that topic have manifested into your overall worldview (libertarian strategy)?
The Digital Utilities Commission should be created for AI auditing & all uses of digital info.
It should include explicit commission appointments from both Reps & Dems, at around 45% of each from a Congressional party caucus (Dems choose Dems, Reps choose Reps).
All minutes recorded & transcribed & publicly available, along with all official letters.
Some 5-10% chosen by the Joint Chiefs of Staff of the military, for military secrets NOT available to the public, usually.
China's govt & business are choosing, & investing, & trying to win the AI race.
With DeepSeek.
https://x.com/itsolelehmann/status/2053934188135584176