Tyler Cowen explains how to read a book using o1:
You don’t have to upload any book into the system. The Great Cosmic Mind is smarter than most of the books you could jam into the context window. Just start asking questions. The core intuition is simply that you should be asking more questions. And now you have someone/something to ask!
I was reading a book on Indian history, and the author reference the Morley reforms of 1909. I did not know what those were, and so I posed a question and received a very good answer, read those here. I simply asked “What were the Morley reforms done by the British in India in 1909?”
Then I asked “did those apply to all parts of India?”
You can just keep on going. I’ll say it again: “The core intuition is simply that you should be asking more questions.”
Most people still have not yet internalized this emotionally. This is one of the biggest revolutions in reading, ever. And at some point people will write with an eye toward facilitating this very kind of dialogue.
I’ve long said that looking things up is a superpower, and I suppose this reduces the friction — but you might imagine I tend to look things up, even if I have to open up a new tab and type in a search term.
It’s a good thing nobody corrupts references in Wikipedia, or elsewhere, that might be used by whatever AI is answering your inquiries. Or authors bothering to go to original sources for new insights or facts overlooked by others.
What Cowen is apparently calling the Great Cosmic Mind really is a great leap forward for search. The AI’s worldview produced by ingestion of man’s corpus is many orders of magnitude more finely detailed than backlinks could ever facilitate. That it does not have understanding in the same sense as we have understanding because it is not embedded in the time stream in the same way does not diminish its impressiveness.
A question to ponder: why doesn’t the oracle take over the world?
“I’ve long said that looking things up is a superpower…”
Agreed. My years in education have further confirmed some of my suspicions, for example:
Investigatory thinking is not function of IQ, rather it is a habit of the mind and personality; IQ merely limits the speed and reach of investigatory thinking, not it’s basic forms and expressions. An idiot with the penchant to ask good questions will still be inclined to question, even if he struggles to make sense of the answers he receives. A genius, prone to investigatory apathy, will still struggle to pose meaningful questions, even if a preponderance of evidence obviously suggests one.
It’s why I have Sub 90 IQ students in my classes who can still make more meaningful observations on the principles contained within a historical scenario than my 110 IQ students whose sole capacity is relegated to high speed recall and regurgitation.
Dietrich Bonhoeffer was right.
“It’s a good thing nobody corrupts references in Wikipedia”. Indeed. Wikipedia has already infected books published since 2001. If I’m consulting Google Books, I always limit search to pre-2000 to avoid Wikipedia (since if I wanted to search Wikipedia, I could do so directly). AI companies will have to offer models trained on pre-2000 data only.
Phileas Frogg says, “function of IQ.”
Oh dear…
“…NOT…”
Yes dear?
I sometimes forget just HOW semi-literate most Americans are. It’s like this…
For example, the Moh scale and Richter scale are not “measures”. Any formula including either of those numbers is nonsense. But they are meaningful. They are qualitative, i.e. their values are arbitrary, what matters is the ordering. Finding the value for a given specimen on Moh scale is useful, but it’s not measuring; it’s appraisal.
Conversely, the Harmony Meters used by Church of Xenulogy are devices that measure actual physical processes. The way it’s done, however, the resulting data was not demonstrated to be meaningful, whatever their users may claim.
IQ (and similar tests for nothing defined) are neither measurements, nor meaningful. As such, IQ is not used in actual science. It’s mostly used by quacks to sound So Very Scientific and by big children to measure virtual peepees.
See also the link in my previous post for Taleb’s article with relevant statistics.
The Richter scale is a measurement (of the amplitude of waves recorded by seismographs). Perhaps you were thinking of the earlier Mercalli intensity scale?
“IQ (and similar tests for nothing defined) are neither measurements, nor meaningful.”
That was the gist of my comment, yes. I suppose I’m failing to see where the disagreement enters between us. If it’s a product of my own error in understanding I’m fully prepared to accept correction on the matter, but I’m not presently seeing it. The reference to Bonhoeffer’s assessment of stupidity is the crux of the whole thing, and I sorta left it hanging without specificity, so that would be my bad if that’s the case.
From Merriam-Webster:
Function: something (such as a quality or measurement) that is related to and changes with (something else)
So, “function,” can of course refer to a quality as well, not merely a measurement, which would put us in concert.
IQ is the best-replicated and most useful finding in psychology and it isn’t even close. Even Jordan Peterson will tell you that.
I meant it as not being meaningful within the context of having an inquisitive temperament or disposition.
Isegoria says:
Right, my bad, Richter is the logarithmic one.
Hmm. To think of it, logarithmic compression may have been an advantage at the time it was introduced, due to resemblance to the old familiar qualitative scales.
Jim says:
…and is not used in actual science.
I mean, there are areas in psychophysiology that require meaningful estimation of mental performance, in particular changes thereof. For example, comparison of different psychostimulants needs some measurement of their effects. I looked into those once. They didn’t use anything like IQ tests. They used correction tests. Note that this yields two values: speed and error rate. It’s quite transparent what exactly these numbers mean, and what relative changes in them will likely mean for real-world tasks (those friendly fire incidents with meth-head pilots come to mind). A correction test is anything but excessively arcane and abstract.
This application also makes it easy to see how even a meaningful test that outputs a single value would be not very useful at best, and dangerously misleading at worst.
See also the link in my first reply for Taleb’s article beating this spherical horse with some math.
Phileas_ Frogg says:
My point is, if IQ is bogus, any and all “studies” using IQ have swamp water for a foundation. Consequently, trying to draw any conclusions from any of this is still building on a swamp, just with a bit of extra rubbish underfoot. Whether one chooses to agree with, or disagree with, or nitpick any of them, it’s no more meaningful than critique of the shapes of clouds.
When the content is already nonsense, it’s only processed into more nonsense. Likewise, one can join the old topic of whether a bee hive is monarchy or republic. That “discussion” was joined by many oh-so-clever people (including Voltaire), at least some of whom had to be serious about it, but obviously it could not pull itself by the hair from the pit of nonsense.
That’s the way of pseudoscience.
Phileas: “I meant it as not being meaningful within the context of having an inquisitive temperament or disposition.”
IQ is meaningful in every context.
Yes, it isn’t the end-all-be-all. Yes, personality characteristics matter a lot.
IQ nevertheless is the single most important psychometric by a huge margin, so important that it’s literally impossible to make a test that isn’t g-loaded to hell and back—and they’ve tried, oh, they’ve tried.