Finally broke down and went to get the famous Pink Drink from Starbucks. Very refreshing but not sweet enough for me. I AM prediabetic, but I deserve a treat once and a while. Will try the Matcha Creme Frappucino next time.
EXCERPTS: Training artificial intelligence (AI) models on AI-generated text quickly leads to the models churning out nonsense, a study has found. This cannibalistic phenomenon, termed model collapse, could halt the improvement of large language models (LLMs) as they run out of human-derived training data and as increasing amounts of AI-generated text pervade the Internet.
“The message is, we have to be very careful about what ends up in our training data,” says co-author Zakhar Shumaylov, an AI researcher at the University of Cambridge, UK. Otherwise, “things will always, provably, go wrong”. he says.” The team used a mathematical analysis to show that the problem of model collapse is likely to be universal, affecting all sizes of language model that use uncurated data, as well as simple image generators and other types of AI.
[...] Language models work by building up associations between tokens — words or word parts — in huge swathes of text, often scraped from the Internet. They generate text by spitting out the statistically most probable next word, based on these learned patterns.
[...] Collapse happens because each model necessarily samples only from the data it is trained on. This means that words that were infrequent in the original data are less likely to be reproduced, and the probability of common ones being regurgitated is boosted. Complete collapse eventually occurs because each model learns not from reality, but from the previous model’s prediction of reality, with errors getting amplified in each iteration. “Over time, those errors end up stacking up on top of each other, to the point where the model basically only learns errors and nothing else,” says Shumailov.
The problem is analogous to inbreeding in a species, says Hany Farid, a computer scientist at the University of California, Berkeley. “If a species inbreeds with their own offspring and doesn’t diversify their gene pool, it can lead to a collapse of the species,” says Farid, whose work has demonstrated the same effect in image models, producing eerie distortions of reality2.
[...] As synthetic data build up in the web, the scaling laws that state that models should get better the more data they train on are likely to break — because training data will lose the richness and variety that comes with human-generated content, says Kempe,,, (MORE - missing details)
INTRO: The American public intellectual and creator of the television series Closer to Truth, Robert Lawrence Kuhn has written perhaps the most comprehensive article on the landscape of theories of consciousness in recent memory. In this review of the consciousness landscape, Àlex Gómez-Marín celebrates Robert Kuhn’s rejection of the monopoly of materialism and uncovers the radical implications of these new accounts of consciousness for meaning, artificial intelligence, and human immortality.
EXCERPTS: [...] The origins of our perplexity in making sense of experience itself can be traced back to Galileo Galilei, who programmatically excluded subjective experience from the purview of science. One can interpret this sagacious move as a means to understand nature in two phases: let us first start with what lends itself to measurement and mathematisation (the “primary phenomena of motion and touch”, in Galileo’s words) and leave for later what resists it. “I think that tastes, odors, colors, and so on (…) reside only in consciousness”, he wrote in The Assayer in 1623.
Such a strategy proved tremendously successful, giving rise to physics, then chemistry, next biology, and finally psychology. ... Four hundred years later, we can’t ignore the elephant in the room anymore: experience is what makes science possible and yet a proper science of consciousness seems unattainable. The Galilean knot remains untied. Today we call it “the hard problem”.
[...] Gathering under the same roof most of the greatest contemporary thinkers of one of the greatest questions one can ever try to answer, Kuhn’s landscape enacts the quasi-extinct art of true scholarship...
[...] The landscape comprises 10 major categories and it is organised in a gradient of “isms”, from die-hard materialist positions to mind-only propositions...
[...] Kuhn’s faithful description of each position without the urge to adjudicate deserves nothing but praise and gratitude. ... Rather than divide and conquer, let us unite and wonder... (MORE - missing details)
INTRO: New research from a collaborative team featuring Texas A&M University atmospheric scientist Dr. Andrew Dessler is exploring the climate impact of the 2022 Hunga Tonga volcano eruption and challenging existing assumptions about its effects in the process.
The remarkable two-day event, which occurred in mid-January 2022, injected vast amounts of volcanic aerosols and water vapor into the atmosphere. Historically, large volcanic eruptions like Tambora in 1815 and Mt. Pinatubo in 1991 have led to significant cooling effects on the global climate by blocking sunlight with their aerosols. However, Hunga Tonga’s eruption presented a unique scenario: As a submarine volcano, it introduced an unprecedented amount of water vapor into the stratosphere, increasing total stratospheric water content by about 10%.
Because water vapor is a powerful greenhouse gas, Dessler says there was initial speculation that it might account for the extreme global warmth in 2023 and 2024. Instead, the results of the team’s research, published in the Journal of Geophysical Research: Atmospheres, reveal the opposite: The eruption actually contributed to cooling the Earth, similar to other major volcanic events... (MORE - details, no ads)
VIDEO EXCERPTS: . . . Someone doing something really, really, breathtakingly stupid and provoking the United States ... This has happened before. I mean you could make the argument that was basically Pearl Harbor...
... When I look around the world at the powers that are in play, I don't think the Chinese are anywhere near dumb enough to do that. At least a few years ago [...they realized...] that if there was a fight with the United States, it would be a fight on the water. China depends upon freedom of the seas in order to keep its people alive.
The entire economic model, the food imports, the energy imports would just stop; and China knows that would be suicide.
So the only country right now where that might be an issue, would be the Russians. And that's because of Russian incompetence.
We've we've learned over the last three years that Russia doesn't have a classic army in the sense that most people think of the word. They basically have a mob, that they put guns into their hands, and throw them at things. And that's a strategy that has never worked for the Russians.
If the fight gets to a point that it's hitting US interest, that's where you get the direct clash. So as long as the United States is at least passively interested in NATO -- should Ukraine fall, then we will be in a more direct fight. But we are not there now, and if the Russians continue making the gains that they've been making in the last year for the next five years, we will still not be there...
VIDEO EXCERPTS: Last year you and I had breakfast together, and you warned me then that Wokeism will generate a level of anti-Semitism I have never seen in my lifetime. As of October 8th, anti-Semitism exploded throughout the Western World. What made you predict this level of anti-Semitism?
I had published in the Wall Street Journal an article comparing Wokeism with Islamism, and the similarities, and the differences. The more I looked into and read about social justice and critical theory, this elaborate intersectionality -- this web of identity politics -- the more it became clear to me that on different levels Wokeism would produce an anti-Israel, anti-Zionist, and anti-Jewish hatred.
The demands that Woke makes for diversity, equity, and inclusion [DEI] creates openings that the Islamists take great advantage of and exploit.
The narrative (precursor: cultural hegemony) that there is a white oppressor, and everyone else is oppressed, and that this white oppressor is wealthy or is the owner of these structures -- that also makes it abundantly clear that Jewish people are considered to be white, mostly family oriented, and therefore heterosexual, privileged, and exploiters.
In our universities of which I have just had a glimpse of some of the courses that are taught seek to either avenge the history of colonialism of Western European countries, or they seek to finish off what they think of as the remaining colonial settler state of Israel...
Ayaan Hirsi Ali on how Islamists exploit DEI to spread anti-Semitism & anti-Westernism
It's the Fourth of July celebration in a little town about 100 miles north of Anchorage called Glacier View
Fireworks aren't really a thing there, since in July sunset is about 11:30 PM and sunrise about 4:30 AM. So instead of fireworks, they drive cars off a cliff! What else do you expect them to do?
(Don't worry, there's nobody in the cars.)
The objective seems to be to reach the pond at the bottom of the cliff (where the crowd is)
Despite Glacier View only having ~200 residents, many thousands of people turn out. It's an Alaska thing...
(not something you would ever see in Manhattan or San Francisco...)
![[Image: bNbYzzP.jpeg]](https://i.imgur.com/bNbYzzP.jpeg)
