Hacker News new | past | comments | ask | show | jobs | submit login
How much things can change (rodneybrooks.com)
152 points by bookofjoe on Aug 7, 2020 | hide | past | favorite | 95 comments



The scientific discovery whose recency most shocks me is plate tectonics.

Here you have massively visible physical manifestations—mountains, earthquakes, and volcanos—that have shaped human history and played prominent roles in culture since the beginning of humankind, and we didn't have agreement on a good explanation for them until after Rubber Soul was recorded by the Beatles.


The story of how (and why) plate tectonics came to be accepted not only as established scientific wisdom but the central organising principle of geology is a fascinating one, studied at length by Naomi Oreskes:

https://www.worldcat.org/search?q=au%3Aoreskes+kw%3Atectonic...

Gross continental geography, radioactivity (energy release, isotopes, clocks), bathymetry, magnetic fields analysis, seismic surveys, nuclear testing (more radioactivity), fossil-record analysis, vulcanology, earthquake analysis, stratigraphy, core sampling, and so much more.

All were parsimoniously explained best by plate tectonics.


I feel that I'm conceptually familiar with plate tectonics as an explanation for mountains and earthquakes. But the grandparent comment and you both mention volcanoes, and the relationship between plate tectonics and volcanoes best known to me is that the Hawaiian islands were all formed by the same "fixed-location" volcanic hot spot as a tectonic plate drifted over it.

This would seem to imply to me that plate tectonics isn't a good explanation for volcanoes. How are volcanoes best explained by plate tectonics?


My limited understanding is that volcanoes in Hawaii are the exception rather than the rule. The rule is that they happen where plates converge:

https://www.britannica.com/place/Ring-of-Fire


Yes, that is also what I was taught. But that suggests the following model to me:

1. Volcanoes are caused by Circumstances X, Y, Z, whatever.

2. Collision of two tectonic plates causes Circumstances X.

That is, I want the model of "volcanoes" to account for all volcanoes, not just most volcanoes. If we describe circumstances that are necessary and sufficient to create a volcano, we're done; the fact that the interaction of tectonic plates often creates those circumstances -- and even more often doesn't -- isn't really part of the model at all.


"Whatever" turns out to be a reasonably limited set of exceptions to tectonic boundary models.

Plate tectonics describes and explains the overwhelming majority of known current volcanos and volcanic regions. Mantle plumes, which themselves drive tectonic activity, much the rest.

The overall structure of the Earth, with a thin crust, plastic mantle, and hot core with temperatures resulting from both latent heat of gravitational formation and ongoing radioactive decay itself explains much and is entirely consistent with plate tectonics.

The principle further exceptions might be simply thin crust, or penetrations by impactors --- the universe is complex, volcanos result from escape or exposure of interior magma to the surface by some mechanism. Gods hurling thunderbolts or smithing armour for other gods, for example, has scant evidence.

I'm unaware of any specific terrestrial volcanoes not explained by tectonic or plume mechanisms. There are minor craters (e.g., Ubehebe crater in Death Valley) resulting from steam rather than magmatic bursts, but still driven by geothermal processes.

https://en.wikipedia.org/wiki/Ubehebe_Crater

Further we've now had evidence of similar mechanisms (or their lack) on other bodies, including the Moon (little tectonic activity, some faulting for other reasons, possibly validating Tethys impact formation hypothesis, possible recent volcanism https://coolinterestingstuff.com/volcanoes-erupted-on-the-mo...), Mars (extinct volcanos, former tectonics), Venus (active volcanism, possible tectonics), Mercury (mixed, some faulting thought due to cooling and contraction), and Io (tidal-force induced volcanism), as well as other moons of Jupiter and Saturn.


Volcanism can occur at both divergent (e.g., Iceland / mid-Atlantic rift, African rift valley) and convergent (e.g., Cascadia, Japan, Chile, subduction zone) plate boundaries, as well as in mid-plate regions subject to mantle plumes (Hawaii and, I believe, Yellowstone).

http://volcano.oregonstate.edu/subduction-zone-volcanism

https://youtube.com/watch?v=DUbX67e5rzI


Continental Drift or something similar had been theorized for hundreds of years, but most famously scientifically proposed by Alfred Wegener in the early 20th century. In a famous survey of Fellows of the Geological Society, in 1961 only 22% accepted it, but by 1978 87% did. Consider that --- in 1978 13% of the most accomplished geologists still rejected it.


That was about when it became directly measurable. One of the early 1980s results came from an earthquake related system in California which used lasers to measure distances between mountaintop stations along the Pacific chain.[1] That could directly detect plate motion.

GPS came later.

[1] https://earthquake.usgs.gov/monitoring/deformation/edm/


Thing I remember was in the early 60's detecting clockwork like magnetic reversals in the Atlantic sea floor parallel and mirrored on each side of the Atlantic ridge.


Nobody should wonder why the public doesn't automatically trust scientists or "experts".


Scientific consensus is the worst possible system for determining the truth, except for all of the others.


Not direct revelation? Works okay between two people. Unless you do controlled blind study on everything you observe and hear.


The only truth I think you could confirm is that you believe you got information X from person Y. You couldn't have a high degree of confidence that X is 'true' in any meaningful sense just from the revelation.


According to your view, I can't have confidence that your statement is "true" in any meaningful sense. But in my view, I can (though I don't, because it's "false").


Similar to democracy and capitalism. Must be a human thing.


The public should trust scientists and experts as long as they (1) know more than they do about some subject and (2) have achieved a rough consensus. Because for most practical applications that will be more than good enough. The few cases where it isn't typically will not move the needle for the general public.

The reason why people will not trust scientists (or journalists for that matter) is because other people, who are also not scientists (or journalists) want to manipulate them into acting against their own self interest.

See also: lobbying, the tobacco industry and many other examples besides.


If the data was inconclusive and other promising theories existed, if scientists had a healthy discourse without politization and an us-vs-them mentality, if science was done ethical and rigorous, then I see no reason for the public to distrust scientists just because not everyone did agree on the theory we know is right in hindsight.


The vast majority of distrust of science and experts has nothing to do with epistemology or history of science or anything else rational. It's because accepting scientific facts often leads to politically uncomfortable conclusions, like we can't keep pillaging the planet, or that trans people should be treated like humans.


Plate tectonics became obvious once the sea floor magnetisation data was declassified. Before then, the US Navy didn't want anyone to know what parts of the seabed they considered important, so there was a fair amount of scientific "parallel construction" undertaken first.

https://en.wikipedia.org/wiki/Seafloor_spreading


I love this SMBC comic that describes the story arc of the acceptance of any radically new idea.


I don't understand what's so interesting about pointing out things you personally dislike in whichever discipline without making any attempt to justify why you think they're wrong. That's simply prejudice of the lazy sort.

I can improve on this.

Cosmology:

    If there isn't any dark matter out there, there must be gremlins pushing stuff around to make it look like as there is. Otherwise there's no sensible way to explain the Bullet Cluster for starters.

    If the Universe is not expanding, there must be much bigger gremlins stretching homogeneously everything everywhere out, otherwise there's no way to explain the observational Hubble's law.

    If the Big Bang is wrong, there must be very naughty and ubiquitous gremlins matching somehow the primordial nucleosynthesis ratios and the CMB distribution everywhere you look at.
I would have continued with the physics section, but I can't picture the gremlins doing theory atm.

It's one thing to have a problem with whatever you think you know after reading pop sci, but there's textbooks and papers. That's way harder, but it's all there, the arguments, the counterarguments and more importantly the observational/experimental facts. You can't allow yourself to have problems with facts, if your intuition gets in the way of accepting them, just get rid of it.


"Gremlins" is a very lazy way of making a strawman. Just because one can't imagine a different universe does not mean gremlins made it. All scientific knowledge we discovered made some previously true piece of knowledge false. It is a pretty bad attitude to make fun of people simply stating we could be wrong. We are most definitely wrong. Imagine somebody from 1000 years in the future read this; how embarrassed would they be that people believed what we believe?


Of course I make fun of it. It's the only way to cope with this kind of stuff after being exposed to it for decades. It's cranks in physics, they don't take the trouble to really look up anything by doing actual work, they just know that everything's wrong. It's all there: dark matter doesn't exist, the universe isn't expanding, the big bang is a lie, string theory is bogus, quantum mechanics is incomplete and/or we don't understand it. The list used to include that Einstein was wrong and Special Relativity was false, luckily that's becoming rather unfashionable nowadays.

I'm sorry, but there's a pattern here. Whatever kind of scientific topic that drives "common sense" folks out of their comfort zone has to be wrong, because well, it has to be. It doesn't matter that there are actual measurements and heaps of relevant technical literature full of excellent reasons for this or that theory (I mentioned a few). Whatever the people working in a field find convincing just can't be right because after having spent a few hours trying to fit the headlines to what they think they know about reality, they just can tell it can't be and that's that. Well, saying that something is wrong in science isn't enough, you have to include your convincing reasons for it or else your proposition turns out to be as valuable as saying that gremlins did it (but less fun).

It's been a while now that new pieces of knowledge in physics fit over the old ones. That's because data exist and it's the result of measurements, not subjective numbers, so new theories can't disagree quantitatively with the confirmed predictions of the old ones. The anomalous magnetic moment of the electron won't change the day we find the mythical theory of everything (if such a thing exists), and the mythical theory of everything will have to include within itself a scenario in which plain QED first and then the full-blown SM happens, otherwise it's unlikely that you could predict its value to such an amazing degree. Again, there are technical reasons for this that you can find in texts about effective theories.

To be honest and IMO true and false belong more in mathematics than they do in science. Newton's gravitation theory isn't false, it's a good approximation to a better theory that won't be false either if we come up with something even better. About true theories as in involving the things in themselves, I don't think we're doing that after Kant.


I think bringing up gremlins in order to lampoon some heterodox hypothesis is intellectually lazy as well.

Dark matter assumes gravity is uniform across the universe. What if it isn't?

Measuring the expansion of the universe relies on the speed of light being a constant. What if it isn't?


It must be noted that both of those hypotheses would be considered highly unlikely to pass Occam’s razor, since they involve a far more radical/fundamental change in our understanding of physics, than sons piddling “dark matter”.


Of course it's lazy, but I said I could improve on it, don't expect me to put in more work than that.

Then it's the job of the alternative theorists to come up with working models with better descriptions and/or predictions. To my knowledge neither MOND nor VSL theories do that, but I wouldn't have any inner conflicts if they did. We're here to learn about the Universe, not to dictate how it works.


"If there is no luminiferous ether then there must be tiny invisible gremlins ferrying light particles around the universe" etc etc.

Actually, I agree with you. "Gremlins" is a great way for unknown unknowns or necessary assumptions that are impossible given current observations. I'll start using the term in that way, if you don't mind.


I have a soft spot for very heterodox, but not-quite-crank-but-seriously-wrong theories that I secretly hope are true for no other reason than a contrary nature and the excitement of the new. The aquatic ape theory of human evolution. Abiogenic origin of petroleum. Cold fusion. Julian Jayne's theory on the origin of consciousness.


Same. It's almost certainly wrong, but it's so eccentric without being crank-y that I just love it.


> There is no dark matter.

If we find proof that there's no dark matter then there must be something else like dark matter that behaves in the way that dark matter has been described (something with gravitational pull).

Since gravity requires mass then a universe where there's no dark matter (conceptually as matter) would completely invalidate Einstein's General Theory of Relativity.

That would completely break modern Physics. Right? I'm not expert and this is conjecture from my interest in cosmology and physics, so could someone with knowledge chime in and explain if what I'm saying make sense?


> there must be something else

No, it could also turn out that Einstein's field equations require modification at certain scales or in certain situations.

"Dark matter" is just a term to describe hypothetical matter that would cause the observed deviations from General Relativity – if it were actually there. People have proposed other ways to explain those deviations, though, like MOND, f(R) gravity, entropic gravity and many others, see https://en.wikipedia.org/wiki/Dark_matter#Alternative_hypoth... .


>If we find proof that there's no dark matter then there must be something else like dark matter that behaves in the way that dark matter has been described (something with gravitational pull).

Or...we completely reconfigure how we understand gravity. That solution seems more likely to me, as we know far less about gravity than we do about matter and relativity.


To pull out a couple of the more provocative statements:

> completely invalidate Einstein's General Theory of Relativity.

Only in the sense that general relativity invalidated the Newtonian model. That is to say that the two theories had to agree, within experimental uncertainties, in domains where the Newtonian model was known to work. Where the Newtonian model was known not work, such as the discrepancy in the precession of Mercury's perihelion, is where general relativity has value. It can be argued that dark matter is general relativity's analog to the precession of Mercury.

> That would completely break modern Physics. Right?

No. It could alter certain parts of modern physics, but it can only do so where the observational evidence contradicts current theory, current theory is supported by weak evidence, or the evidence is incomplete. The important thing to realize is that physical theory is based upon mathematical models. Mathematical models are predictive, and observations can be made to support the model or disprove it. Even then, claiming that observations can disprove a model is too strong. Contradictory evidence in one domain does not automatically invalidate supporting evidence in another domain (though it may encourage people to go back to examine the validity of that evidence or improve the accuracy and precision of measurements).

It is also important to consider that a mathematical model is an approximation that must agree with observations within uncertainties. As such it is possible that multiple models may fit the observations. Part of the reason why we look for models that predict something new, such as Eddington observing the deflection of light as predicted by Einstein, is seek an understanding that accomplishes something more than fitting the data. The ability to come up with new theories that fit the data is also why we are hesitant to adopt new theories when there is a discrepancy. When it was discovered that the precession of Mercury did not fit the Newtonian model, there were efforts to seek out explanations that fit the Newtonian model (e.g. is there a body perturbing the orbit that hasn't been previously discovered). General relativity's ability to predict the discrepancy in the perturbation was wonderful, but the deflection of light by gravity (a previously unknown effect) is one of the bits of evidence that allowed us to say general relativity is true since we are now confident that the inability to find a body perturbing Mercury's orbit is because no such body exists.

It should be noted that none of this is meant to claim that a better theory will replace Einstein's, simply that any new theory cannot overturn what we already know from general relativity.


It's important to remember that (at least as I understand it) gravity is a unique force in that we don't have a quantum explanation for it, unlike the other fundamental forces. Figuring this out will probably be important to or related to understanding the mystery of dark matter.


> Figuring this out will probably be important to or related to understanding the mystery of dark matter.

The physicist in me would replace "probably" with "hopefully" here. :)


Hi codethief! Sorry I'm hijacking your reply. I can't seem to reach you out from your bio. About 3 months ago I posted that I'm working on a pomodoro timer app with analytics—Session. You replied and wanted to know when I released it.

Just added Show HN on https://news.ycombinator.com/item?id=24085041. The product is on https://www.stayinsession.com

Cheers!


I believe you are making assumptions the rule out alternative explanations that don't break physics.


Obligatory link to a paper from one of the "Big Bang Bashers" http://web2.utc.edu/~tdp442/RedshiftEssay.pdf


During his children's lifetime we learned that we exist thanks to a meteorite that killed off every land creature bigger than a cat.

And only in the last decade did we establish that another, only 12800 years ago, killed off most American megafauna, and much of nascent American cultural development.

It is somewhat likely we will learn that agriculture originated before it, with all the evidence now 100+ ft below sea level. Along with a lot else.


> And only in the last decade did we establish that another, only 12800 years ago, killed off most American megafauna, and much of nascent American cultural development.

Could you provide a source for that? Everything I'm finding says this is anything but well-established.


http://deadlyvoyager.net/References/

A layer all over the Americas and out to Syria, Chile, and South Africa, right at 12800 years ago, the start of the Younger Dryas freeze that lasted 1200+ years. It is packed with soot from hemisphere-wide wildfires, melt glasses, hexagonal nanodiamonds, and elevated platinum.

Bones of American mammoths, mastodons, horses, camels, giant sloths, dire wolves, glyptodonts, and many others appear only up to the layer. Also, all Clovis culture artifacts end immediately below the layer.

Current hypothesis is a broken-up comet and multiple Tunguska-style airbursts scattered over a wide area.

(The only naysayers, last published in 2017, cite one study that sampled only one site, and not at the layer in question.)


Doesn't look truish based on this:

https://en.wikipedia.org/wiki/Younger_Dryas_impact_hypothesi...

but I retain some small probability that the jury is still out.


Imagine Wikipedia being wrong...

The opposition gave up publishing in 2017. There is a great Nova episode on Youtube where they sponsored a trip to Greenland to collect ice samples from that period, and had a camera in the room when the first SEM images came up. It's all-over nanodiamonds.

https://www.pbs.org/wgbh/nova/video/megabeasts-sudden-death/

That was in, what, 2010? Makes me wonder what motivated everything published against, since.


Last I read, it seems most likely that agriculture originated independently in at least 3 places around the world (the Levant, somewhere in China, somewhere in south America).


That would be consistent with it originating somewhere now submerged, then carried by the refugees to wherever they fetched up.

I was astonished to learn that wheat was farmed in China for thousands of years before they took up rice.


Nope. The date of origination in all these places and the very different type of agriculture in each location argues strongly against "carried by refugees".

Just consider the rather immense period of time between whenever the Americas were settled (currently looking like at least 30k years ago) and the emergence of agriculture.


That period without actually argues for a common origin. The places we know started around 10-12 thousand years ago. The idea of doing it at all, like the idea of writing language at all, is more important than the local details.

But the New Guinea highlands do appear to have invented agriculture, in isolation, 7000 years ago.

We don't know when agriculture in the Amazon basin started, just that it was much, much longer ago than anyone had imagined. The entire Amazon was a manicured garden when the Spanish got there. Then smallpox wiped it off the map.


Maybe I'm being dense, but given that you had effectively isolated populations in the Americas by the time agriculture was discovered/invented/developed there, and it wasn't a generation or two since they were last "connected" to any possible common origin, I don't see how this supports the idea of a common origin.

Unless you're arguing that they actually were not isolated, and there was already cultural exchange going on between other continent's populations and those in the Americas.


Just that you can't rule out things you don't know anything about. Archaeologists nowadays enforce a rule where anything they don't have firm evidence about must be absolutely assumed not to exist. But that's not science, and we have no need to toe that line.

12,000 years ago there were very different kinds of technology at different places. We are certain that Australia and New Guinea (at the time, a continent joined by the now submerged Sahul region) was settled by boat across open sea. The entire Pacific rim, from New Guinea to Chile, could have been connected by boat traffic operated by a now-vanished seafaring culture, with any remaining evidence now deep under water. Thousands of years is enough time for a very great deal to happen, and then the artifacts rot.

Absent evidence, there is no certainty. Maybe ideas arose independently all around the same time period entirely by accident, and maybe they didn't. But generally it is unwise to insist that simultaneous parallel events are entirely independent when there is possibility of contact.


> that another, only 12800 years ago, killed off most American megafauna

Hopefully his children or grandchildren don't experience another one of these.


notably a surprising number of his predictions are how current concepts are wrong, and not that new concepts exist.

yet most of the discoveries of the past, DNA, semiconductors, relativity have nothing to do with prior information being wrong,

just that prior knowledge was incomplete or insufficiently accurate.


I think that is because he can't tell you what the new idea is, but suggest the old ideas the new idea might replace. So relativity replaced aether. Some new idea could maybe replace dark matter, but he can't know what it is.


I had the same initial reaction, but I think it's because the first section of predictions (his cosmology section) only enumerates things that might be wrong, which makes a strong impression.

The predictions that best satisfy my intuition's classifier (not exactly a scientific process) are the ones that suggest an expanding of our current knowledge through the discovery of some new, fundamental thing—like a new piece of quantum mechanics, a new systems-level understanding of neural activity, new life forms that evolve our models of life, etc.


Michelson-Morley categorically refuted and falsified the aether hypothesis. DNA was confirmation of and predicted by the inheritence component of natural selection. Relativity invalidated fixed- and universal-frames-of-reference models of mechanics.

New models imply falsification or rejection of prior models.

Semiconductors were more an emergent materials-properties discovery, though with analogues in earlier systems (diodes, vacuum tubes, physical gates and valves, etc.). Not so much a model as a discovered behaviour.


>In just the last few years we have realized that human bodies contain ten times more cells that are bacteria, than they contain cells that have our DNA in them, though the bacterial mass is only about 3% of our body weight. Before that we thought we were mostly us.

I am really interested to see where this goes. How much of this bacteria determines which diseases we get?


Me too. But I was distracted by the author saying at the end "Before that we thought we were mostly us."

Idk what conventional opinion would be, but the mass part is a lot more important in my mind than the number of cells part. We're still be 97% 'us' by mass. As an off the top of my head example, Jupiter is mostly the big gas planet thing, not the 79 or so moons that orbit it.


There’s a big meta-question, pertaining to pace of progress: How will the process of scientific research itself change?

There are drastic changes on the horizon, both positive and negative.


>There’s a big meta-question, pertaining to pace of progress: How will the process of scientific research itself change?

I suspect that progress on certain kinds of science, especially physics, will slow. The reason is that confirming things at finer and finer grains of detail experimentally gets multiple orders of magnitude more expensive over time.

For example, Newton managed to empirically test many of his ideas by himself, in a cottage, with minimal equipment. Later discoveries have required fully-staffed laboratories, later still and we needed to construct multi-billion dollar supercolliders and deep space missions to empirically test our ideas.

As a layman, I've read that it could take solar-system sized experiments to test theories like String Theory. Eventually, the cost becomes too great in absolute terms and "real" discovery slows. (Theory will continue to develop, but it will be based on shakier and shakier foundations over time. Eventually we'll need experimental confirmation or what we are doing is no longer "science").


I don't believe it will be "no longer science". Physics is still underpinned by mathematics. It'll just be a branch of pure mathematics.

From what I understand, quite a lot of pure math research has begun thanks to theoretical physics. Here is such a Quanta article: https://www.quantamagazine.org/how-quantum-theory-is-inspiri...


We seem to be in an anti-science phase right now - death threats against Dr. Fauci are a poignant example. Hopefully this won't last too long, but it could definitely slow down the pace of discovery. Back in the 50s-70s era the US recognized that they needed to emphasize science training in order to win the Cold War. A lot of progress then was driven by the space race. Public funding for science was lavish in that era as compared to the current era.

Hopefully our current anti-science, anti-intellectual phase is just that: a phase. But it's also possibly a harbinger of a longer dark age to come, at least for the US.


> We seem to be in an anti-science phase right now - death threats against Dr. Fauci are a poignant example.

As always, people are on a spectrum. There will always be vocal anti-science minorities (flat-earthers, anti-vaxxers, etc.) no matter how enlightened the civilization. But hopefully they largely stay that way - minorities.

Furthermore, there are people who are quick to trust new science and people that are slow to trust new science (think people that instantly embrace the latest study vs. people that are instantly skeptical of the latest study). Both sets of people may be completely pro-science and trust in the scientific method, but the disagreement arises on how fast the scientific method converges on truth, the perceived strength of the presented evidence, etc. Sometimes I think we are too quick to label people in the second set as "anti-science".


Are they minorities though? I think current anti-science stances are far more widespread than your post seems to give credit for.

I argue that the more abstract science is, the easyer is for the laymen to disagree with it. Sometimes in _highly_ unproductive ways.


> Are they minorities though? I think current anti-science stances are far more widespread than your post seems to give credit for.

It is all speculation until we have real hard numbers. It's easy to argue in both directions.

People believed all kinds of conspiracies when I was a kid. And people still do. The only difference is that it's more visible now. In the pre-WWW days you'd only hear it if you were in the room with them. Now you can see it even if they said it days ago.

Carl Sagan wrote The Demon Haunted World in the 90's, and it is mostly about exactly this: Whereas science had made tremendous progress in his life, the attitudes and crazy beliefs people have had not shifted much. The whole book is a lament about this, and a certain amount of bitterness on this being the case despite him spending most of his life promoting science.

Probably the only real thing that has changed is the role and type of media people consume. In the 80's if a whole bunch of people believed something that was clearly wrong, and were geographically spread out, they didn't have an easy means to make that an issue for elections. You would need to form a lobby group which takes time and money. The newspapers usually did not entertain them.

Now you don't need newspapers and news channels to get your message out. You have Twitter, Youtube, Facebook, etc. The cost to organize and spread your message is much lower. So tiny fringe groups have the potential to be much more influential. The politicians can now exploit that.


> Are they minorities though? I think current anti-science stances are far more widespread than your post seems to give credit for.

There definitely are a group of hardcore anti-science people, but I think it's far less widespread than you credit. It may appear more widespread because there is a tendency by some to present public policy preferences as 'science'. Since public policy choices always involve specific intention (what is our desired outcome) and trade-offs (what policies are we precluding by choosing this one) they are not 'science', they are rational proposals based on scientific evidence, and in some cases they can be rationally disagreed with in terms of both intents and trade-offs.


Again though, people are on a spectrum. It's not strictly "I trust all science" vs. "I don't trust any science".

It could be "I trust 'pure' science but I can't tell if this particular science has been tainted by politics or money or corporate interests so I'm not sure I trust it". So any given pro-science person might hold one or more "anti-science stances" for various reasons.


I agree that anti-science bias is more widespread than some here seem willing to admit (perhaps they just haven't noticed or hang around in circles where it's rampant?). Anti-vaxxes, climate change deniers, anti-maskers - it seems pretty rampant in parts of the US (especially in the red states). We like to think that society moves linearly "up and to the right" - towards more enlightenment, but history shows that that's not always the case and we seem to be living through a period of history where it is not the case.

Related article: Combating antiscience: Are we preparing for the 2020s https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7141687/


Those "anti-science" people are a minority and they aren't really anti-science. They are only anti-science for the few things that conflict with the way they -want- the world to be. They are fine with the other 99% of science, though they won't likely be contributing much to it.


There is going to be a huge shift in scientific practice as a result of the success of GPT-3 and other similar breakthroughs in NLP. The methodology that produced GPT-3 is going to work for many other domains, not just text. The process is basically this:

1. Compile a huge dataset of observations relevant to your domain

2. Using the cross-entropy loss or some close variant, train a raw data statistical model of the observations

3. Make the model immensely complex (GPT-3 has 175 billion parameters); this can be done without overfitting because of the huge size of the dataset.

4. Once the raw data model is available, reuse it for various related tasks of practical interest. The immense sophistication of the raw data model will make the practical task much easier.

GPT-3 demonstrates clearly that this methodology works, and produces far more powerful results than other more mainstream approaches in NLP that depend on laborious human annotation. In my view it is nearly certain to produce comparably dramatic results in other domains (I'd be happy to bet on this if anyone is interested).

Here's a longer essay on this topic: http://danburfoot.net/research/RediscoveryInteriority.html


what practical task in science will gpt-3 like systems improve?


Applied genetics. The mapping from gene to phenotype is not, shall we say, injective. Bring in the complexity of epigenetics, biochemistry and you have a complex mess. Biological systems are tremendous black boxes with many second and third order effects that cannot be efficiently simulated and modelled (you can get pretty good approximations within a very narrow scope with supercomputers but those tend to have little practical use outside of basic scientific research).

Another use case would be for pharmaceutical research, there are a number of companies already in the ML -> drug pipeline space but GPT has brought new insight to the table.


"Progress disappoints in the short run, surprises in the long run." -- Jim Keller


> We discover life elsewhere in the solar system and it is clearly not related to life on Earth.

I still remember a cartoon one of my father's colleagues had posted on his office door: a UFO crashed in the desert, with a little green man crawling away under a blazing sun, crying out "ammonia! ammonia!"

https://i.imgur.com/NLy7Klz.jpeg


> All animals, plants, and fungi, belong to the latter class. But in fact there are two very distinct sorts of prokaryotes, both single celled, the bacteria and the archaea. The latter were completely unknown until the first ones were found in 1977.

Wait, I can't parse that sentence. Does it mean to say bacteria (`the first ones`) were only found in 1977 ?


The latter (archaea) were completely unknown until the first (archaea) were found in 1977.

It's not a well-written sentence.


You're the only one of (currently) four replies to correctly identify that the sentence is confusing because it's ambiguous what "first one" refers to, rather than because the GP doesn't know what archaea are or what "latter" means.


> The latter (archaea) were completely unknown until the first (archaea) were found in 1977.

Aha ! Thanks.

Quite frankly the rest of the article isn't easy to decipher either (but English isn't my first language).

The list of beliefs/theories that could change throws me off because some statements are about some topics that I thought most were in agreement (eg: the universe is expanding)

> Things will continue to change. Below I have put a few things that I think could change from now into the beginning of the next century.

> - The Universe is not expanding.

The cosmology list is more about fact waiting to happen that expansion of knowledge but I am just a layman.


By the time we experimentally confirm that the entire universe isn't expanding, we may have transitioned into the phase of universal contraction.

Or something.


What's the other interpretation of this sentence? "The latter" comes directly after an enumeration, with arachea as the last term. It seems pretty clear?


"First ones" could mean "former", as in comparison with latter:

> ...the bacteria and the archaea. The latter (archaea) were completely unknown until the first ones (bacteria) were found in 1977.


Latter of the two, "Bacteria and Archaea" being Archaea

"In 1977, Carl Woese, a microbiologist studying the genetic sequences of organisms, developed a new comparison method that involved splitting the RNA into fragments that could be sorted and compared with other fragments from other organisms."

[1] https://en.wikipedia.org/wiki/Archaea



That's how I read it too... equally confused...


His list seems pretty reasonable to me, and in agreement with speculation I've heard from people in some of those fields. Unfortunately, I think that while our understanding of some of these fields will change, the practical applications will be limited (maybe with the exception of some developments in biology)


Its easy to lose sight of social change too.

In 2008, when Joe Biden debated Sarah Palin, they both supported civil unions over gay marriage because that was the centrist position at the time.

When the author was born, the USA was over 90% christian, and less than 5% of people had no religion. Today the country is closer to 70/20.

When the author was born, black people were still being lynched.


> When the author was born, black people were still being lynched.

The BLM movement would argue that they effectively still are.


After the "Red Summer" of 1919, in which black people fought back against attacks by white people, whites blamed the violence on "communists" inciting the black community to riot.


Then they would be wrong.


Unfortunately you can't ask Oscar Grant, Ahmoud Arbery, Atatiana Jefferson, Breonna Taylor, Jordan Edwards or the other countless POC who are killed disturbingly regularly by our police whether they'd count their deaths as the modern equivalent of lynching.


> Lynching is a premeditated extrajudicial killing by a group. It is most often used to characterize informal public executions by a mob in order to punish an alleged transgressor, punish a convicted transgressor, or intimidate. (Wikipedia)

These deaths can hardly be called premeditated extrajudicial mob executions, with the sole exception of Ahmoud Arbery.


But is it the equivalent of lynching? Since premeditated execution by cop is in several ways worse, and in most ways about the same, I would say yes.


Which of these deaths were planned out in advance?


Will we get flying cars in active use? :)


The change in how much the average person knows must have been equally drastic. Is there a statistic of how many words the average person has at least heard of?


Lately I've been having more conversations with older people. It's fascinating hearing stories about such a different world.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: