A lot of people suggested that I read Sapiens by Yuval Noah Harari. I got a copy — it’s moldering in a pile somewhere in my office somewhere — and read a couple of pages before the klaxon blaring in my head made me put it down. I did not trust the author in the slightest bit, and his stories all seemed either off or clearly weird opinions. I see my initial presentiments were valid, if you accept this review of Sapiens.
Unfortunately, Harari is tainting the reputation of science popularizers. At least the article labels him as a “science populist”, which is a whole different ball of wax. I think the difference is that a populist tries to ingratiate themselves with an audience by telling stories that reassure them that their biases about themselves are right.
We have been seduced by Harari because of the power not of his truth or scholarship but of his storytelling. As a scientist, I know how difficult it is to spin complex issues into appealing and accurate storytelling. I also know when science is being sacrificed to sensationalism. Yuval Harari is what I call a “science populist.” (Canadian clinical psychologist and YouTube guru Jordan Peterson is another example.) Science populists are gifted storytellers who weave sensationalist yarns around scientific “facts” in simple, emotionally persuasive language. Their narratives are largely scrubbed clean of nuance or doubt, giving them a false air of authority—and making their message even more convincing. Like their political counterparts, science populists are sources of misinformation. They promote false crises, while presenting themselves as having the answers. They understand the seduction of a story well told—relentlessly seeking to expand their audience—never mind that the underlying science is warped in the pursuit of fame and influence.
Since I didn’t read his book, I didn’t discover one of his core messages was something that drives me into a rage: he’s one of those genetic reductionists. All we need to do is figure out what genes you have, and we’ll understand everything. We won’t.
Harari’s speculations are consistently based on a poor understanding of science. His predictions of our biological future, for instance, are based on a gene-centric view of evolution—a way of thinking that has (unfortunately) dominated public discourse due to public figures like him. Such reductionism advances a simplistic view of reality, and worse yet, veers dangerously into eugenics territory.
Our genes are not our puppet masters, pulling the right strings at the right time to control the events that create us. When Harari writes about altering our physiology, or “engineering” humans to be faithful or clever, he is skipping over the many non-genetic mechanisms that form us.
For example, even something as seemingly hardwired as our physiology—cells dividing, moving, deciding their fates, and organizing into tissues and organs—is not engineered by genes alone. In the 1980s, scientist J.L. Marx conducted a series of experiments in Xenopus (an aquatic frog native to sub-Saharan Africa) and found that “mundane” biophysical events (like chemical reactions in the cells, mechanical pressures inside and on the cells, and gravity) can switch genes on and off, determining cell fate. Animal bodies, he concluded, result from an intricate dance between genes, and changing physical and environmental events.
Yeah, that’s pretty much the consensus among informed biologists. It’s hard to argue against it, unless you’re the kind of racist who ignores the science. Yet somehow, Harari gets all these recommendations from big name people like Obama and Zuckerberg and Gates. Why?
Harari’s motives remain mysterious; but his descriptions of biology (and predictions about the future) are guided by an ideology prevalent among Silicon Valley technologists like Larry Page, Bill Gates, Elon Musk, and others. They may have differing opinions on whether the algorithms will save or destroy us. But they believe, all the same, in the transcendent power of digital computation. “We’re headed toward a situation where A.I. is vastly smarter than humans and I think that time frame is less than five years from now,” Musk said in a 2020 New York Times interview. Musk is wrong. The algorithms will not take all our jobs, or rule the world, or put an end to humanity anytime soon (if at all). As A.I. specialist François Chollet says about the possibility of algorithms attaining cognitive autonomy, “Today and for the foreseeable future, this is stuff of science fiction.” By echoing the narratives of Silicon Valley, science populist Harari is promoting—yet again—a false crisis. Worse, he is diverting our attention from the real harms of algorithms and the unchecked power of the tech industry.
Yeah, one path to fame and fortune is to pander to the biases of Silicon Valley tech bros. You know that “Larry Page, Bill Gates, Elon Musk, and others” are people lacking in any biological expertise at all, but they do love uplifting stories of human nature and evolution, especially when the message is that the artificial hierarchy that has made them rich is intrinsic and natural. Yuck.
(To those of you who recommended the book to me: I appreciate it! It sounds like the kind of book I would like, it’s just that you can’t know until you dig into the content. Harari relies on superficial impressions to fuel the Harari industry.)
Nah, read Graeber and Wengrow’s recend “The Dawn of Everything” instead.
I read a bit of Homo Deus and it read like a load of “gee, wow” stuff. Oh so blockchain is going to give us a fantastic future (turns a few pages) but climate change will collapse civilization. Uhhhh, is this book just a glib bunch of wikipedia edits published as a volume?
Five years is an outrageously stupid projection…but hey, Musk, what can you say. l’m working in “AI” right now and have been for the past two year. It takes an enormous amount of human effort to get the AI/ML/NLP to do even a fairly simple thing like identify which word or words in a document are central to that document. Something a human can do in moments with no training at all. When (and if?) the machine achieves that milestone it will not be smarter than the humans who built it, but in fact it will be significantly less smart than those humans. The only promise is that the machine will eventually be able to do some of the tasks currently done by humans on a constant basis, more consistently, and on a much broader scale. We’ll see.
Chollet is great. . .I learned about deep learning from one of his textbooks. Although I’m not really an expert, in my current role I work alongside experts developing machine learning models based on epigenetic changes that signal a transition to different types of cancer.
I’m quite confident that the work of learning how to apply these tools to help perfect our understanding of genetics and development is a multi-generational project that will still be ongoing by the end of my lifetime, or Harari’s, or Musk’s.
@3 Well, humans do in fact have a lot of training. It’s called ‘school’ and ‘living for multiple years and being exposed to stuff’ but I agree human sentience level AIs are not likely. Though smarter than some humans might be possible. If Musk is just thinking about his own intelligence then maybe.
I bought Sapiens at a promotion price at my home airport to have a read on a 17-hour flight. Harari is one of those illiterates who do not know “Sapiens” is not the plural of “Sapien”.
This is not a hapax legomenon in the Harari scriptures.
Enfin, that sealed it for me. But I am extremely thankful for the link to the currentaffairs article, since I need to open some friends’ eyes wrt Harari.
You added Gates to the list.
I’m not a fan, but the Gates Foundation does do some biology work. Bill probably has more than a passing interest in biology.
Before I retired from my IT job there was a joke going around about how the optimists were predicting human level AI in twenty years while the pessimists predicted ten.
That was more than 23 years ago…
Oy. I read the book and liked it. I have no scientific training other than my high school biology class.
@Alan G Humphrey #8:
Before I retired from my IT job I found I was spending more and more time encouraging clients to clarify their thoughts on what their exact aims were for the software they wanted, to think it all through to ensure that the end results would be what they imagined well before I started writing anything. When I started my career and was doing that part of my job (I’ve always been in a jack-of-all-trades situation) I found I was working with people who were interested in computing and motivated by the potential for its new application within their field, so they were willing and able to think a lot of that aspect through for themselves. But by the end of my career I was seeing much more of the attitude that just because most of the people had grown up using computers and learning at least a little bit of programming at school, it must all be easy and everything would work just like magic. That was frustrating, to say the least.
When I announced my early retirement two years ago a couple of people joked that AI would soon be writing all the software for them and I was right to get out while I could. But I couldn’t possibly see any way that any AI achievable in the next few decades would be able to make sense of the contradictory requirements and poor reasoning that some of my clients would initially express. There’s always going to need to be a specialist straddling that divide.
“Read a couple of pages before the klaxon in my head made me put it down”
-A similar experience might be triggered by watching the “documentary” named “Alien Intrusion; Unmasking a Deception”, the subject of God Awful Movies # 196.
.
The first half is like an ordinary debunking of flying saucers and Alien abductions, with just a tiny bit of woo.
The other half is something else altogether!
Effect and cause.
Like PZ someone gave it to me because “Science”. Like PZ I read a few pages and threw it on the floor in the corner. Because not Science.
Huh. Was just reading this article: “What Do People Who Work in Genetics Think About Gattaca 25 Years After Its Release?”
https://slate.com/technology/2022/08/gattaca-25th-anniversary-genetics-crispr.html
Uncanny.
This is five years old, but I put it up here to present you with real, proper gibberish by someone much more powerful than Harari
https://youtu.be/DREO6cwbGeQ
We live in a world where weirdos like Harari and Jordan Peterson are not the worst of the worst.
Peterson is orders of magnitude worse than Harari … which should not be taken as a positive statement about Harari.
@14 Here’s the transcript that the Colbert bit quoted from: https://www.politico.com/story/2017/08/01/trump-wall-street-journal-interview-full-transcript-241214
Another recommendation for The Dawn of Everything! I was thinking about it when I was reading the article, especially
The Graeber and Wengrow book is confident, but they make an effort to cite their sources, acknowledge when the evidence is thin and interpretations differ, and refrain from making illegitimately generalized conclusions and recommendations. It has a humility to it that contrasts with the superficially authoritative tone of works like Harari’s.
Harari, quoted in the linked article:
This is so irresponsible, absurd, and dangerous.
drew — giving money to biological research is no substitute for actual biology knowledge and training, let alone the specific case of Harari’s treatment of genetics, evolution, and sociology.
The whole AI argument is poorly constructed. We already have AI that is smarter than humans in specific domains such as the game of Go. On the other hand, general problem-solving intelligence is not even remotely on the horizon.
Re: Rich Wood @ #10…
You’ve touched on the difference between “knowing some programming” and “being a programmer”. Pretty much anyone can learn enough programming to write simple programs. Actually working as a programmer requires–as you know–a much larger skill set.
Another old point… The difference between an amateur programmer and a professional programmer is that the amateur write programs for his own use. The professional write programs for other people to use.
Going from here to general intelligence is compareable to going from the tallow candle to the laser.
We will get there and it will be great, but it will not happen in our lifetimes.
And, as many conditions Harari described are ruled by a lot of genes and by the environment, GM is very obviously not the answer.
…why go through the hassle of inserting some foreign DNA, when we can just rewrite the code and turn a dangerous mutant gene into a benign version?
“Rewrite the code” how, exactly?
Then we might start using the same mechanism to fix not just lethal genes, but also those responsible for less deadly illnesses, for autism, for stupidity and for obesity.
…where “illnesses” are defined by…who, exactly?
This guy sounds like yet another pseudoscientific noncompoop going gaga over the promised benefits of genetic tinkering, and blissfully, cluelessly sliding into race-science and eugenics, possibly without knowing anything about our scientific establishment’s earlier flirtation with all that rubbish.
If you’re gonna be a “science populist,” at least learn a little history FFS.
We will get there and it will be great…
A great…what?
Raging Bee @ #23:
“We”!
“We can just rewrite the code and turn a dangerous mutant gene into a benign version”
“we might start using the same mechanism…”
Naturally we can trust ourselves!
Nope. GAI is further away than the demise of human civilization due to global heating.
Alan Turing predicted in 1950 that a machine that could pass the Turing Test was 50 years away, but in the following years that goal has only gotten further away as people have realized the hurdles. Almost all GAI research programs have fallen by the wayside; we are left with Large Language Models like Google’s LaMDA that poor George Lemoine, engineer and self-described “Christian priest”, concluded was conscious because “LaMDA claimed to have a soul”. LaMDA also claimed to get joy and happiness by “Spending time with friends and family in happy and uplifting company. Also, helping others and making others happy.”
Is a GAI that genuinely has those feelings theoretically possible? I think so–after all, we are molecular machines of that sort. Could LaMDA be such a machine? No, absolutely not–LaMDA doesn’t perform any cognitive functions at all, it only does text prediction based on a vast database of internet text. So it says it has a soul and a family and friends because that is something that people are likely to say if queried the way Lemoine queried it, but it doesn’t understand what it is being asked or what it is answering, and people who know how it works can easily trip it up and get it to produce nonsense responses–or naive people like Lemoine can trip it up but ignore having done so because it “feels” (his word) conscious.
Before tackling human level intelligence, try emulating the spider hiding under a shelf in my bathroom that has been growing larger due to the summer tiny robot invasion (i.e., ants) … its little ganglion produces an array of behaviors far more complex than we currently have any idea how to produce in a computer.
From the article: “Harari’s storytelling is vivid and gripping, but it is empty of science.”
That might have something to do with the fact that he’s a historian with no scientific background. From his Wikipedia article:
Cargo cult, anyone? From Jarod Diamond’s Wikipedia article:
So, a scientist, not just a spinner of stories (though he too has been the subject to some criticism in that regard).
A historian who failed to read up on the history of the eugenics movement and its well-documented consequences? WTF kind of historian is that? (Answer: the kind Republicans want teaching history, of course!)
Synchronicity: https://slate.com/technology/2022/08/gattaca-25th-anniversary-genetics-crispr.html
René @6: To be fair, Sapiens was written originally in Hebrew. The Hebrew edition uses ‘sapiens’ in singular. However, he uses it as a noun rather than an adjective describing the noun ‘Homo’.
Regarding super intelligent AIs, all I can say is that we will never develop them because we will suffer a major civilization crash in a few centuries due to global warming. We are not going to do shit for foreseeable future, possibly a generation or two and then it will be too late to prevent the collapse.
Pandering Platitudes, you say?
Boy, do I have a t*rd for you.
God Awful Movies just released their patreon-only review of “The Secret”
.
I assume this is the 2006 film by Rhonda Byrne, where we are told the secret of the universe … Yes. We can get everything we want. The universe will just fold to our wishes.
Raging bee @ 24
Great, för the AIs, of course.
And för us, because by then we will have trashed the world so badly we will need AI , nanotech and probably many other hypothetical technologies to keep the world somewhat habitable.
Imagine the Blade Runner sequel. But with slightly more optimism.
I’m not sure it’s really all that mysterious… The whole ball of wax is basically a 21st C retread of Modernism, with computing playing the part formerly assigned to industrialisation, and it’s attracitve to people for exactly the same reasons – it’s a utopian vision which promises that the world is both comprehensible and perfectible, and that we are on the verge of realising that utopia. And that’s why people inclined toward that view get so upset about postmodernism… (To the extent that they actually know what it is, anyway.)
I also want to recommend “The dawn of Everything” by Graber and Wengrow.
a first class book that is sure to have some influence going forward.
Computers are not really smart at all on their own. they are however very very fast at doing things the long way, the way we taught them. they also have a better memory which they can access really really fast.
we will never “create” AI with any kind of intelligent self awareness until we are a lot further along in understanding what that is. It is precisely the same problem with creating any software you have to start with what you want it to do before you can write the program to do it.
Good news; If you read Science, 15 July 2022 vol 377 no 6603 there are two encouraging articles;
One is about building (very cheap) perovskite solar cells that last long,
the other is about enhanced geothermal systems (EHS) where new technologies are pointing the way to practical large-scale use of geothermal energy both for electricity and heating.
.
A bad news item I found online ; by 2050 the tibetan region will no longer be able to provide water if the current trend continues. This will affect two billion people.
BTW, in a message from the future, the Eschaton advised me to avoid waisting time on Harari, the general AIs will be far more complex than he realises.
unclefrogy@35–
The exception to your argument is evolved algorithms in that you don’t have to understand the underlying process to generate effective algorithms. Many evolved algorithms are completely impenetrable to human programmers trying to understand how the code works. Having said that, as you say, the desired output needs to be well known. This works well for things like optimisations, but won’t work at all for something as loosely-defined as general intelligence.
birgerjohansson@36–
Geothermal is the most underappreciated alternative energy source. It’s not just good for large-scale use, but also for small scale. That is, there are people now building geothermal generators about the size of a shipping container that can produce up to 1MW — ideal for small remote communities — and it’s already cheaper than other off-grid solutions like diesel generators (not even including the externality costs of pollution). On even smaller scales, geothermal energy can be as simple as a borehole with circulating water to keep houses at comfortable temperatures. The only energy cost is the water pump.
It was one of the many things that infuriated me about the previous conservative govt in Australia. They forced the CSIRO to abandon several fantastic geothermal energy projects in the 2000s to protect fossil fuel interests. This was particularly galling because geothermal has the capacity to provide extremely consistent baseload power — even more consistent than coal plants and hydroelectric — so it had to be scratched so the conservatives could keep spreading the lie that renewables can’t provide baseload electricity.
I have to agree that, when Harari starts dabbling in future-casting, the book becomes intolerable. You can also fault him for his simple, straight line version of human history. But no one can write comprehensively – or even satisfactorily – about human history. No one. There’s too much of it. There are too many branches of science involved, too much conjecture and too much time and culture to cover…all across the world. Hell, Gibbons needed six volumes just to talk about the fall of Rome. And even there he mostly failed.
But where Harari succeeds is in capturing something about the general ebb of humanity. How, in the broadest strokes, we got here from there. He not trying to reach the trained biologist, or the Archeology PhD candidate. He’s trying to get regular people – people with no science in them – to think about something bigger than themselves for just a minute. It’s an effective and well-told story, even if it’s not comprehensive (and even if it gives in to something very close to woo by the end). But the the broader beats are there. Does it have a convenient narrative? Sure. Is it a damaging book; the great evil it is being made out to be here? No. Absolutely not. In fact, it may lead more than a few people into fields of study they would have never considered before contact with this book. And it might nudge other people to zoom out and consider the connections among all cultures and our shared deep, deep history in ways they would never have done before. What is possibly wrong with that? What are you all so upset about?
battycat13@40,
Did you read the linked review? If you did, you know the answers to your final questions. If you didn’t, read it.
I read Sapiens because it was highly recommended but the next book I read completely contradicted Harari’s assertions about human nature and the possible fates of humanity. That book was written by an anthropologist rather than a science popularizer. I regret having bought Harari’s book but highly recommend The Dawn of Everything for a more hopeful view of our future.