> Aahh! Never mind!
@@ -187,6 +220,72 @@
It's conceivable before long that there will, at some point, be robots that can run 100 metres faster than Usain Bolt or shoot lower scores on the golf course than Tiger Woods, even at their best. But would we be interested in this? You might well be if you are fascinated by robotic performance. But most of us were thrilled by Bolt and Woods in their prime precisely because they were flesh-and-blood humans like us... When we read great literature or listen to fine music or view superb paintings, part of the thrill is precisely that another human has been involved in the work - striving, communicating, creating, and, in turn, inspiring, stimulating, and elevating our lives. Again, an indispensable and intrinsic part of that experience is the knowledge that another human is at the other end... And so, no matter how capable our systems are, it's likely that many forms of human expression, not least live performance, will continue to be valued by humans for their own sake. |
+
+
+
+ | The AI Delusion |
+ Gary Smith |
+ Academic |
+
+
+ | The average brain has nearly 100 billion neurons, which is far, far more than can be replicated by the most powerful computers. On the other hand, compared to humans, the African elephant has three times as many neurons and one dolphin species has nearly twice as many neurons in the cerebral cortex. Yet, elephants, dolphins, and other creatures do not write poetry and novels, design skyscrapers and computers, prove theorems and make rational arguments. So it isn't just a numbers game. |
+
+
+
+
+
+ | Ethical Machines |
+ Reid Blackman |
+ Casual |
+
+
+ | Does your head of HR know about Amazon's biased hiring AI? Do they know how it happened? Do they know how bias can creep into an AI? Do they know the ethical, reputational, and legal implications of using biased AI? Does your chief medical officer know Optum's AI recommended paying more attention to white patients than to sicker Black patients? Are your doctors and nurses familiar with it? Does your advertising agency know about Facebook's AI that advertised houses for sale to white people and houses for rent to Black people? |
+
+
+
+
+
+ | Co-Intelligence |
+ Ethan Mollick |
+ Casual |
+
+
+ | Another consequence is that we could reduce the quality and depth of our thinking and reasoning. When we use AI to generate our first drafts, we don't have to think as hard or as deeply about what we write. We rely on the machine to do the hard work of analysis and synthesis, and we don't engage in critical and reflective thinking ourselves. We also miss the opportunity to learn from our mistakes and feedback and the chance to develop our own style. |
+
+
+
+
+
+ | AI Snake Oil |
+ Arvind Narayanan & Sayash Kapoor |
+ Casual |
+
+
+ | To generate a single token - part of a word - ChatGPT has to perform roughly a trillion arithmetic operations. If you asked it to generate a poem that ended up having about a thousand tokens (i.e., a few hundred words), it would have required about a quadrillion calculations - a million billion. To appreciate the magnitude of that number, if every single person in the world together performed arithmetic at the rate of one calculation per minute, eight hours a day, a quadrillion calculations would take about a year. All that to generate one single response. |
+
+
+
+
+
+ | If Anyone Builds It, Everyone Dies |
+ Eliezer Yudkowsky & Nate Soares |
+ Casual |
+
+
+ And how many families would still own an original biological dog, we wonder, if with biotechnology you could make a synthetic sort of dog that was just as bouncy and cuddly and cheerful, and never threw up on your couch or got sick and tragically died? If it's just an option being offered in pure imagination and theory, it's easy to say no, when you don't have to pay for that in stained couches and crying children. But we wouldn't bet on conventional dogs being popular a hundred years later if those sorts of dogs come onto the market.
Similarly, human beings are not likely to be the best version of whatever the AI wants - if those preferences even involve keeping something vaguely human-shaped around, if it even has any preferences like that at all.
We would not be its favorite things, among all things it could create. |
+
+
+
+
+
+ | The Myth of Artificial Intelligence |
+ Erik J. Larson |
+ Academic |
+
+
+ | We can summarize these positions about AI and people as follows. Kurzweilians (mythologists about AI, full-stop) wax mystical about machines after the Singularity having consciousness, emotions, motives, and vast intelligence... Russellians want to keep Ex Machina in movies, downsizing talk about superintelligence to more mathematically respectable ideas about general computation achieving "objectives." Unfortunately, Russellians tend to lump human beings into restricted definitions of intelligence, too. This reduces the perceived gap between human and machine, but only by reducing human possibility along with it. |
+
+