5 ms·

Digital Reality: A Conversation with Neil Gershenfeld

gbog 11y ago
Just half-way through it, I hope oh I hope he is getting to small self-assembling digital parts. In fact just like in New Heroes 6: tiny bots assembling themselves into bigger things. [edit: I hope he is going there because I dreamt about this 6 months ago...]
bawana 11y ago
Back in the day, reading and writing was the privilege of an elite class. Did the world measurably improve when everyone could do it? Yes of course. We have more knowledge, more diversity of things, we are healthier, live longer. Our lives are 'richer. We can pack more people into smaller spaces. We can overstress our ecology, pollution is a human invention. We progressed from the middle ages' definition of a witch to a the Nazi's definition of a subhuman. We traded clubs for guns. We are still humans with an intelligence that has evolved for survival and predation. Beneath the veneer of civilization we designed to thwart the animal, we still embody a beast that is conflicting us daily, and sometimes rages to the surface when hormonally fueled. In using digital communication and digital computation, we enable our imaginations to roam vast landscapes of information but thankfully we are sandboxed in the real world. The new ideas we generate usually can only be implemented by changing our own behavior or immediately local environment. If we happen to discover a meme that is popular, we get a lot of people to copy us, maybe get VC funding and make an entity to powertool it. Or maybe get a rich government to fund it in the name of defense. But can you imagine what might happen if we blindly empower savage humanity with fabs that can make tools that can alter the fabric of life. I have no doubt that it will happen. And we will see wonderful things. I always wanted gills - to be able to swim underwater for hours. I'd settle for that mouthpiece from Star Wars. But how can we protect ourselves from the mutants that will misuse this? Why empower Al-Quaeda?? Our military already deploys these fabs to FOBs to print weapon parts, etc. Are Islamic militants not educated that they could not operate a fab liberated from an African village? When IEDs start looking like MREs, it will be too late. Neil was tasked with answering this question veiled in the civilized form of 'measuring social impact'. But he did not answer it. Some might say this is an argument for integrating an AI. A superhuman one that could decide when to shut the machine off if it was doing bad things. might someone hack the code and insert NOP instructions to disable the AI but keep the function of the code running the machine? Maybe, but an AI would detect that change and delete itself before that happens. And maybe in designing a safe AI, we'll develop a new definition of sentience and intelligence. But the fabs are here now and deep learning isnt there yet. And we havent even begun integrating Asimov's laws into the code that runs AI. Do these machines have a remote disable switch? Will Neil eat his words like Einstein, who tried to take back his suggestion of using nuclear weapons to end WW2?
- christianbryant 11y ago
  Yes, tech is moving faster than human social evolution and we continue to push ourselves hard to speed up the one while neglecting the other, or simply hoping that tech will solve our social problems. However, does it make this exploration or yearning to make possible the seemingly impossible wrong? Here's a task: For each of the very valid points you raise, propose a solution that would allow these incredible new technologies to exist while preventing the terrible consequences of the scenarios you outline from being an issue. Now you're on to something...

gbog 11y ago

Just half-way through it, I hope oh I hope he is getting to small self-assembling digital parts. In fact just like in New Heroes 6: tiny bots assembling themselves into bigger things. [edit: I hope he is going there because I dreamt about this 6 months ago...]

bawana 11y ago

Back in the day, reading and writing was the privilege of an elite class. Did the world measurably improve when everyone could do it? Yes of course. We have more knowledge, more diversity of things, we are healthier, live longer. Our lives are 'richer. We can pack more people into smaller spaces. We can overstress our ecology, pollution is a human invention. We progressed from the middle ages' definition of a witch to a the Nazi's definition of a subhuman. We traded clubs for guns. We are still humans with an intelligence that has evolved for survival and predation. Beneath the veneer of civilization we designed to thwart the animal, we still embody a beast that is conflicting us daily, and sometimes rages to the surface when hormonally fueled. In using digital communication and digital computation, we enable our imaginations to roam vast landscapes of information but thankfully we are sandboxed in the real world. The new ideas we generate usually can only be implemented by changing our own behavior or immediately local environment. If we happen to discover a meme that is popular, we get a lot of people to copy us, maybe get VC funding and make an entity to powertool it. Or maybe get a rich government to fund it in the name of defense. But can you imagine what might happen if we blindly empower savage humanity with fabs that can make tools that can alter the fabric of life. I have no doubt that it will happen. And we will see wonderful things. I always wanted gills - to be able to swim underwater for hours. I'd settle for that mouthpiece from Star Wars. But how can we protect ourselves from the mutants that will misuse this? Why empower Al-Quaeda?? Our military already deploys these fabs to FOBs to print weapon parts, etc. Are Islamic militants not educated that they could not operate a fab liberated from an African village? When IEDs start looking like MREs, it will be too late. Neil was tasked with answering this question veiled in the civilized form of 'measuring social impact'. But he did not answer it. Some might say this is an argument for integrating an AI. A superhuman one that could decide when to shut the machine off if it was doing bad things. might someone hack the code and insert NOP instructions to disable the AI but keep the function of the code running the machine? Maybe, but an AI would detect that change and delete itself before that happens. And maybe in designing a safe AI, we'll develop a new definition of sentience and intelligence. But the fabs are here now and deep learning isnt there yet. And we havent even begun integrating Asimov's laws into the code that runs AI. Do these machines have a remote disable switch? Will Neil eat his words like Einstein, who tried to take back his suggestion of using nuclear weapons to end WW2?

christianbryant 11y ago
Yes, tech is moving faster than human social evolution and we continue to push ourselves hard to speed up the one while neglecting the other, or simply hoping that tech will solve our social problems. However, does it make this exploration or yearning to make possible the seemingly impossible wrong? Here's a task: For each of the very valid points you raise, propose a solution that would allow these incredible new technologies to exist while preventing the terrible consequences of the scenarios you outline from being an issue. Now you're on to something...