<html><body><div style="color:#000; background-color:#fff; font-family:times new roman, new york, times, serif;font-size:14pt"><div><span><var id="yui-ie-cursor"></var>John Grigg wrote:</span></div><div style="font-family: times new roman, new york, times, serif; font-size: 14pt;"><div style="font-family: times new roman, new york, times, serif; font-size: 12pt;"><div id="yiv1803978647"><div style="color: rgb(0, 0, 0); font-family: times new roman, new york, times, serif; font-size: 14pt; background-color: rgb(255, 255, 255);"><div><span></span> </div><div><span>> Why could an <span id="yiv1803978647misspell-1" class="yiv1803978647mark">Mbrain</span> not continually engage in </span></div><div><span>> technological progress to improve it's defenses </span></div><div><span>> and keep pace with potential enemies? </span></div><div><span></span> </div><div><span>The cost of <span id="yiv1803978647misspell-2"
class="yiv1803978647mark">WoMD</span> is much less than the cost of</span></div><div><span>defenses against them.</span></div><div><span></span> </div><div><span><span>John Grigg wrote:</span></span><span><br><br>> Why not a combination of the two concepts, </span></div><div><span>> where you have an <span id="yiv1803978647misspell-4" class="yiv1803978647mark">Mbrain</span> that in a time of </span></div><div><span>> crisis can disperse into much
smaller units,<br>> activate stealth mode for all of them, and then </span></div><div><span>> flee/fight if necessary to various points, until </span></div><div><span>> it is safe to reassemble and regain<br>> the benefits of <span id="yiv1803978647misspell-5" class="yiv1803978647mark">Mbrain</span> computational power.</span></div><div><span></span> </div><div><span>If you have time to disperse and go stealth is the </span></div><div><span>question. A speed of light attack with no</span></div><div><span>warning would negate any such planning. Only</span></div><div><span>existing in <span id="yiv1803978647misspell-6" class="yiv1803978647mark">SND</span> mode all the time can limit such</span></div><div><span>attacks.</span></div><div><span></span> </div><div><span>Even under <span id="yiv1803978647misspell-7" class="yiv1803978647mark">SND</span> huge computational power will</span></div><div><span>be
available. I am not certain we can project</span></div><div><span>advantages more than
a few magnitudes beyond</span></div><div><span>what the human brain and currently technology</span></div><div><span>can do. Does a cluster of scattered AI say 10**4 </span></div><div><span>beyond human </span><span>really have a disadvantage over</span></div><div><span>a centralized AI of say 10**6 beyond human.</span></div><div><span>At what point does adding more and more</span></div><div><span>capability produce diminishing returns? Is there</span></div><div><span>something a 10**20 brain can do that that many</span></div><div><span>lesser </span><span>brains cannot?</span></div><div><span></span> </div><div><span>I know that at our scale better and bigger brains</span></div><div><span>do prove themselves. We also see that many of</span></div><div><span>our best brains have mental health issues.</span></div><div><span>How does mental health in AI scale with size</span></div><div><span>and capability? It is
really a question of stability</span></div><div><span>during scaling. If you put everything into a single</span></div><div><span><span id="yiv1803978647misspell-8">Mbrain</span> and it is mentally unstable is that a foolish</span></div><div><span>investment?</span></div><div><span></span> <var id="yiv1803978647yui-ie-cursor"></var></div><div><span>Dennis May</span></div><div><span></span><span> </span></div><div> </div></div></div><br><br></div></div></div></body></html>