1<!DOCTYPE html> 2<html> 3<head> 4 5 <meta charset="UTF-8" /> 6 7 <link rel="icon" type="image/png" href="/assets/images/rooster-thumbnail.png"> 8 9 <title>Philosophical Multicore | Michael Dickens</title> 10 11 <meta name="viewport" content="width=device-width" /> 12 13 <link href="/assets/css/colors-templatized.css" rel="stylesheet" /> 14 <link href="/assets/css/style.css" rel="stylesheet" /> 15 <link href="/assets/css/table.css" rel="stylesheet" /> 16 17 <!-- jekyll-seo-tag automatically inserts OpenGraph tags here --> 18 <!-- Begin Jekyll SEO tag v2.8.0 --> 19<title>Philosophical Multicore | Donât just not do bad things. Do good things.</title> 20<meta name="generator" content="Jekyll v4.3.4" /> 21<meta property="og:title" content="Philosophical Multicore" /> 22<meta name="author" content="Michael Dickens" /> 23<meta property="og:locale" content="en_US" /> 24<meta name="description" content="Donât just not do bad things. Do good things." /> 25<meta property="og:description" content="Donât just not do bad things. Do good things." /> 26<link rel="canonical" href="http://mdickens.me/" /> 27<meta property="og:url" content="http://mdickens.me/" /> 28<meta property="og:site_name" content="Philosophical Multicore" /> 29<meta property="og:type" content="website" /> 30<link rel="next" href="http://mdickens.me/page2/" /> 31<meta name="twitter:card" content="summary" /> 32<meta property="twitter:title" content="Philosophical Multicore" />
33<script type="application/ld+json"> 34{"@context":"https://schema.org","@type":"WebSite","author":{"@type":"Person","name":"Michael Dickens"},"description":"Donât just not do bad things. Do good things.","headline":"Philosophical Multicore","name":"Philosophical Multicore","url":"http://mdickens.me/"}</script>
34 35<!-- End Jekyll SEO tag --> 36 37</head> 38 39<body> 40 41 42<div id="page"> 43 44 45 <header id="header"> 46 <h1><a href="/"><i class="icon-sun"></i> 47 48 Philosophical Multicore | Michael Dickens 49 50 </a></h1> 51 <p id="site-description"> 52 53 Don't just not do bad things. Do good things. 54 55 </p> 56 </header> 57 58 <nav> 59 <div id="sidebar" role="navigation"> 60 <ul> 61 <li><a href="/">Home</a></li> 62 <li><a href="/about">About</a></li> 63 <li><a href="/archive">Archive</a></li> 64 <li><a href="/reading-notes">Notes</a></li> 65 <li><span class="sidebar-divider"><hr width="50%" align="right"></span></li> 66 <li><a href="https://buttondown.com/mdickens">Subscribe</a></li> 67 <li><a href="/feed.xml">RSS</a></li> 68 </ul> 69 </div> 70</nav> 71 72 73 74 <div id="content"> 75 <div id="postfeed"> 76 <nav> 77 <ul id="main-tabs"> 78 <a class="tab" href="/"> 79 80 <li>All Posts</li> 81 82 </a> 83 <a class="tab" href="/research"> 84 85 <li>Research</li> 86 87 </a> 88 <a class="tab" href="/blog"> 89 90 <li>Blog</li> 91 92 </a> 93 </ul> 94 </nav> 95 96 97 <article class="post"> 98 99 <h1><a href="/2026/09/21/campaing_spending_limits/">Campaign spending limits would reduce extinction risk (but spending limits might not accord with Constitutional principles)</a></h1> 100 101 102 <div class="post-content"> 103 104 105 <img src="/assets/images/stacked-coins.jpg" style="max-height:500px; height:auto;" /> 106 107 108 <!-- Ignore content after a 'more' tag --> 109 110 <p>Many people support stricter regulations on frontier AI, and some of those people donate to political campaigns for pro-regulation candidates. Not many people oppose AI regulations, but that group includes AI companies and execs who have giant piles of cash to throw at super PACs.</p> 111 112<p>This situation is made possible by three court decisions:</p> 113 114<ol> 115 <li>Buckley v. Valeo (1976) held that limits on independent political spending violate the First Amendment (while limits on direct contributions to candidates do not).</li> 116 <li>Citizens United (2010) extended those First Amendment protections to corporations and unions, thus allowing <em>them</em> to spend unlimited money on political ads.</li> 117 <li>SpeechNow.org v. FEC (2010) built on Citizens United to allow unlimited donations to independent-expenditure-only committeesâi.e., super PACsâas long as they donât directly work with political candidates.</li> 118</ol> 119 120<p>Donations straight to a candidate are still capped at a few thousand dollars per person; donations to a super PAC supporting that candidate are not. If SpeechNow were overturned, or a law passed overriding it, the $5,000 cap on super PAC contributions would apply again, which means the popularity of a position would matter more than how rich its supporters are.</p> 121 122<p><em>(I often hear âOverturn Citizens Unitedâ as the slogan, but SpeechNow allows super PACs to exist, and Buckley v. Valeo enables unlimited spending by individuals. What the advocacy groups in the space really want to do is restore the power of Congress and the states to enforce spending limits.)</em></p> 123 124<p>Restoring spending limits would prevent unpopular but rich anti-regulation special interests (like <a href="https://en.wikipedia.org/wiki/Leading_the_Future">Leading the Future</a>) from dominating the political funding landscape. This would make the AI regulation situation better, but it would also improve the situation on any policy issues where unpopular, wealthy special interests wield disproportionate (and arguably un-democratic) influence.</p> 125 126<p>It would also <em>worsen</em> any situation in which rich special interests are in the right, and the majority is in the wrong. But there arenât many of those, and anyway thereâs good reason to dislike an un-democratic process even if it produces a good outcome in a particular case.</p> 127 128<p>An important caveat: Before writing this post, I was confident that imposing campaign spending limits would be a good thing. After doing some more research, I find it hard to dismiss the free speech defense of unlimited campaign spending (Iâll say more about this in the next section). Both sides of the debate have points in their favor. My thesis in this post should be taken as the <em>narrow</em> claim that campaign spending limits would have positive first-order effects on AI extinction risk, and theyâre worth exploring on that basis. I donât have an established view on whether stricter spending limits would be good or bad all-things-considered.</p> 129 130 131 132 <a href="/2026/09/21/campaing_spending_limits/">Continue reading</a> 133 134 135 </div> 136 137 <div class="meta"> 138 <div class="alignleft">Posted on <span class="postdate">Sep 21, 2026</span></div> 139 </div> 140 141 <br /> 142 </article> 143 144 145 <article class="post"> 146 147 <h1><a href="/2026/09/14/population_ethics_quiz_results/">Results from the Population Ethics Quiz</a></h1> 148 149 150 <div class="post-content"> 151 152 153 <img src="/assets/images/crowd.jpg" style="max-height:500px; height:auto;" /> 154 155 156 <!-- Ignore content after a 'more' tag --> 157 158 <p>Two weeks ago, <a href="https://mdickens.me/2026/08/31/population_ethics_quiz/">I released a population ethics quiz</a>. 582 people took it.<sup id="fnref:2" role="doc-noteref"><a href="#fn:2" class="footnote" rel="footnote">1</a></sup> Hereâs what I learned from reading the responses.</p> 159 160 161 162 <a href="/2026/09/14/population_ethics_quiz_results/">Continue reading</a> 163 164 165 </div> 166 167 <div class="meta"> 168 <div class="alignleft">Posted on <span class="postdate">Sep 14, 2026</span></div> 169 </div> 170 171 <br /> 172 </article> 173 174 175 <article class="post"> 176 177 <h1><a href="/2026/09/07/journal_entries/">Some things I wrote in my journal that might be fit for public consumption</a></h1> 178 179 180 <div class="post-content"> 181 182 183 <img src="/assets/images/sockpuppet.jpg" style="max-height:500px; height:auto;" /> 184 185 186 <!-- Ignore content after a 'more' tag --> 187 188 <p>Iâm still analyzing the results from the <a href="https://mdickens.me/pop-ethics/">population ethics quiz</a> from last week, so in the meantime here are some notes Iâve written to myself that you might find kinda neat.</p> 189 190 191 192 <a href="/2026/09/07/journal_entries/">Continue reading</a> 193 194 195 </div> 196 197 <div class="meta"> 198 <div class="alignleft">Posted on <span class="postdate">Sep 07, 2026</span></div> 199 </div> 200 201 <br /> 202 </article> 203 204 205 <article class="post"> 206 207 <h1><a href="/2026/08/31/population_ethics_quiz/">Population ethics is a big deal, which is why I made this population ethics quiz</a></h1> 208 209 210 <div class="post-content"> 211 212 213 <img src="/assets/images/mere-addition-ladder.png" style="max-height:500px; height:auto;" /> 214 215 216 <!-- Ignore content after a 'more' tag --> 217 218 <p>TLDR: Take the population ethics quiz here: <a href="https://mdickens.me/pop-ethics/">https://mdickens.me/pop-ethics/</a></p> 219 220<p>Population ethics is an oft-overlooked subfield within ethics. Many people hold views that they donât realize contradict each other, or that have strange implications that they wouldnât endorse if they thought about it more.</p> 221 222<p>Not just thatâpopulation ethics is a BIG DEAL. A lot of ethical decisions hinge on how you think about changes in future populations.</p> 223 224<p>Iâll give concrete examples below the fold, but reading the examples might change your answers, so you should <a href="https://mdickens.me/pop-ethics/">take the quiz first</a>. It has 13 to 17 short questions. (Questions are skipped when they canât affect the results.)</p> 225 226 227 228 <a href="/2026/08/31/population_ethics_quiz/">Continue reading</a> 229 230 231 </div> 232 233 <div class="meta"> 234 <div class="alignleft">Posted on <span class="postdate">Aug 31, 2026</span></div> 235 </div> 236 237 <br /> 238 </article> 239 240 241 <article class="post"> 242 243 <h1><a href="/2026/08/24/don't_treat_ai's_skillset_as_static/">Don't Treat AI's Skillset as Static</a></h1> 244 245 246 <div class="post-content"> 247 248 <!-- Ignore content after a 'more' tag --> 249 250 <p><em>or: <a href="https://mdickens.me/2026/03/29/future_will_be_weirder_than_that/">The Future Will Be Weirder Than That</a>, Part II</em></p> 251 252<p>People often treat AI as if it will improve on its current capabilities, but wonât gain any new ones. Supposedly, AI will continue to get better at coding and at assisting with research, but it wonât broadly replace human labor. It wonât be able to do good strategic planning, and certainly wonât know how to operate in the physical realm.</p> 253 254<p><em>Maybe</em> thatâs true. But we donât have strong reason to believe it. Just looking at the past few years, AI has rapidly acquired new skills. The simplest extrapolation of recent trends is that it will continue to do so.</p> 255 256<p>Some people think of AI capabilities growth like this, with each colored area representing another few years of progress:<sup id="fnref:1" role="doc-noteref"><a href="#fn:1" class="footnote" rel="footnote">1</a></sup></p> 257 258<p><img src="/assets/images/AI-deepening.png" alt="" /></p> 259 260<p>But itâs more likely to look like this:</p> 261 262<p><img src="/assets/images/AI-broadening.png" alt="" /></p> 263 264 265 266 <a href="/2026/08/24/don't_treat_ai's_skillset_as_static/">Continue reading</a> 267 268 269 </div> 270 271 <div class="meta"> 272 <div class="alignleft">Posted on <span class="postdate">Aug 24, 2026</span></div> 273 </div> 274 275 <br /> 276 </article> 277 278 279 <article class="post"> 280 281 <h1><a href="/2026/08/17/unawareness_reply/">Reasons to Act on Your Best Guess: A Reply to the Challenge of Unawareness</a></h1> 282 283 284 <div class="post-content"> 285 286 287 <img src="/assets/images/Conway-spaceship.png" style="max-height:500px; height:auto;" /> 288 289 290 <!-- Ignore content after a 'more' tag --> 291 292 <p><em>This post is a submission to the <a href="https://forum.effectivealtruism.org/posts/ZQhThndgHNXEKFdSx/cluelessness-critiques-week-discussion-thread">Cluelessness Critiques Essay Competition</a>.</em></p> 293 294<p><em><a href="https://mdickens.me/confidence_tags/">Credence</a>: Likely.<sup id="fnref:27" role="doc-noteref"><a href="#fn:27" class="footnote" rel="footnote">1</a></sup></em></p> 295 296<p>Anthony DiGiovanniâs sequence, <a href="https://forum.effectivealtruism.org/s/rHqdsrieinyhM5KDv">The challenge of unawareness for impartial altruist action guidance</a> (henceforth âthe sequenceâ), argues that impartial altruists have no non-arbitrary reason to prefer any decision over any other. The central problem is <strong>unawareness:</strong> âmany possible consequences of our actions havenât even occurred to us in much detail, if at all.â</p> 297 298<p>I disagree. Impartial altruists should act <em>as if</em> we have precise probabilities for beliefs and precise expected utilities of decisions, even if we (usually) do not explicitly state them. Itâs okay to act on heuristics, but those heuristics are <em>in service</em> of maximizing the expected value of a (non-literal) utility function, rather than being âterminalâ heuristics.</p> 299 300<p>âAct as ifâ means: either I give an explicit precise credence; or I donât, but I operate under the assumption that I <em>could</em> give a precise credence if I spent enough time reasoning through my beliefs.</p> 301 302<p>This post starts by <a href="#my-interpretation-of-the-central-arguments-for-suspending-judgment">
302explaining the sequenceâs central arguments in my own words.</a> Then it offers <a href="#five-defenses-of-acting-on-your-best-guess">five defenses of acting on your best guess.</a></p> 303 304 305 306 <a href="/2026/08/17/unawareness_reply/">Continue reading</a> 307 308 309 </div> 310 311 <div class="meta"> 312 <div class="alignleft">Posted on <span class="postdate">Aug 17, 2026</span></div> 313 </div> 314 315 <br /> 316 </article> 317 318 319 <article class="post"> 320 321 <h1><a href="/2026/08/10/on_democratizing_ASI/">On Democratizing ASI to Preserve Civil Liberties</a></h1> 322 323 324 <div class="post-content"> 325 326 327 328 <!-- Ignore content after a 'more' tag --> 329 330 <p><em>I continue to believe <a href="https://intelligence.org/briefing/">we should pause frontier AI development.</a> Any discussion of alternative strategies should be thought of as planning for contingencies.</em></p> 331 332<p>A unifying driver behind many <a href="https://mdickens.me/2026/04/11/pause_for_post-alignment_problems/">post-alignment risks</a>âcatastrophic risks that remain even if we solve the alignment problemâis that <a href="https://mdickens.me/2026/04/06/by_strong_default_ASI_will_end_liberal_democracy/">by strong default, ASI would end liberal democracy</a>. Liberalismâin which people have individual rights, autonomy, and the ability to choose their own destinyâis an important force protecting human welfare.<sup id="fnref:5" role="doc-noteref"><a href="#fn:5" class="footnote" rel="footnote">1</a></sup> When people are free, we are reasonably good at making our lives better of our own volition.</p> 333 334<p>Many post-alignment risks have a certain flavor. AI-empowered terrorism; coups; permanent dictatorships; concentration of power. Those risks already exist today (and existed 20 years ago), but theyâre mitigated by the fact that power is relatively evenly distributed across people. The most powerful person in the world doesnât have an extraordinary advantage over the 10th-most-powerful person. ASI could change that.</p> 335 336<p>If people still have civil liberties post-ASI, that will only be because the controllers of ASI allow us to have them.<sup id="fnref:9" role="doc-noteref"><a href="#fn:9" class="footnote" rel="footnote">2</a></sup></p> 337 338<p>One way of thinking goes: AI will be extremely powerful. If everyone had their own personal AI, we could each use it to protect our own interests, and things will turn out okay for us. But how do you get there? Itâs not going to happen automatically, but it may be possible to set up a gradual process to keep power balanced.</p> 339 340 341 342 <a href="/2026/08/10/on_democratizing_ASI/">Continue reading</a> 343 344 345 </div> 346 347 <div class="meta"> 348 <div class="alignleft">Posted on <span class="postdate">Aug 10, 2026</span></div> 349 </div> 350 351 <br /> 352 </article> 353 354 355 <article class="post"> 356 357 <h1><a href="/2026/08/03/links_for_august/">Links for August</a></h1> 358 359 360 <div class="post-content"> 361 362 363 <img src="/assets/images/links.webp" style="max-height:500px; height:auto;" /> 364 365 366 <!-- Ignore content after a 'more' tag --> 367 368 <p>Hereâs some cool stuff Iâve enjoyed recently.</p> 369 370 371 372 <a href="/2026/08/03/links_for_august/">Continue reading</a> 373 374 375 </div> 376 377 <div class="meta"> 378 <div class="alignleft">Posted on <span class="postdate">Aug 03, 2026</span></div> 379 </div> 380 381 <br /> 382 </article> 383 384 385 <article class="post"> 386 387 <h1><a href="/2026/07/27/notes_on_moral_progress/">Notes on the possibility of moral progress</a></h1> 388 389 390 <div class="post-content"> 391 392 393 <img src="/assets/images/utopia-thing.jpg" style="max-height:500px; height:auto;" /> 394 395 396 <!-- Ignore content after a 'more' tag --> 397 398 <p>To achieve the best possible future, we must know <em>what that future looks like</em>. In other words, we need to solve ethics.<sup id="fnref:11" role="doc-noteref"><a href="#fn:11" class="footnote" rel="footnote">1</a></sup></p> 399 400<p>The problem of solving ethics is so large and abstract that itâs difficult to say useful things about. In lieu of any structured analysis, herein lies a collection of thoughts about the problem.</p> 401 402 403 404 <a href="/2026/07/27/notes_on_moral_progress/">Continue reading</a> 405 406 407 </div> 408 409 <div class="meta"> 410 <div class="alignleft">Posted on <span class="postdate">Jul 27, 2026</span></div> 411 </div> 412 413 <br /> 414 </article> 415 416 417 <article class="post"> 418 419 <h1><a href="/2026/07/20/letting_myself_look_foolish/">Letting myself look foolish</a></h1> 420 421 422 <div class="post-content"> 423 424 425 <img src="/assets/images/fool.jpg" style="max-height:500px; height:auto;" /> 426 427 428 <!-- Ignore content after a 'more' tag --> 429 430 <p>I prefer to avoid saying things that I suspect might make me look
430stupid. Over the last couple of years, however, Iâve made a conscious effort to write publicly about what Iâm thinking about, even if Iâm afraid it might sound dumb.</p> 431 432<p>Sometimes I have this feeling, like:</p> 433 434<blockquote> 435 <p>I have an observation about the world, about something that doesnât seem to add up. Probably Iâm missing something, and everyone but me can easily see whatâs going on, and if I bring it up, then everyone will know how foolish I am for missing this obvious thing.</p> 436</blockquote> 437 438<p>Sometimes Iâm right, and I really was missing something obvious. But some of my most well-received writings turned out to be ones where I had this feeling, and I was ambivalent about whether it was worth saying anything.</p> 439 440<p>Lately, Iâm making more of an effort to overcome that feeling, and say the stupid thing anyway, because it might turn out not to be stupid.</p> 441 442 443 444 <a href="/2026/07/20/letting_myself_look_foolish/">Continue reading</a> 445 446 447 </div> 448 449 <div class="meta"> 450 <div class="alignleft">Posted on <span class="postdate">Jul 20, 2026</span></div> 451 </div> 452 453 <br /> 454 </article> 455 456 457 458 459<div class="postnavigation"> 460 461 462 <span class="nope left">← Newer</span> 463 464 465 <span class="pages">Page 1 of 24</span> 466 467 468 <a class="next right" href="/page2/">Older →</a> 469 470 471</div> 472 473 474 475</div> 476 477 478 </div> 479 480 481 <!-- comments --> 482 483 <footer id="footer"> 484 <p class="copyright">Copyright © 2026 Michael Dickens. Powered by <a href="http://jekyllrb.com">Jekyll</a>, adapted from Solar theme by <a href="http://www.webmaster-source.com">Matt Harzewski</a></p> 485</footer> 486 487 488</div> 489 490 491 <!-- google_analytics_enabled --> 492 493 <!-- MathJax for rendering LaTex --> 494
494<script type="text/javascript" async 495 src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-MML-AM_CHTML"> 496 </script>
496 497 498</body> 499</html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.