Intelligence vs Virtue

Are we feeding the good wolf?

I have been using a tool called Warp for quite some time now, and these days I am rebuilding my homelab. So I’m giving it a workout. And I am very pleased with how Warp, which is a terminal emulator with AI, is helping me reconfigure and debug a lot of systems I use from the command line. It is now equipped with Claude 3.7 which seems to be benchmarking nicely - and there are things it still can’t do, but it is a highly evolved tool. It falls into that category I have as an expert system.

Now if you’ve been a coder for any period of time, you surely know three or four repositories of expert testimony upon which the corpus of LLMs are based. Some of them are IRCs, some of them were usenet groups, one was Slashdot. The most recently popular of these are Reddit and the two kings of them all are Stack Overflow and Github itself. One of the reasons that LLMs are powerful is that they can vacuum up all of that text and make sense of it. But in the context of the Warp product, it only cares about programming and systems administration. And this week Claude 3.7 is the programming champion.

I’m glad I used that term champion, because I cannot think about Muhammad Ali without thinking also about Howard Cosell and Angelo Dundee. For those of you who don’t know, Dundee was Ali’s coach & trainer and Cosell was the greatest announcer in the sport of boxing. I say that Ali would not be The Greatest if it were not for their contributions to the sport. I don’t want to leave this as the last word on Ali but the analogy fits in the same way that Deep Blue only beat Kasparov because the IBM team had their chess experts working with their programming experts.

Muhammad Ali was an expert boxer surrounded by the best. He watched every fight ever filmed, multiple times. Kasparov is an expert chess player, surrounded by the best. He has studied every chess opening and end game, multiple times. That made them approach and attain greatness - even though a computer team could beat Kasparov and a machine gun could tear Ali to shreds.

I’m not sure anyone cares about boxing or chess as much as we did when those two were in their primes. So the extent to which we might rely on an artificial intelligence to vacuum up the tapes and moves would be susceptible to abuse and misinterpretation at the hands of a retail consumer, not counting the AI’s likelihood to make errors.

---

Next Generation Jobs
This is part of my thinking around autonomous systems that use various kinds of expert systems in order to make sense of their environments and take appropriate actions. My new T50 entrant (read intellectual hero) and hopefully future boss is Palmer Luckey. I binged about 7 hours of YouTube of interviews with him in the past week or so and he strikes me as one of those people who goes incredibly deep in one thing, and correspondingly seeks that level of expert understanding in a wide variety of other things. I can tell just by listening to the way he speaks. In that way, I think he most reminds me of the person I remember as Bruce Sterling from an ACM seminar I attended at UCLA in the late 80s. [Bruce] was this improbable-looking dumpy kid who was extraordinarily well-informed. I kept asking myself where did he learn all that stuff? He was absolutely confident and his arguments were blisteringly coherent. He spit laser beams of facts in bursts of conceptual volleys. He was like me, if I was an only child. The kid who read all of the books in his parents libraries and spoke only to adults - never having to slow down for his kid brothers. My ethics are deeply burned in, I am learning. I have satisfied myself be being curious and my curiosity is only loosely coupled to my ability to strafe tactically. I’m the AWACS. I loiter.

Luckey is Bruce Sterling with a billion dollars. He is as literate and articulate as Steve Jobs and he uses the same line with the engineers he hires as Jobs did with Pepsi CEO John Sculley in 1983. Do you want to work on the future or sell sugar water to consumers? Luckey asks Silicon Valley techbros if all they want to build are shite apps with Javascript that are ruining the social skills of tweens or build the next generation of defense technology for the USA.

What Luckey understands very well, unlike the hype sellers of AGI is that there are already a great number of intelligent agentic machines out there doing work. He has read enough science fiction, and presumably having the appropriate clearances, enough government publications to have a fairly comprehensive understanding of the history of super soldier and and brainy weapon dreams. He readily acknowledges that these ideas are not new. Importantly, he has the understanding that machines do decide and discriminate without human interference, and do a damned good job as well. One brilliant example of that is the Phalanx CIWS. You don’t aim it, it aims itself. You simply arm it and say go.

There is no man in the middle, just as with a nail gun, there are no fingers holding the nails. You build discriminatory intelligence into the tool and you clear it to engage. The human is responsible for giving that order, that’s what matters. The brain in that machine doesn’t go HAL 9000. Its role isn’t so comprehensive. It’s not a super-intelligence and it doesn’t need super intelligence. It just has to be quick enough and accurate enough to shoot down enemy missiles, boats and aircraft who get too close to an American warship, period. We’ve been using them for years. Their brains were designed in 1969. Its gun was built in 1959. All said, it could probably use an upgrade, but the philosophy of the thing is not at issue. It’s a thinking system, an expert system.

---

When I was at Xerox in the early and mid 80s, most of the world was enthralled by the inventor named Ray Kurzweil. He famously made electronic synthesizers used by Stevie Wonder. Less well known was his work on text-to-speech; a reader for the blind who no longer needed braille was his invention. Artificial intelligence was basically all those things that our 70s and 80s sensibilities told us that computers could not do, with the exception being HAL 9000 and Skynet and other pop sci-fi. Passing the Turing Test is now boring. In fact, I took a crack at beating ELIZA when I was in highschool, in a sophomoric exercise that said maybe I can make somebody believe it’s a crazy person on the other end of the line. NUTZ was the name of my first big program. More recently, however Kurzweil became famous again for ‘The Singularity’, which I never put much credibility in, even before I read Iain M. Banks.

The Post-Scarcity Leviathan

If I haven’t made it clear before, I have no truck with transhumanism and radical life extension. People who can’t deal with the inevitability of death are not to be trusted. Which brings me around, as always, to the matters of virtue.

Stephen L. Carter
I don’t think I have ever given Carter the props he deserves over here at Stoic Observations. I gather he would belong to that class of writers I found most intriguing in my years as a young father in the early 2000s. Speaking of early, I should also mention Gerald Early. It seems to me that if you are not familiar with these gentlemen’s writing, then you may want to consider their context in what I consider chips off the old block of the Old School. Specifically to Carter, I read two of his books on virtue which I don’t think have reached the critical acclaim I think they deserve. That could be because the first one was a blockbuster, Reflections of an Affirmative Action Baby, which I refused to read for years until I finally went for it. It was a breath of fresh air, surprising as all get out. I can actually recall the day I decided to pick it off the shelf of a bookstore in NYC one block away from the Cage at West 4, just past the Mr Softee ice cream truck. As well, I remember the days I was reading the massive Emperor of Ocean Park in Houston when I was first considering going Republican. My review is here among the brainspew.

• Integrity (1996)
  ”We the people of the United States, who a little over two hundred years ago ordained and established the Constitution, have a serious problem: too many of us nowadays neither mean what we say nor say what we mean. Moreover we hardly expect anybody else to mean what they say either.”

• Civility (1998)
  ”The idea that we should use our freedom for the common good, rather than to seek our own pleasures, has long been at the center of Christian and Jewish ethics. On this point, at least Augustine and Maimonides were in full agreement. It goes back at least to Aristotle and was central to the defense of human freedom by such enlightenment philosophers as John Locke. But it is not one of those ideas so old and venerable and self regarding, like the profit motive that we cannot imagine life without it. It is and has always been an idea so fragile that, if not, carefully nurtured it will certainly die.”

It is not at all astonishing that some of the most hotly discussed items in today’s America is how much it will cost to get our next generation of computing tools, AIs to stop hallucinating. We are ready to throw hundreds of billions of dollars at that problem. One should be troubled, but not astonished that our society wants these things from machines that they have failed to demand from citizens.

We have been living in a vacuum of civility and integrity for so long that we have failed to reconcile ourselves to anything other than the tireless game of gotcha. Remember the old adage that behind every good man is a good woman? It seems we have college professors arguing that there is nothing inherently biologically good. More postmodern shenanigans.

Warren Smith, who looks fairly young to me, seems to be one of those who is discovering that ‘everything they told me was a lie’. He’s almost gracious about it. But notice his shrill titles, and that he must be in the Secret Scholars Society. I only make the example of Warren because so many have decided to shut up, or even pile on Joe Rogan. It speaks poorly of our society, even as the vibe pendulum swings.

So Stephen L Carter and certainly millions of others saw this mess in the 90s. I was busy being a bohemian and stepping out of my yuppy costume at the time, exercising my upwardly mobile sensibilities in the relative shadow of heterodox thinking. But I knew things were missing, which was why I picked up Carter in the first place. The answers are in here, (points at heart), so we seek validation of them out there.

Alas, Carter’s books on virtue are not available on Kindle. Nor are there many of Gene Hackman’s films available outside of the streaming market. We live in a cultural desert, will we go feral from dehydration of our soul sustenance?

---

I’m sure that I know good people. I’m not sure that we are getting what we deserve or even if it’s proper for us to have great expectations. Yet I am supremely confident that as human beings, we possess, whether or not we nurture it, a fine grained ability to suss out good and evil in ourselves and others. It is never a good time to sustain cynicism or abandonment of virtue in a world full of evil. Some are so inured to virtue that they expect our thinking machines will be evil. Be patient. Somebody out there has the mind to do what’s right.

Sooner or later, you will be faced with a challenge that you cannot think your way out of. You will not have enough information to make the right decision. Your instincts may not save you either. The kind of thinking you do will determine your fate and the fate of others. I think virtue wins in the end.

I leave you with Clint Bruce. Remember that guy?