The Tragedy of the Laser Pointer

Posted on by pharpend

I have a lot of cats. They like playing laser.

After a while the cats figure out a few things. Among these are:

  1. the red dot appears immediately following the clicky sound, usually in the late evening when the benevolent hairless ape that dispenses food and cleans up my poop is watching the glowing box
  2. the benevolent hairless ape that dispenses food and cleans up my poop controls the red dot
  3. I’m never going to catch the red dot
  4. I still like chasing the red dot

As the benevolent hairless shit-scooping ape, I sometimes feel it’s cruel to let the cats chase the laser pointer around knowing they’ll never catch it. But if you pay attention for long enough, you’ll start to notice subtle clues that the cats are actually totally in on the game. They know exactly what’s going on, and they’re actually totally cool with it. I’m the one being played for a fool here.

Cats are a lot smarter than they seem on first glance. You can really only appreciate their intelligence if you observe them over a long period of time with an open mind. The little furry purr monsters want us to think they’re stupid little shit-rats larping as apex predators. But if you observe them long enough it will slowly dawn on you that cats are actually psychological masterminds that have cleverly manipulated us low-IQ humans into feeding them, watering them, and cleaning up their shit.

The theses of this article are that

  1. the median cat is operating at a higher wavelength of cognition than the median human
  2. the size of this gap is actually several orders of magnitude.

In an acute sense, I was inspired to write this article because this video titled “L8 Principal’s Agentic Engineering Workflow” got recommended to me by the Evil Eyeball Succubus Robot Demon of Death (the youtube algorithm). I clicked out of curiosity.

It’s not worth your time to watch. I will instead list some direct quotes from the video just to give you a taste of what you’re missing:

00:29
    In the recent couple of years, I have been building frontier
    coding agents at Atlassian, and helped many engineering teams
    figure out how to use them effectively.  And I have been building
    heavily with agents myself, and shipping almost 40-50 PRs almost
    every day, sometimes more.

09:54
    The next one I use a lot is Codex CLI. It's written in Rust, and
    you can feel it's a little bit smoother than Claude Code when
    you use it. It's also open-source so if you run into some
    problems, you can often just have Codex inspect its own source
    code and figure out a workaround by itself.

11:02
    I have been very strict about making my workflow agent-agnostic,
    because the landscape is changing very very fast. Who knows which
    model or agent will be the best-performing one next month?

18:44
    You should generally avoid installing random skills from the
    internet, even the ones that have a lot of GitHub stars.

29:38
    What I do here, when the agent says the work is done, is not to
    start reviewing the diffs or start manually testing the changes.
    That's too much overhead on myself. I send the change into a
    pipeline I built called No Mistakes.

I enjoy watching bullshit artists for the same reason I like going to the zoo and watching chimps throw their feces at the glass. In a weird twisted way it’s art.

I have a bad habit of

  • being right about Thing in private when everyone else is wrong
  • being proven right about Thing years later
  • having no public-facing evidence to show I called it

For instance, I spotted that FTX was a scam immediately as soon as I learned of its existence. This was maybe 2022. You know how I knew? I was watching a baseball game and I saw the logo on the umpires’ uniforms.

Cryptocurrency in its natural setting is a revolt against every mainstream institution and everything they stand for. Anything crypto-related that had that degree of buy-in from mainstream institutions just had to be a scam. I didn’t know how it was a scam or how the scam worked. But there was no possible way that MLB was publicly endorsing something that was actually counter-regime. The much more likely explanation was that it was some sort of way for elites to siphon off the remaining funds of retarded boomers who after living through

  • 1971
  • the 1980s worldwide banking collapse
  • NAFTA
  • 9/11 and the War on Terror
  • 2008
  • COVID

still haven’t figured out that if some hot new thing is being promoted by every major institution, then it’s almost certainly a scam.1

Unfortunately, I only shared these thoughts in my workplace’s Slack chat, which I no longer seem to have access to, so I can’t even share totally real screenshots to prove I was right.

I’ve had this experience many times.

See there’s a feeling I get in the back of my head that something about what I’m looking at just doesn’t add up. But like the low-IQ boomer depicted above, there’s some part of me that still relies on my mammalian herd servival instincts, despite the herd’s overwhelming track record of running directly off the edge of clearly visible oncoming cliffs.

I tend to keep my heterodox opinions to myself because like obviously there must be something I’m missing, right?. It can’t be the case that what my eyeballs and common sense are telling me is right and everyone else is wrong.

I’m getting that feeling about AI. I had the same feeling about crypto when I first learned about it. I had the same feeling about Large Data… sorry Big Data when I first learned about it. The feeling is something like

  • a profound lack of interest in Topic originating from System I2
  • some logical part of my brain in System II telling me that Topic is objectively important and I should learn about it
  • some weird feeling inside that I just am not smart enough to understand Topic and there must be something I’m missing

Invariably once I learn enough about the topic, it turns out that actually there was nothing I was missing, it’s just that everyone else was foolish/lying.

So I’m going to go out on a limb here and call bullshit on AI. The whole thing. I might be wrong. Of course, since this was the topic where I decided to break my pattern and call bullshit on everything in public, this will be the one time I actually turn out to be wrong. But that video pushed me over the edge.

This is all bullshit. The only jobs that AI replaces are jobs of people who were already useless.

I can give like a precise logical case with facts and evidence and whatnot. I might do that in another post. But that’s not why I think AI is bullshit. It’s just a gut feeling. Something about what I’m looking at just doesn’t add up. I can point to specific pieces of evidence and I can make any number of logical cases to argue this point. But that’s not why I think this is all bullshit. It’s just a feeling.

And like I said, I’ve been around long enough to know I should trust this feeling.

I don’t mean to say that there is no positive utility in AI at all. Of course that’s not true. I don’t know a single person who doesn’t use AI at least a little bit. There’s some things it’s really good at. AI is great at solving initial reference problems (I don’t know anything about welding, where should I start? What questions should I be asking?). It’s really useful to point Claude Code at a codebase I’m not familiar with and tell it to find a call path to get that software to do XYZ, where XYZ is something that is obviously possible, but it would take me hours-at-minimum to read the codebase and documentation enough to figure out how to do it.

But we all have used AI enough to recognize exactly what it’s good for and what its limitations are. And you should’t be fooled into thinking there’s anything more to it than that. Use your fucking eyeballs.

I occasionally make the mistake of looking directly into the Eye of Sauron, known in the common tongue as “Twitter”, which in official contexts we now have to pretend is called “X”. The tweets I see about AI tend to come in the following flavors:

  • “It’s amazing how good AI is at <thing outside my domain of expertise> given how bad it is at <thing inside my domain of expertise>”
  • “I asked AI and it told me exactly what I wanted to hear, so therefore …”
  • “People at <AI company> are freaking out about the capabilities of <new AI model which will be the free tier in 3 months>”

I’ve been around the computing industry long enough and I’ve seen trends come and go. I work ironically in the industry which was the hot Topic of yesteryear: cryptocurrency. Crypto got BTFOd by the very problem it was invented to solve: money is too cheap. When money is too cheap (meaning money invested loses value over time at a greater rate than strategic investments can generate returns), long-term strategic investment is heavily disincentivized. Capital instead flows into short-term laser-pointer projects that generate very rapid returns immediately, but don’t meaningfully move the envelope forward in terms of the technical fundamentals.

The Tragedy of the Laser Pointer is the reason why the crypto industry has poured millions of dollars into NFTs and crypto kitties and fully anonymous off-chain autonomous on-chain agents with zkp L2-squared scaling hyper whateverthefucks, and meanwhile there’s nowhere on the planet where you can use cryptocurrency to purchase a banana. Money is too cheap, and so there’s no incentive for anyone to fund really boring basic nuts-and-bolts projects like Gajumaru (where I work, slightly shameful plug).

In a past life I spent many hours volunteering for a project called Snowdrift. It’s basically kickstarter or patreon but only if your project is free (not as in price… remember the adjective free has nothing to do with whether or not the noun its attached to is free as the term means in common parlance… remember folks, good terminology is terminology where every time you use it, you have to stop and spend 8 minutes explaining to the audience that the term doesn’t mean what it sounds like it means). Today, Snowdrift.coop is in more or less exactly the same place it was when I stopped paying attention to it 10 years ago. The name Snowdrift refers to the Snowdrift problem in game theory. The idea is that there’s a snowdrift in the middle of the road blocking everyone’s usage of the road. Everyone is paying the cost for the snowdrift blocking the road, but nobody wants to be the sucker that just goes out there and clears the snowdrift. Because now I’ve paid a huge cost, for which I receive only a marginal benefit, and everyone else paid no cost and gets free use of the road that I spent all my effort cleaning.

This is why I don’t tell people that I work in the crypto industry, by the way. As soon as I say the word crypto, immediately their brain fills in exactly the opposite message of what I am trying to communicate. It’s really a tragedy. It’s the Tragedy of the Laser Pointer.

Framed differently, we call this the Tokyo Millionaire vs. Beijing Billionaire problem. China has more billionaires than any other country in the world, but weirdly Japan (with less than 10% of the population) has more millionaires. And, counterintuitively, a millionaire in Japan has a much higher quality of life than a billionaire in China. He just has less social status relative to his neighbors.

In China you cannot put your child into a stroller and go for a walk in the park. The market for body parts of babies in China is so lucrative that there’s a strong possibility your baby will be kidnapped and you will never see your baby again. This sounds completely insane to westerners, but I promise you this is true. You’re welcome to fact check me if you like.

The water in Beijing is so polluted that you cannot drink it. The air is so polluted that you cannot breathe. The CCP destroyed all the great monuments of ancient Chinese civilization that existed at the time in an effort to manifest what could be unburdened by what had been. You live in a hellhole. But you are better than your neighbors. That’s all you have. The great civilization that you live in the ruins of is long gone. The only remains of it to be found are either in Taiwan or were only discovered after Mao died.

Meanwhile in Japan you might not have any reason to think that you’re inherently better than the retard who lives in the slightly smaller house down the hill. And maybe there’s someone else you know who lives in a slightly bigger house than you and drives a slightly nicer car. But the air is clean. There’s no trash in the streets. There’s no crime. You can go take your children (plural) for a walk in the park and they’re not going to get kidnapped so their organs can be harvested. (I mean, I’m sure that happens sometimes, but it’s not a concern you worry about when you leave the house).

In China everything is zero-sum. Everything in life is about maximizing your own status relative to others no matter what the cost is. Externalities are not even an afterthought. In Japan, the game is to build a beautiful society in which we all live in harmony.

Obviously I’m oversimplifying. Japan has very serious economic and social problems. But your child isn’t going to get kidnapped for organ harvesting.

This is the solution to the snowdrift problem. And I would wager that this is the only solution to the snowdrift problem.

Just get in your truck and put down the plow and clear the fucking snowdrift. Even if you bear the cost locally, you still will benefit overall, because you get to live in a society that bears the fruits resulting from that snowdrift being cleared.

So much of western thought is too materialistic and focused on what is locally provable. Galileo, Newton, and Darwin killed the western notion of God. Nietzsche wrote the eulogy, and then Cantor and Gödel sealed the coffin shut. Our entire notion of what things mean became “falsified” due to our newfound knowledge of what things are and how they work. Ironically that knowledge was only discoverable becuase Christianity holds truth as its most central value (nothing means more than the truth). Christ gave western civilization a very very long rope, so much so that it took 2000 years for western civilization to figure out how to use it to hang itself. Quite impressive but still tragic.

My parents were biological anthropologists and I grew up as an only child raised as a reddit atheist in a heavily Mormon neighborhood in Utah. I had as negative a perception of religion as one could possibly have. And then life happened to me. It’s a very long and very private story. But a few years ago I realized that I could no longer call myself an atheist. I don’t know what I would call myself now. But when I explain it to people I always circle back to John chapter 1:

In the beginning was the Word, and the Word was with God, and the Word was God. He was in the beginning with God. All things were made through Him, and without Him nothing was made that was made. In Him was life, and the life was the light of men. And the light shines in the darkness, and the darkness did not comprehend it.

The word “Word” is a bad approximation of the greek “Logos”, root word behind “technology”, “logic”, “anthropology”, etc. It refers to some combination of the natural order of the universe and our ability as humans to reason about it using the power of language and the light of human reason.

Why is it that 2+2=4 just makes sense to us, but 2+2=5 doesn’t? What is that fundamental order to the universe that makes it so that two plus two is equal to four and not equal to five? I have a strong enough background in mathematics that I can “prove” that 2+2 is equal to 4. But to “prove” such a statement only means to show that its truth is a consequence of more “fundamental” axioms which are themselves less obviously true than the statement I’m trying to “prove”. And it still begs the question of why those rules of inference are valid in the first place, or how we can be sure we’re applying them correctly.

There’s no way to escape this problem. At some point you have to take a leap of faith. You have to believe that there is some natural order to the universe. You have to believe that some statements are true, and some statements are false. You have to believe that human reason is a powerful enough (although extremely limited) tool to reason about such things. That idea is the idea of the logos. But it generalizes outside stupid twitter questions like whether or not 2+2=4 or whether or not homeless people should be allowed to shoot heroin on the children’s playground.

In some sense it feels cooky to try to write this sort of stuff. It somehow feels like I need to justify everything I’m saying within some materialistic framework, in which every obviously true statement is derivable from logic and physical mathematical laws of the universe. I don’t actually think it’s impossible to do so. But we need to stop being afraid to call a spade a spade.

In order to survive in today’s world, you do have to have some faith in your own ability to make sense of it. And the world is constantly assaulting you with reasons to doubt yourself. And you have to just box it out. There’s no substitute for analyzing things according to their fundamentals. And there’s no substitute for your sense of smell.3

The cat never catches the laser pointer. The difference between the cat and the human is that the cat is in on the joke.

  1. Weirdly when I talk to boomers (by “boomers” I mean anyone who reads The New York Times and takes it at face value), they’re quite reticent to believe anything unless it has some endorsement from mainstream institutions. There’s no peer-reviewed study that says the sky is blue after all.

    My favorite piece of Nassim Taleb wisdom is “don’t read the news.” Even if everything you read on the news was true (it’s not), it still would be a bad idea. News definitionally is a log of outlier/extreme events. If you wake up every morning and read the news, you are subconsciously assembling a model of the outside world that is composed entirely of extreme outlier events. Again this is in the best possible case.

    And you might counter “ah I’ve read Taleb! only outlier events contain useful information”. Yes, that’s true, but what I’m talking about is your brain’s tendency to hallucinate patterns in these outlier events. The patterns that your brain will detect will almost all be wrong.

    Add in the variable that whomever controls what appears on the screen in front of you has this massive asymmetric ability to influence your perception of reality, and it’s just really bad news.

    We are monkeys on a rock in space who mostly do monkey things.↩︎

  2. The super helpful and very descriptive System I/System II terminology comes from Kahneman. System I is the “fast” part of your brain that makes very instinctual hipfire decisions, and System II is the slow, methodical part of your brain that analyzes things slowly and logically.

    Kahneman is worth reading but only with a heavy dose of salt. Readers of zxq9 probably won’t struggle with this idea. But if you have a low IQ, you should probably read Taleb first. I would suggest only reading The Black Swan, Antifragile, and Skin in the Game, in that order.

    Fooled by Randomness is ok, but it’s definitely the weakest book in the series, and most of what’s written in there has aged very poorly. For instance there’s (to my memory, I will not go check) many long passages in which Taleb prosily fellates the profound genius of George Soros. It’s kind of like most of the Old Testament where it’s nice to read in order to get some background context, but the New Testament has all the cool shit like walking on water and healing blind people and glowing horses.

    I’m reminded of a recent interaction I had with my friend Mike (fake name). Mike is a lot more financially successful than me, but I’m (on paper) a lot smarter than him. Mike and I have a very deep mutual respect for each other, although you would have no idea if you watched us interact face-to-face.

    My non-mathematician friends seem to have a thing where they ask me to explain mathematical concepts, and then turn around and argue with me as if math is something that can be argued with. For instance, zxq9 (the owner of this website) and I have a years-long fewd over the definition of the word “convex” (I am right. It is a mathematical term with a precise mathematical definition.). We recently got into a similar dispute regarding whether or not the concept of “order of a function” (the O(n) notation you see programmers casually throw around without knowing what it means) is a well-defined concept.

    zx seems to think that if technical terminology leaks out into common parlance and then loses its precise technical meaning (this is literally a concept in linguistics) that somehow the term inherently loses its precise technical meaning.

    Let’s not get distracted by other people being wrong, and let’s stay focused on the subject of this discussion, which is that I am always right.

    Mike and I were having dinner a while ago and he asked me to explain the Monty Hall Problem to him. For those who don’t know, the Monty Hall Problem is a classic example of how academic problems in conditional probability violate natural intuitions regarding chance.

    The problem setting is as follows: – you’re on a game show – there are three doors A/B/C – behind two doors there are goats – behind one of the door there is a car – you have to pick a door – let’s say you pick door A – I (the host, Monty Hall) open door C and show you that there’s a goat behind it – I now ask you if you want to change your guess to door B

    Counterintuitively, it’s to your advantage to change your guess to door B. The easiest way to see this is by simply enumerating the outcomes:

    ABC     Guess     Shown     ResultIfStay    ResultIfChange
----------------------------------------------------------
GGC     A         B         Lose            Win
GCG     A         C         Lose            Win
CGG     A         B         Win             Lose

    The phenomenon remains the same irrespective of which door you guess (A/B/C). You win 2/3 of the time if you change your guess.

    Mike was trying to argue with me that this was wrong. And his point was that if he applied this sort of thinking to his trading, he would be broke. Obviously, once you see that there’s a goat behind door B, you can just completely eliminate door B from your thinking, and there’s a 50/50 shot the goat is behind doors A or C.

    Logic has a funny way of colliding with reality sometimes.↩︎

  3. Consider that small/taste is both the most rapidly-evolving sense in the human genome, and the only sense where evolution is still leaps and bounds ahead of material science.↩︎

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.