humancode.us

Copyright will not save us from AI

May 15, 2024

Much as I dislike the theft of human labor that feeds many of the generative AI products we see today, I have to agree with Cory Doctorow that copyright law is the wrong way to address the problem.

To frame the issue concretely: think of whom copyright law has benefited in the past, and then explain how it would benefit the individual creator when it is applied to AI. (Hint: it won’t.)

Copyright law is already abused and extended to an absurd degree today. It already overreaches. It impoverishes society by putting up barriers to creation and allowing toll-collectors to exist between citizen artists and their audience.

Labor law is likely what we need to lean on. Unions and guilds protect creators in a way that copyright cannot. Inequality and unequal bargaining power that lead to exploitation of artists and workers is what we need to address head-on.

Copyright will not save us.

I would go further to say that applying copyright law to AI will take us further from the equitable future we want. If copyright is successfully applied to AI, what we will see after the dust settles is a handful of media behemoths that profit mightily from AI, without slowing down the damage that AI does to the value of creative human labor.

I always return to this pithy guide by Emily Bender when thinking about this topic: we need to think of AI as automation, albeit one that is more effective at displacing a wide variety of human labor than ever. We can’t use copyright to stop automation; it will just enrich a different set of kingpins without stopping its effects.

https://www.youtube.com/watch?v=eK0md9tQ1KY

How should regulators think about "AI"?

Written version available on Medium: https://bit.ly/3TdpyMgOn Thursday 9/28, I had the opportunity to speak at a virtual roundtable convened by Congressman B...

The Relentless Monetization of Everything

April 12, 2024

I think the framing of everything as a Means of Monetization™ has killed a lot of the excitement and optimism that we used to feel back in the PC decades. It’s no longer sufficient to produce software and hardware for their own utility; you must also sell some subscription or disruptive network-effect technology behind it. You can’t make a living by filling a niche need any more; you must also conduct arbitrage.

The relentless monetization of everything means you can’t make a living just selling software or hardware any more; you must eventually sell a service. Not everything should be sold as a service (I’m looking at you, video games), and not everyone wants to run a service business. Unless you are already wealthy and can afford to give your work away for free (e.g. open source), there’s really not many viable ways to be rewarded for honing your computer arts these days.

Is there a market for non-monetized, well-crafted products? Yes. But that market is shrinking, and increasingly focused on the wealthy.

Only the wealthy can afford to pay artisans to create small-batch products that do nothing but what they claim to do, without constant subscriptions, upsells, nags, and arbitrage. Turns out, people actually like products that don’t try to constantly show ads, but only few can afford them any more, thanks to our current product culture.

You can still make artisanal hardware and software. But they’ll probably go into some billionaire’s home entertainment system or private jet.

Bet on human trust

March 3, 2024

Techbros will keep attempting to invent ways for people to exchange information without relying on a network of human trust, and again and again, human trust will prove to be an essential part of a functioning system.

Systems that mainly attempt to replace human trust always work great until they don’t, then you fall off a lonely cliff from which no one can help you out.

Never bet against a network built on human trust and a shared sense of personal reputation, because that’s where all human collaboration eventually gets done, and because that’s where every hare-brained “don’t trust humans” network will eventually end up, lest they perish.

A ranking system for vanity plate inscrutability

February 3, 2024

I propose these levels of understandability for (California1) vanity license plates, from most inscrutable to most understandable:

(Most inscrutable)

  1. Physical comedy, where you have to manipulate the plate to get it: X32TTU8 (read it in a mirror), 3SI73 (flip over to read ELISE), BLONDE mounted upside-down.
  2. Inside jokes. Only 4 people will get it: N4BLOOP. What the hell does it even mean?
    • 2.a. Uncommon language plates. I think this one is a subcategory of Inside Jokes because most people have no idea what they mean.
  3. Topical memes for the Chronically Online: COVFEFE, SKIBIDI.
  4. Nerd references: NCC1701, GANDALF, FFFF00 on a yellow car.
  5. Car People Things That Only Car People Get: 997C2S, RMR6MT.
  6. Car People Things That Non Car People Get: GO2FAST, REVVIT.
  7. Major sports teams: 40NINRS, KCHIEFS.
  8. Description of car ownership: WIFEYV8 or BOBSLEX.
  9. Names and nicknames: SLAPPY,JIMBOB.
  10. Nouns: 3LATTES, SHIHTZU.
  11. Common phrases: BRUHH, OHWOW, LOLOMG.

(Most understandable)

Did I miss anything?

PS: I think EV jokes like GIGWATT, LOLOIL, HIGHI2R belong to Car People Things that Only Car People Get. It was suggested that Tesla plates should get their own tier, but I think they belong on an axis of insufferability that is orthogonal to this ranking system.

PPS: There are two other orthogonal axes that exist for vanity plates. The first is getting it which is how well the plate combines with the car in a satisfying way, and the other is insufferability which is how insufferable the car owner looks because of the plate. A plate that says IMRICH on an old beater is funny and not insufferable, but the same plate on a million-dollar car is not funny and very insufferable.

  1. This ranking is based on my experience with California license plates. For what it’s worth, California license plates are typically limited to seven alphanumeric characters, and maybe a space or half-space if it fits. 

Always be making

January 13, 2024

Magnetic domains in a ferromagnetic material aligned by an external magnetic field

By MikeRun - Own work, CC BY-SA 4.0

I like to think of myself as a maker. I like to learn skills and construct things. But all too often, I get so discouraged by my daily busyness that I forget to want to make something. When I get an unexpected stretch of discretionary time, I sometimes don’t know what to do with it: I spend it doing random chores, playing games, or just resting—all very good things to do—but I leave a lot of potential on the table.

We often complain we don’t have time to do what we want. But the truth is, even the busiest adult has some time at their discretion; it just comes in tinier chunks. An hour here, 15 minutes there… but taken together, it adds up to a lot. The question is: when one of these precious opportunities falls in your lap, would you know what to do with it?

Read more…

< Newer Posts