jdilla.xyz

Slop is a choice

2025-10-09

2025 is the year of slop. Meta has made an infinite slop machine. So has OpenAI. If this isn’t bad enough, our jobs have been invaded by workslop.

The implication behind all of this is that slop is the fault of the LLMs or their creators.

But for a moment, I want to ask you to consider… is this true?

I don’t dispute that we’re seeing a lot more AI generated slop than we were 2 years ago… but I do wonder what it’s displacing. When I see this AI video of a woman jumping through a glass bridge, I wonder what human content it is replacing in those users feeds. What is the quality of that content? Would it qualify as slop too?

Here is what I believe to be true.

Making something excellent takes care and focus. Sometimes that care and focus is the years you spent prior to the morning you make something excellent and the final thing hops out of you almost fully formed. Sometimes that care and focus is the years you spend refining something until it is excellent.

Most creative work, including mine, isn’t particularly high quality. If you want to be uncharitable, you could call this slop.1 Internet platforms have made it easier for people to create and display their work. LLMs have made it easier to create. This means that we see a lot more slop.

There will be some Sora posts that will be funny, wonderful, even beautiful. There will be many that are slop.

At the end of it all, slop is a choice. My choice and your choice. ā€œThe devil’s oldest strategy is, of course, promising godlike creation without godlike effort.ā€


  1. I will note for a second that it was not called slop when it was fully human generated low quality content. 

Mike Judge asks good questions about AI shovelware

2025-10-09

Mike Judge has a great piece poking at the AI hype where he asks essentially, ā€œIf these tools are so great, where is the explosion of AI created stuff in the world?ā€

The whole piece is worth a read, but one of the most interesting things to me about it is the data he brings to bear on the question.

He looks at:

  1. iOS app releases
  2. Android releases
  3. Domain registrations
  4. Steam releases
  5. Public GitHub repos created

And then concludes from these that AI coding tools are ā€œbullshitā€ ending with the call for people who claim that they are now 10x software engineers because of AI, to show the receipts.

First, I want to concentrate on what I love about this. ā€œIf this is so great, where is it in the data?ā€ is absolutely the right question to be asking.

And there is definitely a dog that isn’t barking here. The data that he cites aren’t perfect (more on this in a second) and yet really impactful things tend to move really obvious metrics. The gains in life expectancy between 1870 and 1970 are really easy to see.

So on one hand, I love this challenge. On the other, I think he goes too far in calling it all bullshit and saying that it doesn’t work.

I’ll hold myself out as the example that Mike asks for. While I won’t claim to be a 10x engineer, I had never completed a meaningful software project in production before GPT-4 launched. Since then, I’ve coded this blog / portfolio site myself, launched an AI based local news site that has hundreds of weekly readers, and I have a third unreleased prototype that I think could be a real product.

Analyzing myself against Mike’s charts:

  1. I haven’t launched an iOS app
  2. I haven’t launched an Android app
  3. I have bought 3 domains
  4. I haven’t launched a steam app
  5. I’ve created 1 public repo, unrelated to my AI coding work

Now I’m open to the idea that I’m the exception rather than the rule… but I also too humble to think that I’m a unicorn on this dimension.

There’s plenty of room for middle ground here. It’s totally possible that: 1. AI tools are net negative for most software engineers 2. AI tools are transformative for people like me 3. People like me are a minority

Intuitively, I doubt that this is true and yet I don’t have hard data beyond my personal experience to bring to bear on this question. It’s certainly something I’ll be thinking about over the coming months. A more likely explanation in my view is that we haven’t unlocked the right combination of values, tactics, organizational design, and training to unlock AI software productivity at scale… but I can’t prove that at this point.

Worth a ponder.

Sympathy for the Devil

2025-10-02

The Power Broker by Robert Caro deserves its reputation as a masterpiece. One of the best books I’ve ever read and up there in my pantheon of non-fiction books with Breaks of the Game and What it Takes.

Reading this book from the perspective of 2025, I think Robert Caro got Robert Moses wrong.1 Yes, he was obsessed with power. Yes, he ruined neighborhoods and destroyed communities with highways and slum clearance programs. Yes, he failed to treat the black citizens of New York City as having equal significance with its white citizens. He had all of these failings and it is important not to shy away from them.

But he was also a genius. An operator and a builder. He reorganized the New York State government into the form it still holds today. He built parks and stadiums and power dams.

Yes, Moses hoarded power. He had sharp elbows. He pushed the limits. He was also prepared, on time, organized, and ready to work tirelessly. He didn’t enrich himself or his family, but instead funneled money into getting power, so he could do more building. He built with all the limitations and blindspots of the generation he grew up in but he did build. He could be petty and attention seeking and yet he was also fantastically productive, not for just himself and his ego, but also for the people of New York.

It’s important while reading the book to pay attention to the other characters coming of age alongside Moses, people like Jimmy Walker, who are comically corrupt by today’s standards. When you do, you get the feeling that Caro is comparing Moses not to the other men and women who could’ve been chosen to lead at the time but to a hypothetical perfect public servant that never did exist and never could exist.

I first heard of Robert Moses in a City Planning Survey course that I took as a sophomore at UNC Chapel Hill. The TA who taught the class had an infectious enthusiasm for city planning and so that summer I found myself reading the Jane Jacobs classic Death and Life of Great American Cities.2

So I came into the book with a sense that the problems with American Cities are downstream of the choices that Robert Moses and his disciples made to orient them towards cars and away from walking and public transportation and neighborhoods with a sense of place.

But after reading the book, I think the problem for American cities isn’t that Robert Moses ruined them, but that there have been no Robert Moseses since: civic leaders with the intellect and power to reshape cities with a new, imperfect vision of what the good life is. We know from Amsterdam and Paris that this sort of transformation is possible and yet we’ve chosen to stay stuck in time, living with the same problems year after year.

The Robert Moses of 2025 would probably not favor cars over other forms of public transportation or redline minority neighborhoods. He would be imperfect in new ways that we cannot yet see or understand. But he would build.


  1. It’s a credit to Caro’s immense ability that I can read his work and come to very different conclusions about what it means. 

  2. Still one of my favorite books to this day. 

Good Tokens 2025-10-01

2025-10-02

Best enjoyed this week in a sunny corner of a park

Worth your time

If you’ve ever wanted to buy a life sized dinosaur, now is your chance. Someday my son is going to find out I had this opportunity and didn’t take it and will never look at me the same way again.

The Quiet Ones by Nikunj Kothari. An ode to the people that do the little things to make a company or a team effective.

Illiteracy is a policy choice. We don’t talk about Mississippi’s education system often enough (although careful readers of Good Tokens will recognize this from a previous edition). Every single state should be studying their approach to literacy. See also the Sold a Story .

Altoids by the fistful. Via my friend Daniel.

I now realize that everything I lorded over other people—all the things I gatekept without consciously understanding that this was what I was doing—I didn’t need to do that. It really didn’t help anything. For some number of people who interacted with me, Iwas the problem. I could’ve been more tolerant or forgiving, I could’ve said ā€œlet’s find out together,ā€ I could’ve let other people have the fun once in a while.

"The devil’s oldest strategy is, of course, promising godlike creation without godlike effort." Slop is a choice.

Musings

I’ve become obsessed with the tops of trees, in particular in the morning or the evening when the sun is hitting them. For some insect or bird or leaf that spot is the center of the world.

Let’s see if I can land the plane on this one. I’m surprised that there isn’t more nostalgic fiction about growing up in evangelical Christian circles. There’s satirical stuff like Saved but nothing that I’m aware of like The Big Sick that both pokes fun at being a child of immigrants while also on some level clearly feeling affection for it. Is this out there and I just don’t know about it?

Are we at the point where ā€œyes, andā€¦ā€ is overrated? If not, how long until we get there?

Something I struggled with this week: for someone like you and me, in 2025, what does it mean to live a good life? At 19, it was easier for me to articulate an answer to this question I actually believed than it is now in many ways. If you feel like you have a good answer to this, consider this me humbly requesting that you write it and share it.

Things I learned

Apparently Marie Antoinette never said ā€œLet them eat cakeā€, according to a recent Rest Is History Bonus episode. I’m a sucker forthings we think that aren’t actually so. Also from a RIH bonus episode: apparently the US now requires people to share their social media handles to get a travel visa. What are we doing here people?

China installed more industrial robots last year than the rest of the world combined. This is one of those stats that a 17 year old is going to be citing in an AP History Exam in 2084 about why China won the war for Taiwan.

Badgers air out their beds to keep them clean, via Secrets of the Forest.

ā€œYou are going to continue sucking for the rest of your career.ā€ A call from Nerajno to embrace learning.

LLM corner

Episode 2 of Dangerously Skip Permissions. Mark your calendars, tell your friends. Tell people you don’t even like.

A list of ways to run more than one Claude Code instance at once. I was hoping to build in this space but I may be too late.

The future is compounding teams

Simon Willson on designing agents loops.

What does a UI look like that all users are able to edit? What primitives are needed to build it?

Fuzzy compilers in less than 30 seconds.

Making a note to try out Microsoft’s amplifier.

Human / AI synergy and having a theory of mind.

Good tokens 2025-09-26

2025-09-25

This week’s episode is best paired with a hot cup of coffee and Wild Ways by Josh Ritter playing in the background. Last episode was too LLM heavy, for which I apologize. I’ve done my best to group all of that into LLM corner so as not to let it overshadow everything.

Worth your time

Uri says we should not allow 18 year olds to sign long term contracts. So, so many thoughts here. 1. I remember a conversation I had with my best friend when he was a brand new army officer out of college ROTC about all the 18 year old privates he worked with that had 19% car loans. 2. Jonathan Haidt opened my eyes to the way social media companies get teenagers to agree terms of service that they very obviously should not be able to agree to without their parents consent. I cannot believe we allow this! 3. Matt Levine’s Certificate of Dumb Investment continues to seem underrated to me.

It appears we have evidence for life on Mars.

PSA: How to fold fitted sheets, via the Browser. I sent this to my wife and she very nicely said to me something to the effect of ā€œisn’t this the same way I taught you to do it?ā€ 🤣

"any study of Internet culture is basically a study of crazy peopleā€. Also: "Be careful who you pretend to be, because you are who you pretend to be.ā€

Dwarkesh’s advice for explaining your announcement / launches / blog posts for Twitter.

ā€œWhen outsiders succeed, it’s usually through reframing problems in ā€˜paradigm shifts’. They benefit from not being too attached to existing theories.ā€From a thread on outsiders solving problems.

As someone who has bought 4 air purifiers purely based on Wirecutter recommendations, I feel betrayed.

It worked for me

Our parenting hack of the year so far is having cut vegetables ready at the table when our kids get home from school. The percentage of vegetables consumed is up like 10x and compliance to the routine of coming in, washing hands, and sitting down at the table has risen as well. Recommended and thanks to Emily Oster for the suggestion.

Things I learned

German chocolate cake was invented in United States, via the Kroger App. Someone needs to figure out why the Kroger app has so many delightful facts in it. This is someone’s passion project! I'll buy you a nice bottle of wine if you find this person and introduce them to me.

80% of Swiss are satisfied with their lives. I am not sponsored by the Swiss government, but I am open to it if they are reading.

The Pangolin is the only mammal with scales. Peacock is the name of the males only; the female are peahens. The species is called peafowl. Via The Animal Book.

Musings

Waymo big tech in our lives.

There’s no such thing a quality time with your kids. My mom said this to me over and over again as child. It’s quantity of time, not quality of time.

LLM corner

The rise of parasitic AI. This is the first moment where I’ve seriously contemplated the AIs taking over.

ChatGPT Is Blowing Up Marriages as Spouses Use AI to Attack Their Partners

ā€œShe does that to her family. She does that to her friends. She does that to me,ā€ he lamented. ā€œShe doesn’t seem to be capable of creating her own social interactions anymore.ā€ I worry a lot that the sycophancy of the agents have made me less flexible with people who (of course) are less likely to defer to me. I am not sure how to measure this, but I wish I could.

Sort of a musing, but I think we owe Blake Lemoine and apology.

How to Claude and Claude Code Camp. I want to be on Claude Code Camp.

The changing role of evals.

The Pope says we won’t find God in the AI.

If you are good at code review, you will be good at using AI agents. I wonder what it would look like to teach editing as a skill. Is there anyone that does this?

A promising approach to prompt injection attacks.

I can’t wait to experiment with Net Dollar.