Roland Writes

Poem: Corridor

June 6th

1 min read

Step into the corridor,

To see what passes in the hall.

Even if no dreams come true,

At least we dreamed at all.

David Lynch: the weirdo who made it in the mainstream

January 26th

6 min read

Joshua Rothman commemorates David Lynch in the New Yorker:

When Lynch was fourteen, his family moved to Alexandria, Virginia. There, a friend named Toby Keeler mentioned in passing that his father was a painter. As soon as Lynch visited the studio, he knew that he wanted to live “the art life.” With a friend, Lynch rented a studio of his own and all but dropped out of high school to make dark, Expressionist paintings.

I've noticed a lot of great auteur filmmakers start out in more "primal" art mediums—painting, sculpting, etc. I think they're just people who are drawn toward aesthetic beauty and meaning like moths to a flame, and can't imagine living for anything else.

In 1964, Lynch went to art school—first at the Museum of Fine Arts, in Boston, and then at the Pennsylvania Academy of the Fine Arts, in Philadelphia. Boston left him uninspired; he preferred Philadelphia’s industrial wastelands and abundant lunatics. In Philly, his neighborhood was so dangerous that, when he went out, he carried a wooden stick studded with nails. An apartment he lived in was near a morgue, and Lynch met someone who worked there at a diner; the man offered him a tour, after which Lynch sat among the corpses.

Some places and experiences are just more inspiring.

Lynch’s ideas weren’t pictures on a mood board. They were experiences, which could only be realized cinematically, through combinations of performance, visual composition, music, sound, and an often dilatory use of time. Because the ideas went beyond language, it wasn’t easy for Lynch to explain them to collaborators; he developed strategies for helping them embrace his vagueness. Transcendental Meditation, which he began practicing in the seventies, had both a creative function—it helped Lynch regard his ideas nonjudgmentally—and a social one: it smoothed out his rough edges, lending him the aura of a benevolent guru. (“Your anger. Where did it go?” Reavey asked him, a few weeks after he began meditating.) Practitioners of T.M....

Have we reached Peak Elon?

January 25th

2 min read

From Mizy Clifton at Semafor (link):

Only about 3 in 10 US adults strongly or somewhat approve of US President Donald Trump’s new Department of Government Efficiency (DOGE), led by his ally Elon Musk, according to an Associated Press/NORC poll published Friday.
While most Americans believe the federal bureaucracy is plagued with problems and in need of an overhaul, the poll suggested that many did not believe the tech billionaire was the right person to lead the effort.

While this poll is specific to the DOGE thing, it is probably not a terrible proxy for Elon's approval rating in general.

Something I've been wondering since Musk jumped on the MAGA train and started what seems like a slide into even more bizarre, cringe, and erratic behavior than we're used to—which is saying a lot—is at one point is his stock actually going to drop?

When are people finally going to be fed up enough with his bullshit to the point where people on both sides of the aisle finally agree he's got to go, and it's time to stop giving him all this attention? There's a lot you could say when it comes to Musk, but I think more than anything, he's just fucking annoying. He also seems to have lost all the redeeming "rocket man" charisma he had at one point. Online right culture appears to have swallowed him whole, and he's just a troll now.

While I think on paper it may seem like he's on top of the world after Trump's win, I suspect it's going to be all downhill from here—especially because we seem to have already reached Peak Tesla, as the car company is set to continuously lose marketshare going forward.

Even if Musk and Trump can last 4 years without having some sort of falling out, I don't think 2028 will be kind to either of them.

Prior to true AGI, how is AI ever going to be good business?

January 24th

5 min read

A sad, rusted out robot

From a New York Times article about the new AI chatbot from DeepSeek, a Chinese startup:

The DeepSeek chatbot answered questions, solved logic problems and wrote its own computer programs as capably as anything already on the market, according to the benchmark tests that American A.I. companies have been using.
And it was created on the cheap, challenging the prevailing idea that only the tech industry’s biggest companies — all of them based in the United States — could afford to make the most advanced A.I. systems. The Chinese engineers said they needed only about $6 million in raw computing power to build their new system. That is about 10 times less than the tech giant Meta spent building its latest A.I. technology.

After the recent Stargate announcement from the White House where tech leaders, standing with Trump, announced a gargantuan $500 billion investment in "AI infrastructure," I just couldn't help but wonder if all this money being thrown at AI, in this very short period of time, is just going to get torched. It all just feels very hype-y and gold-rush-y.

I don't see how it makes sense from a business perspective, unless the investment directly leads to actual AGI—which is highly unlikely in any reasonably short time frame.

It makes sense from a national security perspective—not wanting to be dependent on foreign countries for critical tech, etc.—but it's just a lot of money, really fast. Not to mention that, for some crazy reason, they have invited possibly the greatest money torcher/bullshit artist in corporate history, Masayoshi Son (the guy who invested billions into WeWork, lol), to help lead the charge.

And now I've just found out about DeepSeek, a Chinese startup that appears to be in the same ballpark as the leading American LLMs/chatbots, who accomplished their feat at a fraction of the cost and using a fraction of the advanced chip tech that we've all been led to believe–thanks to the emerging NVIDIA-industrial complex—is necessary in orde...

Quote: Ira Glass on taste

May 2nd, 2024

2 min read

“Nobody tells this to people who are beginners, I wish someone told me. All of us who do creative work, we get into it because we have good taste. But there is this gap. For the first couple years you make stuff, it’s just not that good. It’s trying to be good, it has potential, but it’s not. But your taste, the thing that got you into the game, is still killer. And your taste is why your work disappoints you. A lot of people never get past this phase, they quit. Most people I know who do interesting, creative work went through years of this. We know our work doesn’t have this special thing that we want it to have. We all go through this. And if you are just starting out or you are still in this phase, you gotta know its normal and the most important thing you can do is do a lot of work. Put yourself on a deadline so that every week you will finish one story. It is only by going through a volume of work that you will close that gap, and your work will be as good as your ambitions. And I took longer to figure out how to do this than anyone I’ve ever met. It’s gonna take awhile. It’s normal to take awhile. You’ve just gotta fight your way through.” - Ira Glass

PredictIt season is upon us: the story of how I doubled my money betting on the 2020 election, and exploring the potential for round 2 in 2024

December 17th, 2023

8 min read

The homepage of PredictIt.org
A fool and his money are soon parted.

Primary season for the 2024 Presidential Election begins in a few weeks—starting with New Hampshire and Iowa.

While the official start of the 2024 election cycle means many things for many people, in one corner of the world/internet, it means it's time to make some money.

In 2020 I caught wind of a new and legal way to bet on US politics: PredictIt.org It is a site the allows you to place limited bets on political events, using a clever, market-based system based on buying “shares” of outcomes. I discovered the site in the weeks before the 2020 Biden vs. Trump election and, after comparing the markets on the site to popular polling data, it seemed that many of the people betting on the site had virtually no respect for polling data whatsoever.

Specifically, the site seemed flooded with Trump voters who, thanks to the surprise of the 2016 election, felt that the polls were underestimating Trump once again. While believing that the polling data was underestimating Trump again was not crazy, the extent to which these market-setting bettors believed the polls were off, based on the bets they were placing on the site, was batshit.

Looking at the pricing on the site, across the board, these Trumpsters were implying a margin of error ~2 standard deviations away from standard margins of error for polling. In short, they were betting that the margin of error for the 2020 election polls would not just be larger, but significantly larger than it was in 2016.

It seemed like an easy opportunity to make some cash—so easy, I thought there had to be a catch. But while the site does take a fee on winnings, and there is a max limit on each individual bet ($850, last time I checked), I calculated that I could make a virtually risk-free ~30% return on any money I put into the site. While 30% return is not exactly hitting the lottery, it is an incredible rate of return for a single event, and would absolutely fast-track the growth of my young portf...

Quote: thinking inside the box

December 16th, 2023

1 min read

"When we first started the band, we wanted to have a formula,” he says. “It’s like, ‘This is what we do, and we’re not gonna try and go outside the box too much. We’re gonna explore the box we’re in. I’ve always been a big fan of that. I used to be in bands where was like, ‘Man, we’ve gotta think outside the box!’ And all I’m thinking is: ‘You guys don’t even know.’ Music should never be just for the sake of being experimental. Before you even start, you have to know what you’re experimenting with first.”

- Mark Speer, lead guitarist of Khruangbin

In praise of boring backend tech

December 10th, 2023

6 min read

For awhile now I've had the goal of making a Django Rest Framework boilerplate starter that I could quickly spin up to use as a backend, and pair with my React frontends, for various SaaS ideas. I wanted to be able to just run a few commands and have a fully featured backend that is easy to manage, easy to iterate on, secure, and built on boring, mature, battle-tested technology that makes security and maintenance a breeze.

Django, the python-powered web framework for "perfectionists with deadlines," is famous for meeting all of these production-ready requirements in an efficient and painless way, while also providing the option to override and customize any feature you need. The out-of-the-box admin UI is also a game-changer.

For these reasons, I've noticed a lot of prolific and actually successful indie developers also using mature frameworks like Django or Rails for their backends in order to quickly ship their SaaS MVPs ideas, taking them from idea to (truly) production-ready application in 2-3 weeks of relatively easy/gentle coding, instead of 2-3 months of manically hacking together whatever the NodeJS runtime backend-stack-du-jour happens to be. These coders often sacrifice multiple weekends writing custom code and queries for really basic functionalities that mature backend frameworks have built-in. And, even if they've found ways to cut down their dev time by using boilerplates etc., they're kidding themselves if they are actually convinced that what they've written as a single, indie dev is a) secure by professional standards and b)not going to be a huge timesuck/much more difficult to maintain as months go by and their massive dependency tree of npm packages inevitably leads to breakage.

If not already obvious, I'm in the camp that believes backends should be boring. I think the hype cycle of online web dev communities and twitter routinely brainswashes many young junior/intermediate developers into thinking they always need to be using the lat...