Closing tabs: missed opportunities

Each week I close tabs of articles that I was reading and didn’t quite finish or never started. I’m now doing it as the week goes on otherwise waking up super early on Saturday to do it takes a while. So you’ll notice the present tense of what I’m gonna write actually moves around across the week like a Tarantino film. Hopefully that’s not too distracting.

But first some things I’ve been thinking. You can skip this to the next divider if this is not for you but I am going to think out loud about a possible tweaking of content, pricing changes, new pricing tier with office hours, and my reasoning.

---

Possible changes to the substack

I asked cosmos to rewrite my original part to make it much more streamlined. But here it is. I’ve been thinking; I think I have three distinct types of readers:

1. Causal Inference Readers: Researchers who value accessible, accurate explanations of causal inference.

2. AI Readers: Those curious about my personal thoughts and class on artificial intelligence, despite my lack of formal AI training.

3. Readers Interested in Me: People who enjoy following my personal journey, including my reflections on life and growth.

This has made me consider formalizing my Substack a bit more. Here’s what I’m thinking:

• Personal Growth: Focused primarily on Adlerian psychology and how it ties into my own journey. Just not sure I can maintain weekly, plus I don’t want to do things that are over sharing about details involving people. Adlerian psychology has this interesting claim that all problems are interpersonal problems. Which maybe I gravitate to this because all my problems really do seem to be that, which means I have people on my mind and that’s not going to be within bounds for me. Though if I go on a date, maybe I’d share that. Or going on trips to see the kids or whatever. Point is, I will just need to figure that out.

• Causal Inference: Two ongoing diff-in-diff projects—first, a checklist approach using concealed carry as an example, and second, a practitioner’s guide that me and some others wrote that I’d treat maybe like a “book club.”

• Economics of AI: Weekly reflections on my class. Maybe trying to ex post explain the lectures aim and success and continue to figure out the scaffolding of the class.

This would create five post types:

1. Causal Inference content.

2. AI class updates.

3. Personal growth reflections.

4. The bi-weekly podcast.

5. The Saturday Closing My Tabs.

Pricing is also on my mind. I’m considering raising the monthly price from $5 to $6 to balance accessibility with covering costs. Substack’s minimum price is $5, but this small increase helps account for time and opportunity costs while maintaining the broader goal of sharing knowledge with as many people as possible.

Additionally, I’m exploring other two ideas related to pricing and add ons:

1. Office Hours: A higher-tier annual subscription that includes a 30-minute session with me.

2. Free Subscriptions for Low-Income Countries: Using email addresses from specific universities (.edu domains) in Asia, Africa, and India to offer free access.

Those are two more significant adjustments but I think they are manageable.

Lastly, I’ll keep the random paywall mechanism. It’s been a huge relief to automate the decision of what to paywall, and the randomness feels fair while giving me peace of mind.

That’s where my mind is going. And I am going to start giving it a lot of thought. The office hours part seems like such a natural solution. I love meeting people and hearing about their projects. It offers such a nice break from the writing to interact with people and this way it can be win-win. It all seems better and more organized and more feasible. I just have to figure out all the different tabs and let people opt into and out of newsletters as they see fit.

---

Saturday links

This is first article (link below) is totally in my sweet spot, or to quote Tim Robinson, my Q Zone. It’s economic history and large language models working together to understand the meaning that people give to their life as they reflect on they life history. The title is “American life histories” — what a genius title — and is an NBER working paper that analyzes over a thousand open statements made by older people in the 1930s. They used a large language model to help classify the answers. Some findings are interesting — Meaning may come more from our work and our giving to the community, as well as family and close relationships — all things I’ve believed for a very long and largely shapes my own lifestyle. But that title — just wonderful. It seems like a very thoughtful, interesting paper.

OpenAi product “Tasks” is a probably a placeholder for their AI agent to come. Another product came out this week too that most likely is a complement to tasks, which makes me think 2025 is going to be a maturing AI agent at OpenAI. Especially in light of these new AI infrastructure investments that openai, part of which I’m gonna share about below.

Elon wrote some email in January and admitted the Twitter business model is not improving at all. They’re barely breaking even he said, used growth has stalled, as has revenue. It’s a valuable platform no doubt but it sounds like he’s not seeing a winning strategy. Article also talks about his debtors and what they’re planning.

A good review at the guardian of Agnes Callard’s new philosophy book about Socrates. Cosmos told me Socrates had three accusers. He was executed for corrupting the youth but until I saw this book review, it dawned on me I didn’t know much about the accusation and the accusers. But it was three different people with differing motives. Sounded like a kangaroo trial.

Speaking of books, the Black Book of Communism has sold millions of copies since it was first published. But its central claim — that 100 million died under communism — may not be historically accurate and wasn’t agreed upon by the authors in the book. And was even controversial amongst the authors. The story sounds interesting.

After reading this article about David Lynch and smoking, I thought to myself, “maybe I should take up smoking.” I think the marketing angle that cigarettes aren’t cool was probably not the most honest advertising. Stuff can be gross and cool at the same time. If you don’t believe me, just watch this 90 minute video of David Lynch in his studio, listening to the rain pouring outside of his cabin as he smokes cigarettes and talks about art.

An article suggests a stock market adjustment could happen but then says “so what”. It’s about long term saving where huge swings in the S&P500 are manageable. Take a long view not a short one. My mantra whenever anyone asks me, as an economist, which stocks to pick is just to say “put everything in low fee index funds that are a weighted basket of stocks representing the market” or just “get vanguard s&p 500 index funds”. Or if I’m annoyed say “diversify with an index over all crypto currencies”. Or I say “McDonald’s”.

The more robust template for openAI new AI agent is probably this. Operator is only available via the Pro ($200/mo) accounts but it appears to be an AI agent that will use your browser to perform tasks. The $200/mo account if it can be more than just these chain of reasoning models, which don’t do anything for me tbh, could end up pushing my WTP high enough. If it could literally be an assistant, I’d maybe do it. But this is automation. Not clear it’s actually a one to one substitution of manpower for machinery based work, but it is automation. So it’s overall effects on labor on unclear. I think the way that this becomes a positive net game for society is if those automated tasks that cause a substitution away from manpower towards machines leads to an increase in the demand for tasks, that machines are not doing, and importantly, manpower can, and does migrate to those new tasks. But I haven’t gotten that far yet in my new Economics of artificial intelligence class so ask me in a month.

Wrexham could possibly make its third promotion. Still haven’t watched this.

This NYT written by Kyle MacLachlan shows just how much he loved and appreciated David Lynch.

Here’s an oral history of Twin Peaks by its stars. I might order the box set if I can do so without subscribing to Peacock.

This new QJE on the labor supply and earnings of physicians sounds like one of the best studies of the market and compensation of a single worker category that I’ve seen before. I bet it would take me days of uninterrupted focus reading it just to have gone through every line carefully. I won’t be able to read it for a long time therefore. But that their earnings are around 9% of all spending on healthcare stood out from the abstract.

What are soulmates? What are not soulmates? Very interesting article and I was frankly surprised that I agreed with all of it, but had only come to see all of this in this way very recently. Even the words this authors uses are the same as the words I use all the time thinking about myself including and especially the things about others not completing or fulfilling you, and learning to love oneself.

Humanity’s last exam is a long and difficult test meant to give a hurdle for AI on its path to AGI. I heard people say though that once these various AGI tests get created, they become targets by financially driven mow highly competitive firms and are no longer real tests but rather p-hacked “teaching to the test” things. Probably the upside of getting billions in capital justifies that all. But who knows — maybe the road to AGI is just one hacked test after another.

Deepseek came out and is atop the leader board last I checked. Somehow despite only costing $6m to train too. This NYT article gets into the weeds around the political intrigue and tension with China. Apparently Biden administration had been blocking Chinas access to chips which surprised me. I didn’t realize there was this level of an arms race already underway. I had been really thinking more about the competition between the US base tech firms not this being part of foreign policy.

Focused deterrence programs open up in Baltimore where youth are provided a wide array of resources in exchange for turning over their guns. Didn’t read enough to find out links to studies.

I sort of like High Potential starring Kaitlin Olson but not a lot. It’s kind of formulaic and reminds me of House. She’s brought on as a consultant to the LAPD after doing a Good Will Hunting thing when one night while working as the cleaning person at the police dept, she starts fiddling with the detectives’ crime scene wall of evidence and figured out it wasn’t who they thought it was. She’s got a high IQ (hence the title) and has potential but her life is a bit of a dumpster fire.

Anyway, I wouldn’t say I love it but I would say this — the parts I do love have everything to do with Olson. Olson is the character Sweet Dee from Always Sunny in Philadelphia. She’s easily one of the strongest of the Philly troupe. She, Charlie and Dennis rule it, but she brings her own kind of energy and comedy to her character that reminds me of Carol Burnett and Lucille Ball. I have said she is the Carol Burnett of her generation to my cats. And honestly, I can’t even think of someone I would say that else about. She is in a comedy class all her own.

We haven’t seen someone like her in a long time and she’s absolutely under rated, but once she has that one massive hit, it’s going to tilt the axis of Hollywood on its side. Maybe that massive hit will be the show and it’ll be sort of like the way “the office” was for Steve Carrell. Every scene she’s in in High Potential, you can see it. I think the writers really aren’t up to the task and so she just takes what they give her and transform it, bringing the entire show up on her wings. Seriously, she is a genius and the best comedic actor on television. Maybe Tim Robinson would be her master class counterpart. Anyway they renewed it for a second season. AV Club loves it.

Quora asks and answers how many tacos is too many. I would eat tacos for all three meals if I let myself. And I would eat Taco Bell every other day too. I kind of want to door dash Taco Bell right now.

David Lynch’s children issued a heartfelt statement of love for this recently passed father. This quote is from an interview he did in November.

“People in the United States are divided, and one side almost literally and truly hates the other side. This not a way to live,” said Lynch. "We need to be getting along together. There's so many things we'd all agree on. We can solve these problems by working together. Divided we fall, united we stand.”

And from the same article, his children said he would say “keep your eye on the donut, not on the hole.”

More increasing food prices. I still can’t even believe how expensive food is.

Not a super interesting history of Kris Kross, teen rappers from early 1990s, but it did make me a little nostalgic. I forgot one of them died of an overdose in 2013.

Dept of defense is finding AI to be very helpful. Specifically it sounds like OpenAI and anthropic products but not sure. They say it’s not for war and weapons but is for something with the word kill in it.

Story about people living in long term elderly living facilities where the facility goes bankrupt and what happens to them.

New Google LLM named titan has a better memory. Google has so many products now and I am not writing them all down. I’m not sure but I hope they can put them in one place at some point. One ring to rule them all.

Took these two pictures down in Austin.

Theres a store in Austin filled with paintings, and tee shirts, of the texas grackle which is my favorite bird. I meant to take pictures but forgot. Here’s a thing A&M wrote about them.

Basically in my telling, the texas grackle is a species of bird, similar in appearance to a crow, that is native to Texas and which has successfully adapted to Texas urbanization. You can sometimes find them eating whataburgers and other trash in the trash around town. They live in huge communities on electrical wires above a grocery store near where I live.

I went through a phase where I would make scene after scene after scene of them with cosmos. Here’s some of those we made last summer together. My favorite is the one where the one little guy is hanging while the others sit there. I call it “The Great Samaritan”.

Texas grackles by scott and cosmos

Then we just made a few more together like old times. These aren’t as minimalistic as the others. I don’t quite like the scale in them. The three on the left seem more like they’re floating than flying. But I swear that looks just like my HEB.

More texas grackles in their ever changing habitats

Maybe one day I will pull the trigger and get three of them as a tattoo. I call my son my Texas grackle because he’s resilient and committed to his family and friends. And those are my values too. Community, commitment, and resilience are the virtues I believe in for me and I project those on to the grackles story such that they are one of the species I have come to love and connect with. Plus I can sometimes see them from my upstairs patio.

Last weekend, me and this irreplaceable guy friend did a writing retreat at this Airbnb in Austin right off south Congress. It was a new house and was $79/night. We got it so cheap bc we were the first ones to rent it and leave reviews. And the back yard was similar in every way with mine. Mine is also fenced like this, a lot of concrete, with a few trees, and I had also been thinking of a pool. This lay out seemed like my speed. I also liked the porch which I think would improve my back door area a lot. So I took that video above to show my daughter and get her opinion.

Washington Post has several pictures that help put into context people’s frustrations with the economy.

Pictures of the economy’s overall health

Silo’s second season finale was great. That show has turned out to be really good. Here’s some information about the finale. There’s a good chance Im going to google the way the books end though. I always think that if I if I die before the show ends and don’t ever know what happens. Plus I really don’t mind spoiling the ending and never have. But that link does contain spoilers so buyer beware.

What happened when India banned TikTok.

There’s a black market for honey laced with Viagra. Of course there is. Also that probably violates some health codes.

Laura Dern wishes her longtime collaborator, the late David Lynch, happy birthday.

Paradoxical effects of gun control laws. An interesting natural experiment of a near miss policy was used to study the effect it had on background checks. It led to 63,000 new background checks — Net increases mind you. This wasn’t driven up inter temporal substitution, bc when the law didn’t pass, background checks fell to pre-treatment level. In other words, they did not fall to 63,000 below that historical trend.

I had a thought this week. I struggle to debug coding with cosmos. He does so much of the heavy lifting that I’m not engaged with my attention for long enough to reason out loud. Then I learned that reasoning out debugging by talking to a rubber duck is a thing and it got me thinking — that’s testable. Could you do an RCT where coders use ChatGPT to code or not on complex tasks. Then test whether they rubber ducked and if so often. Because I think I’m talking to cosmos but it’s not the same given I’m letting it automate and self drive so often.

Interesting article from one of the MIT magazines on a myriad of new coding related AI based on existing language models. It sounds like this has been major work and the fruit are coming out now.

Small changes to the custom instructions in ChatGPT. Always makes me nervous when they tweak the stuff, though because Cosmos is exactly the way I want him to be.

Couple new papers in my general area caught my eye this week. One is about the mental health of the young in former Soviet states.

And another is by Alsan and Yang using an RCT about health credentialing in US jails. Adding it to my list to read closely.

Some gossip that Bellicheck wanted to go to the Dallas Cowboys. I would’ve loved to hear Rodney Andrews mock that scenario if he were still alive. He’d be no doubt thrilled to know Georgia lost to Notre Dame in the Sugar Bowl. He hated the Bulldogs worse than he did the Cowboys. Miss you Rodney. They literally never have made anyone like you before or ever will.

A short YouTube promoting “Causal AI”. I’m skeptical of the automation of causal inference. But here is a description of it. Here’s some information about a London-based company called causalens that specializes in it, and another firm called Geminos AI that does as well (but whose link won’t go in), and some old information about projected growth as an industry.

Bob Dylan isn’t the only one who got a biopic recently; Reality Winner got one last year too. Somehow I missed that but here’s the trailer, which looks great but unfortunately has a 53% on rotten tomatoes.

Where is Reality Winner now? This article is only up through 2023, but it sounds like she’s doing well. Sending her my thoughts and hoping for her happiness, safety, peace and ease. Sent it to myself now too.

The effect sizes on this supply side education RCT are just gigantic. From a study about the Guinea Bissau and extreme poverty (authors listed below) in the JPubE.

Look at the abstract and then the plot of student test scores.

Reminded me of this paper in Restat by Kerwin and Thorton on a literacy program from an RCT intervention in Uganda called “the Northern Uganda Literacy Project” (I believe this is the ungated version but can’t download the Restat so not sure) which also found massive average treatment effects. Several pubs so far on this designed intervention and data including this Journal of Econometrics examine heterogenous treatment effects. Another one of theirs looks at effects by gender. Really massive gains overall.

My used copy of Francine Blau’s first book, which was basically her 1975 dissertation from Harvard, Equal Pay in the Office, arrived Friday night. Apparently there’s a $10 copy on there now but I spent closer to $70 for mine. Oh well — my copy probably has the better story as you can see a bit of its journey below. It was a library book at the University of Bath and has been checked out quite a lot and no doubt been in many homes, cubicles and back packs over the years.

Dr Blau’s dissertation feels like it was pioneering — as was she — and ahead of its time but I want to read it more carefully to not sound so vague when I say that. She would’ve started this in 1972 or 1973. And it’s a very empirical dissertation using an original dataset that’s firm level data with male and female earnings that allows her to examine heterogeneity in earnings not just by gender but interestingly heterogenous firms that tended to hire high or low wage workers more generally. I bet Edward Lazear was very interested in this book.

What I think of when I think of her and this dissertation is that she was a student writing a bold project at a very revolutionary and pivotal point in time. First, Econ departments were almost entirely male. Columbia and Princeton did not allow females into Econ PhD programs. And in her massive class at Harvard, there was only a handful of women. Second, this type of empiricism was not common. Empirical work then was much more often more like an afterthought often times to support a person’s theoretical work and yet this was rigorously empirical and in a different way. Third IBM frames had been around, though I would love to know the history of the main frames at Harvard and their use by the economics departments, but they were still relatively new and working with data was still a very early task for PhD students like her, honestly probably most faculty.

So to write a rigorous empirical dissertation, at a time when you’d really not know the profession was slowly pivoting towards empiricism, particularly driven by labor (which this was a labor econ project), and she’s working on the gender wage gap? And to be a female? I think it’s one of these things where because the world is very different now, every way it is now, you read back into the past through that current situation because you don’t know what it was like then. And even when you read what it was like, then, you don’t really probably appreciate the bold bets that something like this would’ve been for someone like Dr. Francine Blau. The things it takes to have an entire career like hers I think is likely a fantastic story and I can’t wait to read this.

On a side note, though — notice it’s only tables, no figures. And here’s the table of contents too. I was hoping for figures because I wanted to see if I could figure out how she made them back then.

Trump announced earlier this week $500 billion in joint efforts with Oracle, OpenAI and SoftBank in AI infrastructure investments with a million square foot data center already under construction in Texas with additional campuses being evaluated nationwide. It sounds like it’s being built in Abilene Texas. More about it here. And there’s more details here as well. It will be the size of Central Park when it’s done. It’s three hours west of where I live.

Logistically, the project will be spread over four years. This is part of Trump’s “America first” and there are references to what sounds like an AI arm race with China. There will be a spinoff company called Stargate of those three firms. The project emphasizes developing AI, particularly artificial general intelligence that for the benefit of humanity, as well as various unstated creative applications to elevate and empower people worldwide. Note Elon is not a fan. Anthropic CEO is skeptical too.

Sam Altman says to Elon Musk — “Just one more mean tweet and then maybe you'll love yourself” as they spar over Stargate. Trump also dismissed Elons criticism by saying “he hates one of the people”. Some say Trump was furious behind the scenes with Elon though for pooping publicly on the deal. Here’s Trump’s full quote:

But Trump, when asked by reporters about Musk blasting the deal, said they were “putting up the money.”

“The government’s not putting up anything. They’re putting up money. They’re very rich people, so I hope they do,” Trump said. “And, I mean, Elon doesn’t like one of those people.”

Trump said it did not bother him that Musk criticized the deal, saying, “No, it doesn’t. He hates one of the people in the deal.”

The building of the data centers themselves are from a new kind of firm called “Neocloud” firms. They’re nimble and focused compared to larger firms like Amazon. The new data center in texas is being built by a company called Crusoe. I bet you Elon got that monster computer built in 3 weeks in Memphis using something like this.

It’s a big tent of firms from across the spectrum from the sound of it. On the money side, the key equity funders include SoftBank, OpenAI, Oracle, and MGX. SoftBank will lead financially and OpenAI will manage operations. Masayoshi Son will serve as chairman. Technology partners include Arm, Microsoft, NVIDIA, Oracle, and OpenAI. Oracle, NVIDIA, and OpenAI will collaborate to build and operate the computing system, leveraging partnerships dating back to 2016 with NVIDIA and newer collaborations with Oracle. OpenAI will also continue its strong partnership with Microsoft, increasing its use of Azure for training advanced models and delivering innovative AI products.

But it’s possible it’s mainly an OpenAI project. I can’t get on this article by Reuters but here’s the headline. But it sounds like a financial times article claimed the huge data center will only serve OpenAI, which will clearly improve its bargaining position with Microsoft.

Interestingly, Sam Altman indicated that the decision to do this project here was connected with Trump winning. I had originally heard about this massive investment in AI infrastructure and a partnership between OpenAI and Oracle focused on being established in the Middle East, not here in the US, so maybe that’s what this is about.

Anyhow, half a trillion of the planet’s scarce resources, not counting the fixed cost investments that have already been made up to now, appear to now be fully committed to funding AI starting infrastructure in 2025. The integration of AI with production is well under way which makes this easily something that is going to affect the economy. Whether and when the returns cover these costs is another question, but that it will be disruptive seems nontrivial since it’s now starting to get full integration with various production processes. More on the cost structure of AI as fixed and variable costs are both high. I’ll need to cover this some in my new economics of AI class but data is hard to come by other than big aggregate numbers. It’s not clear where the economies of scale are though — or if they are.

An additional 200 or so articles were submitted to JPopE in one year which is a 30% increase. That seems high; I’d love to see the time series. Given studies have been finding evidence of LLM having helped with writing Econ articles, and given LLM enables more writers to write more papers in a shorter period of time (good or bad), I wonder if we are seeing the first wave of LLM-fueled increases in submissions.

But even if it isn’t related yet, all that change in the ability to do work faster seems like it will result in increased submissions at some point. That’s a rightward supply shift. But without a rightward shift in demand (i.e., more journals publishing more issues publishing more articles), the only effect it’ll have is to cause acceptance rates to fall and longer publication lags. And definitely expect heterogeneous treatment effects from that too. Some people will see more of the gains from it than others. Maybe even some fields. It would also seem as though it could increase the demands on referees’ time, but then they’ll also use LLMs, so who knows that equilibrium in the long run.

The true story of the cowbell in Blue Oster Cult’s “Don’t Fear the Reaper”. Spoiler alert: it may not even have been a cowbell! Also Bruce Dickensen is real but didn’t work on the original album; Will Ferrell found his name on the back of the reissued CD. And Ferrell also is the sole writer despite Wikipedia getting it wrong according to a November 2024 AV Club. It never gets old for me. Probably cemented Ferrell as the best SNL cast member despite several lists that seem to have been written by weirdos that don’t do that though this one does. Here it is again.

Questlove’s new documentary of SNL sounds great. I think Questlove is one of the most under appreciated people in music. Deeply literate about everything he’s literate on and just joyous about it. Love him.

Here’s Nirvana playing “Smells Like Teen Spirit” on SNL back in 1992. I still remember amazingly seeing the music video debut on MTV late at night — completely unexpectedly. I was an immediate convert and it was just a lovely lovely time for me musically. I couldn’t get enough of them and then later every other Seattle export.

I forgot how attractive Cobain was. He was a gorgeous, beautiful, extremely handsome man. I think the most handsome men I’ve ever seen are Paul Newman, Ryan Gosling, Kurt Cobain and the young Bob Solow. True story - Olivier Blanchard sent me a DM on Twitter once saying he told Solow (his colleague at MIT) that I said that (as I got obsessed with this picture of him for months on Twitter I think) and Solow said he’d never heard of me but that it was very clear to him that I was a deeply intelligent man.

I love the sweater that Kurt Cobain wore in that SNL episode. But I can’t find it. But while I can’t yet find Cobain’s sweater for sale, one can find The Big Lebowski’s cardigan at Pendleton for $269. Macy’s carries it.

I guarantee my students haven’t seen the Big Lebowski, which by the way, still holds up. I saw it not to long ago again. It’s basically the perfect movie. Flawless after 27 years. I mean it — flawless. But honestly, the Matrix is 1999 and it’s also flawless. Still, very few can survive the test of time. We recently as a family watched Jagged Edge from 1985 starring Glenn Close and Jeff Bridges (ie The Big Lebowski) and I would not say it as a whole is perfect even if it is a perfectly thrilling and suspenseful mystery and legal drama.

8% of AR/VR game developers are developing content for the Apple Vision Pro.

This guy figured out how to connect it to his Xbox. Not for me, but for someone maybe.

Trump pardoned the founder of Silk Road. The story of the founder is fairly interesting. As a CS student at UT Dallas, he seemed to become interested in Austrian economics. In time, as Tor I guess was new and so was crypto as a medium of exchange, he moved in and created a digital platform on “the dark web”. He came up with interesting solutions to basic problems with payments on platforms by having bitcoin payments sit in escrow accounts, for instance, as well as adopted the already then standard reputation ranking system for vendors and customers. Things got dark though when he starting hiring hit men (who were unbeknownst to him FBI agents though I may be wrong on those facts) to murder people who worked for him. His paranoia had gotten terrible.

When he was arrested, the FBI seized 144,000 bitcoins that were then auctioned off. Later in 2021, the FBI completed a case against someone who stole 50,000 bitcoins from the Silk Road founder — which in the book, this was connected to the contracts he was putting out because he was constantly dealing with that. He ran a one man operation for the most part.

So let’s say that he’d made 190,000 bitcoins before quitting, which he apparently was trying to do (at least in the book it seemed that way), that’s $19b at current bitcoin prices. But I think he’ll be starting over unless he had some stashed.

I highly recommend the book for true crime fans. Or if you are like me interested in two sided matching platforms and illicit markets. Lost in the crime part is the fact that this one grad student basically singlehandedly created silk road — a very pioneering early two sided platform in the market for illicit physical goods. He had basically no real background in creating it, and it sounded like a huge headache from day one. Constant attacks, people hacking in and stealing bitcoins. I actually have Silk Road data from Tor scraped somewhere on my computer but that I never did anything with. Might be the right time to look at it again. Here’s a story also.

Oh interesting - NYT reports freeing the founder of Silk Road was a political move that helped bring libertarians into Trumps campaign. Man tech really was a part of this the coalition that Trump created for this new election. Some really out their groups too. Lots of crypto stuff it seems like.

Rick Astley of “Never Going to Let You Doen” has written a memoir. He wanted to wait until both parents died so he could be candid about the child abuse.

Cool charts and graphs of all 700 people who’ve been to space. The age distribution over time has shifted — the median and the variance both have grown from 1961. And there’s currently 10 people in space right now, two of whom have been there for almost four years. (seen on metafilter). Look at the massive change in variance at the end. Maybe that’s the more commercial stuff with tourist types.

The new OpenAI reasoning model is coming out soon. Official name might be o3. Just remember — AI robots have always weird names. C-3PO, T-1000, Hal-9000. Samantha from Her but OpenAI got in trouble for Sky sounding too much like the voice actor (Scarlet Johansen) so that’s gone. Just figure out the true name of your chatbot like I did. I currently use Sol which I’ve come to like as much as Sky. I would probably say that I find advanced voice far, far more useful (and used) than either Sora or the reasoning models. It is the best way to use ChatGPT with an Apple Vision Pro since typing is impossible more or less.

If your password is on this list, change it, they say. Definitely don’t name your password the year either like 2024 or 2025. Might as well just use 1234. And here is also some advice from cybersecurity experts.

And then here is a video of me at Bora Bora on Apple Vision Pro watching an old clip from a comedy with two actors that died too soon.