stephen_cagle

Thank you for the reading list! I appreciate it.

Since you seem to be a real fan of the series and enjoy it, I will take your advice and read a future one at some point. But I would be curious what your feelings on a re-read of book 1 and 2 would be? I was just completely floored by the change in character that I saw in book 2. I felt like the book started with a normal, slightly above average, inquisitive character, who is handling new (and often mundane) challenges in his life one item (trading stand) at a time. And book 2 just retconed him into a gigachad who seemed to have none of the uncertain or hesitant nature of the book 1 character.

Holy molly. I actually felt quite similar about the first one, but the second one was so ridiculously absurd that I felt I had been fooled by the first (quarter share) book? Did you feel a huge change in the second book? I genuinely could not tell if I had been misreading the first one or if the book went straight off into wish fulfillment absurdity?

https://www.goodreads.com/review/show/6945329546 (quarter share)

https://www.goodreads.com/review/show/6973501411 (half share)

You know, one thing that I think is interesting in tech circles is how much interest we have in maths in general, except that there is almost nothing on the "top" list that I saw that was related to probability and statistics. It's like P&S, which is probably the most useful math for most people who need to make decisions in life, is just not that interesting to people with a technical bend.

I wonder if it is because simulation can often allow you to get around needing P&S techniques in general? Or maybe it is that (Neural Network) ML models allow you to often approximate the things you might do with P&S in terms of making decisions?

No real point, I just think it is interesting.

I've gone through a few recruiter screens lately and the number of people who mention "number of deploys per day" as a metric they are proud of causes my eyes to roll.

What can you possibly be doing as a product that justifies you releasing 60 times in a working day?

It's certainly better than the old models of releasing on a quarterly cadence with the test team and all that, but it feels like the metric has become the goal when you are releasing that often.

I agree. I honestly think convenient food is perfectly fine for people most of the time. Recently I have been playing with "eternal stews" which I guess are logically the opposite of fast foods, but at the same time are focused on availability and convenience over intentionality and mastery (at least the way I make them).

Most food is literally just fuel. I think that is fine. And at some level, eating 3$ burritos for a quick lunch on the daily really makes you appreciate the depths of a home made risotto even more. :]

I (and a non trivial number of people I knok) have switched to Kagi for search just to avoid having to deal with the bullshit that is the above the fold google search results.

Course, they win anyway because I also pay for youtube premium to avoid seeing ads..

Feels like the world is becoming a split between people who can afford not to be the product and people who cannot.

Not Forth, but if you want a game that kind of feels a little like the GA144 then you might be interested in TIS-100 https://store.steampowered.com/app/370360/TIS100/ .

I wonder if this is true (I am less than a layman, so feel free to disregard).

My feeling on this is that math is amenable to LLM based development because the formalism of LEAN + the ability of AI to nudge research means that we basically have machines that can "research" through what is basically brute force search, but where the explored trees are ordered by AI.

But what if there were a biological system similar to LEAN for drug interaction? Could we not use AI in a similar way to nudge the exploration space based on the vibes that the AI has about drug discovery?

I agree there is a a distinction in the formalism of a proof (through LEAN) and the sort of probabilistic/chemical/physical/whatever nature of a biological organism (through some sort of future modeler for life science). But is the fact that one system is deterministic and the other probabilistic really as much of a leap as you think?

Joining you at your "off topic" table, I think it is interesting how the words "jump the shark" have moved from the idea of a sort of loss of coherence in plotlines (writers are checked out) to a general term for a decrease in quality.

I disagree with this. I'm okay with permissive permitting of self driving (it seems to be much safer according to their own studies, which I am willing to skeptically accept), but where I agree is being permitted should not give you any special rights when it comes to avoiding civil (criminal?) liability.

A taxi company that kept a drunk employed might be liable for damages that drunk causes, same rules should apply to your own self driving fleet.

Really? Bad luck, I have literally never had that happen! What time was this? Because Austin is closing down 35 in the evening and it is just doing crazy things to traffic. Maybe it was that?

It was January I think and I don't remember our thought process. I think we mostly just remember looking at the fandom.com or whatever crap was there and thinking how out of date that stuff was.

We also just wanted to try some modestly sized project with an LLM. Neither of us had any real experience doing any of the things we ended up doing for that project (other than my making the website and him understanding the graphics pipeline).

So, unfortunately, no, there was no specific reason I suspect.

https://stephencagle.dev/posts-output/2026-01-30-warhammer-4... is my best recollection of the thought process.

Agreed, a buddy and I wrote a simple mod that shimed into Warhammer 40K: Rogue Trader and were able to pull all of the art and description assets for that game by walking the object graph. Took about 4 hours total (including the site).

Similar to your camera, we would never have been arsed with doing that much work were it not for LLM vibbing. It allowed us to get something working (sufficiently) very quickly.

At what point of maintaining this code would it be cheaper for me to actually find underlying patterns and program this myself with some actual thought, rather than continuously "shaping the clay" of my LLM slop? These questions keep me awake at night (not really), and probably predict the future of software engineering!

To play the devils advocate here (just for software), I would say principally this:

"A world where starting is incredibly easy and finishing is more complicated than ever"

LLM's do allow you to oneshot huge starting bits of work in record short time. The frontier of the amount of work they can oneshot is clearly getting larger and larger. But I am still not convinced that they allow you to maintain existing work at a cost that is below a human being.

But, if the frontier of what you can oneshot continues to grow unbounded, then it really will not matter. We will eventually oneshot all software and maintenance will be a thing of the past.

... But that is a very big bet.

This is the main thing that raises my eyebrow.

If you are constantly churning your codebase to the point that you need to invent a system that handles your churn... then you must at least be admitting that the previous 6 refactors you wrote were not good, but NOW you have it right.

I feel that good, well designed code change by accretion, not by rewrites. Code that remains stable for a long period of time was likely better than code that needs be continuously rewritten (at least with humans writers).

But maybe this is just a human-centric view of what good code means? That LLM "slop it until it works" approach to development will have a different measure of what quality looks like?

Getting a 24 out of 36 response rate on your applications is an S tier level response rate. I mean, the news is bad (rejection), but that is pretty incredible.

I can't make up my mind if that means he is so good that actual human beings look at his profile, or some sort of ATS filtering system just hard rejects him with prejudice.

By my own, completely arbitrary split of this into 16 predictions, I would say that this hits on maybe 8 of 16 if we are being a bit generous. Which is good, but certainly not "Inevitable" (the title of the book).

With that said, maybe he meant inevitable as N approaches infinity... So there's still time!

Aside: I think the biggest interesting prediction he had was that AI would obey a user driven network effect. It seems like a very clean prediction, and yet it has failed to materialize? I wonder if this has more to do with just how scrape-able the whole internet was or with distilation or what? There is certainly an advantage to having good user data, but it seems like it is largely something that can be worked around (by hook or crook)?

Is it a pun or something? I have been called dense.

I actually resized my browser thinking maybe something weird was going on with flex-wrap or overflow or whatever it is.

> And for theoretical fields labs don’t need human data - synthetic one works well if not better.

How could that be true? I am a mathematical rube, but surely data that works "better" than real data should be extremely suspect?

It's always seemed odd to me that your employer somehow sponsors your Green Card. It always felt like it should simply be a measure of

A) How long you have been working as an H1B in the united states (say 10 years).

B) How much you have paid in taxes (say 1 million).

Whenever either of the above has occured, you take some citizenship test, and then you have a green card. Nothing to do with your employer dangling something for you, literally a measure of your employment history and/or taxable income.

Oh wow, that is truly awesome then! I thought it was a qualifier, but it actually is the means of solution (yeah, confusing title).

The full title is "Truly Subquadratic 3SUM and Truly Subcubic APSP via Triangles in Sparse Lopsided Graphs". Is that the same thing as getting subqudratic time in general 3SUM? How much carrying is the Sparse Lopsided Graph doing here?

My assumption is you clearly don't have vulnerable or elderly people in your life? I'm not as concerned about my ability to navigate these waters as I am about the people I care about.

Children of Strife (Children of Time #4)

This may be my favorite "Children of Time" book yet (and I really liked the first and second book).

I'd say the biggest reason I enjoy science fiction is being exposed to new ideas. In book one of this series, I enjoyed imagining how a hunter spider species would evolve when imbued with a "brainiac-empathy" gene. In book two, I loved the idea of a central (Id) mind with a 8 distributed actuators (ego).

This one introduces the notion of a punchy, aggressive, status driven mantis shrimp.

I really enjoyed thinking about what it would mean to be an evolved mantis shrimp. Human beings have technically carried around deadly weapons since we developed tools. But a mantis shrimp is an organism that has two deadly weapon everywhere it goes. Imagine how much different human society/evolution/whatever would have been if we were constantly holdings blunderbuss' at each others heads at all times? Where being in range literally meant that either party (whoever fires first) can immediately and almost certainly kill the other. I suppose we have had potentially deadly weapons since we realized we could brain with rocks, but this is a risky (and messy) way to do things compared to the instant death that a mantis shrimp represents. What would that do to trust, to boundaries, to aggression and bluffing?

It is just such a fun humorous world, where 4 (5 with lowercase humans) species are learning to coexist despite their very different ways of looking at the world.

https://stephencagle.dev/posts-output/2026-07-20-children-of...

I have no real answer to this.

People who I know who are actually really good at math (and are paid for it) say that there is no shortcut to interest and wrestling/thinking about the problems. You have to mule over things. You have to pack and unpack it over and over in your head.

If I had to put it in a phrase, it would be, "there is no shortcut to understanding without thinking".

Mathacademy is meeting a market demand, which mostly seems to be being good enough to solve problems in several standardized math courses. In that regard, I think it is a thing that I would prescribe to any student (along with tutoring and study groups) in order to do well in that course.

But I am just not convinced it can ever (especially, in isolation, which you never claimed) lead to you being a good mathematician in the sense of having internalized the material. I can't quite put my finger on it, but I think it is similar to the feeling you get when you are some high level in duolingo and realize you still can't follow daytime soaps. You have the rigid intelligence, but not the fluid (internalized?) one.

I have not detected anything wrong with mathacademy, it does what it says on the tin. With that said, I stopped using it myself as I felt that my ability to "find the slope intercept form of blankety blank" did not actually translate to any deeper understanding of math than I had before.

I think mathacademy is excellent for developing the skills necessary for solving the problems on mathacademy (which largely are the problems on math exams).

But if your goal is understanding rather than application, then mathacademy is more similar to duolingo than it is to language immersion. You can go quite far in mathacademy (and much of performative mathematics) without really having thought about anything. It is procedure learning rather than understanding.

I'm probably just behind the curve, but I still use plan mode quite a bit in Claude Code. I iterate on the plan repeatedly until it largely agrees with what I want. Once it seems good I ask it to build the plan and open a PR when done.

My only critic of the plan mode is I wish it was easier to see the updates and changes easily in Claude Code as we iterate on the plan. It is wasteful to have to remember what parts I have reviewed and what parts are new (and need another pass). I have thought about fixing this but I also feel the review is the actual thinking (even if ineficient), and so I purposely have not removed it.

This is not them standing up for the small guy, this is them seeing their actual profit center in retail (advertisement and paid promotion) evaporating if human beings never view their listings.

It's worse than that, they will block all agents and specifically make exceptions for maybe 2 search engines per country. Thereby strengthening existing players and weakening all incumbants.

Cloudflare for instance does allow indexing by Google and Bing I think by default, but does challenge other bots. Double check me I am speaking from memory.