steve_adams_86

"The Rees algebra is an algebra over Z[t^−1]"

Such a small sentence and yet it means very little to me. I understand some constituent pieces, but I don't understand what Z is here other than a 'ring' and I don't really grasp how t^-1 converts this into a generalized family of algebra. It would take me a lot of effort to understand this and use it practically. I find that fascinating because it really is such a small statement that seems perfectly cromulent, but there's a lot packed in there that someone like me is totally missing.

I suppose there may be similar concepts in computer science, but nothing comes to mind that ever stumped me. To be frank, the field has been relatively accessible to me because it hasn't been too challenging. Not sure if that's a personal aptitude thing or it is genuinely simpler.

Does V2 finally imminentize the eschaton? V1 has completely failed at this so far

Unfortunately not a choice at my organization

A couple of my coworkers think I’m nuts for watching cost explorer so closely but

1. The time it takes to look and notice costs that don’t make sense easily pays for itself, and then some (in my experience). I doubt you spent $7k of your time tracking this down, and you probably noticed optimization opportunities that saved you even more

2. I hate the idea of wasting money on buying Jeff Bezos a bigger yacht

It’s a niche, yes, but there could be others like it. No idea about how the company should be valued. We pay them chump change for our services, but enough that with any scale it could be meaningful. And their reach is pretty incredible, so, there is a lot of potential there.

Whenever I leave a company I make sure everything that belongs to the company goes back to them

Right. I noticed a coworker who recently left the organization was still running some of our software on his personal computer (evident in the access logs) and notified him that I could see, he should be more careful, etc. We agree to these contracts because compliance matters, not just because we need the job.

Here along the BC Coast, the organization I work for has an expansive sensor network. Weather stations, CTDs, custom equipment in watersheds, research facilities with all kinds of equipment to monitor, and so on. There is no broadband or fiber on remote islands along the coast. We used to use satellite internet, and getting data off of our main hubs (everything is relayed to the hubs by radio) was very slow and precarious. Since starlink it's a breeze. We will finally be able to get video feeds off of some of the stations; a totally untenable concept before.

If you ever watch these guys in an aquarium, you notice they're basically constantly chewing on things. I've wondered many times how they keep such tiny teeth in good condition if they never given them a rest, but, here's why. Nature creates such cool creatures

In their defence, google searching anything about plants these days leads to awful results. It’s saturated with slop. A response from an LLM might be slightly more reasoned and targeted. It’s hard to tell. This is a category of knowledge that’s being destroyed by people gaming google and dumping huge amounts of bad LLM and image generation onto the web.

Fascinating, I am remarkably bad at this and consistently fail at the 8th word. This might be limitation caused by whatever ADHD does to me; in particular, holding the sets of letters in my memory as I try to compose words in my mind. It's a struggle. Super fun, though. I should probably practice this.

It's so absurdly sensitive. It bailed out earlier today working on a TypeScript client for a sensor network API which happens to include some temperature and pH sensors for tanks, which yes, are used for biology experiments. But wow, we're degrees of separation from the actual biology work.

It's making it very hard to justify even trying to use Fable. When it works, awesome; it's legitimately good. But I can't trust it to do a task without deferring to Opus and that's really annoying at times. I want to know what I'm getting up front, not after the fact.

I’ve been noticing that it will test hypotheses for evidence frequently. I really like that. With Opus I have to instruct it with a skill to build prototypes, but Fable will do it and do it really well by default. Really nice detail. I’m guessing it’s part of the system prompt, but the higher capability paired with this experimentation lean makes it far better at planning and verifying.

It’s literally the price of lost opportunities. It’s not understood to be a financial cost. It’s exactly what it says on the tin. Lost opportunities may not be a financial cost, but they’re certainly a cost.

I think this depends a lot on two things: how well-defined your requirements are up front, and how well-defined the problem space/domain are.

If there's some clarity and structure with clear expectations, it seems to produce fairly reliable, consistent results, and the process of validating specs with Claude can yield all kinds of unexpected dead-ends or holes in the idea you had. I find this very helpful.

But I'm curious too, what are the skills you'd use instead? superpowers are just skills themselves.

I didn't mean to paint with such a broad stroke, there. I've worked with people who are capable of writing exceptional software without type systems or other abstractions. I'm not that smart, though. I need a lot of guard rails to keep myself from doing stupid things. I was speaking more generally about people who didn't produce exceptional code yet still weren't open to means of improving their work if it meant having to open their minds to new ways of working.

This is totally fine sometimes though, because like you say, there are shops where this type of execution is suitable. It's why vibe coding is actually okay in many cases; it's not a lot different from what it replaces in these cases. There are a lot of situations where this is the bar, it produces enough utility an value, and that's great.

Cool, it worked!

https://drop-1e1a536f-10d.honeysuckle-gull.workers.dev/

It's minesweeper, but the logic uses xstate/store. The link in the bottom is broken; it's supposed to go to `building-minesweeper-with-xstate-store.html`

I have no need for this but I love that my friends could vibe out a website, drop it here, claim it, and host it for pennies. This is great.

"Your site is reachable within ~32ms of 95% of the world’s Internet-connected population" isn't new but it's cool to see that achieved so trivially.

The solar system diagram doesn't work for me. When I click on the planets, it will center on them. When I click on the sun, nothing happens. When I click on a planet next, it goes to the sun.

Absolutely, TypeScript is remarkably expressive in my opinion. The inference and option to bail out with `any` is nice for some teams in some cases, too. They did an excellent job of making it accessible.

I worked with people who would consider themselves serious, and are still in the industry and doing fine. A few have certainly gone on to be more prominent and get paid a lot more than I am—not that it's a perfect measure of seriousness.

In the early days they would often say things like "but we have prop types, why use TypeScript", "why not use JSDoc" (this made no sense at the time), or "it's an exercise in needless complexity". It was really tough to sell them on TypeScript for years.

I think there are developers who are very goal-oriented with a narrow perspective on getting from point A to point B, and their understanding of the process isn't particularly holistic, rigorous, or geared towards external or knock-on factors like maintainability, performance, bugs, etc. They deal with it when circumstances force them to, and no sooner. Defining types is a complete waste of time to someone like that.

These people thrive where teams are primarily expected to just ship things, and in my experience they often hate needing to think about things like types, tests, or code quality beyond running a linter.

So, they're serious people in one school of thought. They contribute meaningfully to projects. I think they're a large constituent of the new class of vibe coders who laugh at you if you look at the code. That's fine, they're doing their thing, and there are more than a few ways to get programs into people's hands. That way just isn't the way I like to.

That’s fine, but I trust him to play the YouTube game as a result, not as an engineer.

He's clearly very knowledgeable about some things, but I think he has harmed his credibility be becoming a 'tuber who prioritizes thumbnails and hot takes over engineering.

I find it very useful. What do you use that’s better?

I like super powers because I can be more in the loop editing and understanding the plan. Letting fable loose is genuinely impressive, and probably perfect for vibe coding, but I can’t be that far away from the plan and steps for code that matters.

As for being great for vibe coding, it’s cool but I can’t justify that kind of cost for throwaway code. At this point I’ve had good experiences using fable to review code, but I’m totally content with opus for all of my workflows still. If fable was the same price I’d switch (for projects not involving biology, at least), but I’d still use something like super powers to stay in the loop.

Anecdotal, but working on a biodiversity-related project has made Fable completely useless to me. It sends literally everything to Opus immediately because it’s related to biology. No word from Anthropic since I sent feedback, so I won’t bother relying on it and I don’t plan to.

I get that it’s biology and there are potential risks in that field, but this is stuff like making sure the name of a worm exists in a database or coordinates where the worm was seen are valid. Fable won’t even inspect the purpose, though.

I totally understand the US admin played a significant role here. I don’t think it was the only role, though. I also think the classifier is way too crude. I doubt I could do better on short notice, or perhaps it’s due to external constraints entirely. Considering the cash my organization dumps on them, we hoped for better. Such a disappointment. We’re all pretty leery now and investigating other models.

I actually posted praising Fable’s code review quality when it first debuted, and someone accused me of being an LLM (ha). This new version gets no praise whatsoever, though. It’s such a contrast.

Lately I wonder if people should learn philosophy because it teaches you how to reason, what to reason about, and why. Just like with programming, you’re constantly forced to interrogate your impressions and reconsider what you took for granted. It’s an extremely useful exercise. Nothing will show you how wrong you constantly are like testing the logic of a program you wrote.

It’s a bit like learning to program, but without a compiler as the referee or the domain constraints. Maybe that’s where we should put more energy if learning to think is the goal, though I don’t know what could replace the purely logical and verifiable qualities of programming. That isn’t so readily available with philosophy, for better or worse.

We do need people to practice thinking and self-interrogation far more than we do today.

We will always find ourselves in situations in which our existing laws allow for people to fall through cracks. I can see pardoning as a way to address that partially, yet it doesn’t explain what an adequate judgment would be instead. It’s simultaneously important in niche situations and yet too coarse of a solution.

What did you expect, though?

dead-dove.jpg

They also offer some degree of curation, so your dollar goes further by volume/weight/unit on average, but also in less quantitative ways quite often. I trust what I buy from Costco, but I’ve completely stopped buying from Amazon (many years ago now) because apart from the poor value prop, I simply don’t like the quality or reliability of what’s offered.

Genuinely one of the most fun performances I've ever been to, as well! The small venue accentuated Al's insanely huge personality.

I’m with jpeeler here and I don’t think there’s a single optimum harness. Pi is really nice in that it allows you to hone in on what suits your needs, though. Some people won’t want to manage their harness though, so for them it’s probably not ideal.

All I can say with certainty is that Claude Code isn’t a panacea, and it’s quite buggy and clunky. I think this is evident just from regular usage, but becomes even clearer when you use harnesses that get in the way less and demonstrate better implementations across the board. Sometimes I think Claude Code feels like software that happened while other harnesses feel like they were engineered intentionally.