The rise of Slow AI - Why devs should stop speedrunning stupid (CodeRabbit)
https://www.coderabbit.ai/blog/the-rise-of-slow-ai-why-devs-should-stop-speedrunning-stupid
Ferns
Latest posts
AI Code Reviews (Hanselminutes)
What if AI could not only speed up your code reviews—but actually make them fairer? In this episode of Hanselminutes, Scott Hanselman chats with Howon Lee, Senior Software Engineer at CodeRabbit, about how AI is changing the game when it comes to code reviews. They dig into the unspoken dynamics that can pop up in traditional reviews—like how seniority or team politics sometimes matter more than…
Diagonalization-Power-Law Conjecture
Conjecture: We should be able to unify the dozen (two dozen, three dozen, go read MEJ Newman’s review ) putative universal causes of power law phenomena by proving that a Cantor-style diagonalization process can come up with them at the probabilistic-logical level.
Why are there no nonexponential nonlinear loans?
The amount paid in interest in the often-taught, rarely-used simple interest regime is $P(1 + rt)$, where $P$ is principal, $r$ is the interest rate and $t$ is the time. The main influence on the behavior of such functions is their functional form - the parameters can be futzed with and are obviously material in practice but don’t fundamentally change the behavior in the typical interest domain…
Suzerainty in Megacorps
I often hear questions about “How is management at FooCorp?” where FooCorp is some massive white-collar firm of 20,000 people. This is a futile question, because like other megacorps, FooCorp is undoubtedly feudal in important ways. Feudalism, at its core, consisted of three parts: an attitude towards the means of production, an attitude towards religion, and an attitude towards violence and…
Translucent Fractal Tteok
This is really a cooked potato starch oobleck, but it’s translucent and the texture is uncannily like tteok so that’s why I’m calling it tteok. Read on for why I mentioned fractals. Liangfen noodles are a not-totally-different recipe but the water concentration is very different and the very different mung bean starch is used. There is also a not-totally-different preparation of translucent potato…
Martingale Project Timing
The previous introduction was pretty arcane, so this one should be a little less. Suppose we start one day with one task, Bingling Zoobleybops. It’s day 0. Unlike in other methods of project timing, especially estimation, in Martingale timing we always rigidly set a next time of 1 day to look at it again. One day passes. Day 1. We look at our progress in Bingling Zoobleybops. Perhaps it’s a…
My Software Estimation Method, Where Every Task Is Estimated At Infinite Time
When I estimate software projects at work, I conform to the way we do it at work. But I can do software estimates the way I want to for my own projects, so I do them in this strange way. Despite the confounding premise, it is blisteringly simple. Estimation Set a minimum time for an issue. I usually use 1 day. Call that t0 . At the end of that time is a decision. I decide whether to do the task or…
Regulating Social Media Addictiveness
There’s a lot of other terrible things about social media, but a specific material terrible thing that might be the subject of regulation is how addictive it is. You don’t get to the usage per day that social media platforms have without at least a little addictiveness. It’s addictive! We regulate liquor and cocaine and casinos, why not regulate the addictive structure of social media? I suggest…
Program Abstraction as Gambling
Note: I hope I don’t insult any gambling folks by the comparison or offend anyone with real gambling problems. Those are of course much more serious. Abstraction exists, even if folks don’t always agree about what it is. It is a stable term, if not a stable thing, which can be talked about. Abstraction also has benefits. Lots of people disagree on the nature and the distribution of the benefits,…