JonathanRaines

*not true of hammers (jk, great post)

I don't think you mean "Alas, ScrollPods was born". Maybe "And thus,..."?

Pangram claims to be super accurate. Do you know how? I understood detection is a hard problem.

I hope they don't start bugging our homes.

I'm setting up my site and (slowly) adding some previous projects. Enjoy.

High altitude balloon launches using weather balloons to get photos of the curvature of the earth.

We estimate that these laws prevented fatalities of 57 children in car crashes in 2017 but reduced total births by 8,000 that year

These two options are not equivalent. I will take 8,000 3rd children not being born over 57 kids dying in car crashes every time.

I've used your library and really like it!

Also check out CADQuery and or build123d (python equivalents)

I agree with what you're worried about - humans using AI to do bad things. However, have you considered also being worried about what AI can do? (can you ever have too much existential dread?) You cast it as a tool, it may be the first thing we've made that goes beyond that.

"According to the laws of thermodynamics, all that exists does so solely to consume, destroy and extinguish, and in this way to accelerate the slide toward cosmic obliteration"

Bit of a leap from "heat cannon, of itself, pass from one body to a hotter body"?

Nice! They provide trained checkpoints on their GitHub. Repeating their results would be a good start. https://github.com/sapientinc/HRM

I advise scepticism.

This work does have some very interesting ideas, specifically avoiding the costs of backpropagation through time.

However, it does not appear to have been peer reviewed.

The results section is odd. It does not include include details of how they performed the assesments, and the only numerical values are in the figure on the front page. The results for ARC2 are (contrary to that figure) not top of the leaderboard (currently 19% compared to HRMs 5% https://www.kaggle.com/competitions/arc-prize-2025/leaderboa...)

I think you are right that once the exercise becomes hunting for a scapegoat it's pointless.

However, it can be a way for everyone to understand the system better. The goal should be making each of the dominoes less likely to fall. Doing so can simplify rather than add complexity.

"You seem stuck on X. Do you want to try investigating Y?"

MS Clippy was the AI tool we should all aspire to build

Zed is worth a look for your IDe and plays nicely with LLMs

Fascinating that you chose to compare OpenAI's culture to Los Alamos. I can't tell if you're hinting AI is as world ending as nuclear weapons or not.