yayitswei

One point for TUIs I haven't seen mentioned: they're faster for coding agents to iterate on. No need for screenshots or computer use. Agents have made GUIs cheap, but have made TUIs even cheaper, both to build and verify.

The linked article is such a gem: You can choose how many bugs you want now https://nolanlawson.com/2026/08/16/you-can-just-choose-how-m...

Love the concept of agents adding epicycles to fix bugs. It maps closely to my experience.

If we're leaning into audio, why not always-on instead of push-to-talk?

Related: Rich Hickey's 'Simple Made Easy' (2011) https://youtu.be/SxdOUGdseq4

Classic talk that still influences my design decisions today.

Try a coding agent for writing and tuning the OpenCV part, and have it explain its choices. That's probably the most practical path to shipping a working system.

Speaking from experience: never used OpenCV before, recently vibe coded a tool that makes supercuts of pool videos, trimming each clip from the cue ball's first strike to when the motion stops.

In case anyone was wondering, the IOCCC specifically permits LLM use in their guidelines.

"The IOCCC has a rich history of remarkable winning entries created by authors who skillfully employed various techniques (often their own tools) to develop their code."

What if you could finish your passion project in an evening? Would that increase not the time, but capacity for deep exploring and side projects?

Expensive because they were perceived as rare. Classic article: https://priceonomics.com/diamonds-are-bullshit/

Synthetic has pretty much eaten their lunch now.

I noticed that Futhark the programming language is also on the front page.

Loved it.

This is one of the few articles where I noticed a bunch of LLM-isms and still read to the end because it was interesting.

I think of Tesla autopilot as sophisticated cruise control. Can perform most driving tasks better than I can, saves a lot of cognitive work, still needs close of my 100% attention.

How ironic that agents were supposed to free us from labor, but instead they're ushering in 996 culture.

Is there some magic lost also when using AI to write your blog post?

Fwiw I was going to make the same comment about the naming, but you beat me to it.

What's up with these?

Genesis 1:13, Eve optimal Replace 'Then' with 'And' in optimal ('And the LORD God said') and poetic_daily to preserve narrative vav-consecutive connective consistently.

Love Tonsky's articles. My favorite is about centering things: https://tonsky.me/blog/centering/

He has a knack for putting words to the vague frustrations I feel but can't quite articulate. How does he find so many perfect examples that nail exactly what's wrong?

Tried it in earnest. Definitely detect some aggression, and would feel stressed if this were an exam setting. I think it was pg who said that any stress you add in an interview situation is just noise, and dilutes the signal.

Also, given that there's so many ways for LLMs to go off the rails (it just gave me the student id I was supposed to say, for example), it feels a bit unprofessional to be using this to administer real exams.

That was my first impression as well.

Looks like the theme changes are part of the arrangement (see lines 135-149).

I'm cash flow positive on my SMS sales agent, it serves just one client and my revenue is at least 3x the cost of inference/hosting.

Imo the key is to serve one use case really well rather than overgeneralize.

Some of those, like horses, are 1% hobbies. But many of the others can be done very affordably. Buying used equipment, learning from YouTube and online resources, starting small and scaling gradually make most of those hobbies accessible at a fraction of the cost.

One of the authors is the Kenton who built this awesome lan party house: https://lanparty.house/

Gen Z takes selfies with the higher quality back-facing camera using 0.5x zoom. You can't see yourself while taking the photo, but that's part of the appeal.

Awesome. I wonder if there is a pulse sequence that maximizes that feeling of randomness?

Management roles have always involved outsourcing cognitive work to subordinates. Are we seeing a cognitive decline there too? Maybe delegation was the original misalignment problem.