ravila4

I think this would’ve been very handy to me a couple years ago when I was learning to use Blender and asking LLMs for help performing certain actions like “how do I display the normals of all the vertices in my mesh?” I spent a lot of time trying to figure out which button the model was talking about.

Thanks, this is helpful!

I used voice dictation to transcribe my rambling thoughts on the article, which was thereafter re-arranged into a still verbose but somewhat more coherent reply. I should have taken the time to reword it properly - my laziness is on me.

I was not aware the rule against AI-assited text at the time, and will be more careful next time. Nonetheless, I find it unfortunate that we've spend more time talking about AI usage in a comment, than the blog which prompted it.

To reiterate, those were legit questions. I used voice dictation to my phone, with an app that transcribed and formatted my message.

Interesting project! I scan 35mm and medium-format film with a mirrorless camera, macro lens, and Lomography film holder, then convert with Negative Lab Pro in Lightroom. I’m happy with the results, but the Lightroom subscription has me considering writing my own converter.

Could you share more about your NumPy pipeline, or publish the code? I’m particularly curious about how you use the blank film reference and roll-level measurements to handle the orange mask, color balance, and tone curves. Do you use camera/scanner calibration profiles or film-stock-specific adjustments, or estimate everything from the scans themselves?

Also, your screenshot appears to show an OpticFilm 7400. Would you recommend it, and roughly what did you pay? Have you tried camera scanning? I’ve been considering automating film advance and capture on my setup, but consistent color conversion is the part I’m least confident about.

Yeah, I get that form follows function, but I’m hoping for more native features of quality control, some sort of dashboard to catch slop. Also looking forward to more domain-specific harnesses.

Why do all harnesses end up looking the same?

Exactly, and the classical lock and key analogy ignores electromagnetic and hydrophobic/hydrophilic interactions, which we can fully model in classical molecular dynamics systems

Why build a whole website for what is essentially a markdown file?

Your brain contains the script of what little people can do, it just takes a compound to activate it.

Fable's ridiculous. It's flagging basic biology research questions as a security risk. I'm talking basic fundamental genetics topics that make working on any genetics-adjacent codebase unusable.

As a workarond, I have noticed that a better estimate of time is number of sessions required to fully implement, review, and test a given feature. You can ask the LLM to help you estimate how complex a feature will be during the planning phase. The number of tokens can easilly be translated to sessions, asumming that you normally aim to keep a session below 60% of the context limit. As a rule of thumb, a simple feature can be fully implemented, tested and reviewed in a single session. Whereas a complex feature might take multiple sessions. Then depending on your workload and attention span, you can have a decent idea of how many full context sessions you can actually plan and review in a single day. Of course, some features can happen in parallel but in general the bottleneck is your ability to understand the context that goes into each session.

As someone with ADHD, it’s a lot more nuanced than that. Coding agents can remove task paralysis, but they also introduce many other distractions. Being one prompt away from zero to one is a double edged sword, because it means any random thought, idea and side project is also a prompt away.

Glad you find it useful!