hankbond

Design is illegible. Needs more contrast between text and background. I'm not even vision impaired and I was squinting immediately.

I feel like it's directionally like telling depressed people to just "stop being depressed". Works for me because of where I'm at currently but idk how generalizable it is.

RAF files contain (or at least can be configured to contain) a full size JPG embedded in them that will have the film sims baked in (as all JPGs do). You can use EXIFTools to extract them. Immich can be configured to make thumbnails off these embedded images instead of the demosaic'd RAF. I'm not sure if they are progressive JPGs but if they are I imagine it would be pretty efficient to directly read the images at a lower res and not even have to generate thumbnails (although that probably is less generalizable as a pipeline).

This is my current understanding and may be inaccurate.

This may be a very tone deaf thing to say, but what really helps me is to just tell myself I am happy. It's certainly not fool-proof when faced with genuine problems, but I often find myself unhappy or annoyed for no particular reason. In those moments I just tell myself I have everything I need and life is great, and then it just feels better.

That said all the findings that these studies seem to gravitate towards like close relationships also align with my experiences as being powerful effectors.

This is one of the cases where I think people really want a personal touch. Your README is the "hook" of a project, it's the first thing we see, and so you want to advertise "this thing was made with thought and care". Also the README does not need to be exhaustive, it should be a form of advertising and you can nest the lower level documentation in a /docs and link the sections from the README.

I would argue that the choice to render the embedded JPG is optimal from the UX of a Fuji user. We tend to use the JPG as the creative reference we want to start from using the film sims.

I appreciate the effort and I think this is a great way to engage a broader audience. This level of software complexity is perfect for non technical users to vibe code.

Non technical users cannot code. If you want non-technical users to customize the product they purchased you have three options. - Add a narrow surface of customizability hardcoded into the application: colors, maybe fonts, picking from the plugins others have created - Provide a bounded method for using LLMs to generate some code from user requirements, plugin documentation (with code/architectural conventions), and probably some guardrails build process. - Tell them to just YOLO using vanilla claude code/codex

Previously most "personal computing" (building custom stuff for your specific needs) was limited to people that either already are intimately aware of how computers and programming work, or people that were interested in starting down that path.

Plenty of people do not have the time/energy to go down that path, but still want custom stuff. Yes you can't zero-shot robust enterprise software, but the models are good enough to few-shot solid results when it's bounded to some interface requirements (like an SDK).

I don't make many predictions but this is one I am comfortable saying is the future of personal computing and what "harnesses" will look like in the future. Models will shrink and post-train to the use case at the enterprise level, and agent harnesses will go from things like codex/claude code to what look like self-assembling software. Give regular users an app with a plugin system, an agent, and a clear set of entry points and thats all they need to run.

the era of wabi-sabi is among us.

I would be fine lauding Gemini models if the only benefit of them was superior understanding of intent (read between the lines). I don't need it to code because other models are tuned for that explicitly, but I would like a model that is tuned to produce less mechanical output.

It's the next day and I'm still thinking about how much of a difference that small touch made.

I read about 80% of this and it ended up being more of a puff piece than anything. I'm not really sure what I was supposed to get out of this other than a description of one persons tenacity in general?

400 bucks for some vibe coded templates to create vibe coded websites?

Is model routing going to become the next "look at this memory system I made" but for enterprise?

Did you animate the cars separately? That really did a lot to add to the immersion (and the sound effects too). If nothing moved I think the effect would be significantly diminished.

Not trying to be pedantic but do you mean half the speed of C?

Enjoying this, starts simple but gets pretty tricky! Really appreciate the simple "teach you a new mechanic" levels. Very intuitive without needing text to explain, well done on that!

Jellyfin is great for my needs, and I appreciate the effort that goes into it. Recently was having rendering issues doing HDR -> SDR tone mapping since I watch my content on a 2008 era Pioneer plasma display. I found that oddly enough while the server is pretty solid, the client experience is a bit hit or miss. I'm using the mpg-shim adapter, and I'm looking forward to them building a native app with better native render compositing.

Did you consider using WebDAV here?

Thank you I can see why it only works on linux. I can see this being successful with the SteamOS variants. Would this be a simpler way to share a gaming box across two different clients simultaneously? Like if my kids only played lightweight titles.

Does anyone have a name for that really neat "the bar chart is a bookshelf, you can hover a spine to see the cover" vis? I've never seen it before.

Yeah this is a miss for me. I've never felt that this was a considerable source of friction. Certainly not enough to justify a dependency.

I guess... just, why?

Dang I'm not really sure what to say but thank you (author) for writing this. This was evidently written with love and care.

Moonshine runs each streaming session in its own isolated environment, separate from your desktop. This means your host PC stays usable while you stream, and it works without an active desktop session.

So does this create like a virtual desktop? How does that work? Is this just a separate "screen" or is this more like an entirely separate compositor instance (if thats the right term here)?

This is a neat project and I don't think I'm aware of any sequencers that allow for arbitrary execution per-step (not that I am that deep into the space). The font choice for the website is really... interesting. I don't think I can recall ever feeling this dyslexic trying to read the copy here.

one of the points of doing psychedelics is to recapture your own imagination

this also is my takeaway as to the primary utility. its so easy to accumulate ideas about how things _must_ be represented over time. every so often (maybe 5-10 years?) its probably good to have a way to shave them off and re-expose the surface.

Lmao whoops, I meant gentle.

Dang that's my bad. Thank you for the gentile direction.

What is the current state of understanding on high cholesterol as a leading or trailing indicator of poor health? I vaguely remember reading one theory that cholesterol was used by the body to mitigate arterial damage from other causes?