agumonkey

Yes a LLM is mostly an information memory de vice, but as a worker you still have to vet it.

I believe the human constraint made us find ways to layer and modularize systems so it is easier to check. Now the LLM produces a lot of code and abstraction have to be enforced. Something is missing.. (resonates with the latest Pocoo blog)

how are things in this industry ? are you doing low level cpp ? python glue code ?

So you're on the engineering side of the animation pipeline ?

oh yeah i remember how maya was architectured so well, and from the start indeed. i don't know the innards of blender but the python layer seems powerful.. yet afaik there's no derived computed attributes like $ball.x = $feet.x / 2 or that would require using some 2nd class objects (transformators or something like this), quite un-ergonomic.

was this all builtin maya features or a lot of commercial extensions for game artists ?

I've always been confused how all the CGI world is helping you get into flow with complex data while IDEs felt so heavy and glitchy.

Interesting, maybe major customers are insisting on other features ?

How long after were these put down as writings ?

yeah i get it too, i'm just flabbergasted that this is today's market

it reminds me of the pre-vulkan game programming days.. drivers were black boxes, game developpers had to resort to magic tricks to do stuff, until everybody got fed up and wanted some logical ground to operate

I'm unable to understand how much value can be in low-definability non deterministic prompts. It feels like the kept the right divinity spell into a chest.

is it common to measure LLMs with landing pages like this ? it seems a bit too simple

Did you work in this space before ?

interesting project, good luck

I still want to try audiobooks more, but I agree that cutting your mind and diving into the words is something quite special. And I was convinced that ebook readers or tablets or whatever from the 2010s would be able to top that but so far I never saw it getting close. Weird.

Funny for some reason the submission title summoned this manga in my mind even though I've never read it only a free chapter in a magazine I was obsessed with.

I'm always surprised by the abstraction level of Sanskrit ideas. Very rapidly they talk about infinite ways of assembling rules. Was there even a strong motivation (economic or logistical value) behind this?

I see strange market forces where companies need to be trendy, so they absorb what's in the air, and then people kinda have no choice but to buy it. Nobody really wanted to but it materializes.

Around me every house now has tall anthracite fully closed fences. It's gloomy as hell, streets went from cute garden with wood fences to lock-down mode in a few years. It's all the same model, all the same vibe..

My usual thought on this is that I don't want to get stuck at ctrlp ctrlf level. I always pick a tool that gives an intermediate expressiveness level even if it means a bit more efforts.. especially if it's not gui because I can reuse and compose it.

For instance jq falls too far on the capabilities curve. It's a nuclear weapon but it's almost a programming language and I never can keep the operators in mind (even though I loved the idea at first).

"invisible" is something I used to describe emacs magit. It's a thin layer over git output, it pops up, it infers parameters from ui state, call the usual git command, and goes out. Lean and fast (unless large project it seems).

Also because lisp core can be reduced to a few "equations" (at least that was my understanding)

I won't disagree but most books and people who made lisp in the early decades were somehow hinting at not manipulating state but trying to represent ideas and operators to raise the level of expressiveness without shooting yourself in the foot (as much as possible)

it's dark side-ish at the ast level mostly, which is not the same as the dark side of mutable state over byte arrays

Is it your first rewrite/migration (with or without llm) ?

good luck nonetheless

I wonder if anybody has an implicit fear that with LLM you're expected to be a 20x engineering all the time, otherwise you're out. Can also lead to people producing shiny apps that impress others (sometimes for legit reasons) even though they have no idea how anything work. A "ship value" culture will not bother with the inner workings or actual skills.

Then you'll have postgrest on postgrust

I do wonder if they print end of page lines for you to cut afterward (sorry couldn't read the whole webpage, maybe they explained it)

he has a way to make it lean and engaging that is very sweet

Thanks a lot

I'm very curious about inhibition failures in brain areas, especially between visual perception and motor control. I'm no neurologist but your brain seems to generate a lot of imaginary interpretation when sensing the visual field, but sometimes there are short circuit like failure that leak those potential imaginary futures with you current real self (leading to strange uncoordinated or overlapping motor control signals)

Just a guess, it may be that he understood distribution more as the static distribution of meaning in fine grain entities (objects) to allow a wide range of context, flexibility. Not necessarily the time aspect of distribution.

How was it ? Useful or mild or dreadful?