Founder @ https://xitter.cf/t.co/wiehRakDRc - Find your Story with Ozu - Computational Cinema. He/Him - Art & Dev.

Where the Thunder Lives
Joined April 2009
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
Replying to @c_valenzuelab
Culture series books by Iain M Banks has some interesting takes on this too. Player of Games comes to mind.
1
99
Replying to @log4j @DavidBGrys
Do you have evidence it’s Samsung and not a supply chain attack by a nation state putting monitoring software clandestinely ?
2
1
800
React and other declarative ui tooling don’t suffer as much because it’s open source. You can look under the hood and best practices can be worked on by a community. Apple needs to reframe its open source stance in 2027 asap.
1
2
49
Replying to @fromjrtosr @mredig
I think it’s clear it’s just due to the declarative paradigm. By definition there’s hidden mechanism that takes instructions of composition and turns it into “mechanical” linkages and data passing. The abstraction is hidden by design.
1
2
56
Replying to @dvsch @cyber__sin
Vision is trained. As is cot data, tdd style dev is absolutely in there. I think the thing we are seeing is how much novelty is left in combinatorics and transfer learning. I’m not knocking it by any means! This is maybe the most important moment in human technological history.
16
Replying to @mredig @fromjrtosr
Eh kind off kind of not? It’s easy to use SwiftUI “wrong” and small changes can have huge performance ramifications, yes. But that’s true with any system. Maybe I’m misunderstanding the complaint?
1
1
17
Replying to @dvsch @cyber__sin
Yeah those folks need to catch up so their arguments track the fast changing reality. That doesn’t change them being very amazing stochastic parrots that I am constantly surprised by. But when you compress all the worlds knowledge into lerpable vectors :D :D
1
1
27
Replying to @dvsch @cyber__sin
I think you may underestimate the diversity of data these modern frontier models train on? It’s absolutely in the data set. But that isn’t the claim. The claim is anything grossly outside the data set distribution it tends to fail on. Not “it regurgitates training data verbatim”
1
23
Replying to @fromjrtosr @mredig
All code can be slow. This is the risk is giving folks composable systems. You can make AVFoundation slow. You can make Metal slow, UIKit slow, etc. You can use tools in ways they were not meant to be used or just poorly. Optimization is a skill like writing code is. Learn it.
1
1
21
Replying to @sarunw
Res motha-fucking-edit baby.
1
42
I don’t think any of the folks who claim these are good work in the space, there’s so much more than a ui that does sort of what you expect. There’s a slew of considerations, diverse codec support, color management, plugin ecosystem, standards compliance and conform.
27
Replying to @doristsao
I have a Q Dr Tsao - is there evidence or speculation that anesthetics drugs modulate the neuronal mode switching - causing them to not attend to conscious mode signals, making one unconscious? Or adrenaline increasing the rate, making one feel time slow down? Thank you!
7
847
Consciousness runs on a realtime os with a very low resolution scheduler
Replying to @doristsao
8/ At the single-neuron level, we found that cells concertedly switched their tuning every 60 ms, from representing the conscious face to the suppressed face. Thus the same population of neurons represent both what you’re aware of and what you’re not, at different moments of time.
1
3
217
Replying to @hishnash
Nice. There’s so many levels to this game.
44
Replying to @hishnash
This caught my attention. Do you have or know any writeups showing those kind of improvements? Curious to learn more.
1
1
300
Replying to @quasimondo @dh7net
That art of holding a fucking grudge for the rest of your life out of spite.
1
28
Replying to @matamalaortiz
How aligned is the visual thinking to what’s actually happening? Is it illustrative (ie the model is reasoning about how to show what it’s doing) or is it coupled to actual reasoning and is descriptive? In other words are both outputs subject to different hallucinations?
156
Replying to @dh7net
Great now i can add another layer of indeterminacy to software I write. No one will know what its going to do, but ship it, fast, we gotta get imaginary internet points!
2
56