Toasty
@toastyraptor.bsky.social
He/Him. I'm pretty flexible on my identity though.
Sometimes I create art, music, and code.
Alexithymia and Mania. Brain is wonky sometimes but I try my best.
Sometimes I create art, music, and code.
Alexithymia and Mania. Brain is wonky sometimes but I try my best.
Right. It annoys me when people use the first amendment as support for cruelty because it means absolutely nothing outside of government involvement and requires a caring government to actually enforce it anyway.
December 1, 2025 at 6:59 PM
Right. It annoys me when people use the first amendment as support for cruelty because it means absolutely nothing outside of government involvement and requires a caring government to actually enforce it anyway.
"i'll always defend trans rights as i have all my career. but"
Ahh the old "I'm not racist but" and "i'm not homophobic but" just with transgender rights.
That But is always followed by contradicting views.
Ahh the old "I'm not racist but" and "i'm not homophobic but" just with transgender rights.
That But is always followed by contradicting views.
December 1, 2025 at 6:51 PM
"i'll always defend trans rights as i have all my career. but"
Ahh the old "I'm not racist but" and "i'm not homophobic but" just with transgender rights.
That But is always followed by contradicting views.
Ahh the old "I'm not racist but" and "i'm not homophobic but" just with transgender rights.
That But is always followed by contradicting views.
*inhales the serotonins* YESH.
December 1, 2025 at 5:57 PM
*inhales the serotonins* YESH.
It's a pointless question to ask if something is conscious. The definition can range from too simple to be meaningful to too complex to be provable anyway.
What's it like to be a cat? No amount of studying my cats seems to get me any closer to the answer. Maybe it's because i'm not a cat...
What's it like to be a cat? No amount of studying my cats seems to get me any closer to the answer. Maybe it's because i'm not a cat...
November 28, 2025 at 11:21 PM
It's a pointless question to ask if something is conscious. The definition can range from too simple to be meaningful to too complex to be provable anyway.
What's it like to be a cat? No amount of studying my cats seems to get me any closer to the answer. Maybe it's because i'm not a cat...
What's it like to be a cat? No amount of studying my cats seems to get me any closer to the answer. Maybe it's because i'm not a cat...
It's built off numerically unstable statistical foundations that require an insane amount of data, computational power, and money than anyone can truly comprehend.
November 28, 2025 at 7:04 PM
It's built off numerically unstable statistical foundations that require an insane amount of data, computational power, and money than anyone can truly comprehend.
"It’s very hard for me to “turn off” from work"
This! So much of this.
Work is work. That switch gets turned off entirely when away from work for the sake of mental sanity.
This! So much of this.
Work is work. That switch gets turned off entirely when away from work for the sake of mental sanity.
November 28, 2025 at 6:58 PM
"It’s very hard for me to “turn off” from work"
This! So much of this.
Work is work. That switch gets turned off entirely when away from work for the sake of mental sanity.
This! So much of this.
Work is work. That switch gets turned off entirely when away from work for the sake of mental sanity.
The standard ANN activation functions assume mostly linear representation when a lot of datasets are absurdly nonlinear and not perfectly continuous.
So you end up having fractured functions that span over multiple layers in hopes that they numerically behave as the dataset needs.
So you end up having fractured functions that span over multiple layers in hopes that they numerically behave as the dataset needs.
November 27, 2025 at 9:23 PM
The standard ANN activation functions assume mostly linear representation when a lot of datasets are absurdly nonlinear and not perfectly continuous.
So you end up having fractured functions that span over multiple layers in hopes that they numerically behave as the dataset needs.
So you end up having fractured functions that span over multiple layers in hopes that they numerically behave as the dataset needs.
For image data, there are some things you can implement as human reasoning. Like throwing away the color data (since it's not usually relevant) and capturing 2.5D structure with an edge projection.
My biggest gripe with ANN's is that their activation function is never ideal to a dataset though.
My biggest gripe with ANN's is that their activation function is never ideal to a dataset though.
November 27, 2025 at 9:20 PM
For image data, there are some things you can implement as human reasoning. Like throwing away the color data (since it's not usually relevant) and capturing 2.5D structure with an edge projection.
My biggest gripe with ANN's is that their activation function is never ideal to a dataset though.
My biggest gripe with ANN's is that their activation function is never ideal to a dataset though.
They have the same trouble that all ANN's do. They are too flexible and can represent anything tractable. Unfortunately that means that in order to do that, they need a crazy amount of data.
Using ANN's is when you absolutely can not inject any human insight at all... which is a terrible idea.
Using ANN's is when you absolutely can not inject any human insight at all... which is a terrible idea.
November 27, 2025 at 8:04 PM
They have the same trouble that all ANN's do. They are too flexible and can represent anything tractable. Unfortunately that means that in order to do that, they need a crazy amount of data.
Using ANN's is when you absolutely can not inject any human insight at all... which is a terrible idea.
Using ANN's is when you absolutely can not inject any human insight at all... which is a terrible idea.
Autoencoders are a hilariously inefficient way to capture feature compression but they certainly are a thing. 🫠
November 27, 2025 at 7:44 PM
Autoencoders are a hilariously inefficient way to capture feature compression but they certainly are a thing. 🫠
A surprising amount of them have a crippling fear of Irrelevance. That all their hard work will mean nothing at all.
November 27, 2025 at 1:03 AM
A surprising amount of them have a crippling fear of Irrelevance. That all their hard work will mean nothing at all.
I thankfully didn't get the something wrong part.
Just the personality divide that I can on occasion feel creeping up rapidly.
Like suddenly smelling food like you've been hungry for days that dulls the inhibitions. A sharpened clarity of the senses and the dissonance that comes with it. 😅
Just the personality divide that I can on occasion feel creeping up rapidly.
Like suddenly smelling food like you've been hungry for days that dulls the inhibitions. A sharpened clarity of the senses and the dissonance that comes with it. 😅
November 27, 2025 at 12:58 AM
I thankfully didn't get the something wrong part.
Just the personality divide that I can on occasion feel creeping up rapidly.
Like suddenly smelling food like you've been hungry for days that dulls the inhibitions. A sharpened clarity of the senses and the dissonance that comes with it. 😅
Just the personality divide that I can on occasion feel creeping up rapidly.
Like suddenly smelling food like you've been hungry for days that dulls the inhibitions. A sharpened clarity of the senses and the dissonance that comes with it. 😅
Closest thing I have to dysphoria is the thin wall between this life and that one. I wouldn't quite call that discomfort or unease though.
More self-fear I suppose. 😅
More self-fear I suppose. 😅
November 26, 2025 at 11:21 PM
Closest thing I have to dysphoria is the thin wall between this life and that one. I wouldn't quite call that discomfort or unease though.
More self-fear I suppose. 😅
More self-fear I suppose. 😅
That checks out I think.
Especially if you have any kind of dysphoria. That would be a lot of conceptual power to change whatever you wanted at a moments notice.
Especially if you have any kind of dysphoria. That would be a lot of conceptual power to change whatever you wanted at a moments notice.
November 26, 2025 at 11:18 PM
That checks out I think.
Especially if you have any kind of dysphoria. That would be a lot of conceptual power to change whatever you wanted at a moments notice.
Especially if you have any kind of dysphoria. That would be a lot of conceptual power to change whatever you wanted at a moments notice.
I just have boring and common kinks. Haha
November 26, 2025 at 11:16 PM
I just have boring and common kinks. Haha
I can't say i'll ever quite understand that kink. 🙃
November 26, 2025 at 11:15 PM
I can't say i'll ever quite understand that kink. 🙃
Had to google that one. That's just uhhh... that's just shapeshifting?
November 26, 2025 at 11:12 PM
Had to google that one. That's just uhhh... that's just shapeshifting?
Uhhhhh. Wut. Hahaha.
November 26, 2025 at 11:09 PM
Uhhhhh. Wut. Hahaha.
Uhhh.... did this dude really just "you can't experience temporary family/friend happiness because only your country celebrates that"?
Huh.
Huh.
November 26, 2025 at 10:56 PM
Uhhh.... did this dude really just "you can't experience temporary family/friend happiness because only your country celebrates that"?
Huh.
Huh.
Sometimes... we recognize it. 🙃
November 26, 2025 at 7:56 PM
Sometimes... we recognize it. 🙃
Nifty. A lot of my recent research is... theoretical.
Optimization methods for gradient based algorithms.
But someone has to build the backbone to make it all happen.
Optimization methods for gradient based algorithms.
But someone has to build the backbone to make it all happen.
November 26, 2025 at 4:23 PM
Nifty. A lot of my recent research is... theoretical.
Optimization methods for gradient based algorithms.
But someone has to build the backbone to make it all happen.
Optimization methods for gradient based algorithms.
But someone has to build the backbone to make it all happen.